diff --git a/.cargo/config.toml b/.cargo/config.toml
index 69740df..c6df80c 100644
--- a/.cargo/config.toml
+++ b/.cargo/config.toml
@@ -4,6 +4,7 @@
 [build]
 # Number of parallel jobs for compilation
 jobs = 4
+target-dir = "/Volumes/Devel/ontoref/target"
 
 # Code generation backend
 # codegen-backend = "llvm"
diff --git a/.gitignore b/.gitignore
index 6ba45c4..ae25bf2 100644
--- a/.gitignore
+++ b/.gitignore
@@ -2,6 +2,8 @@ CLAUDE.md
 .claude
 logs
 logs-archive
+.ncl-cache
+data
 utils/save*sh
 .fastembed_cache
 presentaciones
diff --git a/.ontology/core.ncl b/.ontology/core.ncl
index 5bbaef5..d1216e9 100644
--- a/.ontology/core.ncl
+++ b/.ontology/core.ncl
@@ -85,9 +85,16 @@ let d = import "../ontology/defaults/core.ncl" in
         "adrs/adr-010-protocol-migration-system.ncl",
         "adrs/adr-011-mode-guards-and-convergence.ncl",
         "adrs/adr-012-domain-extension-system.ncl",
+        "adrs/adr-013-vcs-abstraction-layer.ncl",
+        "adrs/adr-014-runtime-service-toggles.ncl",
+        "adrs/adr-015-mcp-tool-inventory-auto-derive.ncl",
+        "adrs/adr-016-component-lift-out-pattern.ncl",
+        "adrs/adr-017-registry-credential-vault-model.ncl",
+        "adrs/adr-018-level-hierarchy-mode-resolution-strategy.ncl",
+        "adrs/adr-019-per-file-recipient-routing-tenant-isolation.ncl",
         "CHANGELOG.md",
       ],
-      adrs = ["adr-001", "adr-002", "adr-003", "adr-004", "adr-005", "adr-006", "adr-007", "adr-008", "adr-009", "adr-010", "adr-011", "adr-012"],
+      adrs = ["adr-001", "adr-002", "adr-003", "adr-004", "adr-005", "adr-006", "adr-007", "adr-008", "adr-009", "adr-010", "adr-011", "adr-012", "adr-013", "adr-014", "adr-015", "adr-016", "adr-017", "adr-018", "adr-019"],
     },
 
     d.make_node {
@@ -214,11 +221,12 @@ let d = import "../ontology/defaults/core.ncl" in
       name        = "Ontoref Daemon",
       pole        = 'Yang,
       level       = 'Practice,
-      description = "HTTP daemon for NCL export caching, file watching, actor registry, and MCP surface. Provides notification barrier, HTTP API (11 pages), MCP server (29 tools, stdio + streamable-HTTP), Q&A NCL persistence, quick-actions catalog, passive drift observation, unified auth/session management, per-file ontology version counters (GET /projects/{slug}/ontology/versions), and annotated API catalog (GET /api/catalog). API catalog populated at link time via #[onto_api] proc-macro + inventory — zero runtime overhead. Launched via ADR-004 NCL pipe bootstrap: nickel export config.ncl | ontoref-daemon.bin --config-stdin. Graph, search, and api_catalog UI pages carry browser-style panel navigation (back/forward history stack). File artifact paths open in external tabs: card.repo (Gitea source URL) for most files, card.docs (cargo docs) for .rs files — no inline file loading. card_repo/card_docs injected into Tera context from insert_brand_ctx; | safe filter required for URL values inside <script> blocks.",
+      description = "HTTP daemon for NCL export caching, file watching, actor registry, MCP surface, and GraphQL API. Provides notification barrier, HTTP API, MCP server (stdio + streamable-HTTP), GraphQL endpoint (async-graphql 8, Apollo Federation v2, WebSocket subscriptions), Q&A NCL persistence, quick-actions catalog, passive drift observation, unified auth/session model (ADR-005), runtime service toggles (ADR-014), and annotated API catalog (GET /api/catalog). API catalog populated at link time via #[onto_api] proc-macro + inventory — zero runtime overhead. Launched via ADR-004 NCL pipe bootstrap: nickel export config.ncl | ontoref-daemon.bin --config-stdin. Optional services (MCP, GraphQL) are feature-gated and runtime-toggleable via PUT /api/services/:service or UI manage page — changes take effect immediately without restart.",
       invariant   = false,
       artifact_paths = [
         "crates/ontoref-daemon/",
         "crates/ontoref-daemon/src/api_catalog.rs",
+        "crates/ontoref-daemon/src/graphql/mod.rs",
         "crates/ontoref-daemon/templates/pages/api_catalog.html",
         "crates/ontoref-derive/",
         "install/ontoref-daemon-boot",
@@ -233,6 +241,8 @@ let d = import "../ontology/defaults/core.ncl" in
         "crates/ontoref-daemon/src/ui/login.rs",
         "crates/ontoref-daemon/src/ui/search_bookmarks_ncl.rs",
         "justfiles/ci.just",
+        "justfiles/build.just",
+        "adrs/adr-014-runtime-service-toggles.ncl",
       ],
     },
 
@@ -479,6 +489,71 @@ let d = import "../ontology/defaults/core.ncl" in
       ],
     },
 
+    d.make_node {
+      id          = "mcp-surface",
+      name        = "MCP Server Surface",
+      pole        = 'Yang,
+      level       = 'Practice,
+      description = "Model Context Protocol server exposing ontology state, reflection modes, ADRs, backlog, config, Q&A, and search bookmarks as structured tools. Two transports: stdio (for AI assistants that launch daemon as subprocess) and streamable-HTTP at POST /mcp (for remote MCP clients). Authentication uses the same unified session model as the REST API (ADR-005) — Bearer token from POST /sessions. Runtime-toggleable without restart (ADR-014). Connect: (1) stdio — run `ontoref-daemon --mcp-stdio`; tool list auto-discovered by the AI client. (2) streamable-HTTP — configure MCP client with url=http://127.0.0.1:7891/mcp, auth=Bearer <session-token>. Session token obtained via POST /sessions with project key. ONTOREF_TOKEN env var injected automatically by CLI. Available tools include: ontoref_help, ontoref_list_projects, ontoref_set_project, ontoref_search, ontoref_get, ontoref_get_node, ontoref_get_adr, ontoref_get_mode, ontoref_status, ontoref_describe, ontoref_guides, ontoref_list_adrs, ontoref_list_modes, ontoref_constraints, ontoref_backlog_list, ontoref_backlog, ontoref_action_list, ontoref_action_add, ontoref_validate_adrs, ontoref_validate, ontoref_impact, ontoref_qa_list, ontoref_qa_add, ontoref_bookmark_list, ontoref_bookmark_add, ontoref_api_catalog, ontoref_notify.",
+      artifact_paths = [
+        "crates/ontoref-daemon/src/mcp/mod.rs",
+        "crates/ontoref-daemon/src/api.rs",
+      ],
+    },
+
+    d.make_node {
+      id          = "graphql-surface",
+      name        = "GraphQL API Surface",
+      pole        = 'Yang,
+      level       = 'Practice,
+      description = "GraphQL endpoint over the ontology graph, impact traversal, project state, backlog, ADRs, cross-project config comparison, and notification stream. Built with async-graphql 8 (Apollo Federation v2 entity resolution built-in). Endpoints: GET /graphql (GraphiQL IDE), POST /graphql (queries/mutations), GET /graphql/ws (WebSocket subscriptions). Auth: if graphql.read_token_hash is set in config.ncl, all endpoints require Authorization: Bearer <token>. Generate token: `just generate-graphql-token` — prints TOKEN and the Argon2id HASH to add to config.ncl. Pre-auth GraphiQL: GET /graphql?token=<token> — pre-injects the Bearer header and WebSocket connectionParam so the IDE works without manual header setup. Runtime-toggleable without restart (ADR-014). Key queries: ontologyNode(slug, id), impactGraph(slug, nodeId), project(slug), backlog(slug), adrs(slug), crossProjectSummaries, crossProjectConflicts. Key mutations: emitNotification, proposeBacklogStatus, ackNotifications, ackAllNotifications. Subscription: notifications(project, eventType) — streams live notifications over WebSocket.",
+      artifact_paths = [
+        "crates/ontoref-daemon/src/graphql/mod.rs",
+        "crates/ontoref-daemon/Cargo.toml",
+        "adrs/adr-014-runtime-service-toggles.ncl",
+      ],
+    },
+
+    d.make_node {
+      id          = "registry-credential-vault",
+      name        = "Registry Credential Vault",
+      pole        = 'Yang,
+      level       = 'Practice,
+      description = "Per-project sops multi-recipient credential vault stored as OCI artifact in ZOT (ADR-017). Each actor role has its own age keypair — the sops DEK is encrypted separately per recipient, enabling revocation without key rotation. vault_key is decrypted into RESTIC_PASSWORD or KOPIA_PASSWORD for the operation duration only; never written to disk. The daemon is structurally excluded: it holds no .kage, cannot decrypt sops files even if it reads them, and cannot be used as a credential amplifier. DOCKER_CONFIG is isolated to a per-call tmpdir; no ambient ~/.docker/config.json is ever consulted. src-vault OCI artifact is cosign-signed on every push and verified before any pull — tamper-evident and substitution-resistant. Access logs are co-located in the vault artifact as an append-only jsonl layer. Two-level authorization gate: assert-actor-authorized validates ONTOREF_ACTOR→role binding + scope.bound_actor + scope.ops; assert-target-in-scope validates the OCI ref against scope.namespaces glob — both fire before any oras call, no cache hit bypasses. Vault lock OCI artifact (src-vault/<id>:lock) coordinates concurrent edits with TTL 60min and admin-only force-unlock auditable. Impact analysis on secrets-close diffs sops files since last snapshot and maps changes to RegistryEntry IDs before push. Per-file recipient routing via project.ncl::sops.recipient_groups + recipient_rules enables tenant/agent isolation in a single vault using sops creation_rules — admin operates as super-role, clientA/clientB/agents decrypt only their own files. cosign 2+ compatibility via vault.cosign.signing_config_path (Rekor-less) when tlog=false. cosign_password 4th field in access.sops.yaml enables non-interactive CI signing. 14/14 named-error tests cover the helper contract. 3 sops adoption templates (single-team, multi-tenant, agent-first) and 3 integration templates (domain-producer, mode-producer, mode-consumer) provide copy-paste starting points. FAQ entries with data-flow diagrams in reflection/qa.ncl.",
+      invariant   = false,
+      artifact_paths = [
+        "justfiles/secrets.just",
+        "justfiles/_secrets_lib.sh",
+        "reflection/modules/vault.nu",
+        "reflection/modules/secrets.nu",
+        "reflection/requirements/base.ncl",
+        "reflection/tests/test_secrets.nu",
+        "install/resources/schemas/ontoref-project.ncl",
+        "install/resources/templates/sops/",
+        "install/resources/templates/integration/",
+        "reflection/migrations/0016-registry-credential-vault.ncl",
+        "reflection/qa.ncl",
+        "adrs/adr-019-per-file-recipient-routing-tenant-isolation.ncl",
+      ],
+      adrs = ["adr-017", "adr-019"],
+    },
+
+    d.make_node {
+      id          = "level-hierarchy-resolution",
+      name        = "Level Hierarchy and Mode Resolution",
+      pole        = 'Yang,
+      level       = 'Practice,
+      description = "Three-level specialization hierarchy with per-mode resolution strategy (ADR-018). Level 1 = ontoref base (generic protocol), Level 2 = project domain (e.g. provisioning-domain), Level 3 = domain instance (e.g. libre-daoshi, libre-wuji). Level identity is declared in manifest.ncl (level.index: Base|Domain|Instance, level.name, level.parent). Each mode at Level 2+ declares strategy: Override (complete implementation, traversal stops), Delegate (no implementation here, traverse to parent), Merge (accumulate bottom-up, lower wins conflicts), Compose (partial inheritance via extends field). Implicit absence at Level 2+ is treated as Delegate with a Soft validation warning from ore validate modes. Strategy transitions are FSM events in state.ncl — moving from Delegate to Override is an observable, git-tracked architectural decision, not a silent refactor. ore mode resolve makes traversal observable: reports which level answered a mode invocation, the strategy applied, the source file, and the reason.",
+      invariant   = false,
+      artifact_paths = [
+        "ontology/schemas/manifest.ncl",
+        "reflection/schema.ncl",
+        "reflection/migrations/0017-level-hierarchy-strategy.ncl",
+        "adrs/adr-018-level-hierarchy-mode-resolution-strategy.ncl",
+      ],
+      adrs = ["adr-018"],
+    },
+
     d.make_node {
       id          = "ci-pipelines",
       name        = "CI/CD Pipelines",
@@ -493,6 +568,15 @@ let d = import "../ontology/defaults/core.ncl" in
       ],
     },
 
+    d.make_node {
+      id          = "read-tensions-first",
+      name        = "Read Tensions First (ondaod)",
+      pole        = 'Spiral,
+      level       = 'Practice,
+      description = "Before architectural analysis or recommendation, read .ontology/core.ncl tensions; identify which the question engages; describe the synthesis state and direction of motion rather than collapse the Spiral by picking one pole. Default reasoning (human or agent) is Yang — choose, decide, recommend. This Practice surfaces that bias and asks the analyst to characterize the continuous flow first. Operationalizes the ontology-vs-reflection tension; balances against formalization-vs-adoption by choosing minimal structural enforcement (one Practice node, qa entry as canonical content, terse CLAUDE.md addendum) over heavier ADR schema constraints that would impose ceremony before adoption matures. Full procedure and forbidden patterns: reflection/qa.ncl::ontoref-dao-discipline.",
+      artifact_paths = [".ontology/core.ncl", "reflection/qa.ncl"],
+    },
+
   ],
 
   edges = [
@@ -639,5 +723,47 @@ let d = import "../ontology/defaults/core.ncl" in
     { from = "config-surface",        to = "adopt-ontoref-tooling",   kind = 'Complements,  weight = 'Medium,
       note = "Consumer projects adopting ontoref can annotate their config structs with #[derive(ConfigFields)] to participate in the coherence registry." },
 
+    # Registry Credential Vault edges
+    { from = "registry-credential-vault", to = "domain-extension-system",    kind = 'DependsOn,    weight = 'High,
+      note = "Provisioning domain uses registry credentials for OCI artifact push/pull — the vault is what makes ontoref OCI discovery possible in domain projects." },
+    { from = "registry-credential-vault", to = "adopt-ontoref-tooling",      kind = 'Complements,  weight = 'High,
+      note = "Migration 0016 propagates the credential vault pattern to consumer projects — adding sops config to project.ncl and credential_sops to registry entries." },
+    { from = "registry-credential-vault", to = "protocol-not-runtime",       kind = 'Complements,  weight = 'Medium,
+      note = "Vault belongs to the reflection layer, not the protocol. Protocol declares the credential_sops pattern; vault.nu and secrets.nu implement it operationally." },
+    { from = "registry-credential-vault", to = "unified-auth-model",         kind = 'Complements,  weight = 'Medium,
+      note = "Both use age encryption. Auth model secures the daemon API; credential vault secures OCI registry access. Independent mechanisms with the same cryptographic foundation." },
+
+    # Level Hierarchy and Mode Resolution edges
+    { from = "level-hierarchy-resolution", to = "domain-extension-system",   kind = 'Complements,  weight = 'High,
+      note = "Domain extensions (Level 2) now formally declare their hierarchy position via level.index=Domain and mode strategy fields — making the implicit traversal explicit." },
+    { from = "level-hierarchy-resolution", to = "manifest-self-description",  kind = 'ManifestsIn,  weight = 'High,
+      note = "level field in manifest_type is the ADR-018 mechanism for level identity — manifest.ncl is the source of truth for hierarchy position, queryable via ore describe project." },
+    { from = "level-hierarchy-resolution", to = "reflection-modes",          kind = 'Complements,  weight = 'High,
+      note = "Mode strategy (Override/Delegate/Merge/Compose) is declared on _ModeBase — every mode file now carries its resolution contract alongside its DAG steps." },
+    { from = "level-hierarchy-resolution", to = "protocol-migration-system", kind = 'Complements,  weight = 'High,
+      note = "Migration 0017 propagates level identity and strategy declarations to all consumer projects — applied this session to provisioning, libre-daoshi, libre-wuji." },
+    { from = "level-hierarchy-resolution", to = "dag-formalized",            kind = 'ManifestsIn,  weight = 'Medium,
+      note = "Level hierarchy is a DAG of specialization: each instance knows its domain parent, each domain knows the base. Traversal is graph traversal, not implicit lookup." },
+
+    # MCP and GraphQL surface edges
+    { from = "mcp-surface",          to = "ontoref-daemon",           kind = 'DependsOn,    weight = 'High,
+      note = "MCP server is compiled into ontoref-daemon; runtime-toggleable via ServiceFlags (ADR-014)." },
+    { from = "mcp-surface",          to = "unified-auth-model",       kind = 'DependsOn,    weight = 'High,
+      note = "MCP Bearer auth uses the same POST /sessions session token as the REST API." },
+    { from = "graphql-surface",      to = "ontoref-daemon",           kind = 'DependsOn,    weight = 'High,
+      note = "GraphQL endpoint compiled into ontoref-daemon; runtime-toggleable via ServiceFlags (ADR-014)." },
+    { from = "graphql-surface",      to = "unified-auth-model",       kind = 'DependsOn,    weight = 'High,
+      note = "GraphQL auth uses Argon2id token from config.graphql.read_token_hash — separate from project keys." },
+    { from = "graphql-surface",      to = "mcp-surface",              kind = 'Complements,  weight = 'Medium,
+      note = "GraphQL and MCP are complementary surfaces: MCP for structured tool invocation, GraphQL for ad-hoc graph queries and subscriptions." },
+
+    # Dao discipline — surface the read-tensions-first Practice from the named tensions it operationalizes
+    { from = "ontology-vs-reflection",    to = "read-tensions-first",  kind = 'ManifestsIn,  weight = 'High,
+      note = "The named Spiral between contract definition and operational mechanism is the primary tension this Practice asks analysts to engage. Yang reasoning (pick a side) collapses it; this Practice asks for synthesis-state characterization." },
+    { from = "formalization-vs-adoption", to = "read-tensions-first",  kind = 'ManifestsIn,  weight = 'Medium,
+      note = "This Practice is itself in tension with adoption friction — minimal structural enforcement is chosen so the discipline doesn't become ceremony. Heavier mechanisms (ADR schema constraints, schema pole field) deferred until adoption proves the discipline lands." },
+    { from = "read-tensions-first",       to = "dag-formalized",       kind = 'Complements,  weight = 'Medium,
+      note = "If protocol knowledge is DAG-formalized then analyses ABOUT the protocol must reference the DAG nodes — specifically the named Tension nodes — rather than reason from outside it. read-tensions-first extends DAG formalization from project state to project reasoning." },
+
   ],
 }
diff --git a/.ontology/manifest.ncl b/.ontology/manifest.ncl
index 9db2fd8..2ba1e08 100644
--- a/.ontology/manifest.ncl
+++ b/.ontology/manifest.ncl
@@ -3,6 +3,7 @@ let m = import "../ontology/defaults/manifest.ncl" in
 m.make_manifest {
   project     = "ontoref",
   repo_kind   = 'Framework,
+  level       = { index = 'Base, name = "ontoref-base" },
   description = "Protocol specification and tooling layer for structured self-knowledge in software projects. Provides schemas, Nushell automation, and Rust crates so projects can describe what they are, record architectural decisions, track operational state, and execute formalized procedures as typed, queryable artifacts.",
 
   content_assets = [
@@ -265,19 +266,26 @@ m.make_manifest {
     },
     m.make_capability {
       id        = "daemon-api",
-      name      = "Daemon HTTP + MCP Surface",
-      summary   = "HTTP UI (11 pages), 33 MCP tools, annotated API catalog, SSE notifications, per-file versioning, and NCL export cache.",
-      rationale = "Agents and developers need a queryable interface to ontology state without spawning nickel on every request. The NCL export cache reduces full-sync from ~2m42s to <30s. MCP tools give agents structured access to every capability without screen-scraping the CLI. ADR-002 records the architectural decision to extract the daemon; ADR-007 covers the #[onto_api] catalog pattern; ADR-008 covers the config override layer.",
-      how       = "crates/ontoref-daemon uses axum for HTTP. #[onto_api(...)] proc-macro + inventory::submit! registers every route at link time; GET /api/catalog aggregates via inventory::collect!. NclCache (DashMap<PathBuf, CachedExport>) keyed on path + mtime. File watcher (notify) triggers cache invalidation and drift detection after 15s debounce. MCP over stdio and streamable-HTTP.",
+      name      = "Daemon HTTP + MCP + GraphQL Surface",
+      summary   = "HTTP UI, annotated REST API catalog, MCP server (stdio + streamable-HTTP), GraphQL endpoint (queries/mutations/subscriptions), SSE notifications, per-file versioning, and NCL export cache.",
+      rationale = "Agents and developers need a queryable interface to ontology state without spawning nickel on every request. The NCL export cache reduces full-sync from ~2m42s to <30s. MCP gives agents structured tool invocation. GraphQL gives ad-hoc graph traversal and live subscriptions. Both surfaces are optional features runtime-toggleable without restart.",
+      how       = "START daemon: `just build-daemon` (compiles with db,nats,ui,mcp,graphql features) then `ontoref-daemon-boot` or `ontoref-daemon-boot --dry-run`. Default bind: http://127.0.0.1:7891. --- MCP (stdio): run `ontoref-daemon --mcp-stdio`; the AI client auto-discovers the tool list. MCP (HTTP): configure client with url=http://127.0.0.1:7891/mcp; auth=Bearer <session-token>. Get session token: POST /sessions {\"project\":\"<slug>\",\"key\":\"<project-key>\"}. CLI auto-injects ONTOREF_TOKEN. --- GRAPHQL: endpoint is http://127.0.0.1:7891/graphql (POST for queries/mutations, GET for GraphiQL IDE, /graphql/ws for subscriptions). If auth is enabled (graphql.read_token_hash in config.ncl), all requests require Authorization: Bearer <token>. Generate token: `just generate-graphql-token` prints TOKEN + HASH; add HASH to config.ncl graphql.read_token_hash. GraphiQL with pre-auth: GET /graphql?token=<TOKEN> — injects Bearer header and WS connectionParam automatically. --- TOGGLE services: PUT /api/services/mcp {\"enabled\":false} (daemon admin Bearer) or UI manage page toggle buttons. --- AUTH: all surfaces use the same session token from POST /sessions. GraphQL has a separate read_token_hash (simpler for read-only agent access without a project key).",
       artifacts = [
         "crates/ontoref-daemon/",
+        "crates/ontoref-daemon/src/mcp/mod.rs",
+        "crates/ontoref-daemon/src/graphql/mod.rs",
         "GET /api/catalog",
-        "GET /projects/{slug}/ontology",
-        "GET /projects/{slug}/config/coherence",
-        "MCP: ontoref_guides, ontoref_api_catalog, ontoref_validate, ontoref_impact",
+        "GET /services",
+        "PUT /services/:service",
+        "POST /graphql",
+        "GET /graphql (GraphiQL IDE)",
+        "GET /graphql/ws (WebSocket subscriptions)",
+        "POST /mcp (streamable-HTTP)",
+        "MCP stdio: ontoref-daemon --mcp-stdio",
+        "justfiles/build.just (generate-graphql-token, generate-admin-key)",
       ],
-      adrs  = ["adr-002", "adr-004", "adr-007", "adr-008"],
-      nodes = ["ontoref-daemon", "api-catalog-surface", "config-surface", "unified-auth-model"],
+      adrs  = ["adr-002", "adr-004", "adr-005", "adr-007", "adr-008", "adr-014"],
+      nodes = ["ontoref-daemon", "mcp-surface", "graphql-surface", "api-catalog-surface", "config-surface", "unified-auth-model"],
     },
     m.make_capability {
       id        = "reflection-modes",
@@ -506,6 +514,46 @@ m.make_manifest {
       impact    = "ontoref-daemon (db/nats features) and ontoref-reflection (nats feature) cannot build. Build with --no-default-features to work without it.",
       provision = "git clone at ../../../stratumiops relative to this repo root.",
     },
+    m.make_requirement {
+      id        = "oci-registry",
+      name      = "OCI Registry (ZOT or compatible)",
+      env       = 'Both,
+      kind      = 'Service,
+      version   = "",
+      required  = true,
+      impact    = "Without a reachable OCI registry, ontoref cannot push or pull domain artifacts, mode artifacts, or src-vault credential archives. OCI discovery — the mechanism by which actors find published modes and domains — is fully blocked. Credential distribution (ADR-017) is also blocked: src-vault:latest cannot be fetched, so no actor can resolve registry credentials.",
+      provision = "ZOT recommended: https://zotregistry.dev — docker run ghcr.io/project-zot/zot-linux-amd64. Compatible with any OCI Distribution Spec v1.1 registry. Endpoint declared in registry_provides.registries[].endpoint in .ontology/manifest.ncl.",
+    },
+    m.make_requirement {
+      id        = "oras",
+      name      = "oras CLI",
+      env       = 'Both,
+      kind      = 'Tool,
+      version   = "1.2.0",
+      required  = true,
+      impact    = "All OCI artifact operations fail: domain-push, domain-pull, mode-push, mode-pull, src-vault push/pull. OCI discovery and credential distribution both blocked.",
+      provision = "brew install oras  # or: https://oras.land/docs/installation",
+    },
+    m.make_requirement {
+      id        = "cosign",
+      name      = "cosign",
+      env       = 'Both,
+      kind      = 'Tool,
+      version   = "2.0.0",
+      required  = true,
+      impact    = "src-vault OCI artifacts cannot be signed on push or verified on pull. An unsigned vault artifact cannot be trusted — a substituted artifact with malicious credentials would be undetectable.",
+      provision = "brew install cosign  # or: https://docs.sigstore.dev/cosign/system_config/installation/",
+    },
+    m.make_requirement {
+      id        = "sops",
+      name      = "sops",
+      env       = 'Both,
+      kind      = 'Tool,
+      version   = "3.8.0",
+      required  = true,
+      impact    = "Registry credential resolution blocked. access.sops.yaml cannot be decrypted — oras calls have no credentials. All registry operations requiring auth fail.",
+      provision = "brew install sops  # or: https://github.com/getsops/sops/releases",
+    },
   ],
 
   critical_deps = [
@@ -535,6 +583,30 @@ m.make_manifest {
     },
   ],
 
+  # ADR-017 — registry credential vault declaration. credential_sops paths are
+  # relative to the synced src-vault root at ~/.config/ontoref/vaults/ontoref/src-vault/.
+  # This entry is the canonical shape that consumer manifests follow via migration 0016.
+  registry_provides = m.make_registry_provides {
+    participant = "ontoref",
+    registries  = m.make_registries_config {
+      default    = "primary",
+      registries = [
+        m.make_registry_entry {
+          id                 = "primary",
+          endpoint           = "reg.librecloud.online",
+          role               = 'primary,
+          tls                = true,
+          namespaces = {
+            own      = ["domains/ontoref/", "modes/ontoref/"],
+            prefixes = ["domains/ontoref/", "modes/ontoref/"],
+          },
+          credential_sops    = "registry/ro.sops.yaml",   # RO  — pull/list (cdci, ontoref, agent)
+          credential_sops_rw = "registry/rw.sops.yaml",   # RW  — push (admin, developer)
+        },
+      ],
+    },
+  },
+
   layers = [
     m.make_layer {
       id          = "protocol",
diff --git a/.ontology/state.ncl b/.ontology/state.ncl
index ce33cf9..abeee9d 100644
--- a/.ontology/state.ncl
+++ b/.ontology/state.ncl
@@ -23,8 +23,8 @@ let d = import "../ontology/defaults/state.ncl" in
         d.make_transition {
           from      = "adoption-tooling-complete",
           to        = "protocol-stable",
-          condition = "ADR-001 accepted, ontoref.dev published, at least two external projects consuming the protocol.",
-          catalyst  = "10 projects consuming the protocol: vapora, stratumiops, kogral, typedialog, secretumvault, rustelo, librecloud_renew, website-impl, jpl_ontology, provisioning. ADR-001 Accepted. Auth model, install pipeline, personal/career schemas, content modes, API catalog (#[onto_api], ADR-007), config surface (ADR-008), manifest self-interrogation (ADR-009), protocol migration system (ADR-010), mode guards and convergence (ADR-011) all complete. Session 2026-03-30: manifest expanded to 19 capabilities; manifest coverage validation (audit + pre-commit + SessionStart); 3 new migrations (0010-0012). Session 2026-04-05: domain extension system (ADR-012) — bash-layer dispatch for repo_kind-conditional CLI domains; personal domain (PersonalOntology: career, cfp + Sessionize integration, opportunities, content) and provisioning domain (DevWorkspace/Mixed: state, connections, gates, card, backlog); short_alias wrappers (personal, prov); ore help and describe capabilities domain-aware. Session 2026-04-06: typed link schema (ontology/schemas/links.ncl — LinkKind enum + Link record) replacing raw string arrays (urls/docs/emails/slides_url/video_url/repository) across personal/career/core schemas and backlog; error UX in domain commands (print --stderr + exit 1 replacing error make); dispatch hardening for bare subcommand arms. Session 2026-04-07: 'Framework RepoKind added — ontoref declares its own kind; no domain activates for the framework itself. VCS abstraction layer (reflection/modules/vcs.nu) — uniform jj/git API used by all modules; jj is opt-in, git is the default. Agent workspace orchestration (reflection/bin/jjw.nu + jjw-ncl-merge.nu) — jj + ontoref + Radicle lifecycle wrapper; jj/rad requirements not propagated to consumer projects. ADR-013: VCS abstraction layer — filesystem detection, single-module contract, opt-in jj/rad. manifest: 21 capabilities (vcs-abstraction, agent-workspace-orchestration added).",
+          condition = "ADR-001 accepted, ontoref.dev published, at least two external projects consuming the protocol. ADR-017 credential vault hardened end-to-end: per-file recipient routing for multi-tenant isolation, vault lock OCI artifact with TTL, impact analysis on close, 14/14 named-error tests, 6 templates (3 sops + 3 integration) for adoption.",
+          catalyst  = "10 projects consuming the protocol: vapora, stratumiops, kogral, typedialog, secretumvault, rustelo, librecloud_renew, website-impl, jpl_ontology, provisioning. ADR-001 Accepted. Auth model, install pipeline, personal/career schemas, content modes, API catalog (#[onto_api], ADR-007), config surface (ADR-008), manifest self-interrogation (ADR-009), protocol migration system (ADR-010), mode guards and convergence (ADR-011) all complete. Session 2026-03-30: manifest expanded to 19 capabilities; manifest coverage validation (audit + pre-commit + SessionStart); 3 new migrations (0010-0012). Session 2026-04-05: domain extension system (ADR-012) — bash-layer dispatch for repo_kind-conditional CLI domains; personal domain (PersonalOntology: career, cfp + Sessionize integration, opportunities, content) and provisioning domain (DevWorkspace/Mixed: state, connections, gates, card, backlog); short_alias wrappers (personal, prov); ore help and describe capabilities domain-aware. Session 2026-04-06: typed link schema (ontology/schemas/links.ncl — LinkKind enum + Link record) replacing raw string arrays (urls/docs/emails/slides_url/video_url/repository) across personal/career/core schemas and backlog; error UX in domain commands (print --stderr + exit 1 replacing error make); dispatch hardening for bare subcommand arms. Session 2026-04-07: 'Framework RepoKind added — ontoref declares its own kind; no domain activates for the framework itself. VCS abstraction layer (reflection/modules/vcs.nu) — uniform jj/git API used by all modules; jj is opt-in, git is the default. Agent workspace orchestration (reflection/bin/jjw.nu + jjw-ncl-merge.nu) — jj + ontoref + Radicle lifecycle wrapper; jj/rad requirements not propagated to consumer projects. ADR-013: VCS abstraction layer — filesystem detection, single-module contract, opt-in jj/rad. manifest: 21 capabilities (vcs-abstraction, agent-workspace-orchestration added). Session 2026-05-01: Registry credential vault (ADR-017) — per-project sops multi-recipient OCI vaults in ZOT; daemon structurally excluded from credential resolution; cosign signs src-vault on push and verifies on pull; DOCKER_CONFIG isolated per oras call; vault_key ephemeral in env only; vault.nu and secrets.nu added to reflection/modules; secrets.just (12 recipes) added; migration 0016 added; oras/cosign/sops/age declared as Hard prerequisites in requirements/base.ncl and manifest.ncl. Level hierarchy and mode resolution formalized (ADR-018) — three-level Base/Domain/Instance hierarchy; per-mode strategy (Override/Delegate/Merge/Compose) declared on _ModeBase schema; manifest.ncl gains level_type and LevelIndex; reflection/schema.ncl gains ResolutionStrategy; migration 0017 applied to provisioning (Domain, parent=ontoref-base, 12 modes), libre-daoshi (Instance, parent=provisioning-domain, 4 modes), libre-wuji (Instance, parent=provisioning-domain, 6 modes) — all modes declared Override. Session 2026-05-01 (cont): ore validate modes --check (level-declared|strategy-declared|delegate-chain|compose-extends|all) implemented in validate.nu — source-grep approach bypasses local NCL normalizers; local/inherited mode split excludes inherited base modes from strategy-declared check. ore mode resolve <id> implemented — reports effective strategy and answered_by level. All three consumer projects pass validate modes with 0 Hard failures. Session 2026-05-01 (cont2): ADR-018 gaps fully closed — parent_path field added to level_type in manifest.ncl enabling cross-project delegate-chain verification (Hard failure if parent_path declared but parent does not have the mode); ore validate modes --self-test implemented with 9 assertions covering all 4 check variants on positive and negative synthetic fixtures. All checks verified end-to-end. Session 2026-05-03: ADR-017 credential vault end-to-end hardening — sentinel pattern + DOCKER_CONFIG isolation across 6 oras commands (domain_client.nu); credentials.nu portable Layer-2 helper in provisioning/core; assert-actor-authorized + assert-target-in-scope two-level enforcement (scope.bound_actor + scope.namespaces); ore secrets {bootstrap,sync,push,open,close,describe,force-unlock,audit,gen-key,add-key,remove-key,rekey} dispatcher wired; vault lock OCI artifact (src-vault/<id>:lock) with TTL 60min and force-unlock auditable; impact analysis on secrets-close (diff sops files since last snapshot, map to RegistryEntry IDs, prompt confirm or ONTOREF_SECRETS_YES skip); cosign 2+ compatibility (signing-config replaces deprecated --tlog-upload); cosign_password 4th field in access.sops.yaml for non-interactive CI; per-file recipient routing via recipient_groups + recipient_rules (sops creation_rules) for multi-tenant isolation without multi-vault; 6 templates added (install/resources/templates/sops/{single-team,multi-tenant,agent-first}/ and integration/{domain-producer,mode-producer,mode-consumer}/); reflection/qa.ncl FAQ — 6 entries with diagrams covering credential vault best practice, templates, troubleshooting (15 named errors), integration what/why/how, integration troubleshooting; reflection/tests/test_secrets.nu with 14/14 named-error tests passing.",
           blocker   = "ontoref.dev not yet published.",
           horizon   = 'Months,
         },
@@ -52,7 +52,7 @@ let d = import "../ontology/defaults/state.ncl" in
           from      = "modes-and-web-present",
           to        = "fully-self-described",
           condition = "At least 3 ADRs accepted, reflection/backlog.ncl present, describe project returns complete picture.",
-          catalyst  = "ADR-001–ADR-006 authored (6 ADRs present). Auth model, project onboarding, and session management nodes added in 2026-03-13. Personal/career/project-card schemas, 5 content modes, search bookmarks, and ADR-006 (Nu 0.111 compat) added in session 2026-03-15. Session 2026-03-23: api-catalog-surface node added (#[onto_api] proc-macro + inventory catalog), describe-query-layer updated (diff + api subcommands), adopt-ontoref-tooling updated (update_ontoref mode + manifest/connections templates + enrichment prompt), ontoref-daemon updated (11 pages, 29 MCP tools, per-file versioning, API catalog endpoint). Session 2026-03-26: config-surface node added — typed DaemonNclConfig (parse-at-boundary pattern), #[derive(ConfigFields)] coherence registry, override-layer mutation API (PUT /config/{section}), NCL contracts (.ontoref/contracts.ncl: LogConfig + DaemonConfig), manifest config_surface with multi-consumer sections. ADR-007 (inventory/onto_api) extended to ConfigFields; ADR-008 (NCL-first config validation + override-layer mutation). Session 2026-03-26 (2nd): manifest-self-description node added. ADR-009. Session 2026-03-29: browser-style panel navigation. Session 2026-03-30: manifest expanded 3→19 capabilities (complete action surface: modes, compose, plans, backlog graduation, notifications, coder pipeline, forms, templates, drift, quick actions, migration, config, search bookmarks, onboarding, web presence). audit-manifest-coverage validator + pre-commit hook + SessionStart hook. Mode schema extended: Guard type (Block/Warn severity pre-flight checks), Converge type (RetryFailed/RetryAll post-execution loops). ADR-011. Migrations 0010-0012. Bug fix: find-unclaimed-artifacts absolute vs relative path comparison. Justfile split (build/test/dev/ci/assets). Anti-slop novelty-check in coder pipeline (Jaccard overlap against published+QA). Health 43%→100%. Session 2026-04-05: domain-extension-system node added. ADR-012. personal domain (jpl_ontology PersonalOntology: cfp/Sessionize integration, opportunities arrays, career, content) and provisioning domain (DevWorkspace/Mixed: state, card, connections, gates, capabilities, backlog). Bash-layer dispatch + ore help + describe capabilities domain-aware. Short alias system (personal, prov). Session 2026-04-06: personal-ontology-schemas node updated — links.ncl typed link schema; typed link migration across personal/career/core/backlog schemas. Session 2026-04-07: vcs-abstraction node (reflection/modules/vcs.nu) and agent-workspace-orchestration node (reflection/bin/jjw.nu + jjw-ncl-merge.nu) added. 'Framework RepoKind: ontoref self-identifies as framework, no domain activates. manifest: 21 capabilities (vcs-abstraction, agent-workspace-orchestration added).",
+          catalyst  = "ADR-001–ADR-006 authored (6 ADRs present). Auth model, project onboarding, and session management nodes added in 2026-03-13. Personal/career/project-card schemas, 5 content modes, search bookmarks, and ADR-006 (Nu 0.111 compat) added in session 2026-03-15. Session 2026-03-23: api-catalog-surface node added (#[onto_api] proc-macro + inventory catalog), describe-query-layer updated (diff + api subcommands), adopt-ontoref-tooling updated (update_ontoref mode + manifest/connections templates + enrichment prompt), ontoref-daemon updated (11 pages, 29 MCP tools, per-file versioning, API catalog endpoint). Session 2026-03-26: config-surface node added — typed DaemonNclConfig (parse-at-boundary pattern), #[derive(ConfigFields)] coherence registry, override-layer mutation API (PUT /config/{section}), NCL contracts (.ontoref/contracts.ncl: LogConfig + DaemonConfig), manifest config_surface with multi-consumer sections. ADR-007 (inventory/onto_api) extended to ConfigFields; ADR-008 (NCL-first config validation + override-layer mutation). Session 2026-03-26 (2nd): manifest-self-description node added. ADR-009. Session 2026-03-29: browser-style panel navigation. Session 2026-03-30: manifest expanded 3→19 capabilities (complete action surface: modes, compose, plans, backlog graduation, notifications, coder pipeline, forms, templates, drift, quick actions, migration, config, search bookmarks, onboarding, web presence). audit-manifest-coverage validator + pre-commit hook + SessionStart hook. Mode schema extended: Guard type (Block/Warn severity pre-flight checks), Converge type (RetryFailed/RetryAll post-execution loops). ADR-011. Migrations 0010-0012. Bug fix: find-unclaimed-artifacts absolute vs relative path comparison. Justfile split (build/test/dev/ci/assets). Anti-slop novelty-check in coder pipeline (Jaccard overlap against published+QA). Health 43%→100%. Session 2026-04-05: domain-extension-system node added. ADR-012. personal domain (jpl_ontology PersonalOntology: cfp/Sessionize integration, opportunities arrays, career, content) and provisioning domain (DevWorkspace/Mixed: state, card, connections, gates, capabilities, backlog). Bash-layer dispatch + ore help + describe capabilities domain-aware. Short alias system (personal, prov). Session 2026-04-06: personal-ontology-schemas node updated — links.ncl typed link schema; typed link migration across personal/career/core/backlog schemas. Session 2026-04-07: vcs-abstraction node (reflection/modules/vcs.nu) and agent-workspace-orchestration node (reflection/bin/jjw.nu + jjw-ncl-merge.nu) added. 'Framework RepoKind: ontoref self-identifies as framework, no domain activates. manifest: 21 capabilities (vcs-abstraction, agent-workspace-orchestration added). Session 2026-05-01: registry-credential-vault node added (ADR-017). level-hierarchy-resolution node added (ADR-018). adr-lifecycle node updated through ADR-018. manifest.ncl gains level_type + LevelIndex export; reflection/schema.ncl gains ResolutionStrategy + strategy + extends on _ModeBase. Migrations 0016-0017 added. 18 ADRs total.",
           blocker   = "none",
           horizon   = 'Weeks,
         },
diff --git a/.ontoref/project.ncl b/.ontoref/project.ncl
index da36af7..2474170 100644
--- a/.ontoref/project.ncl
+++ b/.ontoref/project.ncl
@@ -8,4 +8,24 @@ s.make_project {
     "/Users/Akasha/Development/ontoref/ontology",
   ],
   keys                = [],
+
+  # ADR-017 — registry credential vault. ontoref participates as a publisher in the
+  # libre-wuji registry under domains/ontoref/ and modes/ontoref/. Until the vault is
+  # bootstrapped (`./ontoref secrets bootstrap`), Layer 2 operations error nominally.
+  # This block is the canonical reference shape that consumer projects copy via
+  # migration 0016.
+  sops = {
+    enabled            = true,
+    vault_id           = "ontoref",
+    vault_backend      = 'restic,
+    registry_endpoint  = "reg.librecloud.online",
+    actor_key_bindings = {
+      developer = "developer",
+      ci        = "cdci",
+      agent     = "ontoref",
+      admin     = "admin",
+    },
+    # master_key_path resolved from ~/.config/ontoref/config.ncl::vault.master_key_path.
+    # Override here only if this project requires a different .kage than the global default.
+  },
 }
diff --git a/CHANGELOG.md b/CHANGELOG.md
index a5300a8..bf1d425 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,6 +7,133 @@ ADRs referenced below live in `adrs/` as typed Nickel records.
 
 ## [Unreleased]
 
+### MCP tool catalog via proc-macro, validate ADR-018 mode hierarchy, UI vault/registry context
+
+**Session 2026-05-12**
+
+**`#[onto_mcp_tool]` proc-macro — link-time MCP catalog registration (`ontoref-derive`)**
+- New attribute macro that registers an MCP tool unit-struct in the catalog at link time via
+  `inventory::submit!(McpToolEntry{...})`. The annotated item is emitted unchanged — `ToolBase`
+  and `AsyncTool` impls remain on the struct. Required keys: `name`, `description`, `input_schema`.
+- All 34 MCP tools migrated from manual `catalog()` wiring to `#[onto_mcp_tool]`. The
+  `catalog()` function is now generated automatically by `inventory::collect!`.
+- Tool count: 29 → 34 (net: `ontoref_list_projects`, `ontoref_search`, `ontoref_describe`,
+  `ontoref_list_ontology_extensions`, `ontoref_get_ontology_extension` added).
+
+**`validate modes` — ADR-018 level hierarchy and strategy compliance (`reflection/modules/validate.nu`)**
+- `validate modes [--check]` — reads `reflection/defaults/workflow.ncl::level_hierarchy` and
+  checks every `.ncl` mode file against the declared levels: level declared, strategy declared,
+  delegate chain coherent, compose `extends` references valid.
+- `mode resolve <id>` — prints which hierarchy level handles the given mode id and why.
+- `validate modes --self-test` — generates synthetic mode fixtures in a temp dir and runs checks
+  against them; fast CI-safe smoke test for the validator itself.
+
+**`validate run-cargo` — two-step Cargo.toml resolution**
+- Resolves `Cargo.toml` in workspace layout first (`crates/<check.crate>/Cargo.toml`), then
+  falls back to single-crate root when `[package].name` matches or `check.crate` equals the
+  repo basename. Lets the same ADR constraint shape apply to workspace repos (stratumiops) and
+  single-crate projects without per-repo special cases.
+
+**UI — vault and registry context in project_picker**
+- Each project card in the project picker now surfaces OCI state inline:
+  - Registry badge (`⟳ <participant>`) when `manifest.ncl::registry_provides` is declared.
+  - Vault badge (`⛁ <vault_id> · N`) coloured green (declarative) or amber (legacy) when
+    `project.ncl::sops.enabled` is true.
+- Expanded project details panel: collapsible **Registry** section listing namespaces, endpoint,
+  and push/pull capability per namespace.
+
+**UI — runtime service toggles in manage page**
+- `manage.html` gains a **Runtime Services** card: MCP and GraphQL can be toggled on/off via
+  HTMX `POST /ui/manage/services/{service}/toggle` without restarting the daemon.
+- `insert_mcp_ctx` helper injects MCP and GraphQL runtime state + daemon version into every
+  Tera template context that needs it.
+
+**Manifest schema — registry topology contract (`ontology/schemas/manifest.ncl`)**
+- New `registry_topology_type` contract: declares multi-registry coordination (which registries
+  the project participates in, push targets, participant scopes, and per-namespace capability).
+- `manifest_type` gains optional `registry_topology` field.
+
+**Requirements — OCI tooling declared in `reflection/requirements/base.ncl`**
+- `oras ≥ 1.2.0` (Hard), `cosign ≥ 2.0.0` (Hard), `sops ≥ 3.9.0` (Hard), `age ≥ 1.1.0`
+  (Hard), `restic` (Soft) added as named requirements with `check_cmd`, `version_extract`,
+  and `install_hint`. Surfaces in `describe requirements` and `ontoref_guides`.
+
+**describe.nu + sync.nu expansion**
+- `describe capabilities` JSON output now includes `registry_topology` and `vault_state` per
+  project when declared.
+- `sync diff --docs` drift check extended to detect `//!` absence on new crates added this
+  session.
+
+---
+
+### Per-file recipient routing for tenant isolation (ADR-019)
+
+**Session 2026-05-03** — credential vault model extended to support per-file
+recipient routing, closing the multi-tenant gap without multi-vault.
+
+**Schema additions** (`install/resources/schemas/ontoref-project.ncl`)
+- `sops.recipient_groups` (record of group → list of age public keys, optional)
+- `sops.recipient_rules` (array of `{ path, groups }` records, optional)
+
+**Bootstrap behaviour**
+- When `recipient_rules` declared, `secrets-bootstrap` generates `<vault_dir>/.sops.yaml`
+  with sops `creation_rules` mapping each path_regex to the union of declared groups.
+- Default `registry/ro+rw.sops.yaml` pair is skipped in declarative mode — the
+  operator defines files matching their tenancy scheme (e.g. `clientA-ro.sops.yaml`).
+- Legacy mode (no `recipient_rules`) keeps `SOPS_AGE_RECIPIENTS` env-var path.
+
+**Operational locks**
+- `secrets-add-key` and `secrets-remove-key` error in declarative mode and
+  direct the operator to edit `project.ncl::sops.recipient_groups` plus `secrets-rekey`.
+- `secrets-rekey` regenerates `.sops.yaml` from `project.ncl` then runs `sops updatekeys`
+  on every `*.sops.yaml` in the vault.
+
+**Adoption templates** (`install/resources/templates/sops/`)
+- `single-team/` — one team, no tenant separation (legacy / Path A)
+- `multi-tenant/` — clientA/clientB groups + path-routed credentials
+- `agent-first/` — admin/developer/agents with `agent-readonly.sops.yaml` for AI access
+
+**Integration templates** (`install/resources/templates/integration/`)
+- `domain-producer/` — contract.ncl + example.json + manifest.ncl scaffold
+- `mode-producer/` — provisioning.ncl + domains.lock.ncl + manifest.ncl scaffold
+- `mode-consumer/` — cabling.ncl scaffold for binding pulled modes to workspace values
+
+**Audit checks** (`secrets-audit`) — three new checks for ADR-019 constraints:
+- `recipient-routing-coherent` — every group referenced by a rule must be declared;
+  every rule must resolve to ≥ 1 recipient
+- `recipient-routing-coverage` — every `*.sops.yaml` in the vault must match at
+  least one declared rule
+- `no-multi-vault` — rejects `sops.vaults =` declaration (multi-vault not implemented)
+
+**FAQ entries** (`reflection/qa.ncl`) — six entries with diagrams:
+- `credential-vault-best-practice` — layered model with data-flow diagram
+- `credential-vault-templates` — three adoption paths (A/B/C)
+- `credential-vault-troubleshooting` — 15 named errors and recovery steps
+- `integration-what-and-why` — federated OCI artifacts (ADR-042) overview with diagram
+- `integration-how-to-implement` — producer / consumer / both, with templates
+- `integration-troubleshooting` — push/pull/invoke common failures
+
+**ADR-019: Per-File Recipient Routing for Tenant Isolation in lieu of Multi-Vault**
+Documents the decision, alternatives (multi-vault explicitly rejected), and 6
+machine-checkable constraints. Each constraint is wired to either `secrets-audit`
+(NuCmd), `Grep`, or `FileExists` — verifiable from CLI without a custom validator.
+
+**Tests** — `reflection/tests/test_secrets.nu` covers 14/14 named errors of the
+helper contract end-to-end (`sops-decrypt-failed` excluded — requires real age
+keypair plumbing).
+
+**Bug fixes**
+- `_find-capabilities-ncl`: glob malformation when walking up to filesystem root
+- cosign `--tlog-upload=false` deprecated in cosign 2+ → use `--signing-config`
+  with `rekorTlogUrls`/`rekorTlogConfig` stripped
+- cosign `sign` needed `DOCKER_CONFIG` to fetch manifest before signing
+- Cache hit in `domain-pull` / `mode-pull` bypassed authorization gate (reordered)
+- Per-file `(...)` interpolation traps in nu strings inside error messages
+- `add-key` / `remove-key` only mutated `access.sops.yaml` instead of all sops
+  files in the vault
+
+---
+
 ### VCS abstraction layer and agent workspace orchestration
 
 #### `'Framework` RepoKind
diff --git a/Cargo.lock b/Cargo.lock
index 38438fe..7da03c5 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -197,6 +197,53 @@ dependencies = [
  "stable_deref_trait",
 ]
 
+[[package]]
+name = "askama"
+version = "0.15.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9b8246bcbf8eb97abef10c2d92166449680d41d55c0fc6978a91dec2e3619608"
+dependencies = [
+ "askama_macros",
+ "itoa",
+ "percent-encoding",
+ "serde",
+ "serde_json",
+]
+
+[[package]]
+name = "askama_derive"
+version = "0.15.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2f9670bc84a28bb3da91821ef74226949ab63f1265aff7c751634f1dd0e6f97c"
+dependencies = [
+ "askama_parser",
+ "memchr",
+ "proc-macro2",
+ "quote",
+ "rustc-hash",
+ "syn 2.0.117",
+]
+
+[[package]]
+name = "askama_macros"
+version = "0.15.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f0756b45480437dded0565dfc568af62ccce146fb6cfe902e808ba86e445f44f"
+dependencies = [
+ "askama_derive",
+]
+
+[[package]]
+name = "askama_parser"
+version = "0.15.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5d0af3691ba3af77949c0b5a3925444b85cb58a0184cc7fec16c68ba2e7be868"
+dependencies = [
+ "rustc-hash",
+ "unicode-ident",
+ "winnow 1.0.2",
+]
+
 [[package]]
 name = "async-channel"
 version = "2.5.0"
@@ -210,10 +257,101 @@ dependencies = [
 ]
 
 [[package]]
-name = "async-nats"
-version = "0.46.0"
+name = "async-graphql"
+version = "8.0.0-rc.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "df5af9ebfb0a14481d3eaf6101e6391261e4f30d25b26a7635ade8a39482ded0"
+checksum = "34a42d615ca2f3557491397f6ed2169dd38fbce9bf5352a5a9f17216cdfeb40f"
+dependencies = [
+ "askama",
+ "async-graphql-derive",
+ "async-graphql-parser",
+ "async-graphql-value",
+ "async-trait",
+ "asynk-strim",
+ "base64",
+ "blocking",
+ "bytes",
+ "futures-util",
+ "http",
+ "indexmap",
+ "mime",
+ "multer",
+ "num-traits",
+ "pin-project-lite",
+ "regex",
+ "rustc-hash",
+ "serde",
+ "serde_json",
+ "serde_urlencoded",
+ "static_assertions_next",
+ "tempfile",
+ "thiserror 2.0.18",
+ "tokio",
+]
+
+[[package]]
+name = "async-graphql-axum"
+version = "8.0.0-rc.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "42398787bd4bf3257c294d6743b228e0b437762cbc686cff1ab4ccab014ed30f"
+dependencies = [
+ "async-graphql",
+ "axum",
+ "bytes",
+ "futures-util",
+ "serde_json",
+ "tokio",
+ "tokio-stream",
+ "tokio-util",
+ "tower-service",
+]
+
+[[package]]
+name = "async-graphql-derive"
+version = "8.0.0-rc.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cdaa2cb6cc70518033cf543e6b735ed03daa71c1ec81eac5ad09fa05e61faa9f"
+dependencies = [
+ "async-graphql-parser",
+ "darling",
+ "heck",
+ "proc-macro-crate",
+ "proc-macro2",
+ "quote",
+ "strum",
+ "syn 2.0.117",
+ "thiserror 2.0.18",
+]
+
+[[package]]
+name = "async-graphql-parser"
+version = "8.0.0-rc.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3ae405016a111a09cd484a973024dcc139ac398a89cb1f4602bebab817298276"
+dependencies = [
+ "async-graphql-value",
+ "pest",
+ "serde",
+ "serde_json",
+]
+
+[[package]]
+name = "async-graphql-value"
+version = "8.0.0-rc.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "15eb1573843563b4518cc43001002512c206d119974b17476a3cbf62d9a1ae61"
+dependencies = [
+ "bytes",
+ "indexmap",
+ "serde",
+ "serde_json",
+]
+
+[[package]]
+name = "async-nats"
+version = "0.48.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "31811585c7c5bc2f60f8b80d5a6b0f737115611dac47567d7f7d94562ebb180b"
 dependencies = [
  "base64",
  "bytes",
@@ -221,20 +359,19 @@ dependencies = [
  "memchr",
  "nkeys",
  "nuid",
- "once_cell",
  "pin-project",
  "portable-atomic",
- "rand 0.8.5",
+ "rand 0.10.1",
  "regex",
  "ring",
- "rustls-native-certs 0.7.3",
+ "rustls-native-certs",
  "rustls-pki-types",
- "rustls-webpki 0.102.8",
+ "rustls-webpki",
  "serde",
  "serde_json",
  "serde_nanos",
  "serde_repr",
- "thiserror 1.0.69",
+ "thiserror 2.0.18",
  "time",
  "tokio",
  "tokio-rustls",
@@ -268,6 +405,12 @@ dependencies = [
  "syn 2.0.117",
 ]
 
+[[package]]
+name = "async-task"
+version = "4.7.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8b75356056920673b02621b35afd0f7dda9306d03c79a30f5c56c44cf256e3de"
+
 [[package]]
 name = "async-trait"
 version = "0.1.89"
@@ -279,6 +422,16 @@ dependencies = [
  "syn 2.0.117",
 ]
 
+[[package]]
+name = "asynk-strim"
+version = "0.1.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "52697735bdaac441a29391a9e97102c74c6ef0f9b60a40cf109b1b404e29d2f6"
+dependencies = [
+ "futures-core",
+ "pin-project-lite",
+]
+
 [[package]]
 name = "atomic-polyfill"
 version = "1.0.3"
@@ -330,6 +483,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "8b52af3cb4058c895d37317bb27508dccc8e5f2d39454016b297bf4a400597b8"
 dependencies = [
  "axum-core",
+ "base64",
  "bytes",
  "form_urlencoded",
  "futures-util",
@@ -348,8 +502,10 @@ dependencies = [
  "serde_json",
  "serde_path_to_error",
  "serde_urlencoded",
+ "sha1",
  "sync_wrapper",
  "tokio",
+ "tokio-tungstenite",
  "tower",
  "tower-layer",
  "tower-service",
@@ -504,6 +660,19 @@ dependencies = [
  "generic-array 0.14.7",
 ]
 
+[[package]]
+name = "blocking"
+version = "1.6.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e83f8d02be6967315521be875afa792a316e28d57b5a2d401897e2a7921b7f21"
+dependencies = [
+ "async-channel",
+ "async-task",
+ "futures-io",
+ "futures-lite",
+ "piper",
+]
+
 [[package]]
 name = "blowfish"
 version = "0.9.1"
@@ -1456,6 +1625,16 @@ version = "0.3.32"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "cecba35d7ad927e23624b22ad55235f2239cfa44fd10428eecbeba6d6a717718"
 
+[[package]]
+name = "futures-lite"
+version = "2.6.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "f78e10609fe0e0b3f4157ffab1876319b5b0db102a2c60dc4626306dc46b44ad"
+dependencies = [
+ "futures-core",
+ "pin-project-lite",
+]
+
 [[package]]
 name = "futures-macro"
 version = "0.3.32"
@@ -1572,7 +1751,7 @@ dependencies = [
  "i_overlay",
  "log",
  "num-traits",
- "rand 0.8.5",
+ "rand 0.8.6",
  "robust",
  "rstar 0.12.2",
  "serde",
@@ -2396,7 +2575,7 @@ dependencies = [
  "p256",
  "p384",
  "pem",
- "rand 0.8.5",
+ "rand 0.8.6",
  "rsa",
  "serde",
  "serde_json",
@@ -2606,9 +2785,9 @@ dependencies = [
 
 [[package]]
 name = "mio"
-version = "1.1.1"
+version = "1.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a69bcab0ad47271a0234d9422b131806bf3968021e5dc9328caf2d4cd58557fc"
+checksum = "50b7e5b27aa02a74bac8c3f23f448f8d87ff11f92d3aac1a6ed369ee08cc56c1"
 dependencies = [
  "libc",
  "log",
@@ -2616,6 +2795,23 @@ dependencies = [
  "windows-sys 0.61.2",
 ]
 
+[[package]]
+name = "multer"
+version = "3.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "83e87776546dc87511aa5ee218730c92b666d7264ab6ed41f9d215af9cd5224b"
+dependencies = [
+ "bytes",
+ "encoding_rs",
+ "futures-util",
+ "http",
+ "httparse",
+ "memchr",
+ "mime",
+ "spin",
+ "version_check",
+]
+
 [[package]]
 name = "ndarray"
 version = "0.17.2"
@@ -2643,7 +2839,7 @@ dependencies = [
  "noisy_float",
  "num-integer",
  "num-traits",
- "rand 0.8.5",
+ "rand 0.8.6",
 ]
 
 [[package]]
@@ -2672,7 +2868,7 @@ dependencies = [
  "ed25519-dalek",
  "getrandom 0.2.17",
  "log",
- "rand 0.8.5",
+ "rand 0.8.6",
  "signatory",
 ]
 
@@ -2736,7 +2932,7 @@ version = "0.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "fc895af95856f929163a0aa20c26a78d26bfdc839f51b9d5aa7a5b79e52b7e83"
 dependencies = [
- "rand 0.8.5",
+ "rand 0.8.6",
 ]
 
 [[package]]
@@ -2760,7 +2956,7 @@ dependencies = [
  "num-integer",
  "num-iter",
  "num-traits",
- "rand 0.8.5",
+ "rand 0.8.6",
  "smallvec",
  "zeroize",
 ]
@@ -2881,6 +3077,8 @@ version = "0.1.0"
 dependencies = [
  "anyhow",
  "argon2",
+ "async-graphql",
+ "async-graphql-axum",
  "axum",
  "axum-server",
  "bytes",
@@ -2958,12 +3156,6 @@ dependencies = [
  "uuid",
 ]
 
-[[package]]
-name = "openssl-probe"
-version = "0.1.6"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d05e27ee213611ffe7d6348b942e8f942b37114c00cc03cec254295a4a17852e"
-
 [[package]]
 name = "openssl-probe"
 version = "0.2.1"
@@ -3189,7 +3381,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "3c80231409c20246a13fddb31776fb942c38553c51e871f8cbd687a4cfb5843d"
 dependencies = [
  "phf_shared 0.11.3",
- "rand 0.8.5",
+ "rand 0.8.6",
 ]
 
 [[package]]
@@ -3280,6 +3472,17 @@ version = "0.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "8b870d8c151b6f2fb93e84a13146138f05d02ed11c7e7c54f8826aaaf7c9f184"
 
+[[package]]
+name = "piper"
+version = "0.2.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c835479a4443ded371d6c535cbfd8d31ad92c5d23ae9770a61bc155e4992a3c1"
+dependencies = [
+ "atomic-waker",
+ "fastrand",
+ "futures-io",
+]
+
 [[package]]
 name = "pkcs1"
 version = "0.7.5"
@@ -3499,7 +3702,7 @@ dependencies = [
  "bytes",
  "getrandom 0.3.4",
  "lru-slab",
- "rand 0.9.2",
+ "rand 0.9.4",
  "ring",
  "rustc-hash",
  "rustls",
@@ -3565,9 +3768,9 @@ dependencies = [
 
 [[package]]
 name = "rand"
-version = "0.8.5"
+version = "0.8.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "34af8d1a0e25924bc5b7c43c079c942339d8f0a8b57c39049bef581b46327404"
+checksum = "5ca0ecfa931c29007047d1bc58e623ab12e5590e8c7cc53200d5202b69266d8a"
 dependencies = [
  "libc",
  "rand_chacha 0.3.1",
@@ -3576,9 +3779,9 @@ dependencies = [
 
 [[package]]
 name = "rand"
-version = "0.9.2"
+version = "0.9.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6db2770f06117d490610c7488547d543617b21bfa07796d7a12f6f1bd53850d1"
+checksum = "44c5af06bb1b7d3216d91932aed5265164bf384dc89cd6ba05cf59a35f5f76ea"
 dependencies = [
  "rand_chacha 0.9.0",
  "rand_core 0.9.5",
@@ -3586,9 +3789,9 @@ dependencies = [
 
 [[package]]
 name = "rand"
-version = "0.10.0"
+version = "0.10.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bc266eb313df6c5c09c1c7b1fbe2510961e5bcd3add930c1e31f7ed9da0feff8"
+checksum = "d2e8e8bcc7961af1fdac401278c6a831614941f6164ee3bf4ce61b7edb162207"
 dependencies = [
  "chacha20",
  "getrandom 0.4.2",
@@ -3884,7 +4087,7 @@ dependencies = [
  "http-body-util",
  "pastey",
  "pin-project-lite",
- "rand 0.10.0",
+ "rand 0.10.1",
  "rmcp-macros",
  "schemars",
  "serde",
@@ -4078,7 +4281,7 @@ dependencies = [
  "borsh",
  "bytes",
  "num-traits",
- "rand 0.8.5",
+ "rand 0.8.6",
  "rkyv",
  "serde",
  "serde_json",
@@ -4123,34 +4326,21 @@ dependencies = [
  "once_cell",
  "ring",
  "rustls-pki-types",
- "rustls-webpki 0.103.9",
+ "rustls-webpki",
  "subtle",
  "zeroize",
 ]
 
-[[package]]
-name = "rustls-native-certs"
-version = "0.7.3"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e5bfb394eeed242e909609f56089eecfe5fda225042e8b171791b9c95f5931e5"
-dependencies = [
- "openssl-probe 0.1.6",
- "rustls-pemfile",
- "rustls-pki-types",
- "schannel",
- "security-framework 2.11.1",
-]
-
 [[package]]
 name = "rustls-native-certs"
 version = "0.8.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "612460d5f7bea540c490b2b6395d8e34a953e52b491accd6c86c8164c5932a63"
 dependencies = [
- "openssl-probe 0.2.1",
+ "openssl-probe",
  "rustls-pki-types",
  "schannel",
- "security-framework 3.7.0",
+ "security-framework",
 ]
 
 [[package]]
@@ -4184,10 +4374,10 @@ dependencies = [
  "log",
  "once_cell",
  "rustls",
- "rustls-native-certs 0.8.3",
+ "rustls-native-certs",
  "rustls-platform-verifier-android",
- "rustls-webpki 0.103.9",
- "security-framework 3.7.0",
+ "rustls-webpki",
+ "security-framework",
  "security-framework-sys",
  "webpki-root-certs",
  "windows-sys 0.61.2",
@@ -4201,19 +4391,9 @@ checksum = "f87165f0995f63a9fbeea62b64d10b4d9d8e78ec6d7d51fb2125fda7bb36788f"
 
 [[package]]
 name = "rustls-webpki"
-version = "0.102.8"
+version = "0.103.13"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "64ca1bc8749bd4cf37b5ce386cc146580777b4e8572c7b97baf22c83f444bee9"
-dependencies = [
- "rustls-pki-types",
- "untrusted 0.9.0",
-]
-
-[[package]]
-name = "rustls-webpki"
-version = "0.103.9"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d7df23109aa6c1567d1c575b9952556388da57401e4ace1d15f79eedad0d8f53"
+checksum = "61c429a8649f110dddef65e2a5ad240f747e85f7758a6bccc7e5777bd33f756e"
 dependencies = [
  "aws-lc-rs",
  "ring",
@@ -4324,19 +4504,6 @@ dependencies = [
  "zeroize",
 ]
 
-[[package]]
-name = "security-framework"
-version = "2.11.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "897b2245f0b511c87893af39b033e5ca9cce68824c4d7e7630b5a1d339658d02"
-dependencies = [
- "bitflags 2.11.0",
- "core-foundation 0.9.4",
- "core-foundation-sys",
- "libc",
- "security-framework-sys",
-]
-
 [[package]]
 name = "security-framework"
 version = "3.7.0"
@@ -4670,6 +4837,12 @@ version = "1.2.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "6ce2be8dc25455e1f91df71bfa12ad37d7af1092ae736f3a6cd0e37bc7810596"
 
+[[package]]
+name = "static_assertions_next"
+version = "1.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d7beae5182595e9a8b683fa98c4317f956c9a2dec3b9716990d20023cc60c766"
+
 [[package]]
 name = "storekey"
 version = "0.11.0"
@@ -4765,6 +4938,27 @@ version = "0.11.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7da8b5736845d9f2fcb837ea5d9e2628564b3b043a70948a3f0b778838c5fb4f"
 
+[[package]]
+name = "strum"
+version = "0.28.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "9628de9b8791db39ceda2b119bbe13134770b56c138ec1d3af810d045c04f9bd"
+dependencies = [
+ "strum_macros",
+]
+
+[[package]]
+name = "strum_macros"
+version = "0.28.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ab85eea0270ee17587ed4156089e10b9e6880ee688791d45a905f5b1ca36f664"
+dependencies = [
+ "heck",
+ "proc-macro2",
+ "quote",
+ "syn 2.0.117",
+]
+
 [[package]]
 name = "subtle"
 version = "2.6.1"
@@ -4862,7 +5056,7 @@ dependencies = [
  "pin-project-lite",
  "quick_cache",
  "radix_trie",
- "rand 0.8.5",
+ "rand 0.8.6",
  "rayon",
  "reblessive",
  "regex",
@@ -4937,7 +5131,7 @@ dependencies = [
  "hex",
  "http",
  "papaya",
- "rand 0.8.5",
+ "rand 0.8.6",
  "regex",
  "rstest",
  "rust_decimal",
@@ -5103,7 +5297,7 @@ dependencies = [
  "percent-encoding",
  "pest",
  "pest_derive",
- "rand 0.8.5",
+ "rand 0.8.6",
  "regex",
  "serde",
  "serde_json",
@@ -5218,9 +5412,9 @@ checksum = "1f3ccbac311fea05f86f61904b462b55fb3df8837a366dfc601a0161d0532f20"
 
 [[package]]
 name = "tokio"
-version = "1.50.0"
+version = "1.52.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "27ad5e34374e03cfffefc301becb44e9dc3c17584f414349ebe29ed26661822d"
+checksum = "110a78583f19d5cdb2c5ccf321d1290344e71313c6c37d43520d386027d18386"
 dependencies = [
  "bytes",
  "libc",
@@ -5235,9 +5429,9 @@ dependencies = [
 
 [[package]]
 name = "tokio-macros"
-version = "2.6.1"
+version = "2.7.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5c55a2eff8b69ce66c84f85e1da1c233edc36ceb85a2058d11b0d6a3c7e7569c"
+checksum = "385a6cb71ab9ab790c5fe8d67f1645e6c450a7ce006a33de03daa956cf70a496"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -5338,7 +5532,7 @@ dependencies = [
  "futures-sink",
  "http",
  "httparse",
- "rand 0.8.5",
+ "rand 0.8.6",
  "ring",
  "rustls-pki-types",
  "tokio",
@@ -5388,7 +5582,7 @@ dependencies = [
  "serde_spanned",
  "toml_datetime 0.6.11",
  "toml_write",
- "winnow",
+ "winnow 0.7.15",
 ]
 
 [[package]]
@@ -5400,7 +5594,7 @@ dependencies = [
  "indexmap",
  "toml_datetime 1.0.0+spec-1.1.0",
  "toml_parser",
- "winnow",
+ "winnow 0.7.15",
 ]
 
 [[package]]
@@ -5409,7 +5603,7 @@ version = "1.0.9+spec-1.1.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "702d4415e08923e7e1ef96cd5727c0dfed80b4d2fa25db9647fe5eb6f7c5a4c4"
 dependencies = [
- "winnow",
+ "winnow 0.7.15",
 ]
 
 [[package]]
@@ -5617,7 +5811,7 @@ dependencies = [
  "http",
  "httparse",
  "log",
- "rand 0.9.2",
+ "rand 0.9.4",
  "rustls",
  "rustls-pki-types",
  "sha1",
@@ -5644,7 +5838,7 @@ version = "1.2.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "470dbf6591da1b39d43c14523b2b469c86879a53e8b758c8e090a470fe7b1fbe"
 dependencies = [
- "rand 0.9.2",
+ "rand 0.9.4",
  "serde",
  "web-time",
 ]
@@ -5729,9 +5923,9 @@ checksum = "06abde3611657adf66d383f00b093d7faecc7fa57071cce2578660c9f1010821"
 
 [[package]]
 name = "uuid"
-version = "1.22.0"
+version = "1.23.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a68d3c8f01c0cfa54a75291d83601161799e4a89a39e0929f4b0354d88757a37"
+checksum = "ddd74a9687298c6858e9b88ec8935ec45d22e8fd5e6394fa1bd4e99a87789c76"
 dependencies = [
  "getrandom 0.4.2",
  "js-sys",
@@ -6405,6 +6599,15 @@ dependencies = [
  "memchr",
 ]
 
+[[package]]
+name = "winnow"
+version = "1.0.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2ee1708bef14716a11bae175f579062d4554d95be2c6829f518df847b7b3fdd0"
+dependencies = [
+ "memchr",
+]
+
 [[package]]
 name = "wit-bindgen"
 version = "0.51.0"
diff --git a/Cargo.toml b/Cargo.toml
index 5ec3de3..697dc8d 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -7,26 +7,28 @@ edition = "2021"
 license = "MIT OR Apache-2.0"
 
 [workspace.dependencies]
-tokio              = { version = "1.50", features = ["full"] }
-async-trait        = "0.1"
-serde              = { version = "1.0", features = ["derive"] }
-serde_json         = "1.0"
-thiserror          = "2.0"
-anyhow             = "1.0"
-tracing            = "0.1"
+tokio = { version = "1.52", features = ["full"] }
+async-trait = "0.1"
+serde = { version = "1.0", features = ["derive"] }
+serde_json = "1.0"
+thiserror = "2.0"
+anyhow = "1.0"
+tracing = "0.1"
 tracing-subscriber = "0.3"
-uuid               = { version = "1", features = ["v4"] }
-chrono             = { version = "0.4", features = ["serde"] }
-regex              = "1"
-bytes              = "1"
-tempfile           = "3"
-tokio-test         = "0.4"
-axum               = { version = "0.8", features = ["json"] }
-tower-http         = { version = "0.6", features = ["cors", "trace", "fs"] }
-notify             = { version = "8.2", default-features = false, features = ["macos_fsevent"] }
-dashmap            = "6.1"
-clap               = { version = "4", features = ["derive"] }
-hostname           = "0.4"
-libc               = "0.2"
-reqwest            = { version = "0.13", features = ["json"] }
-inventory          = "0.3"
+uuid = { version = "1", features = ["v4"] }
+chrono = { version = "0.4", features = ["serde"] }
+regex = "1"
+bytes = "1"
+tempfile = "3"
+tokio-test = "0.4"
+axum = { version = "0.8", features = ["json"] }
+tower-http = { version = "0.6", features = ["cors", "trace", "fs"] }
+notify = { version = "8.2", default-features = false, features = [
+    "macos_fsevent",
+] }
+dashmap = "6.1"
+clap = { version = "4", features = ["derive"] }
+hostname = "0.4"
+libc = "0.2"
+reqwest = { version = "0.13", features = ["json"] }
+inventory = "0.3"
diff --git a/Dockerfile b/Dockerfile
new file mode 100644
index 0000000..c0f90be
--- /dev/null
+++ b/Dockerfile
@@ -0,0 +1,32 @@
+ARG IMAGE_BASE=reg.librecloud.online/lamina
+ARG RUST_VERSION=1.89
+
+# Pull nickel binary from the lamina nickel layer — named stage makes ARG expansion reliable
+FROM ${IMAGE_BASE}/nickel:${RUST_VERSION} AS nickel-layer
+
+# Builder: lamina rust layer provides cargo, rustc, sccache, cargo-chef
+FROM ${IMAGE_BASE}/rust:${RUST_VERSION} AS builder
+
+WORKDIR /build
+ENV CARGO_TARGET_DIR=/build/target
+
+# stratumiops source injected via --build-context stratumiops=../stratumiops
+# Resolves to /stratumiops — matches ../../../stratumiops relative path from crates/ontoref-daemon/
+COPY --from=stratumiops . /stratumiops
+
+COPY Cargo.toml Cargo.lock ./
+COPY crates ./crates
+
+RUN cargo build --release --bin ontoref-daemon
+
+# Runtime: minimal base + nickel (for NCL export at runtime) + daemon binary
+FROM alpine:3.21
+
+RUN apk add --no-cache ca-certificates
+
+COPY --from=nickel-layer /usr/local/bin/nickel /usr/local/bin/nickel
+COPY --from=builder /build/target/release/ontoref-daemon /usr/local/bin/ontoref-daemon
+
+EXPOSE 7891
+
+ENTRYPOINT ["ontoref-daemon"]
diff --git a/JUST.md b/JUST.md
new file mode 100644
index 0000000..d20dc6b
--- /dev/null
+++ b/JUST.md
@@ -0,0 +1,2 @@
+Hay que tener en cuenta que existe una descoordinación real y absoluta entre los Justfiles de los proyectos.
+Esto debería de ser armonizado o sincronizado: elementos comunes en un template o un modo específico ampliable según cada proyecto.
diff --git a/README.md b/README.md
index 1b417e2..383c46d 100644
--- a/README.md
+++ b/README.md
@@ -36,8 +36,8 @@ crates/            Rust implementation — typed struct loaders and mode executo
 | --- | --- |
 | `ontoref-ontology` | `.ontology/` NCL → typed Rust structs: Node, Edge, Dimension, Gate, Membrane. `Node` carries `artifact_paths` and `adrs` (`Vec<String>`, both `serde(default)`). Graph traversal, invariant queries. Zero deps. |
 | `ontoref-reflection` | NCL DAG contract executor with guards (pre-flight Block/Warn checks) and convergence loops (RetryFailed/RetryAll). ADR lifecycle, step dep resolution, config seal. `stratum-graph` + `stratum-state` required. |
-| `ontoref-daemon` | HTTP UI (11 pages), actor registry, notification barrier, MCP (33 tools), search engine, search bookmarks, SurrealDB, NCL export cache, per-file ontology versioning, annotated API catalog, Agent Task Composer. |
-| `ontoref-derive` | Proc-macro crate. `#[onto_api(...)]` annotates HTTP handlers — `description` is optional when a `///` doc comment exists (first line used as fallback). `#[derive(OntologyNode)]` + `#[onto(id, name, paths, description, adrs)]` auto-registers nodes via `inventory::submit!` at link time, merged into `Core` by `merge_contributors()`. `#[derive(ConfigFields)]` + `#[config_section(id, ncl_file)]` registers config struct fields. All three aggregate via `inventory::collect!`. |
+| `ontoref-daemon` | HTTP UI (11 pages), actor registry, notification barrier, MCP (34 tools), search engine, search bookmarks, SurrealDB, NCL export cache, per-file ontology versioning, annotated API catalog, Agent Task Composer. |
+| `ontoref-derive` | Proc-macro crate. `#[onto_api(...)]` annotates HTTP handlers — `description` is optional when a `///` doc comment exists (first line used as fallback). `#[onto_mcp_tool(name, description, input_schema)]` registers MCP tool unit-structs in the catalog at link time via `inventory::submit!(McpToolEntry{...})`; the annotated item is emitted unchanged and `ToolBase`/`AsyncTool` impls remain on the struct. `#[derive(OntologyNode)]` + `#[onto(id, name, paths, description, adrs)]` auto-registers nodes via `inventory::submit!` at link time, merged into `Core` by `merge_contributors()`. `#[derive(ConfigFields)]` + `#[config_section(id, ncl_file)]` registers config struct fields. All four aggregate via `inventory::collect!`. |
 
 `ontoref-daemon` caches `nickel export` results (keyed by path + mtime), reducing full sync
 scans from ~2m42s to <30s. The daemon is always optional — every module falls back to direct
@@ -55,7 +55,8 @@ automatically.
 **Q&A Knowledge Store** — accumulated Q&A entries persist to `reflection/qa.ncl` (typed NCL,
 git-versioned). Not localStorage. Any actor — developer, agent, CI — reads the same store.
 
-**MCP Server** — 29 tools over stdio and streamable-HTTP. Categories: discovery, retrieval, project
+**MCP Server** — 34 tools over stdio and streamable-HTTP, all registered at link time via
+`#[onto_mcp_tool]` (no manual catalog wiring). Categories: discovery, retrieval, project
 state, ontology, backlog, validation, Q&A, bookmarks, API surface. Representative subset:
 
 | Tool | What it does |
@@ -167,6 +168,19 @@ work as both `ore prov <cmd>` and standalone `prov <cmd>`. `ore help` and `descr
 surface the active domain automatically. New domains require only three files — no changes to the Nu
 dispatcher. ([ADR-012](adrs/adr-012-domain-extension-system.ncl))
 
+**Mode Hierarchy Validation** — `validate modes [--check]` reads `reflection/defaults/workflow.ncl::level_hierarchy`
+and checks every `.ncl` mode file for: level declared, strategy declared, delegate chain coherent,
+compose `extends` references valid. `mode resolve <id>` prints which hierarchy level handles the
+given mode and why. `validate modes --self-test` generates synthetic fixtures in a temp dir for
+fast CI smoke-testing of the validator itself. ([ADR-018](adrs/adr-018-workflow-layer-model.ncl))
+
+**Project Picker — vault and registry badges** — each project card surfaces OCI state inline:
+registry participant badge (`⟳ <participant>`) when `registry_provides` is declared; vault badge
+(`⛁ <vault_id> · N`) coloured green (declarative) or amber (legacy) when `sops.enabled` is true.
+Expanded project panel shows a collapsible **Registry** section with namespace, endpoint, and
+push/pull capability. The manage page adds **Runtime Services** toggles — MCP and GraphQL can be
+switched without a daemon restart via HTMX `POST /ui/manage/services/{service}/toggle`.
+
 **VCS Abstraction Layer** — `reflection/modules/vcs.nu` exposes a uniform API over jj and git:
 `detect`, `show-committed`, `restore-file`, `remote-url`, `current-branch`, `uncommitted-files`,
 `commit-count`. All ontoref modules consume `vcs.nu` — never hardcoded `^git`. Detection is
@@ -227,12 +241,62 @@ full ontology enrichment in 8 phases.
 
 `ONTOREF_PROJECT_ROOT` is set by the consumer wrapper — one ontoref checkout serves multiple projects.
 
+## Credential vault and registry federation
+
+ontoref ships a credential model for projects publishing or consuming OCI artifacts
+(domain contracts and integration modes) on a self-hosted registry like ZOT.
+The model is layered, declarative, and avoids ambient docker config:
+
+- **Layer 0** — master age private key (`.kage`) per actor, declared in
+  `~/.config/ontoref/config.ncl::vault.master_key_path` (override per-project
+  in `<project>/.ontoref/project.ncl::sops.master_key_path`)
+- **Layer 1** — `access.sops.yaml` per project, multi-recipient encrypted; carries
+  `zot_username`, `zot_password`, `vault_key`, `cosign_password`
+- **Layer 2** — operation credentials (RO/RW per registry entry) under
+  `src-vault/registry/`, referenced by `manifest.ncl::registry_provides[].credential_sops*`
+
+Tenant isolation within a single vault uses sops `creation_rules` driven by
+`sops.recipient_groups` + `sops.recipient_rules` in `project.ncl` — different
+clients/agents get disjoint recipient sets per file, all in one vault. Multi-vault
+is explicitly out of scope ([ADR-019](adrs/adr-019-per-file-recipient-routing-tenant-isolation.ncl)).
+
+**Adoption** — copy a template from `install/resources/templates/sops/`:
+
+| Template | When to use |
+|---|---|
+| `single-team/` | One team, no tenant separation |
+| `multi-tenant/` | Multiple clients with isolated credentials |
+| `agent-first/` | AI agents (MCP) read a single read-only credential |
+
+For integration artifacts (publishing domain contracts or consuming someone's mode),
+templates in `install/resources/templates/integration/`: `domain-producer/`,
+`mode-producer/`, `mode-consumer/`.
+
+**Day-to-day**:
+
+```sh
+ore secrets bootstrap         # create vault for a new project (admin only)
+ore secrets sync              # pull latest src-vault from ZOT
+ore secrets open              # acquire OCI lock + edit access.sops.yaml
+ore secrets close             # impact report → push → release lock
+ore secrets describe          # full inventory: groups, rules, scopes, ops
+ore secrets audit             # 6 ADR-017 + ADR-019 constraint checks
+```
+
+See FAQ entries in `reflection/qa.ncl` for diagrams, troubleshooting, and the
+15 named errors. ADRs: [017](adrs/adr-017-registry-credential-vault-model.ncl)
+(vault model) and [019](adrs/adr-019-per-file-recipient-routing-tenant-isolation.ncl)
+(per-file recipient routing).
+
 ## Prerequisites
 
 - [Nushell](https://www.nushell.sh/) >= 0.110.0
 - [Nickel](https://nickel-lang.org/) (for schema evaluation)
 - Rust toolchain (for building crates)
 - [Just](https://just.systems/) (for CI recipes)
+- [age](https://github.com/FiloSottile/age) + [sops](https://github.com/getsops/sops) (credential vault, ADR-017/019)
+- [oras](https://oras.land/) + [cosign](https://github.com/sigstore/cosign) ≥ 2 (OCI artifact federation)
+- [restic](https://restic.net/) or [kopia](https://kopia.io/) (vault snapshots)
 
 To build `ontoref-daemon` and `ontoref-reflection` with NATS/SurrealDB support, the
 stratumiops repo must be checked out at `../../../stratumiops`. Without it, build without
diff --git a/adrs/adr-014-runtime-service-toggles.ncl b/adrs/adr-014-runtime-service-toggles.ncl
new file mode 100644
index 0000000..eaab5d5
--- /dev/null
+++ b/adrs/adr-014-runtime-service-toggles.ncl
@@ -0,0 +1,98 @@
+let d = import "adr-defaults.ncl" in
+
+d.make_adr {
+  id     = "adr-014",
+  title  = "Runtime Service Toggles — AtomicBool Flags for MCP and GraphQL",
+  status = 'Accepted,
+  date   = "2026-04-26",
+
+  context = "ontoref-daemon exposes optional services (MCP, GraphQL) compiled in via Cargo feature flags. Once compiled, these services were always active for the lifetime of the process. Two scenarios require disabling them at runtime without restart: (1) security incident — temporarily disable a surface while investigating a compromise; (2) operator choice — enable graphql during a debug session, disable it in production. The alternative of restarting the daemon is disruptive because it drops all in-memory sessions, actors, and notification queues. Compile-time feature flags solve the binary presence problem but cannot address runtime availability.",
+
+  decision = "Introduce ServiceFlags (pub struct in api.rs) holding one AtomicBool per toggleable service, gated by the corresponding feature flag. ServiceFlags::new() initialises all flags to true (enabled). AppState holds Arc<ServiceFlags> shared across all clones. The toggle check lives in a route_layer middleware on the sub-router for each service — not inside the service handlers themselves — so the check is enforced regardless of which handler a request reaches. Two toggle surfaces are provided: (1) REST API: PUT /api/services/:service {\"enabled\": bool} — daemon admin Bearer required; (2) UI: POST /ui/manage/services/:service/toggle — AdminGuard (cookie session). Both surfaces return the new state. The manage page shows each compiled-in service with an HTMX toggle button; the navbar badges reflect runtime state on every page render.",
+
+  rationale = [
+    {
+      claim  = "AtomicBool is the correct primitive for a hot-path toggle",
+      detail = "A toggle check runs on every request to /mcp/* and /graphql/*. AtomicBool::load(Relaxed) is a single CPU instruction with no lock, no allocation, and no blocking — identical cost to a null pointer check. RwLock<bool> would introduce lock contention under concurrent requests with no benefit, since the only invariant needed is 'consistent within a single request', which Relaxed ordering satisfies.",
+    },
+    {
+      claim  = "Middleware layer placement enforces the toggle at the router boundary",
+      detail = "Placing the AtomicBool check inside each handler would require every handler to duplicate the guard, and a new handler added to the sub-router would silently bypass the toggle. A route_layer on the sub-router wraps all handlers uniformly — the check cannot be missed regardless of how many handlers the sub-router gains in the future.",
+    },
+    {
+      claim  = "Arc<ServiceFlags> (not Arc<AtomicBool> per flag) is the correct sharing primitive",
+      detail = "AtomicBool is not Clone. Wrapping the whole ServiceFlags struct in one Arc means: (1) the middleware closure captures one Arc clone instead of N; (2) new flags can be added to ServiceFlags without changing the sharing structure; (3) the UI and REST handlers access all flags via one field on AppState.",
+    },
+    {
+      claim  = "Two toggle surfaces (REST + UI) serve distinct operator workflows",
+      detail = "The REST endpoint (daemon admin Bearer) is suitable for scripted automation and CLI pipelines (curl -X PUT). The UI endpoint (AdminGuard cookie) enables toggle from the browser manage page without exposing the raw admin token. Both require daemon admin credentials — this is not a project-level operation.",
+    },
+    {
+      claim  = "Toggle is volatile — it does not survive daemon restart",
+      detail = "ServiceFlags are initialised to enabled=true on every startup regardless of the pre-restart state. This is intentional: a flag disabled for incident response should not silently persist across restarts, which would hide the disabled state from new operators. If persistent toggle state is needed it belongs in config.ncl as a compiled-in default, not in volatile memory.",
+    },
+  ],
+
+  consequences = {
+    positive = [
+      "MCP and GraphQL can be disabled in under 1ms without dropping in-memory state",
+      "Toggle surfaces require daemon admin credentials — the same identity used for project management",
+      "New optional services added to the daemon gain toggle support by adding one AtomicBool field to ServiceFlags",
+      "Middleware placement means the toggle cannot be bypassed by adding new handlers to the sub-router",
+    ],
+    negative = [
+      "Toggle state is lost on restart — disabling a service for incident response must be re-applied after restart or captured in config",
+      "ServiceFlags is not Clone (AtomicBool is not Clone) — AppState clones the Arc, not the flags; callers that pattern-match on AppState fields must be aware of this",
+    ],
+  },
+
+  alternatives_considered = [
+    {
+      option       = "Restart daemon with different feature flags or config",
+      why_rejected = "Restart drops SessionStore (all active logins), actor registry, notification queue, and NATS subscriptions. A 1-second outage is acceptable for planned maintenance but not for rapid incident response.",
+    },
+    {
+      option       = "RwLock<bool> per service in AppState",
+      why_rejected = "RwLock introduces lock contention on every request. The toggle check does not need mutual exclusion with writes — a store and a load never run concurrently in a way that would corrupt state. Relaxed AtomicBool is sufficient and faster.",
+    },
+    {
+      option       = "Dynamic axum Router rebuild — swap out the sub-router entirely",
+      why_rejected = "axum Router is not live-rebuildable without replacing the entire tower Service. This would require Arc<RwLock<Router>>, a custom Service wrapper, and would still incur a lock per request. The middleware approach achieves the same result with orders of magnitude less complexity.",
+    },
+  ],
+
+  related_adrs = ["adr-002", "adr-005"],
+
+  ontology_check = {
+    decision_string    = "runtime toggle AtomicBool service mcp graphql middleware",
+    invariants_at_risk = [],
+    verdict            = 'Safe,
+  },
+
+  constraints = [
+    {
+      id        = "c-014-1",
+      claim     = "ServiceFlags::new() must initialise all AtomicBool flags to true — a compiled-in service is always enabled at startup",
+      scope     = "crates/ontoref-daemon/src/api.rs",
+      severity  = 'Hard,
+      rationale = "Starting with flags disabled would silently hide services from operators who have not read this ADR. Explicit runtime toggle (documented in UI and REST API) is the correct mechanism to disable a service.",
+      check = { tag = "Grep", pattern = "AtomicBool::new(true)", paths = ["crates/ontoref-daemon/src/api.rs"], must_be_empty = false },
+    },
+    {
+      id        = "c-014-2",
+      claim     = "Any new optional service added to the daemon must add an AtomicBool to ServiceFlags and a route_layer toggle middleware on its sub-router",
+      scope     = "crates/ontoref-daemon/src/api.rs",
+      severity  = 'Hard,
+      rationale = "A service without a toggle violates the operator contract established by this ADR. Future services must be consistently operable.",
+      check = { tag = "Grep", pattern = "pub struct ServiceFlags", paths = ["crates/ontoref-daemon/src/api.rs"], must_be_empty = false },
+    },
+    {
+      id        = "c-014-3",
+      claim     = "Service toggle endpoints require daemon admin credentials — project-level auth is insufficient",
+      scope     = "crates/ontoref-daemon/src/api.rs, crates/ontoref-daemon/src/ui/handlers.rs",
+      severity  = 'Soft,
+      rationale = "Service availability is daemon-wide, not per-project. Allowing a project admin to disable MCP or GraphQL for all projects would be a privilege escalation.",
+      check = { tag = "Grep", pattern = "AdminGuard", paths = ["crates/ontoref-daemon/src/ui/handlers.rs"], must_be_empty = false },
+    },
+  ],
+}
diff --git a/adrs/adr-015-mcp-tool-inventory-auto-derive.ncl b/adrs/adr-015-mcp-tool-inventory-auto-derive.ncl
new file mode 100644
index 0000000..24b7658
--- /dev/null
+++ b/adrs/adr-015-mcp-tool-inventory-auto-derive.ncl
@@ -0,0 +1,88 @@
+let d = import "adr-defaults.ncl" in
+
+d.make_adr {
+  id     = "adr-015",
+  title  = "MCP Tool Catalog via #[onto_mcp_tool] Proc-Macro + Inventory",
+  status = 'Accepted,
+  date   = "2026-04-26",
+
+  context = "ontoref-daemon exposes 33 MCP tools via the rmcp ToolRouter in crates/ontoref-daemon/src/mcp/mod.rs. The authoritative tool implementation lives in each tool struct's `ToolBase`/`AsyncTool` impls (name, description, schema). However, the agent-facing tool catalog returned by `ontoref_help` was a hand-typed JSON literal (~100 lines, mod.rs:1129-1226) duplicating every tool's name, description, and parameter shape. This is the same failure mode ADR-007 documents as the original motivation for `#[onto_api]`: the previous-session bug where `insert_mcp_ctx` listed 15 tools while the router had 27. ADR-007 fixed that drift for the HTTP API surface (inventory-collected `ApiRouteEntry`) but did not extend the pattern to MCP. As of 2026-04-26 the manifest claim 'daemon exposes 33 MCP tools' is also a hand-maintained string in `.ontology/manifest.ncl`. With the rate of MCP tool additions through 2026-Q1 (qa, bookmarks, actions, config, ontology extensions all added in separate sessions), drift was becoming inevitable.",
+
+  decision = "Every MCP tool struct in ontoref-daemon must carry `#[onto_mcp_tool(name, description, category, params)]`. The proc-macro (in `crates/ontoref-derive`) emits `inventory::submit!(ontoref_ontology::McpToolEntry{...})` at link time and leaves the annotated struct unchanged — the existing `ToolBase` and `AsyncTool` impls are untouched. A new pure function `ontoref_daemon::mcp::catalog()` walks `inventory::iter::<McpToolEntry>()`, sorts by name, and returns `Vec<&'static McpToolEntry>`. `HelpTool::invoke` now serializes `catalog()` instead of holding a hand-typed JSON literal. `McpToolEntry` lives in `ontoref-ontology` next to `ApiRouteEntry` and reuses `ApiParam` for parameter metadata — both are protocol surfaces, the type is generic. `tool_router()`'s compile-time `with_async_tool::<T>()` list is left as-is; the Rust type system requires the type list at compile time and a separate macro architecture would be needed to derive it from inventory (out of scope for this ADR).",
+
+  rationale = [
+    {
+      claim  = "Same drift class as ADR-007 — same fix pattern",
+      detail = "ADR-007 cites the insert_mcp_ctx (15 listed) vs router (27 actual) drift as proof that hand-maintained registries decay. The HelpTool JSON literal had identical mechanics: every new tool struct required two coordinated edits (with_async_tool registration + JSON entry) with no compiler enforcement that they stayed in sync. Applying the same inventory linker pattern that fixed it for HTTP routes closes the equivalent gap for MCP.",
+    },
+    {
+      claim  = "Co-location of annotation and implementation",
+      detail = "`#[onto_mcp_tool]` sits directly on the tool struct, immediately above its `ToolBase` impl. Adding a tool that compiles but has no inventory entry requires actively skipping the attribute — a much higher-friction error than forgetting to update a JSON list 100 lines away in the same file. Removing a tool struct removes its inventory entry automatically.",
+    },
+    {
+      claim  = "Zero new dependencies",
+      detail = "ontoref-derive and inventory are already workspace dependencies (introduced by ADR-007). McpToolEntry mirrors ApiRouteEntry's shape and reuses ApiParam — no schema duplication. The change is additive: existing #[onto_api] macros are unaffected.",
+    },
+    {
+      claim  = "Preserves runtime cost guarantees",
+      detail = "inventory::iter walks a linker-built linked list — no HashMap, no Arc, no allocation. catalog() is a pure function returning &'static references. HelpTool's per-call cost is unchanged (still O(n) over the tool list, no cache invalidation, no daemon state required).",
+    },
+  ],
+
+  consequences = {
+    positive = [
+      "Adding a tool struct without #[onto_mcp_tool] makes the tool invisible to ontoref_help — drift is detectable on first agent invocation",
+      "Removing a tool struct removes its catalog entry automatically; no orphaned help entries",
+      "Manifest claim '33 MCP tools' is verifiable at runtime via catalog().len() rather than asserted by hand",
+      "Future tool tiering (essential vs extended, mentioned but deferred from this ADR) becomes a single-field addition to McpToolEntry",
+      "Same proc-macro infrastructure as #[onto_api] — no new linker semantics, no platform-specific concerns introduced",
+    ],
+    negative = [
+      "Per-tool params are stringly-typed in the `params = \"name:type:constraint:desc; ...\"` attribute argument — a typo in `required` vs `optional` is not caught at compile time (mirrors the same limitation in #[onto_api])",
+      "tool_router()'s with_async_tool::<T>() list still requires manual maintenance — the inventory does not eliminate it. A drift between router registrations and inventory entries is possible (caught only at runtime by the regression test, not at compile time)",
+      "33 tool structs require a one-time annotation pass — non-trivial diff size, but mechanical and reviewable",
+    ],
+  },
+
+  alternatives_considered = [
+    {
+      option       = "Generate the help JSON from ToolBase::name() + description() directly via reflection",
+      why_rejected = "ToolBase exposes name and description but not parameter metadata in a structured form (input_schema returns a JsonObject blob, not the per-param required/values/note shape that ontoref_help emits). Walking the JSON schema and reverse-engineering the per-param hints is brittle. The inventory entry lets us declare the agent-facing param documentation in a stable, typed shape next to the ToolBase impl.",
+    },
+    {
+      option       = "Replace tool_router()'s with_async_tool::<T>() list with macro-driven iteration over inventory",
+      why_rejected = "rmcp's ToolRouter requires the tool type at compile time (generic associated types over each tool's Parameter/Output types). Driving registration from inventory entries — which are runtime values of `&'static McpToolEntry` — would require either a build.rs that emits the registration list or a macro that takes the full type list as input. Either is feasible but is a larger architectural change with no immediate reliability win beyond what the inventory already provides for the help/catalog surface. Deferred.",
+    },
+    {
+      option       = "Skip the macro and write an `inventory::submit!` block manually next to each tool",
+      why_rejected = "Eliminates the macro infrastructure but loses the validation that #[onto_mcp_tool] applies (key spelling check, param string parsing reuse). Manual blocks also bypass the file!() capture for source-file traceability that ApiRouteEntry already uses.",
+    },
+  ],
+
+  constraints = [
+    {
+      id       = "onto-mcp-tool-on-all-tools",
+      claim    = "Every MCP tool struct in ontoref-daemon must carry #[onto_mcp_tool(...)]",
+      scope    = "ontoref-daemon (crates/ontoref-daemon/src/mcp/mod.rs)",
+      severity = 'Hard,
+      check    = { tag = 'Grep, pattern = "#\\[onto_mcp_tool", paths = ["crates/ontoref-daemon/src/mcp/mod.rs"], must_be_empty = false },
+      rationale = "catalog() is only as complete as the set of annotated tool structs. Unannotated tools are invisible to agents calling ontoref_help — equivalent to undocumented tools and a re-introduction of the ADR-007 drift class.",
+    },
+    {
+      id       = "mcp-catalog-router-parity",
+      claim    = "The number of #[onto_mcp_tool] annotations must equal the number of with_async_tool::<T>() calls in tool_router()",
+      scope    = "ontoref-daemon (crates/ontoref-daemon/src/mcp/mod.rs)",
+      severity = 'Soft,
+      check    = { tag = 'NuCmd, command = "let a = (open crates/ontoref-daemon/src/mcp/mod.rs | lines | where ($it | str starts-with '#[ontoref_derive::onto_mcp_tool') | length); let b = (open crates/ontoref-daemon/src/mcp/mod.rs | lines | where ($it | str contains 'with_async_tool::<') | length); if $a != $b { error make {msg: $'annotation count ($a) != router registration count ($b)'} } else { 'ok' }" },
+      rationale = "Without compile-time enforcement, a tool could be registered with the router but lack the annotation (or vice versa). This soft constraint catches the divergence at validate time. Future work (deferred alternative above) can lift this to compile time.",
+    },
+  ],
+
+  related_adrs = ["adr-007", "adr-001"],
+
+  ontology_check = {
+    decision_string    = "Use #[onto_mcp_tool] proc-macro + inventory linker registration as the single source of truth for the MCP tool catalog; HelpTool renders from catalog() instead of a hand-typed JSON literal",
+    invariants_at_risk = ["protocol-not-runtime"],
+    verdict            = 'Safe,
+  },
+}
diff --git a/adrs/adr-016-component-lift-out-pattern.ncl b/adrs/adr-016-component-lift-out-pattern.ncl
new file mode 100644
index 0000000..ecf1d5f
--- /dev/null
+++ b/adrs/adr-016-component-lift-out-pattern.ncl
@@ -0,0 +1,101 @@
+let d = import "defaults.ncl" in
+
+d.make_adr {
+  id     = "adr-016",
+  title  = "Component Lift-Out Pattern: Four-Criterion Gate for Standalone Extraction",
+  status = 'Accepted,
+  date   = "2026-05-01",
+
+  context = "Ontoref itself was extracted from stratumiops (ADR-001). The same pattern recurred in the provisioning project: buildkit-launcher (build substrate) and backup-manager (backup orchestration) grew inside provisioning but serve a broader class of consumers — vapora, workspace infras, CI pipelines, and third-party projects. No formal criterion existed for deciding when a component is ready to become a standalone peer project. Without a criterion, the decision is arbitrary and either premature (decomposition overhead before consumer plurality) or delayed (host coupling accretes, extraction becomes expensive).",
+
+  decision = "A component is extracted as a standalone peer project when it passes all four criteria: (1) Orthogonal concern — the component's core domain is not the host project's core domain; (2) Consumer plurality — at least two distinct callers exist or are immediately planned; (3) Release cadence divergence — the component's evolution is not gated by the host project's release cycle; (4) Config path-agnostic — the component can receive its config from any caller without importing host infrastructure (workspace crates, host config loaders, host schema registries). The extracted project registers in ontoref before any code moves. The host project retains extension-side artifacts (schemas, defaults, component declarations) that allow workspace infras to declare directives for the extracted tool.",
+
+  rationale = [
+    {
+      claim  = "Orthogonality is the primary criterion",
+      detail = "A build substrate (lian-build) and a backup orchestrator (cloudatasave) do not belong to provisioning's core domain (workspace lifecycle management). Provisioning calls them; it does not own their concerns. Orthogonality is the structural signal that extraction is correct, not premature.",
+    },
+    {
+      claim  = "Consumer plurality prevents premature extraction",
+      detail = "A component with a single caller has no demonstrated need for independent lifecycle. Two or more distinct callers (e.g. provisioning + vapora, or provisioning + a workspace CI pipeline) prove the component's value is general, not host-specific. ADR-001 precedent: ontoref serves stratumiops, typedialog, vapora, kogral.",
+    },
+    {
+      claim  = "Config path-agnostic is the coupling criterion",
+      detail = "If a component imports the host's config loader, schema registry, or workspace crates to function, it cannot be independently deployed or tested. Severing this import at extraction time is not simplification — it is restoration of the correct dependency direction. The extracted project has its own config infrastructure; callers pass directives in the extracted project's own NCL vocabulary.",
+    },
+    {
+      claim  = "Ontoref registration before code move establishes ontological identity",
+      detail = "Creating the .ontology/ and ADRs in the new project before any code moves ensures the project exists as a declared concept before it exists as an implementation. This prevents the common failure mode where a project's identity is derived post-hoc from accumulated code rather than declared from design.",
+    },
+    {
+      claim  = "Host retains extension-side artifacts",
+      detail = "The host project (provisioning) retains the extension schemas, defaults, and component declarations that allow workspace infras to configure the extracted tool. This is not coupling — it is the host's side of the protocol contract. The extracted project defines its capabilities in its own ontology; provisioning extensions expose those capabilities to workspace consumers.",
+    },
+  ],
+
+  consequences = {
+    positive = [
+      "Decisions to extract are justified, not arbitrary — four criteria provide a checkable gate",
+      "Extracted projects are immediately usable by non-provisioning callers (vapora, CI, workspace infras)",
+      "Host projects remain focused on their core domain; extension schemas handle the integration surface",
+      "The pattern is self-documenting: ADR-001 (ontoref from stratumiops) and ADR-016 instances (lian-build, cloudatasave from provisioning) form a corpus of precedent",
+      "Ontoref registration before code move ensures the project has identity before it has implementation",
+    ],
+    negative = [
+      "Extraction requires a deliberate severing of host infrastructure imports — this is work that must be done correctly, not deferred",
+      "Two registration surfaces per project: the extracted project's own .ontoref/ and the host's extension schemas",
+      "Consumer projects (workspace infras) must import from the extracted project's schema vocabulary, not from the host's workspace schema",
+    ],
+  },
+
+  alternatives_considered = [
+    {
+      option       = "Extract only when a second consumer exists and requests it",
+      why_rejected = "Reactive extraction means coupling has already accreted. By the time a second consumer requests the component, host infrastructure imports are deep. Proactive evaluation at the four-criterion gate is cheaper than reactive untangling.",
+    },
+    {
+      option       = "Keep all tools inside provisioning as internal workspace crates, expose via provisioning CLI",
+      why_rejected = "This routes all callers through provisioning's release cycle and binary. vapora and workspace CI pipelines cannot use the tool without depending on provisioning. The four-criterion test exists precisely to identify when this constraint becomes incorrect.",
+    },
+    {
+      option       = "Publish extracted crates to crates.io immediately",
+      why_rejected = "Published crates require stable API contracts before the consumer relationship is proven. Path-based standalone project references allow simultaneous development of the extracted project and its consumers without publication ceremony.",
+    },
+  ],
+
+  constraints = [
+    {
+      id       = "ontology-before-code",
+      claim    = "An extracted project must have .ontology/core.ncl, .ontology/state.ncl, and adrs/adr-001 committed before any code from the host is moved",
+      scope    = "any project applying the lift-out pattern",
+      severity = 'Hard,
+      check    = {
+        tag     = 'FileExists,
+        path    = ".ontology/core.ncl",
+        present = true,
+      },
+      rationale = "Ontological identity precedes implementation. A project that exists only as code has no declared purpose, axioms, or constraints — the extraction is indistinguishable from a rename.",
+    },
+    {
+      id       = "no-host-infrastructure-import",
+      claim    = "The extracted project must not import workspace crates, config loaders, or schema registries from the host project",
+      scope    = "extracted project source code",
+      severity = 'Hard,
+      check    = {
+        tag          = 'Grep,
+        pattern      = "platform_config|stratum-config|provisioning-config",
+        paths        = ["src/", "crates/"],
+        must_be_empty = true,
+      },
+      rationale = "Importing host infrastructure is the definition of host coupling. If the import cannot be severed, the component fails criterion 4 and must not be extracted until the coupling is resolved.",
+    },
+  ],
+
+  related_adrs = ["adr-001-protocol-as-standalone-project"],
+
+  ontology_check = {
+    decision_string    = "components are extracted as standalone peer projects when they pass a four-criterion gate; ontoref registration precedes code move; host retains extension-side schemas",
+    invariants_at_risk = ["protocol-not-runtime", "self-describing"],
+    verdict            = 'Safe,
+  },
+}
diff --git a/adrs/adr-017-registry-credential-vault-model.ncl b/adrs/adr-017-registry-credential-vault-model.ncl
new file mode 100644
index 0000000..d90c291
--- /dev/null
+++ b/adrs/adr-017-registry-credential-vault-model.ncl
@@ -0,0 +1,217 @@
+let d = import "defaults.ncl" in
+
+d.make_adr {
+  id     = "adr-017",
+  title  = "Registry Credential Vault Model: src-vault, Multi-Recipient sops, and Actor-Scoped Access",
+  status = 'Accepted,
+  date   = "2026-05-01",
+
+  context = "The provisioning registry (zot OCI) became the primary coordination hub for domain and mode artifacts (ADR-016 consequences, migration 0015). This introduced registry credentials as a first-class concern: every project that pushes or pulls artifacts needs credentials, and those credentials must be distributed across actors (developers, CI pipelines, the ontoref daemon, AI agents, ops tooling) with different access levels. The naive model — environment variables or ambient ~/.docker/config.json — has three compounding failure modes: (1) credentials are ambient and unscoped, so an operation against project A can use credentials that belong to project B; (2) there is no distribution mechanism — adding a new team member or CI key requires manual redistribution of a shared secret; (3) there is no audit trail for credential access or changes. A supply-chain attack via misconfigured registry namespace endpoints is the concrete risk: if the daemon proxied registry calls, any actor with MCP access could redirect them to an arbitrary endpoint using ambient credentials.",
+
+  decision = "Registry credentials are managed through a per-project src-vault stored as an OCI artifact in the same ZOT registry it protects. The src-vault uses sops with age multi-recipient encryption: each actor role has its own age keypair, and credential files are encrypted for the exact set of recipients that need access. The vault backend (restic or kopia) handles versioning and local copies; the ZOT registry handles distribution. Each project has a local access.sops.yaml in ~/.config/ontoref/vaults/<project-slug>/ containing three fields: zot_username, zot_password, and vault_key. All three are encrypted by the actor's master age private key (.kage), which lives at an external path the actor controls (hardware key, encrypted disk, or declared path in config.ncl) — never inside the vault directory. At operation time, sops decrypts access.sops.yaml in memory: zot_username and zot_password are used to pull the src-vault OCI artifact via a DOCKER_CONFIG tmpdir that is deleted immediately after; vault_key is passed as RESTIC_PASSWORD or KOPIA_PASSWORD env var for the duration of the vault operation and never written to disk. No plaintext credential of any kind persists beyond the operation scope. Credential resolution runs exclusively in the ontoref CLI — the daemon is structurally excluded. All oras invocations use an isolated DOCKER_CONFIG tmpdir; no ambient ~/.docker/config.json is consulted. Access logs are appended to a jsonl file stored as a layer in the same src-vault OCI artifact and mirrored locally in ~/.config/ontoref/vaults/<project-slug>/logs/access.jsonl.",
+
+  rationale = [
+    {
+      claim  = "src-vault in ZOT closes the distribution gap structurally",
+      detail = "Storing the vault as an OCI artifact in the same registry that protects means distribution is the same operation as registry access. Any actor with registry RO credentials can pull the vault; only admin can push. No separate key distribution channel is needed. When a developer joins, their public key is added to src-vault and they can immediately pull their copy — no coordinator required.",
+    },
+    {
+      claim  = "Multi-recipient sops eliminates shared secrets",
+      detail = "Each actor has its own age keypair. The sops DEK is encrypted separately for each recipient public key. Removing a recipient requires only updating .sops.yaml and running sops updatekeys — no key redistribution. The private key of the removed actor becomes useless for future vault operations the moment updatekeys completes. This is the only model that supports revocation without rotation of all other credentials.",
+    },
+    {
+      claim  = "Daemon structural exclusion is not a policy, it is an architectural property",
+      detail = "The daemon process has no age private key and no access to project .kage files. It cannot decrypt sops files even if it reads them. The daemon surfaces only the declarative topology from manifest.ncl (registry_provides, credential_sops paths as strings). Credential resolution happens in the CLI process of the authenticated user. This cannot be bypassed without modifying the daemon to acquire keys — which would be an explicit, reviewable change, not a configuration drift.",
+    },
+    {
+      claim  = "Lock state in ZOT with admin-only write ACL makes the lock structural",
+      detail = "The src-vault namespace has write ACL restricted to admin credentials in ZOT config. The lock artifact (src-vault/<project>:lock) can only be pushed or deleted with admin credentials. Other roles can read the lock state but cannot modify it. This converts a cooperative lock into a structurally enforced one — bypassing the lock requires admin credentials, which is an auditable action, not an accident.",
+    },
+    {
+      claim  = "Audit log co-located with vault enables tamper-evident access history",
+      detail = "The audit.jsonl layer is part of the same OCI artifact as the vault content. Every push of src-vault:<project>:latest includes the updated log. Since only admin can push, and the log is append-only within a vault session, the log cannot be modified without admin credentials and a vault open/close cycle — both of which are themselves logged. This does not prevent a malicious admin from editing the log, but it makes any edit detectable via the OCI manifest digest history.",
+    },
+    {
+      claim  = "Role scope enforcement at two independent layers",
+      detail = "Scope is enforced first by the CLI (pre-check against scopes/<role>.ncl before calling oras) and second by the ZOT ACL (the token itself carries the permissions issued by the registry). These layers are independent: a misconfigured scope file does not grant registry permissions, and a misconfigured registry ACL does not bypass the CLI pre-check. The two layers must be kept in sync via secrets-audit in CI.",
+    },
+    {
+      claim  = "credential_env is excluded by design, not by convention",
+      detail = "Environment variables are visible in ps aux, inherited by all child processes, logged by CI systems unless explicitly masked, and cannot be scoped to a specific operation. Excluding credential_env from the schema means it is impossible to accidentally use it — there is no field to populate. The only credential reference fields are credential_sops (current) and credential_oidc (declared for future OIDC workload identity support).",
+    },
+    {
+      claim  = "vault_key in sops ensures the restic/kopia key is never plaintext at rest",
+      detail = "The restic/kopia encryption key (vault_key) lives inside access.sops.yaml, encrypted by the actor's age master key. It is decrypted into RESTIC_PASSWORD or KOPIA_PASSWORD env var only for the duration of the vault operation — never written to a key file on disk. This means the only plaintext secrets on disk are the actor's .kage (which the actor is responsible for protecting) and the sops-encrypted access.sops.yaml (unreadable without the .kage). The pattern is: one master key protects everything else; everything else is encrypted at rest.",
+    },
+    {
+      claim  = "restic and kopia are equivalent vault backends behind a thin abstraction",
+      detail = "Both restic and kopia provide encrypted, versioned snapshots with content-addressed storage. The choice between them depends on operational preference (kopia has a richer UI and faster incremental backups; restic has broader ecosystem support). The vault-backend.nu abstraction wraps init, backup, restore, and snapshot-list — switching backends requires changing vault_backend.tool in config.ncl and re-initializing the local repo. The OCI artifact in ZOT is backend-agnostic.",
+    },
+  ],
+
+  consequences = {
+    positive = [
+      "Credential distribution is the same operation as registry access — no separate channel",
+      "Revoking a developer's access requires one command (secrets-remove-key) with no coordination",
+      "The daemon cannot be used as a credential amplifier regardless of MCP actor permissions",
+      "Every vault open, edit, and close is logged in the artifact that contains the credentials",
+      "Adding a new role (e.g. backup-agent) requires only a new keypair and a .sops.yaml update",
+      "Local copy in ~/.config/ontoref/vaults/<slug>/ provides offline fallback for admin",
+      "Bootstrap creates the first consistent vault state before any live registry interaction",
+      "impact analysis on secrets-close identifies services affected by credential changes before confirming",
+    ],
+    negative = [
+      "Bootstrap requires admin to collect all public keys before first vault push — cannot be fully automated",
+      "sops updatekeys must run after every recipient change — forgetting it leaves old recipients in files",
+      "Zot ACL and scope NCL files must be kept in sync manually — drift produces security theater",
+      "Loss of .kage means access.sops.yaml cannot be decrypted — vault_key and ZOT credentials are unrecoverable without the master key; admin must keep the .kage backed up independently",
+      "ore secrets open/close adds a ritual to any secrets change session — acceptable overhead, but real",
+      "Direct sops invocations bypass the lock and audit log — CI audit detects but does not prevent",
+    ],
+  },
+
+  alternatives_considered = [
+    {
+      option       = "Environment variables (REGISTRY_TOKEN, DOCKER_CONFIG) per CI job",
+      why_rejected = "Visible in process list, inherited by subprocesses, logged by CI systems, cannot be scoped to a specific registry endpoint. No revocation without rotating all consumers. The primary attack surface for supply-chain credential leakage.",
+    },
+    {
+      option       = "Shared age keyring — one private key distributed to all developers",
+      why_rejected = "Revocation requires rotating the shared key and redistributing to all remaining members — coordination cost scales with team size. No per-actor audit trail. A leaked key compromises all actors simultaneously.",
+    },
+    {
+      option       = "HashiCorp Vault or similar external secrets manager",
+      why_rejected = "Introduces a network dependency for every credential resolution. Requires operating a separate service with its own HA, backup, and auth model. The OCI registry is already the coordination hub — using it as the vault distribution backend reuses existing infrastructure and auth.",
+    },
+    {
+      option       = "Daemon resolves credentials on behalf of CLI actors",
+      why_rejected = "The daemon is a long-lived process accessible to multiple actors (developer, agent, CI via MCP). Giving it credential resolution capability means any actor with daemon access can trigger registry operations using credentials they do not personally hold. The structural exclusion of the daemon from credential resolution is a load-bearing architectural property.",
+    },
+    {
+      option       = "Single credential file per registry (no RO/RW split)",
+      why_rejected = "A single credential with RW access distributed to read-only actors (cdci, ontoref, agent) violates least-privilege. A compromised CI pipeline with RW credentials can push malicious artifacts. The RO/RW split means a compromised read-only credential cannot alter the artifact namespace.",
+    },
+  ],
+
+  constraints = [
+    {
+      id       = "vault-access-credentials-independent",
+      claim    = "The credentials required to pull the src-vault OCI artifact from ZOT must be stored outside the vault itself — each src-vault has its own access credentials, held in the actor's local .kage, not inside any vault it protects",
+      scope    = "all projects with src-vault in ZOT",
+      severity = 'Hard,
+      check    = {
+        tag  = 'NuCmd,
+        cmd  = "ore secrets audit --check bootstrap-credentials",
+      },
+      rationale = "A vault whose access credential is inside itself cannot be opened. The bootstrap credential (registry RO for the src-vault namespace) must preexist in the actor's local .kage. Two projects on the same ZOT instance have independent src-vault namespaces with independent access credentials — access to one does not imply access to the other.",
+    },
+    {
+      id       = "no-credential-env",
+      claim    = "RegistryEntry must not use credential_env — only credential_sops or credential_oidc are valid credential reference fields",
+      scope    = "all manifest.ncl files declaring registry_provides",
+      severity = 'Hard,
+      check    = {
+        tag          = 'Grep,
+        pattern      = "credential_env",
+        paths        = [".ontology/manifest.ncl"],
+        must_be_empty = true,
+      },
+      rationale = "credential_env is excluded from the schema. Its presence indicates a manual edit bypassing the contract.",
+    },
+    {
+      id       = "multi-recipient-mandatory",
+      claim    = "Every *.sops.yaml credential file must include at minimum the admin and the bound_actor recipients for its access class",
+      scope    = "all projects with registry_provides.registries[].credential_sops declared",
+      severity = 'Hard,
+      check    = {
+        tag  = 'NuCmd,
+        cmd  = "ore secrets audit --check recipients",
+      },
+      rationale = "A credential file encrypted for a single recipient is indistinguishable from a shared secret. Multi-recipient is the mechanism that enables revocation without rotation.",
+    },
+    {
+      id       = "uses-registry-declared",
+      claim    = "Any domain or mode that pushes or pulls from a registry must declare uses_registry referencing the RegistryEntry id in manifest.ncl",
+      scope    = "domain and mode NCL declarations",
+      severity = 'Hard,
+      check    = {
+        tag  = 'NuCmd,
+        cmd  = "ore secrets audit --check registry-deps",
+      },
+      rationale = "Without the explicit dependency declaration, secrets-close cannot compute the impact of credential changes. Undeclared dependencies produce silent breakage after a credential rotation.",
+    },
+    {
+      id       = "ore-secrets-exclusive-wrapper",
+      claim    = "All sops operations on registry credential files must go through ore secrets — direct sops invocations bypass lock state and audit log",
+      scope    = "CI pipeline and developer workflow",
+      severity = 'Soft,
+      check    = {
+        tag  = 'NuCmd,
+        cmd  = "ore secrets audit --check lock-compliance",
+      },
+      rationale = "ore secrets enforces lock state verification before any sops edit. Direct sops invocations are detectable in CI audit but not preventable at the filesystem level. The constraint is soft because enforcement is post-hoc.",
+    },
+    {
+      id       = "docker-config-isolation",
+      claim    = "Every oras invocation must use DOCKER_CONFIG pointing to a tmpdir containing only the credential for the target registry endpoint — no ambient ~/.docker/config.json",
+      scope    = "nulib/cli/integration.nu and any code calling oras",
+      severity = 'Hard,
+      check    = {
+        tag          = 'Grep,
+        pattern      = "DOCKER_CONFIG",
+        paths        = ["provisioning/core/nulib/"],
+        must_be_empty = false,
+      },
+      rationale = "Ambient credentials allow operations against project A to use credentials that belong to project B if both resolve to the same registry hostname. Isolation is enforced by constructing the config at call time and deleting it immediately after.",
+    },
+    {
+      id       = "vault-key-never-plaintext",
+      claim    = "vault_key must never be written to disk in plaintext — it is decrypted from access.sops.yaml into RESTIC_PASSWORD or KOPIA_PASSWORD for the duration of the operation only",
+      scope    = "nulib/platform/vault-backend.nu and all ore secrets callers",
+      severity = 'Hard,
+      check    = {
+        tag          = 'Grep,
+        pattern      = "vault_key.*save\\|vault_key.*write\\|restic.*--key-file",
+        paths        = ["provisioning/core/nulib/"],
+        must_be_empty = true,
+      },
+      rationale = "Writing vault_key to a file defeats the purpose of encrypting it in sops. The env var approach (RESTIC_PASSWORD, KOPIA_PASSWORD) keeps the key in process memory only. The .kage master key is the single plaintext secret the actor manages — everything derived from it must be ephemeral.",
+    },
+    {
+      id       = "vault-backend-abstracted",
+      claim    = "All vault snapshot operations must use vault-backend.nu — direct restic or kopia invocations are not permitted in recipes or modules",
+      scope    = "justfiles/secrets.just and nulib/platform/vault-backend.nu callers",
+      severity = 'Soft,
+      check    = {
+        tag  = 'Grep,
+        pattern      = "^\\s*restic\\|^\\s*kopia",
+        paths        = ["justfiles/", "provisioning/core/nulib/"],
+        must_be_empty = true,
+      },
+      rationale = "Switching between restic and kopia must require only changing vault_backend.tool in config — not hunting direct invocations across recipes and modules.",
+    },
+    {
+      id       = "src-vault-cosign-signed",
+      claim    = "Every push of src-vault/<project>:latest to ZOT must produce a cosign signature; every pull must verify the signature before the artifact is trusted",
+      scope    = "justfiles/secrets.just secrets-push and secrets-sync recipes",
+      severity = 'Hard,
+      check    = {
+        tag  = 'NuCmd,
+        cmd  = "ore secrets audit --check cosign-signature",
+      },
+      rationale = "The src-vault artifact contains access.sops.yaml with encrypted registry credentials. An unsigned artifact can be substituted without detection — a malicious actor with write access to the registry could replace it with a version encrypted for attacker-controlled recipients. COSIGN provides tamper evidence independent of ZOT ACLs: even if ACLs are misconfigured, the signature check prevents a substituted vault from being consumed. cosign is now a Hard prerequisite for ontoref (declared in requirements) for this reason.",
+    },
+  ],
+
+  related_adrs = [
+    "adr-005-unified-auth-session-model",
+    "adr-012-domain-extension-system",
+    "adr-016-component-lift-out-pattern",
+  ],
+
+  ontology_check = {
+    decision_string    = "registry credentials are managed via per-project sops multi-recipient vaults stored as OCI artifacts in ZOT; the daemon is structurally excluded from credential resolution; actor-role binding is declared in project.ncl; vault backend is restic or kopia behind a thin abstraction; access logs are co-located with the vault artifact",
+    invariants_at_risk = ["protocol-not-runtime", "no-enforcement"],
+    verdict            = 'Safe,
+  },
+}
diff --git a/adrs/adr-018-level-hierarchy-mode-resolution-strategy.ncl b/adrs/adr-018-level-hierarchy-mode-resolution-strategy.ncl
new file mode 100644
index 0000000..f36dc7b
--- /dev/null
+++ b/adrs/adr-018-level-hierarchy-mode-resolution-strategy.ncl
@@ -0,0 +1,148 @@
+let d = import "defaults.ncl" in
+
+d.make_adr {
+  id     = "adr-018",
+  title  = "Level Hierarchy and Mode Resolution Strategy — Observable Boundary Traversal",
+  status = 'Accepted,
+  date   = "2026-05-01",
+
+  context = "ADR-012 introduced a domain extension system that enables project-specific specialization of ontoref. In practice this created a three-level hierarchy: (1) ontoref base — generic protocol and reflection operations; (2) project domain — specialization of ontoref for a specific project type (provisioning, personal, ...); (3) domain instance — a concrete project derived from a domain (a workspace, an infra, a team ontoref). The hierarchy was not formalized: no level declares its identity, no mode declares whether its implementation is complete or delegates to the level above, and no mechanism exists to determine which level answered a given operation. The result is implicit traversal — a caller invoking 'build-docs' has no way to know whether the operation ran at level 1, 2, or 3, whether it merged contributions from multiple levels, or whether it silently fell through to the base because the domain had not implemented it yet. The discussion that produced this ADR also identified that resolution strategies cannot be uniform: the correct strategy depends on the mode, the project context, and the adoption phase. A domain in early adoption may Delegate all documentation modes to the base; the same domain at maturity Overrides them with project-specific implementations. The transition between strategies must itself be observable — crossing a level boundary is an architectural event, not an implementation detail.",
+
+  decision = "Formalize the three-level hierarchy with four mechanisms: (1) Level identity declaration — each ontoref instance declares level.index (1=base, 2=domain, 3=instance), level.name, and level.parent (name of the level above; absent at level 1) in manifest.ncl. This makes level identity explicit and queryable. (2) Per-mode resolution strategy — each reflection mode declares a strategy field using one of four values: 'Override (implementation is complete at this level, traversal stops), 'Delegate (no implementation here, traverse to parent), 'Merge (accumulate fields bottom-up across all levels; lower level wins on conflicts), 'Compose (declare explicit partial inheritance via an extends field naming specific steps or fields to inherit from the parent). Strategy is required at level 2+; absent at level 1 (base is always Override by definition). Implicit absence at level 2+ is treated as 'Delegate with a Soft validation warning. (3) FSM-bound strategy transitions — when a mode's strategy is expected to change as the project matures, declare a state dimension in state.ncl whose current_state tracks the strategy value. The transition from 'Delegate to 'Override (or any other pair) is then an observable FSM event: tracked in state.ncl, visible in ore describe state, and triggerable by the same transition conditions as any other dimension. Strategy changes are architectural events, not silent refactors. (4) Observable traversal via ore mode resolve — the command 'ore mode resolve <id>' reports which level answered the operation, the strategy applied, the source file, and the reason (declared strategy or FSM state). No mode resolution is ever silent.",
+
+  rationale = [
+    {
+      claim  = "Implicit traversal makes level boundaries invisible and coupling complaints undiagnosable",
+      detail = "The coupling discussions that preceded this ADR (provisioning workspace management appearing coupled to ontoref, credential management appearing to require ontoref) were symptoms of invisible level traversal. When it is impossible to determine whether an operation ran at level 1 or 2, it is also impossible to determine whether the coupling is load-bearing (level 2 deliberately uses level 1) or accidental (level 2 forgot to implement something and fell through). Explicit level declaration and strategy declaration make this distinction mechanically checkable.",
+    },
+    {
+      claim  = "Per-mode strategy is the correct granularity — global strategy per level does not hold",
+      detail = "A domain in active use may Override its core modes (workspace management, infra deploy) while still Delegating peripheral modes (documentation generation, ADR validation) to the base. A global strategy per level would force a false choice: either all modes are overridden (blocking base adoption) or all modes delegate (making the domain invisible). Per-mode strategy reflects the actual adoption curve: domains specialize incrementally, not all at once.",
+    },
+    {
+      claim  = "FSM-bound transitions make strategy changes architectural events",
+      detail = "Without FSM binding, a mode moving from 'Delegate to 'Override is a silent code change with no recorded rationale. With FSM binding, the transition requires updating current_state in state.ncl, which is a deliberate action visible to ore describe state, tracked in git history, and subject to the same transition conditions (catalyst, blocker) as any other dimension. The architectural significance of 'we now own this mode' is formally recorded.",
+    },
+    {
+      claim  = "'Compose is necessary for surgical specialization that preserves base infrastructure",
+      detail = "Merge and Override are extremes: Merge accumulates everything (field-level conflicts resolved by lower level winning), Override discards everything from above. Compose covers the real case: a domain wants to keep the base's generate-api step and the base's lint-docs step, but replace the publish step with a domain-specific one. Without Compose, the domain must either fully override (losing base infrastructure it wants) or Merge (risking unintended inheritance of base steps it does not want).",
+    },
+    {
+      claim  = "Level parent as name reference decouples identity from location",
+      detail = "parent = 'ontoref-base' in a domain's level declaration is a name reference, not a file path or OCI artifact coordinate. The resolution of that name to a concrete implementation is context-dependent: local checkout for development, OCI artifact for distributed domains. The ADR defines the declaration contract; resolution is implementation-dependent and may itself evolve (ADR-010 migration system handles protocol evolution).",
+    },
+    {
+      claim  = "ore mode resolve makes the system self-describing at the operational layer",
+      detail = "ontoref's core identity (ADR-001, core.ncl) is self-describing: it consumes its own protocol. ore mode resolve extends this to the operational layer: the resolution mechanism is itself queryable. An agent or developer can always answer 'which level is handling this mode right now and why' without reading source code. This is the concrete expression of the 'always know where you are and when you cross the frontier' requirement.",
+    },
+  ],
+
+  consequences = {
+    positive = [
+      "Level identity is always queryable — ore describe project shows level.index, level.name, level.parent",
+      "Mode resolution is always observable — ore mode resolve <id> shows level, strategy, source, reason",
+      "Strategy changes are FSM events — visible in git, queryable in state, subject to transition conditions",
+      "Delegate chains that break (no Override anywhere in the parent chain) are caught by ore validate modes",
+      "Phased domain adoption is formally supported — early phases Delegate, mature phases Override",
+      "Coupling complaints become diagnosable — is provisioning using ontoref base because it Delegates (load-bearing) or because it forgot to Override (accidental)?",
+      "Compose enables surgical specialization without full Override — base infrastructure is reusable explicitly",
+    ],
+    negative = [
+      "All existing modes at level 2+ require strategy field addition — migration 0017 needed",
+      "ore mode resolve is not yet implemented — constraint delegate-chain-complete cannot be checked until it is",
+      "ore validate modes is not yet implemented — constraints are declared but not executable at ADR acceptance",
+      "Compose requires extends field with precise step/field references — mismatches produce runtime errors, not schema errors",
+      "FSM dimensions for strategy transitions add state.ncl surface area — domains with many modes in transition accumulate many dimensions",
+    ],
+  },
+
+  alternatives_considered = [
+    {
+      option       = "Implicit inheritance — current state, no declaration required",
+      why_rejected = "Makes level traversal invisible. Coupling is undiagnosable. Adoption phase is unknown. Boundary crossing is unobservable. This is the state that produced the architectural confusion this ADR resolves.",
+    },
+    {
+      option       = "Global strategy per level — one strategy applies to all modes at a given level",
+      why_rejected = "Domains specialize incrementally. A global Override for level 2 blocks early adoption (all modes must be implemented before any work). A global Delegate for level 2 makes domains invisible (nothing is ever implemented). Per-mode strategy reflects the real adoption curve.",
+    },
+    {
+      option       = "Override-only — levels either implement fully or do not appear",
+      why_rejected = "Eliminates phased adoption. A domain cannot exist in the system until it has implemented every mode — a steep entry cost that contradicts ADR-001 (voluntary coherence, no enforcement). Delegate and Compose are specifically needed for incremental adoption.",
+    },
+    {
+      option       = "Separate level.ncl file instead of level field in manifest.ncl",
+      why_rejected = "manifest.ncl already holds structural self-description (requirements, capabilities, registry topology, layers). Level identity is structural metadata of the same kind. A separate file adds coordination overhead (two files to keep in sync, two files for describe to read) for no additional expressiveness.",
+    },
+  ],
+
+  constraints = [
+    {
+      id       = "level-declared-at-domain",
+      claim    = "Any ontoref instance at level 2 or 3 must declare level.index, level.name, and level.parent in manifest.ncl",
+      scope    = "all manifest.ncl files in domain and instance projects",
+      severity = 'Hard,
+      check    = {
+        tag  = 'NuCmd,
+        cmd  = "ore validate modes --check level-declared",
+      },
+      rationale = "Level identity is the precondition for all other mechanisms in this ADR. An undeclared level cannot participate in observable traversal or FSM-bound strategy transitions.",
+    },
+    {
+      id       = "strategy-declared-at-domain",
+      claim    = "Every reflection mode at level 2+ must declare strategy explicitly; implicit absence is a Soft violation",
+      scope    = "reflection/modes/*.ncl in domain and instance projects",
+      severity = 'Soft,
+      check    = {
+        tag  = 'NuCmd,
+        cmd  = "ore validate modes --check strategy-declared",
+      },
+      rationale = "Implicit Delegate (mode exists at level 2 without strategy field) is functionally equivalent to explicit Delegate but invisible. The Soft severity allows gradual adoption — existing modes that predate this ADR are not hard-blocked, but the warning drives migration.",
+    },
+    {
+      id       = "delegate-chain-complete",
+      claim    = "No Delegate chain may terminate without an Override at some level — every mode invocation must resolve to a concrete implementation",
+      scope    = "all level 2 and 3 modes with strategy = 'Delegate",
+      severity = 'Hard,
+      check    = {
+        tag  = 'NuCmd,
+        cmd  = "ore validate modes --check delegate-chain",
+      },
+      rationale = "A Delegate chain that ends without an Override means the operation has no implementation — invoking it produces a silent no-op or an opaque error. This is the failure mode that makes implicit traversal dangerous.",
+    },
+    {
+      id       = "strategy-state-coherent",
+      claim    = "If state.ncl declares a dimension for a mode's strategy, its current_state must match the strategy field declared in the mode NCL",
+      scope    = "state.ncl dimensions whose id matches the pattern '<mode-id>-strategy'",
+      severity = 'Soft,
+      check    = {
+        tag  = 'NuCmd,
+        cmd  = "ore validate modes --check strategy-state",
+      },
+      rationale = "Strategy drift between the FSM state and the mode declaration means the state dimension has become decorative — it records an intention that is no longer reflected in the actual implementation. Soft severity allows the transition to be in-flight (state says Override, code not yet complete).",
+    },
+    {
+      id       = "compose-extends-valid",
+      claim    = "A mode with strategy = 'Compose must declare an extends field; every step or field reference in extends must exist in the named parent mode",
+      scope    = "reflection/modes/*.ncl with strategy = 'Compose",
+      severity = 'Hard,
+      check    = {
+        tag  = 'NuCmd,
+        cmd  = "ore validate modes --check compose-extends",
+      },
+      rationale = "A Compose mode with a broken extends reference silently inherits nothing from the parent — equivalent to an unintentional Override. Since Compose is specifically chosen to preserve parent infrastructure, a broken reference is always a bug.",
+    },
+  ],
+
+  related_adrs = [
+    "adr-001-protocol-as-standalone-project",
+    "adr-010-protocol-migration-system",
+    "adr-011-mode-guards-and-convergence",
+    "adr-012-domain-extension-system",
+  ],
+
+  ontology_check = {
+    decision_string    = "three-level hierarchy (base, domain, instance) with per-mode resolution strategy (Override, Delegate, Merge, Compose); strategy transitions are FSM events in state.ncl; ore mode resolve makes traversal observable; ore validate modes enforces chain completeness",
+    invariants_at_risk = ["protocol-not-runtime", "no-enforcement"],
+    verdict            = 'Safe,
+  },
+}
diff --git a/adrs/adr-019-per-file-recipient-routing-tenant-isolation.ncl b/adrs/adr-019-per-file-recipient-routing-tenant-isolation.ncl
new file mode 100644
index 0000000..34cca21
--- /dev/null
+++ b/adrs/adr-019-per-file-recipient-routing-tenant-isolation.ncl
@@ -0,0 +1,151 @@
+let d = import "defaults.ncl" in
+d.make_adr {
+  id     = "adr-019",
+  title  = "Per-File Recipient Routing for Tenant Isolation in lieu of Multi-Vault",
+  status = 'Accepted,
+  date   = "2026-05-03",
+
+  context = "ADR-017 established per-project credential vaults as OCI artifacts in ZOT, encrypted with sops + age multi-recipient. The model held one recipient set per vault: every actor who could decrypt access.sops.yaml could decrypt every credential file inside the vault. Real projects (libre-wuji, the canonical multi-tenant example) need stronger separation — a single project hosts multiple clients with distinct services, AI agents that operate with restricted access, and developer/admin roles that occasionally overlap. Three concrete failure modes appear: (1) credential-of-clientB is visible to anyone who can open the vault even when only working on clientA — recipient lists are coarse and indiscriminate; (2) AI agents with read-only intent receive credentials whose blast radius exceeds their declared scope; (3) blast-radius limitation requires per-file recipient sets, which the original single-set model does not express. A naive answer is multi-vault — one vault_id per tenant or per environment — but it cascades through every layer (schema, helpers, recipes, dispatcher, migration) and creates an architectural debt without justifying a real isolation requirement (separate master keys, separate restic repos) for the typical case.",
+
+  decision = "Tenant isolation within a single project is expressed via sops creation_rules, declared in project.ncl::sops as recipient_groups (named lists of age public keys) and recipient_rules (path_regex to group-union mappings). The bootstrap recipe generates <vault_dir>/.sops.yaml from these declarations and sops natively encrypts each file with the union of declared groups. One vault_id per project remains the unit. Multi-vault is explicitly NOT implemented and remains out of scope until a project requires HARD isolation (separate master keys, separate restic repos, compliance-grade separation surviving accidental cross-decryption); such a case requires a future ADR. When recipient_rules are declared, project.ncl is the single source of truth: secrets-add-key and secrets-remove-key error out and direct the operator to edit project.ncl plus run secrets-rekey, which regenerates .sops.yaml and re-encrypts every *.sops.yaml file. Three adoption templates (single-team, multi-tenant, agent-first) ship in install/resources/templates/sops/ as copy-paste starting points; a project may adopt any pattern or none.",
+
+  rationale = [
+    {
+      claim  = "Use sops creation_rules natively — do not invent a parallel routing layer",
+      detail = "sops already provides per-file recipient routing via creation_rules in .sops.yaml. Using it directly inherits all of sops' tooling (updatekeys, --decrypt with .sops.yaml discovery, --filename-override) and avoids a competing convention that would double the surface and break composability with sops itself.",
+    },
+    {
+      claim  = "Preserve the single-vault structural invariants of ADR-017",
+      detail = "Single vault_id, single OCI artifact, single restic repo, single lock, single dispatcher subcommand surface. The schema delta is two optional fields (recipient_groups + recipient_rules). Most code paths require zero modification — additivity over existing helpers is the migration story for projects already on ADR-017.",
+    },
+    {
+      claim  = "Project.ncl is the single source of truth for recipient sets",
+      detail = "Direct sops mutations via secrets-add-key and secrets-remove-key are forbidden in declarative mode (recipes error explicitly). The reason: any direct mutation diverges from project.ncl, and the next secrets-rekey would silently revert. Forcing edits through git via project.ncl makes recipient changes auditable and reproducible across machines.",
+    },
+    {
+      claim  = "Honest trade-off: per-file routing protects against accidental cross-decryption, not against encrypted-byte visibility",
+      detail = "ClientA's lead with their .kage cannot decrypt clientB-*.sops.yaml files — sops rejects decryption when the recipient set excludes the actor. But all encrypted files are layers in the same OCI artifact; clientA SEES that clientB-* files exist. For HARD isolation (separate master keys, separate restic repos, compliance-grade separation surviving accidental cross-decryption), multi-vault remains the future option — a separate ADR captures that work when a real case requires it.",
+    },
+    {
+      claim  = "Three adoption templates anchor common patterns without forcing them",
+      detail = "single-team, multi-tenant, and agent-first templates ship in install/resources/templates/sops/ as copy-paste starting points. The schema fields are optional with sensible defaults — a project may adopt any pattern, mix them, or skip templates entirely while still being a valid consumer of ADR-017 + ADR-019.",
+    },
+  ],
+
+  consequences = {
+    positive = [
+      "Tenant isolation in a single vault, single master key, single OCI artifact — adoption cost minimal.",
+      "Compatible with all existing ADR-017 enforcement (assert-actor-authorized, assert-target-in-scope, vault lock, impact analysis).",
+      "Migration from legacy single-set mode is additive: existing projects keep working, new fields opt them into per-file routing.",
+      "Defense in depth: actor scope (ops + namespaces) + recipient routing — both must permit an operation.",
+    ],
+    negative = [
+      "Operators must understand sops creation_rules ordering (first match wins) — surfacing rule conflicts requires care.",
+      "secrets-add-key and secrets-remove-key behavior diverges between legacy and declarative modes, introducing a mode-aware UX.",
+      "Cross-tenant visibility of encrypted file paths in the OCI manifest — operationally clientB knows clientA exists in the vault.",
+    ],
+  },
+
+  alternatives_considered = [
+    {
+      option = "Multi-vault: project.ncl::sops.vault_id (single string) becomes sops.vaults (record of named SopsConfigs)",
+      why_rejected ="Cascades through every layer (schema, helpers, recipes, dispatcher), forcing a 12-component refactor and a migration for every existing project to express what one optional schema field accomplishes via sops creation_rules. The HARD isolation it provides (separate master keys, separate restic repos) is rarely required; per-file routing covers the common case while leaving the door open for a future multi-vault ADR if a project genuinely needs filesystem-level separation.",
+    },
+    {
+      option = "Single recipient set + role-based decryption gating in helper code",
+      why_rejected ="Would require a custom layer over sops and reinvent recipient routing. sops already does it natively via creation_rules. Inventing a parallel mechanism doubles the surface and breaks composition with sops tooling (sops --decrypt, updatekeys).",
+    },
+    {
+      option = "Externalize tenant credentials to an external secret manager (e.g. HashiCorp Vault)",
+      why_rejected ="Adds an external runtime dependency that contradicts the ADR-017 invariant of self-contained, distribution-via-OCI credentials. Reasonable for projects that already operate such a system, but inappropriate as the default ontoref pattern.",
+    },
+  ],
+
+  constraints = [
+    {
+      id       = "rules-imply-groups-defined",
+      claim    = "Every group referenced in recipient_rules must be declared in recipient_groups",
+      scope    = "all projects with sops.recipient_rules non-empty",
+      severity = 'Hard,
+      check    = {
+        tag = 'NuCmd,
+        cmd = "ore secrets audit --check recipient-routing-coherent",
+      },
+      rationale = "An undeclared group resolves to an empty recipient list, producing files encrypted to nobody. Catch at audit time before push.",
+    },
+    {
+      id       = "no-empty-group-on-active-rule",
+      claim    = "A rule whose group union resolves to zero recipients is rejected at bootstrap and rekey",
+      scope    = "all projects with sops.recipient_rules non-empty",
+      severity = 'Hard,
+      check    = {
+        tag = 'NuCmd,
+        cmd = "ore secrets audit --check recipient-routing-coherent",
+      },
+      rationale = "Encrypting to zero recipients silently produces an unrecoverable file. Reject at the source rather than waiting for sops to fail at use time.",
+    },
+    {
+      id       = "declarative-mode-locks-direct-mutations",
+      claim    = "secrets-add-key and secrets-remove-key recipes must error when project declares recipient_rules; canonical workflow is edit project.ncl + secrets-rekey",
+      scope    = "all projects with sops.recipient_rules non-empty",
+      severity = 'Hard,
+      check    = {
+        tag         = 'Grep,
+        paths       = ["justfiles/secrets.just"],
+        pattern     = "HAS_RULES.*declarative",
+        must_be_empty = false,
+      },
+      rationale = "Direct sops mutations would diverge from project.ncl, and the next rekey would silently revert them. Forcing the rekey path keeps git as the single source of truth.",
+    },
+    {
+      id       = "every-vault-file-matches-a-rule",
+      claim    = "Every *.sops.yaml under <vault_dir>/ must match at least one declared rule when recipient_rules is non-empty",
+      scope    = "all projects with sops.recipient_rules non-empty",
+      severity = 'Hard,
+      check    = {
+        tag = 'NuCmd,
+        cmd = "ore secrets audit --check recipient-routing-coverage",
+      },
+      rationale = "sops fails encryption with 'no matching creation rules found' for unmatched paths. Catch the mismatch at audit time, surface which path needs a rule (or which rule needs broadening).",
+    },
+    {
+      id       = "multi-vault-not-implemented",
+      claim    = "Projects must not declare a multi-vault structure (e.g. sops.vaults map). Multi-vault adoption requires a new ADR superseding or extending this one",
+      scope    = "all projects with .ontoref/project.ncl",
+      severity = 'Hard,
+      check    = {
+        tag           = 'Grep,
+        pattern       = "sops.vaults *=",
+        paths         = [".ontoref/project.ncl"],
+        must_be_empty = true,
+      },
+      rationale = "Premature multi-vault implementation without a justifying use case generates schema and helper debt across 12+ components. The constraint is structural: any project that hits a real HARD-isolation requirement must capture it in a new ADR before adopting multi-vault.",
+    },
+    {
+      id       = "templates-discoverable",
+      claim    = "Three adoption templates (single-team, multi-tenant, agent-first) live under install/resources/templates/sops/ and are referenced by qa.ncl::credential-vault-templates",
+      scope    = "ontoref protocol layer",
+      severity = 'Soft,
+      check    = {
+        tag   = 'FileExists,
+        paths = [
+          "install/resources/templates/sops/single-team/project.ncl.snippet",
+          "install/resources/templates/sops/multi-tenant/project.ncl.snippet",
+          "install/resources/templates/sops/agent-first/project.ncl.snippet",
+        ],
+      },
+      rationale = "Without templates, adoption requires reading the schema, sops docs, and the FAQ — too high a friction. Templates are not mandatory but their absence is a soft signal of protocol decay.",
+    },
+  ],
+
+  related_adrs = [
+    "adr-017-registry-credential-vault-model",
+    "adr-015-mcp-tool-inventory-auto-derive",
+  ],
+
+  ontology_check = {
+    decision_string    = "tenant isolation within a single credential vault is expressed via sops creation_rules driven by project.ncl::sops.recipient_groups + recipient_rules; multi-vault is explicitly out of scope; project.ncl is the single source of truth for recipient sets and direct sops mutations are forbidden in declarative mode",
+    invariants_at_risk = ["protocol-not-runtime"],
+    verdict            = 'Safe,
+  },
+}
diff --git a/adrs/adr-020-three-layer-ontoref-instance-model.ncl b/adrs/adr-020-three-layer-ontoref-instance-model.ncl
new file mode 100644
index 0000000..799c19f
--- /dev/null
+++ b/adrs/adr-020-three-layer-ontoref-instance-model.ncl
@@ -0,0 +1,221 @@
+let d = import "defaults.ncl" in
+
+d.make_adr {
+  id     = "adr-020",
+  title  = "Three-Layer Model for Project Ontoref Instances",
+  status = 'Proposed,
+  date   = "2026-05-03",
+
+  context = m%"
+Multiple projects now adopt the ontoref protocol (lian-build is the explicit
+external case under bl-002 / bl-008). Field experience across ontoref +
+lian-build + provisioning shows that an ontoref-onboarded project's
+repository carries TWO distinct ontoref-shaped layers, while a third layer
+exists outside the project (in caller repositories). Without codification,
+adopters re-derive the model per project, mix layers accidentally, and lose
+boundaries silently:
+
+  - Layer 3 content (caller-side cabling) drifts into a project's qa.ncl,
+    making the FAQ a how-to-deploy guide for one specific caller.
+  - Layer 2 schemas live under reflection/ as 'project notes', drifting
+    from the binary because they're not on the contract path.
+  - Layer 1 architectural rationale ends up in catalog/domains/<id>/
+    contract.ncl, where consumers expecting a typed shape get prose.
+
+The model has been observed (lian-build/reflection/qa.ncl::lian-build-what-
+and-why), described (ontoref/reflection/qa.ncl::ontoref-three-layer-model),
+and queued for codification (bl-009). This ADR commits the model with
+machine-checkable constraints. Acceptance is gated on the constraints
+running clean across the existing ontoref-onboarded projects (ontoref
+itself, lian-build, provisioning).
+"%,
+
+  decision = m%"
+A project's ontoref instance has THREE distinct layers — but only the first
+two live in the project's own repository:
+
+  LAYER 1 — Self-management ontoref (about the project itself)
+    paths     .ontology/  reflection/  adrs/
+    audience  this project's developers and maintainers
+    purpose   describe the project to itself — axioms, FSM dimensions,
+              binding decisions, open questions, accepted knowledge
+    presence  MANDATORY on every ontoref-onboarded project
+
+  LAYER 2 — Specialized domain/mode ontoref (the integration surface)
+    paths     schemas/  catalog/{domains,modes}/
+              manifest.ncl::registry_provides
+    audience  OTHER projects that want to integrate this project
+    purpose   the contract surface other projects bind to — typed domain
+              artifacts, orchestration mode artifacts, registry-namespace
+              claim
+    presence  OPTIONAL but BICONDITIONAL — a project either has all of
+              {schemas/, catalog/, registry_provides} or none. Half-Layer-2
+              is a contract violation.
+
+  LAYER 3 — Caller-side implementations (NOT in this project)
+    paths     <caller>/extensions/<this-project>/
+              <caller>/catalog/components/<...>/  (when consuming)
+              <workspace>/infra/<ws>/integrations/
+    audience  operators and CI of caller projects
+    presence  PER CALLER, NEVER in this project's repo. Cross-references
+              from this project to Layer 3 are explicit pointers, never
+              copy-paste.
+
+The three-layer axis is ORTHOGONAL to ADR-018's level hierarchy
+(Base/Domain/Instance). A project at any level may have any combination of
+Layer 1 (always) and Layer 2 (sometimes). Layer 3 is the boundary outward,
+not a property of the project. The 3-layer × 3-level matrix is navigable in
+both axes: 'where in the protocol hierarchy' (level) is independent of
+'where in the project's repo' (layer).
+
+Cross-layer references inside a project carry an explicit layer-N tag on
+qa entries. A qa entry whose primary topic is a Layer N concern wears
+the layer-N tag; satellite entries (operational how-to, troubleshooting)
+inherit the layer of their anchor without re-tagging.
+"%,
+
+  rationale = [
+    {
+      claim  = "Audience separation requires lifecycle separation",
+      detail = "Layer 1 evolves with the project's decisions (ADR cadence). Layer 2 evolves with the binary (schemas in lock-step with code that produces and consumes them). Layer 3 evolves with caller infrastructure (workspace cabling, deployment changes). Three lifecycles in one namespace produces drift in two of them — observed in pre-codification cases where schemas drifted because they were treated as project notes.",
+    },
+    {
+      claim  = "Layer 2 biconditional is what makes a federated peer detectable",
+      detail = "If catalog/ exists without manifest.ncl::registry_provides, the project has artifacts but no namespace claim — consumers can't resolve where to pull from. If registry_provides exists without catalog/, the project claims a namespace it doesn't fill. Either half-state breaks integration. Treating the pair as biconditional makes 'is this a federated peer?' a single boolean check.",
+    },
+    {
+      claim  = "Layer 3 isolation is project-specific in spelling, universal in principle",
+      detail = "What counts as caller-specific differs per project (lian-build forbids `provisioning_workspace|vapora_|woodpecker_` per its adr-001; another project would forbid different patterns). The protocol-level constraint cannot enumerate; instead, each project carries its own constraint-bearing ADR and the protocol verifies presence of SUCH a constraint indirectly via cross-layer-tag-discipline. Project-specific spelling, ecosystem-wide presence.",
+    },
+    {
+      claim  = "Cross-layer tag discipline makes filtering scalable",
+      detail = "An ontoref instance accumulates dozens to hundreds of qa entries over its life. 'Filter the qa for what other projects can integrate' (Layer 2) becomes a tag query rather than a full-text search. Tagging only anchors keeps the result set small and authoritative; satellites don't dilute the filter.",
+    },
+    {
+      claim  = "Orthogonality with ADR-018 lets the model expand without conflict",
+      detail = "ADR-018's level hierarchy describes WHERE a project sits in the protocol specialization (Base = ontoref itself; Domain = a project-type domain; Instance = a concrete workspace). The three-layer model describes WHERE inside a project's repo content sits (self / integration-surface / caller-side). Different axes; no terminology collision (level vs layer); a 3x3 matrix populated empirically. Treating them as orthogonal avoids re-litigating ADR-018 every time the layer model evolves.",
+    },
+  ],
+
+  consequences = {
+    positive = [
+      "ore describe project can statically answer 'is this a federated peer?' (Layer 2 biconditional) and 'is the project minimally onboarded?' (Layer 1 mandatory)",
+      "Adopters discover the model from documentation (qa::ontoref-three-layer-model) plus this ADR, not by osmosis from existing projects",
+      "qa-tag-based filtering (layer-1, layer-2) becomes a navigable surface as ecosystem corpus grows",
+      "The 3-layer × 3-level matrix is explicit, removing a recurring source of design ambiguity (e.g. 'should this go in .ontology or in catalog?')",
+      "ontoref setup gains a clear acceptance criterion: it produces a Layer 1-compliant scaffold by construction",
+    ],
+    negative = [
+      "Existing projects must audit their structure for Layer 1 compliance — typically reflection/qa.ncl and reflection/backlog.ncl are missing on early adopters",
+      "Six new constraints add per-project CI overhead (4x FileExists, 2x NuCmd) — running cheap but additive across the ecosystem",
+      "Terminology discipline required: 'layer' (this ADR) vs 'level' (ADR-018) are different concepts and prose must not collapse them",
+      "Layer 2 biconditional rejects half-states that some projects might want as an interim — projects mid-Layer-2-rollout must either complete or not start until ready",
+    ],
+  },
+
+  alternatives_considered = [
+    {
+      option       = "Single 'ontoref content' namespace, no layering",
+      why_rejected = "Observed drift outcomes (Layer 3 in qa, Layer 2 schemas treated as notes, etc.) are caused precisely by absence of layering. The single-namespace alternative is what we're correcting; rejecting it is the entire decision.",
+    },
+    {
+      option       = "Two-layer model collapsing self-management with integration-surface",
+      why_rejected = "lian-build's adoption explicitly experienced the schema-vs-rationale split: schemas/build_directives.ncl (Layer 2 contract) and adrs/adr-001-lian-build-as-standalone.ncl (Layer 1 rationale) have different audiences, different change cadences, different validation rules. Collapsing them produced the FAQ-vs-contract confusion that retiring lian-build/.ontology/FAQ.md addressed. The two-layer alternative re-creates that confusion.",
+    },
+    {
+      option       = "Make Layer 3 (caller-side) optionally co-resident with Layer 2 in the producer's repo",
+      why_rejected = "Creates ambiguity about who owns the cabling. If a producer's repo carries `extensions/<self>/`, who maintains it when a caller's workspace evolves? The producer doesn't know about the caller's infrastructure; the caller can't depend on the producer to update the cabling on its schedule. Co-residence inverts the integration arrow.",
+    },
+    {
+      option       = "Numbered layers (Layer 1 / 2 / 3) vs named layers ('self' / 'integration-surface' / 'caller-side')",
+      why_rejected = "Named layers are more descriptive but verbose; numbered layers are more precise but flat. The compromise (this ADR): use 'Layer N — descriptive name' in prose, 'layer-N' in tags. Tag economy wins; prose retains the descriptive name.",
+    },
+  ],
+
+  constraints = [
+    {
+      id       = "layer-1-self-ontology-core",
+      claim    = "Every ontoref-onboarded project has .ontology/core.ncl",
+      scope    = ".ontology/",
+      severity = 'Hard,
+      check    = {
+        tag     = 'FileExists,
+        path    = ".ontology/core.ncl",
+        present = true,
+      },
+      rationale = "Layer 1's irreducible minimum: a project with no axioms, tensions, or practices has no ontoref instance. Overlaps with adr-016's lift-out check by intention — same observable, two architectural decisions resting on it.",
+    },
+    {
+      id       = "layer-1-reflection-qa",
+      claim    = "Every ontoref-onboarded project has reflection/qa.ncl",
+      scope    = "reflection/",
+      severity = 'Hard,
+      check    = {
+        tag     = 'FileExists,
+        path    = "reflection/qa.ncl",
+        present = true,
+      },
+      rationale = "Layer 1's accepted-knowledge surface. A project without qa.ncl cannot accumulate verified Q&A — adopters re-derive answers each time. ontoref setup must scaffold this; existing projects backfill.",
+    },
+    {
+      id       = "layer-1-reflection-backlog",
+      claim    = "Every ontoref-onboarded project has reflection/backlog.ncl",
+      scope    = "reflection/",
+      severity = 'Hard,
+      check    = {
+        tag     = 'FileExists,
+        path    = "reflection/backlog.ncl",
+        present = true,
+      },
+      rationale = "Layer 1's open-question surface. Projects without backlog.ncl bury unresolved alternatives in chat threads, defeating the protocol's routing mechanism (graduates_to). Required so every undecided question has a place to land.",
+    },
+    {
+      id       = "layer-1-adrs-directory",
+      claim    = "Every ontoref-onboarded project has an adrs/ directory",
+      scope    = "adrs/",
+      severity = 'Hard,
+      check    = {
+        tag     = 'FileExists,
+        path    = "adrs/",
+        present = true,
+      },
+      rationale = "Layer 1's binding-decisions slot. The directory may be empty for a brand-new project, but its presence is what makes 'where do ADRs go?' a settled question.",
+    },
+    {
+      id       = "layer-2-biconditional",
+      claim    = "A federated-peer catalog (catalog/domains/ or catalog/modes/) exists if and only if manifest.ncl declares registry_provides",
+      scope    = "catalog/domains/, catalog/modes/, manifest.ncl, .ontology/manifest.ncl",
+      severity = 'Hard,
+      check    = {
+        tag         = 'NuCmd,
+        cmd         = "let fed = (('catalog/domains' | path exists) or ('catalog/modes' | path exists)); let m_root = (try { open manifest.ncl | str contains 'registry_provides' } catch { false }); let m_onto = (try { open .ontology/manifest.ncl | str contains 'registry_provides' } catch { false }); let prv = ($m_root or $m_onto); if $fed == $prv { 0 } else { error make {msg: $'Layer 2 biconditional violated: federated_catalog=($fed) registry_provides=($prv) — half-states advertise integration the consumer cannot complete'} }",
+        expect_exit = 0,
+      },
+      rationale = "Federated-peer status is detectable by a single biconditional. The marker is catalog/domains/ or catalog/modes/ specifically — not catalog/ as a whole, because some projects (provisioning) carry catalog/components/, catalog/providers/, etc. that serve workspace composition rather than federated artifact publication. The manifest may live at root or under .ontology/ depending on project convention; bl-010 (forthcoming) tracks canonicalization of that path. Half-states (one side without the other) are the violation.",
+    },
+    {
+      id       = "cross-layer-tag-discipline",
+      claim    = "qa entries whose primary topic is a Layer N concern carry the layer-N tag (anchors only; satellites inherit by association)",
+      scope    = "reflection/qa.ncl",
+      severity = 'Soft,
+      check    = {
+        tag         = 'NuCmd,
+        cmd         = "let entries = (nickel export reflection/qa.ncl --format json | from json | get entries); let suspicious = ($entries | where {|e| (($e.answer | str contains 'LAYER 1') or ($e.answer | str contains 'LAYER 2')) and ($e.tags | where {|t| ($t | str starts-with 'layer-')} | is-empty) and ($e.id | str contains 'what-and-why')}); if ($suspicious | is-empty) { 0 } else { error make {msg: $'Anchor entries discussing layers must carry layer-N tag: ($suspicious | get id | str join \", \")'} }",
+        expect_exit = 0,
+      },
+      rationale = "Tag-based filtering scales as the corpus grows. Without discipline, querying for 'Layer 2 concerns' returns either too few entries (false negatives because anchors aren't tagged) or too many (false positives because every entry that mentions a layer gets tagged). Anchors-only is the sustainable middle path; this constraint enforces it for the most common drift case.",
+    },
+  ],
+
+  related_adrs = [
+    "adr-001-protocol-as-standalone-project",
+    "adr-016-component-lift-out-pattern",
+    "adr-018-level-hierarchy-mode-resolution-strategy",
+  ],
+
+  ontology_check = {
+    decision_string    = "An ontoref-onboarded project's repository carries Layer 1 (mandatory self-management) and optionally Layer 2 (integration surface, biconditional with registry_provides). Layer 3 lives in caller projects, not in the producer. The three-layer axis is orthogonal to ADR-018's level hierarchy.",
+    invariants_at_risk = [],
+    verdict            = 'Safe,
+  },
+}
diff --git a/assets/presentation/docs/2-week-sprint-knowledge_done.md b/assets/presentation/docs/2-week-sprint-knowledge_done.md
new file mode 100644
index 0000000..a587632
--- /dev/null
+++ b/assets/presentation/docs/2-week-sprint-knowledge_done.md
@@ -0,0 +1,184 @@
+# A Two-Week Sprint for Knowledge
+
+## From Agile to Design Thinking
+
+ ---
+  In 1986, Hirotaka Takeuchi and Ikujiro Nonaka published "The New New Product Development Game" in the Harvard Business Review, describing how Honda, Canon, and Fuji-Xerox organized teams
+  to build copiers and automobiles. Jeff Sutherland read that article, combined it with lean manufacturing principles and object oriented programming, and in 1993 adapted it for software
+  development at Easel Corporation. Ken Schwaber formalized the method at the OOPSLA conference in 1995. By 2001, seventeen software developers gathered at a ski lodge in Snowbird, Utah,
+  and signed the Agile Manifesto, declaring that they were "uncovering better ways of developing software." Not knowledge or meaning. A manufacturing framework for software.
+
+  The Agile Manifesto's own language is explicit about its scope, yet the industry has applied Agile methodology to virtually everything in a technological environment — including knowledge
+   work, semantic systems and now artificial intelligence. And within the same breath, organizations and experts regale the industry with talk of innovation and the emergence of a new
+  paradigm in ways of working. Yet those ways of working are locked inside of Agile methodologies and Scrum, designed for the factory floor and the codebase — not for the messy, recursive,
+  deeply human work of making meaning out of complexity.
+
+  Scrum Is a Manufacturing Schedule
+
+  Let's return to the Takeuchi and Nonaka paper, where the title—"The New New Product Development Game" describes how product teams at Japanese firms developed physical goods such as the
+  FX-3500 copier and the Honda City car. They draw on the rugby metaphor because the team "tries to go the distance as a unit, passing the ball back and forth." The unit of delivery is a
+  working prototype of a material artifact. The feedback loop is a machine that works or fails.
+
+  Sutherland's 2014 book, Scrum: The Art of Doing Twice the Work in Half the Time, reinforces this with military and manufacturing case studies such as the F-86 fighter production and
+  Toyota. Sutherland's book rebuilds the argument with the same assumption—that the thing being built has discrete, inspectable, testable parts that, when finished, are finished.
+
+  The 2020 Scrum Guide defines the Increment as "a concrete stepping stone toward the Product Goal," one that is "additive to all prior Increments and thoroughly verified, ensuring that all
+   Increments work together." These words are sensible if you are building a mobile application, with a login screen and a settings page. But for knowledge systems, these words are
+  destructive, because the components of a semantic architecture are not additive in the way Scrum dictates. A controlled vocabulary is not a subcomponent of a taxonomy the way a login
+  screen is a subcomponent of an app. A taxonomy without disciplined, vocabulary control as its underpinnings does not function well, if at all. You cannot ship half a taxonomy and call it
+  an Increment.
+
+  You see, Scrum assumes tame problems that can be time-boxed, estimated with planning poker, decomposed into user stories, tracked on a velocity chart and declared done against a
+  checklist. If only knowledge were this tame. Knowledge organization is anything but tame. More than fifty-years-old, in 1973 Horst Rittel and Melvin Webber warned that applying the
+  engineering paradigm uniformly to all problems was "bound to fail." And yet here we are, applying a manufacturing schedule and methodologies to knowledge, expecting that knowledge will
+  fit into the confines of old ways of working — an artifact of the Third Industrial Revolution imposed on the demands of the Fourth.
+
+  In 2016, Klaus Schwab wrote about this historical pattern. The First Industrial Revolution, beginning around 1760, gave us steam and mechanical production. The Second, in the late
+  nineteenth century, gave us electricity and the assembly line — the world that produced Toyota, lean manufacturing, and the production logic that Scrum still carries as its core ethos.
+  The Third, beginning in the 1960s, gave us semiconductors, mainframe computing, and eventually the internet — the world in which Agile was born, by software developers, for software
+  development. The Fourth Industrial Revolution, Schwab argued, is fundamentally different, in kind, as it is defined by the fusion of artificial intelligence, cyber-physical systems and
+  knowledge technologies that blur the boundaries between the physical, digital and biological worlds.
+
+  The problems of this era are not necessarily engineering problems with stable requirements. We are wrestling with problems such as how to ground machine intelligence in meaning, how to
+  organize evolving knowledge and how to build systems that can explain themselves. These are wicked problems that demand new epistemologies and frameworks for problem solving. Agile may
+  have been the right methodology for the Third Industrial Revolution but we are no longer in the Third Industrial Revolution. The frameworks have not caught up, and the cost of that lag is
+   exacerbated every time an organization forces Third Industrial Revolution principles applied to Fourth Industrial Revolution work.
+
+  In other words, this is not a product problem or a platform problem—it is learning to embrace new ways of working. This is an infrastructure problem that is governed by epistemology—a
+  fundamental philosophy pertaining to knowledge.
+
+  Why Knowledge Cannot Be Sprinted
+
+  An ontology is, per Thomas Gruber's foundational definition, "a specification of a conceptualization." A conceptualization is a whole, not parts of a whole. Specifying half of it produces
+   logical incoherence, not a minimally viable product. For example, ANSI/NISO Z39.19, the standard for constructing controlled vocabularies, requires resolving synonymy, homonymy and scope
+   before terms enter production, because any inconsistency propagates downstream into every subsequent use. SKOS, the W3C standard for encoding knowledge organization systems, requires
+  that hierarchical relationships form coherent structures with transitive closure and OWL 2 requires that class axioms be satisfiable. These are preconditions, necessary for logical
+  systems to reason and express meaning.
+
+  In practice, I have watched organizations attempt Agile delivery of knowledge systems in two patterns and neither ends well. In the first, the team holds stand-ups, ceremonies and
+  retrospectives. The taxonomist brings tickets and adds fifteen terms to the product vocabulary. Tickets close and a velocity chart goes up. Nothing is integrated, because integration is
+  the hard part and integration does not fit into two weeks so that goes to the backlog. Six months later, the vocabulary is a collection of inconsistently scoped terms, the ontology has
+  orphan classes and the knowledge graph — if it exists — cannot reliably answer queries. But the burndown chart is immaculate and leadership is blindly satisfied because boxes were
+  checked.
+
+  In the second pattern, the team borrows from other methodologies while still calling it Agile. The taxonomist conducts multi-week domain analysis and stakeholder interviews — that is
+  ethnography. She does card sorting and concept modeling — that is design research. She iterates with subject matter experts — that is participatory design. She ships, eventually, a
+  coherent model. The sprint cadence was irrelevant to the work. Martin Fowler named this phenomenon in his 2018 keynote: "faux-agile" — the Agile Industrial Complex.
+
+  Ron Jeffries, another original signatory of the Agile Manifesto, told developers the same year, 2018, to abandon the word entirely. Dave Thomas declared in 2014 that "Agile" had become "a
+   magnet for anyone with points to espouse, hours to bill or products to sell." When three of the seventeen authors of a framework say it has been hollowed out, one has to wonder if Agile
+  is appropriate for all technology work and if it is time to revisit ways of working.
+
+  Innovation Does Not Fit in a Backlog
+
+  The deeper issue is that Agile and Scrum were never designed to inspire or foster innovation. They were designed to deliver known requirements with increasing efficiency. The Scrum
+  Guide's entire apparatus — product backlog, sprint planning, Definition of Done — presupposes that the problem has been formulated by a product owner who can write it in a user story. But
+   knowledge work, like design, is wicked. Rittel and Webber's first property of a wicked problem is that "There is no definitive formulation." The second property states that there is no
+  stopping rule. And the fifth property illuminates that every solution is a "one-shot operation" because there is no opportunity to learn by trial-and-error without consequence.
+
+  Given the nature of wicked problems, it would seem that Agile is ill fitted for wicked problems. And for that matter, ill fitted for the dynamic nature of knowledge, where knowledge
+  infrastructures are never one-shot operations, necessitate stopping rules and without a doubt defies definitive formulations.
+
+  Therefore, knowledge organization in an enterprise domain is wicked in exactly this sense. The problem of how to represent "customer" in a telecommunications company's semantic model has
+  no pre-existing correct answer. Its formulation depends on who is doing what with the representation, and that changes as you build. The scope of the vocabulary shifts as you interview
+  stakeholders. The taxonomy reorganizes as you encounter edge cases. The ontology's axioms tighten or loosen as the reasoner surfaces contradictions. Don't be mistaken—this does not bend
+  to the Agile principle of welcoming changing requirements, as the requirements are constituted through the work. The problem is under construction, often evolving, during the solving.
+
+  Organizations that manage all work through Scrum are, in effect, optimizing for predictability in domains that demand exploration. The results resist innovation and take on the appearance
+   of productivity. Velocity charts rise while the underlying intellectual architecture is flattened into bit sized deployments, uncoupled from the continuum that is knowledge.
+
+  Design Thinking Belongs in the Swamp
+
+  Donald Schön drew the distinction that "In the varied topography of professional practice, there is a high, hard ground where practitioners can make effective use of research-based theory
+   and technique, and there is a swampy lowland where situations are confusing 'messes' incapable of technical solution." Agile lives on the high ground while knowledge work lives in the
+  swamp. Schön's term for how competent practitioners navigate the swamp was "reflection-in-action" — the iterative, tacit recalibration that happens while the work is happening.
+
+  Design thinking, in the lineage of Herbert Simon, Richard Buchanan and Schön, is the methodology of the swamp. It assumes the problem is under construction. Its five phases — empathize,
+  define, ideate, prototype, test — are not a linear pipeline but a recursive, iterative process that expects the definition of the problem to change as understanding deepens.
+
+  Research in the Handbook of Human-Centered Artificial Intelligence confirms this alignment—integrating design thinking into AI development makes systems more user-focused, iterative and
+  impactful, emphasizing data governance, transparency, explainability and bias mitigation. Carnegie Mellon University's Tepper School of Business now teaches a dedicated program on design
+  thinking with AI, arguing that the combination addresses "complex, systemic problems" that linear frameworks cannot manage successfully. IBM's enterprise design thinking practice,
+  developed over two decades, has been applied across client engagements precisely it keeps human experience at the center, while technology evolves underneath — design thinking is
+  "timeless," while technology is "timestamped."
+
+  A 2026 study in Scientific Reports examined the relationship between higher order thinking, generative AI chatbot use and engineering creativity, finding that higher order thinking had
+  significantly greater importance for creative outcomes than AI tool use alone — reinforcing that cognitive depth, not tool velocity, drives innovation. Research in the Proceedings of the
+  Design Society further argues that generative AI is shifting the designer's role from executor to paradigm-setter, making design thinking more essential as AI handles production tasks
+  while humans must guide framing, ethics and meaning.
+
+  To be clear, these frameworks are not different flavors of the same methodology. Scrum and design thinking represent totally different epistemologies. Scrum assumes the problem was handed
+   to you while design thinking assumes the problem must be found and wrestled with.
+
+  Is AI a Knowledge Tool?
+
+  This brings us to the question the industry avoids and I, for one, have been itching to ask. Is AI a knowledge tool? How do you classify it?
+
+  A large language model generates plausible sequences of text. Returning to the Bender, Gebru, McMillan-Major, and Mitchell paper mentioned earlier, the authors defined a language model as
+   a system for stitching together linguistic forms according to probabilistic patterns, "without any reference to meaning." Without reference to meaning, as LLMs out-of-the-box are absent
+  of external grounding. They optimize for plausibility, not truth and therefore as a standalone, can be best categorized as an information-processing tool.
+
+  It becomes a knowledge tool only when grounded in a semantic layer — controlled vocabularies, taxonomies, ontologies, knowledge graphs — that provides reference and provenance. The
+  industry's own corrective measures and associated marketplace confirms this as fact. Retrieval-augmented generation, knowledge graphs for LLMs and what recent discourse calls "context
+  engineering" are, underneath the packaging, the semantic infrastructure that should have been built first. But it was bolted on after the fact to prevent hallucination that was pretty
+  much guaranteed, the moment the model was deployed without one.
+
+  The RAG paper from 2020 called out this problem, identifying that parametric memory is opaque and cannot be updated and that retrieval from structured sources is how provenance is
+  captured and encoded. But it seems the industry has a memory problem or it was too early for most to hear the message.
+
+  Ironically, provenance is a core element of knowledge management and ontology work, as provenance is a principled tenant of the discipline and problem space. Without provenance, knowledge
+   cannot exist. Organizations are only now discovering they need it, because their generative systems cannot explain themselves. Call it lineage, call it traces. But if you are building
+  for knowledge, call it provenance, as provenance includes traces, lineage, citations, the temporal and the spatial.
+
+  So if generative AI is an information processing tool, I posture that for AI to become a knowledge tool, there must be provenance and descriptive logic. And to be frank, nobody sprints
+  their way into provenance as it is architected and sustained.
+
+  Knowledge Is a Spiral, Not a Sprint
+
+  The irony does not escape the moment. The same Nonaka who co-authored the 1986 paper that inspired Scrum, went on to develop the SECI model of knowledge creation in 1995, describing
+  knowledge as a spiral between tacit and explicit forms, moving through socialization, externalization, combination and internalization.
+
+  The spiral does not have an end, as is true with spirals, emblematic of life. Each rotation creates new knowledge that becomes the substrate for the next rotation. There is no Definition
+  of Done because the definition of what is being done is itself being constructed. Scrum took the team structure from Nonaka's first paper and discarded his next decade of work because it
+  did not fit neatly with Scrum. The mechanisms of knowledge work, as Nonaka spent a decade arguing, are metaphor, tacit skill, apprenticeship, shared space and the patient conversion of
+  hunches into models and models back into practice.
+
+  The Ontology Pipeline™ mirrors this spiral. You start with a controlled vocabulary — a disciplined process involving agreements, as to what terms mean. You mature it into a taxonomy by
+  introducing hierarchy, structuring terms into parent-child relations. Thereafter, the taxonomy is matured into a thesaurus by adding associative relationships and synonymy, encoded in
+  SKOS. The thesaurus is baseline work from which an ontology is developed, adding formal semantics using OWL and RDF. You architect a knowledge graph, assembling components to include the
+  knowledge assets developed through the Ontology Pipeline™ iterative framework.
+
+  Each stage prepares the ground for the next and stages cannot be skipped or bypassed. You cannot ship the knowledge graph without an ontology, and the ontology must be logically
+  consistent and coherent to make sense of data. A knowledge graph without a vocabulary control risks becoming a confident liar or failing to work with natural language. Because knowledge
+  assets are dimensional interconnected, knowledge is indeed a spiral, unable to be contained by manufacturing logic.
+
+  The Pattern and the Cure
+
+  Organizations budget for platforms — Confluence, SharePoint, LLM subscriptions, vector databases, a RAG vendor. But they do not budget for descriptive logic and knowledge, perhaps because
+   the problem is not solved by a single tool and is seemingly too amorphous in an Agile, data centric world. The controlled vocabulary is cut unless it delivers value for the database.
+  Often, the taxonomy is cut because it does not demo well and its value sits in the plumbing, delivering its punch over time. The result is an ever-growing pile of documents with a chatbot
+   stapled to the front, confidently answering questions in the voice of a system that has no reference for any of the terms sprinkled throughout the documents.
+
+  Then the system hallucinates. It contradicts itself across sessions. It surfaces outdated policy as current and surfaces private outlook messages from deactivated employees. I have seen
+  this IRL more times than I can count. And the optics are worse when a vendor tool is responsible for revealing private chats and email transactions. Usually the executive sponsor asks why
+   the investment has not worked, ignoring the evidence of faux pas such as the surfacing of private communications. All in a days work, when burndown charts and velocity set the precedence
+   for ways of working and reward systems.
+
+  Executives ignore the fact that the system was never grounded. It was never grounded because grounding requires the slow, expensive, disciplined work of knowledge organization — work that
+   does not fit into a sprint, does not demo well and fails to produce a velocity chart. The organization chooses to bypass investment in the descriptive knowledge apparatus because the
+  lines are too blurry to clearly define according to Agile speak. Because knowledge is infrastructure, not a product or a platform. Knowledge is the gas in the tank, the fuel for meaning,
+  the material that justifies logical reasoning.
+
+  Design thinking gives that knowledge work a home. It is comfortable with ambiguity. It treats problem discovery as an equal to problem solving. It certainly does not demand a shippable
+  increment every fourteen days. Instead, design thinking demands that you understand what you are building before you build it — and that you keep revising that understanding as the work
+  proceeds. Library and information science, with over a century of standards for knowledge organization, provides the methods while design thinking provides the epistemology. Together,
+  they offer what Agile never could—a framework suited to the nature of knowledge itself.
+
+  A two-week sprint to deliver what? If you are building knowledge systems, two weeks will get you nothing coherent. Knowledge is not built fast and broken into shape. It is built
+  carefully, with provenance, governance and patience. Knowledge is not a product to be manufactured and no—you cannot buy off-the-shelf knowledge that perfectly matches a domain or
+  organization. Knowledge is nuanced, specific to each organization, domain and person.
+
+  In the spirit of VC-speak, knowledge is the moat, and that moat is not Agile.
+
+  ---
+  By Jessica Talisman — Intentional Arrangement (Substack)
diff --git a/assets/presentation/2026-02-17-notas_voz.md b/assets/presentation/docs/2026-02-17-notas_voz.md
similarity index 100%
rename from assets/presentation/2026-02-17-notas_voz.md
rename to assets/presentation/docs/2026-02-17-notas_voz.md
diff --git a/assets/presentation/2026-02-27-notas_voz_rust.md b/assets/presentation/docs/2026-02-27-notas_voz_rust.md
similarity index 100%
rename from assets/presentation/2026-02-27-notas_voz_rust.md
rename to assets/presentation/docs/2026-02-27-notas_voz_rust.md
diff --git a/assets/presentation/docs/2026-04-24-234910-me-planteo-estas-dos-propuestas-para-un-open-spac.txt b/assets/presentation/docs/2026-04-24-234910-me-planteo-estas-dos-propuestas-para-un-open-spac.txt
new file mode 100644
index 0000000..c281fd7
--- /dev/null
+++ b/assets/presentation/docs/2026-04-24-234910-me-planteo-estas-dos-propuestas-para-un-open-spac.txt
@@ -0,0 +1,4668 @@
+╭─── Claude Code v2.1.119 ─────────────────────────────────────────────────────╮
+│                                                    │ Recent activity         │
+│                 Welcome back Jesús!                │ No recent activity      │
+│                                                    │ ─────────────────────── │
+│                       ▐▛███▜▌                      │ What's new              │
+│                      ▝▜█████▛▘                     │ `/config` settings (th… │
+│                        ▘▘ ▝▝                       │ Added `prUrlTemplate` … │
+│   Sonnet 4.6 · Claude Max · jpl@jesusperez.pro's   │ Added `CLAUDE_CODE_HID… │
+│   Organization                                     │ /release-notes for more │
+│  /…/Akasha/Development/ontoref/assets/presentation │                         │
+╰──────────────────────────────────────────────────────────────────────────────╯
+
+❯ me planteo estas dos propuestas para un open space y quiero unas slides sobre 
+   ontologia y reflection como conceptos base del proyecto ontoref              
+  es para Desarrolladores, puntos base para tener en cuenta                     
+  Dos propuestas                                                                
+  ▎ "Ontología en mis proyectos: para que la IA sepa lo que no sabe y yo me     
+  aclare"                                                                       
+                                                                                
+    Tesis: La ontología no es para la IA. Es para ti. La IA se beneficia como   
+  efecto secundario.                                                            
+                                                                                
+    Puntos clave:                                                               
+    - Sin modelo explícito, tus proyectos viven en tu cabeza — y lo que vive    
+  solo en tu cabeza no se                                                       
+    puede validar.                                                              
+    - Formalizar invariantes, tensiones y estado actual te fuerza a saber qué   
+  es verdad ahora mismo.                                                        
+    - La IA consume ese modelo y deja de inventarse contexto — opera sobre lo   
+  que tú ya clarificaste.                                                       
+    - El valor real: el proceso de escribir la ontología es donde te aclaras    
+  tú, no la IA.                                                                 
+                                                                                
+    Para el debate:                                                             
+    - ¿No es suficiente con un buen README?                                     
+    - ¿Qué diferencia una ontología de una lista de decisiones?                 
+    - ¿Quién mantiene esto actualizado cuando el proyecto cambia rápido?        
+   más que ver con ontref                                                       
+  ▎ "Apocalipsis: la IA gestiona tu infra, ¿sabes lo que está pasando?"         
+                                                                                
+    Tesis: La IA puede declarar tu infra. No puede entenderla por ti.           
+                                                                                
+    Puntos clave:                                                               
+    - Declarativo != comprensión. Describes el estado deseado, no el por qué.   
+    - La IA opera sobre forma, no sobre significado. Genera YAML correcto para  
+  el problema equivocado.                                                       
+    - Sin ontología de tu infra — qué son las cosas, cómo se relacionan, qué no 
+   puede romperse — no                                                          
+    tienes control, tienes ilusión de control.                                  
+    - Cuando falla a las 2am, el Terraform generado no te dice qué asumió.      
+                                                                                
+    Para el debate:                                                             
+    - ¿Es diferente a copiar Stack Overflow?                                    
+    - ¿Dónde está el mínimo de comprensión necesario?                           
+    - ¿Las plataformas internas resuelven esto o lo esconden?                   
+   más que ver con /Users/Akasha/project-provisioning/provisioning/             
+                                                                                
+  La idea es generar las slides en                                              
+  /Users/Akasha/Development/ontoref/assets/presentation/                        
+  siguiendo el patrón actual de slides.md que puedes sobreescribir              
+
+⏺ Voy a leer los archivos existentes de presentación para entender el patrón.
+
+  Read 1 file, listed 1 directory (ctrl+o to expand)
+
+⏺ Tengo el patrón claro — Slidev dark, español, componentes Vue, speaker notes.
+  Voy a escribir las slides ahora.
+
+  ★ Insight ─────────────────────────────────────
+  El archivo usa Slidev con v-click para animaciones incrementales y el patrón
+  box-highlight border-l-4 como elemento visual recurrente. Las speaker notes en
+   <!-- --> son parte del contrato de presentación — no decoración. El flujo
+  narrativo real va del problema a la tesis, no al revés.
+  ─────────────────────────────────────────────────
+
+⏺ Write(slides.md)
+  ⎿  Added 443 lines, removed 846 lines
+       1  ---
+       2  theme: default
+       3 -title: Ontology and Reflection                                 
+       3 +title: Ontología y Reflection — lo que la IA no puede entender 
+         +por ti                                                         
+       4  titleTemplate: '%s - OntoRef'
+       5  layout: cover
+       6 -keywords: Nickel,Nushell,Ontology,Reflection,StratumIOps       
+       6 +keywords: Ontología,Reflection,IA,Infraestructura,OntoRef,Nushe
+         +ll,Nickel                                                      
+       7  download: true
+       8 -exportFilename: OntoRef                                        
+       8 +exportFilename: ontoref-openspace                              
+       9  monaco: true
+      10  remoteAssets: true
+      11  selectable: true
+     ...
+      27  class: 'justify-center flex flex-cols'
+      28  ---
+      29  
+      30 -<h1 class="font-medium">Ontology & Reflection</h1>             
+      30 +<h1 class="font-medium text-3xl leading-tight">La IA gestiona t
+         +u infra.<br><span class="text-orange-400">¿Sabes lo que está pa
+         +sando?</span></h1>                                             
+      31  
+      32 -<h2 class="flex justify-center mt-5 font-medium text-orange-400
+         -">                                                             
+      33 -    <img class="-mt-2 h-32" src="/ontoref-text.svg">           
+      34 -</h2>                                                          
+      35 -                                                               
+      36 -<div class="flex justify-center mt-3 text-gray-400 text-lg ital
+         -ic">                                                           
+      37 -  Typing the invisible. Verifying the obvious. Sealing what mat
+         -ters.                                                          
+      32 +<div class="flex justify-center mt-6 text-gray-400 text-lg ital
+         +ic">                                                           
+      33 +  Ontología y Reflection como base de un proyecto que se entien
+         +de a sí mismo.                                                 
+      34  </div>
+      35  
+      40 -<div class="flex justify-center">                              
+      41 -  <img class="mt-8 w-40" src="/jesusperez_w.svg">              
+      36 +<div class="flex justify-center mt-8">                         
+      37 +  <img class="w-40" src="/jesusperez_w.svg">                   
+      38  </div>
+      39  
+      44 -<div class="flex justify-center">                              
+      45 -  <img class="mt-8 w-40" src="/ontoref_img.svg">               
+      46 -</div>                                                         
+      47 -                                                               
+      40  <style scoped>
+      41  h1, h2, p { z-index: 10; }
+      42  </style>
+     ...
+       44  <Footer />
+       45  
+       46  <!--
+       55 -Open with silence. Let the subtitle land.                     
+       56 -This talk is not about tools — it is about the epistemic contr
+          -act between code and understanding.                           
+       47 +Abrir con pausa. El título es una pregunta directa.           
+       48 +No es retórica — la mayoría no sabe. Y el punto es exactamente
+          + ese.                                                         
+       49 +Esta charla no es sobre herramientas. Es sobre qué pasa cuando
+          + externalizas comprensión.                                    
+       50  -->
+       51  
+       52  ---
+       60 -layout: two-cols                                              
+       53 +layout: default                                               
+       54  ---
+       55  
+       63 -# ¿Qué estamos resolviendo?                                   
+       56 +# El problema no es la herramienta                            
+       57  
+       58 +<div class="grid grid-cols-2 gap-8 mt-4">                     
+       59 +                                                              
+       60  <div>
+       61  
+       67 -**Los problemas**                                             
+       62 +**Lo que la IA hace bien**                                    
+       63  
+       64  <div class="mt-3 space-y-3">
+       65  
+       66  <div class="flex items-start gap-2">
+       72 -  <span class="text-red-500 font-bold text-sm mt-1">🔴</span> 
+       73 -  <div><span class="text-gray-200 font-semibold">Epistemic dri
+          -ft</span><br><span class="text-gray-400 text-sm">El código hac
+          -e X, el equipo cree que hace Y.<br> Silencioso. Acumulativo.</
+          -span></div>                                                   
+       67 +  <span class="text-green-400 font-bold text-sm mt-1">✓</span>
+       68 +  <div class="text-gray-300 text-sm">Genera YAML correcto sint
+          +ácticamente</div>                                             
+       69  </div>
+       70  
+       71  <div class="flex items-start gap-2">
+       77 -  <span class="text-red-500 font-bold text-sm mt-1">🔴</span> 
+       78 -  <div><span class="text-gray-200 font-semibold">Multi-actor c
+          -ollision</span><br><span class="text-gray-400 text-sm">Humano 
+          -+ agente IA escriben al mismo fichero,<br> sin coordinación.</
+          -span></div>                                                   
+       72 +  <span class="text-green-400 font-bold text-sm mt-1">✓</span>
+       73 +  <div class="text-gray-300 text-sm">Completa Terraform que nu
+          +nca rompe el plan</div>                                       
+       74  </div>
+       75  
+       76  <div class="flex items-start gap-2">
+       82 -  <span class="text-orange-400 font-bold text-sm mt-1">🟠</spa
+          -n>                                                            
+       83 -  <div><span class="text-gray-200 font-semibold">Decision amne
+          -sia</span><br><span class="text-gray-400 text-sm">¿Por qué se 
+          -tomó esta decisión? ¿Quién la autorizó?</span></div>          
+       77 +  <span class="text-green-400 font-bold text-sm mt-1">✓</span>
+       78 +  <div class="text-gray-300 text-sm">Rellena gaps de configura
+          +ción con defaults razonables</div>                            
+       79  </div>
+       80  
+       81  <div class="flex items-start gap-2">
+       87 -  <span class="text-orange-400 font-bold text-sm mt-1">🟠</spa
+          -n>                                                            
+       88 -  <div><span class="text-gray-200 font-semibold">Config schizo
+          -phrenia</span><br><span class="text-gray-400 text-sm">Staging 
+          -funciona, prod no. ¿Qué campo cambió y cuándo?</span></div>   
+       82 +  <span class="text-green-400 font-bold text-sm mt-1">✓</span>
+       83 +  <div class="text-gray-300 text-sm">Traduce intención en form
+          +a declarativa</div>                                           
+       84  </div>
+       85  
+       86  </div>
+       87  
+       88  </div>
+       89  
+       95 -::right::                                                     
+       96 -                                                              
+       90  <div>
+       91  
+       92 +<v-click>                                                     
+       93 +                                                              
+       94 +**Lo que la IA no puede hacer**                               
+       95 +                                                              
+       96 +<div class="mt-3 space-y-3">                                  
+       97 +                                                              
+       98  <div class="flex items-start gap-2">
+      100 -  <span class="text-orange-400 font-bold text-sm mt-1">🟠</spa
+          -n>                                                            
+      101 -  <div><span class="text-gray-200 font-semibold">Docs como art
+          -efactos muertos</span><br><span class="text-gray-400 text-sm">
+          -Escritas una vez, olvidadas, nunca actualizadas.</span></div> 
+       99 +  <span class="text-red-400 font-bold text-sm mt-1">✗</span>  
+      100 +  <div class="text-gray-300 text-sm">Saber <em>por qué</em> es
+          +e servicio no puede escalar a N réplicas</div>                
+      101  </div>
+      102  
+      104 -<div class="flex items-start gap-2 mt-2">                     
+      105 -  <span class="text-yellow-400 font-bold text-sm mt-1">🟡</spa
+          -n>                                                            
+      106 -  <div><span class="text-gray-200 font-semibold">Agent con ceg
+          -uera</span><br><span class="text-gray-400 text-sm">Los agentes
+          - IA empiezan sin contexto,<br> violan invariantes establecidas
+          -.</span></div>                                                
+      103 +<div class="flex items-start gap-2">                          
+      104 +  <span class="text-red-400 font-bold text-sm mt-1">✗</span>  
+      105 +  <div class="text-gray-300 text-sm">Conocer la tensión que ex
+          +iste entre latencia y consistencia en <em>tu</em> sistema</div
+          +>                                                             
+      106  </div>
+      107  
+      109 -<v-click>                                                     
+      108 +<div class="flex items-start gap-2">                          
+      109 +  <span class="text-red-400 font-bold text-sm mt-1">✗</span>  
+      110 +  <div class="text-gray-300 text-sm">Recordar la decisión que 
+          +tomaste en diciembre y por qué descartaste la alternativa</div
+          +>                                                             
+      111 +</div>                                                        
+      112  
+      111 -<div class="mt-2 text-xs text-gray-400 space-y-2 font-mono">  
+      113 +<div class="flex items-start gap-2">                          
+      114 +  <span class="text-red-400 font-bold text-sm mt-1">✗</span>  
+      115 +  <div class="text-gray-300 text-sm">Entender qué no puede rom
+          +perse sin coordinación entre equipos</div>                    
+      116 +</div>                                                        
+      117  
+      113 -**Ejemplo: desincronización real**                            
+      118 +</div>                                                        
+      119  
+      115 -```bash                                                       
+      116 -# Documentación (2024-01)                                     
+      117 -config.max_retries = 3                                        
+      120 +</v-click>                                                    
+      121  
+      119 -# Código actual (después de cambio sin registrar)             
+      120 -config.max_retries = 7                                        
+      122 +<v-click>                                                     
+      123  
+      122 -# Equipo cree que está usando 3 ✗                             
+      123 -# Sistema usa 7 silenciosamente ✓                             
+      124 -# Nadie sabe cuándo cambió ni por qué                         
+      125 -```                                                           
+      126 -                                                              
+      124 +<div class="mt-6 border-l-4 border-orange-500 pl-4 text-sm tex
+          +t-gray-300">                                                  
+      125 +  Declarativo ≠ comprensión.<br>                              
+      126 +  Describes el estado deseado, no el <em>por qué</em>.        
+      127  </div>
+      128  
+      129  </v-click>
+      130  
+      131  </div>
+      132  
+      133 +</div>                                                        
+      134 +                                                              
+      135  <Footer />
+      136  
+      137  <!--
+      136 -Pause after each severity badge. The audience will recognize t
+          -hese — they have lived this.                                  
+      137 -Goal is recognition, not education. Ask silently: "how many of
+          - these are in your project right now?"                        
+      138 +La IA genera forma correcta para el problema equivocado.      
+      139 +El YAML válido es el problema mínimo. El problema real es que 
+          +nadie sabe qué asumió.                                        
+      140  -->
+      141  
+      142  ---
+      143  layout: default
+      144  ---
+      145  
+      144 -# El coste de ignorarlo                                       
+      146 +# Las 2am — cuando el Terraform no habla                      
+      147  
+      148  <div class="grid grid-cols-2 gap-8 mt-4">
+      149  
+      150  <div>
+      151  
+      150 -<div class="box-highlight mt-2 font-mono text-xs">            
+      152 +**Lo que ves cuando falla**                                   
+      153  
+      154 +<div class="mt-3 font-mono text-xs text-gray-300">            
+      155 +                                                              
+      156  ```text
+      153 -Deuda técnica      → arreglable                               
+      154 -Deuda epistémica   → sistémica                                
+      155 -Deuda de confianza → letal                                    
+      157 +Error: Error creating Instance:                               
+      158 +  googleapi: Error 400: Invalid value                         
+      159 +  for field 'resource.machineType'                            
+      160 +                                                              
+      161 +# El plan pasó. El apply falló.                               
+      162 +# ¿Qué asumió el agente?                                      
+      163 +# ¿Qué región eligió? ¿Por qué ese tipo?                      
+      164 +# ¿Hay una restricción de org que no sabía?                   
+      165  ```
+      166  
+      167  </div>
+      168  
+      169  <v-click>
+      170  
+      162 -<div class="mt-4 space-y-5 text-xs text-gray-300">            
+      171 +<div class="mt-4 font-mono text-xs text-gray-400">            
+      172  
+      164 -<div class="border-l-2 border-red-500 pl-2">                  
+      165 -  <div class="font-semibold text-red-300">Técnica</div>       
+      166 -  <div class="text-gray-400">Refactoriza, reescribe, corrige.<
+          -/div>                                                         
+      167 -  <div class="text-gray-500 text-xs">1-3 sprints</div>        
+      168 -</div>                                                        
+      173 +```yaml                                                       
+      174 +# Lo que generó la IA (hace 6 semanas)                        
+      175 +resource "google_compute_instance" "api" {                    
+      176 +  machine_type = "e2-medium"   # "razonable"                  
+      177 +  zone         = "us-central1-a" # "por defecto"              
+      178 +}                                                             
+      179  
+      170 -<div class="border-l-2 border-orange-500 pl-2">               
+      171 -  <div class="font-semibold text-orange-300">Epistémica</div> 
+      172 -  <div class="text-gray-400">Acumula silenciosamente. No sabes
+          - dónde.</div>                                                 
+      173 -  <div class="text-gray-500 text-xs">Crece cada semana</div>  
+      174 -</div>                                                        
+      180 +# Lo que tu política de org requiere                          
+      181 +# → solo europe-west1, solo n2-standard-*                     
+      182 +# → la IA no lo sabía porque tú tampoco lo escribiste         
+      183 +```                                                           
+      184  
+      176 -<div class="border-l-2 border-yellow-600 pl-2">               
+      177 -  <div class="font-semibold text-yellow-300">Confianza</div>  
+      178 -  <div class="text-gray-400">Se disuelve cuando se descubre el
+          - drift.</div>                                                 
+      179 -  <div class="text-gray-500 text-xs">Imposible reparar rápido<
+          -/div>                                                         
+      185  </div>
+      186  
+      182 -</div>                                                        
+      183 -                                                              
+      187  </v-click>
+      188  
+      189  </div>
+     ...
+      192  
+      193  <v-click>
+      194  
+      192 -<div class="mt-2 text-gray-300">                              
+      195 +**El patrón que se repite**                                   
+      196  
+      194 -**Escenario: 6 meses de ignorar**                             
+      197 +<div class="mt-3 space-y-3 text-sm">                          
+      198  
+      196 -<div class="mt-3 font-mono text-xs space-y-2 text-gray-400">  
+      199 +<div class="border-l-2 border-red-500 pl-3">                  
+      200 +  <div class="text-red-300 font-semibold">Sin modelo explícito
+          +</div>                                                        
+      201 +  <div class="text-gray-400">La IA inventa contexto plausible.
+          + Tú asumes que sabe. Falla en prod.</div>                     
+      202 +</div>                                                        
+      203  
+      198 -```yaml                                                       
+      199 -Mes 1: "Solo refactorizar esto"                               
+      200 -Mes 2: "Documentación está desactualizada"                    
+      201 -Mes 3: "¿Por qué prod falla y staging no?"                    
+      202 -Mes 4: "Nadie sabe quién cambió qué"                          
+      203 -Mes 5: Pánico. Rollback arriesgado.                           
+      204 -Mes 6: Equipo rechaza cambios nuevos.                         
+      205 -         "No confiamos sin entender"                          
+      206 -```                                                           
+      204 +<div class="border-l-2 border-orange-500 pl-3">               
+      205 +  <div class="text-orange-300 font-semibold">Con Stack Overflo
+          +w</div>                                                       
+      206 +  <div class="text-gray-400">Copias código sin entender el sup
+          +uesto. Al menos tú lo buscaste.</div>                         
+      207 +</div>                                                        
+      208  
+      209 +<div class="border-l-2 border-yellow-500 pl-3">               
+      210 +  <div class="text-yellow-300 font-semibold">Con la IA</div>  
+      211 +  <div class="text-gray-400">El agente genera con confianza. L
+          +a confianza es contagiosa. El error también.</div>            
+      212  </div>
+      213  
+      214  </div>
+     ...
+      217  
+      218  <v-click>
+      219  
+      216 -<div class="mt-8 text-gray-300 text-base leading-7 border-l-4 
+          -border-orange-500 pl-4">                                      
+      217 -  Estos problemas no tienen solución técnica.<br>             
+      218 -  Tienen solución de <em>arquitectura</em>.                   
+      220 +<div class="mt-6 text-base font-semibold text-gray-200 border-
+          +l-4 border-orange-500 pl-4">                                  
+      221 +  La diferencia con Stack Overflow:<br>                       
+      222 +  <span class="text-gray-400 font-normal text-sm">Al menos ahí
+          + buscabas activamente.<br>Con la IA, delegas también la búsque
+          +da.</span>                                                    
+      223  </div>
+      224  
+      225  </v-click>
+     ...
+      228  
+      229  </div>
+      230  
+      231 +<Footer />                                                    
+      232 +                                                              
+      233 +<!--                                                          
+      234 +Este slide es el que el público siente. Todos tienen una histo
+          +ria de "la IA lo generó y pareció bien".                      
+      235 +La pregunta es: ¿cuándo lo descubriste? ¿Staging? ¿Prod? ¿6 me
+          +ses después?                                                  
+      236 +-->                                                           
+      237 +                                                              
+      238  ---
+      239  layout: default
+      240  ---
+      241  
+      231 -<h1 class="-mt-5"> Verdad y operación — dos fuerzas que deben 
+          -coexistir </h1>                                               
+      242 +# Tu proyecto vive en tu cabeza                               
+      243  
+      244  <div class="grid grid-cols-2 gap-8 mt-4">
+      245  
+      235 -<div class="border-l-4 border-blue-500 pl-4">                 
+      246 +<div>                                                         
+      247  
+      237 -**Yin — la capa formal**                                      
+      248 +**Lo que vive solo en tu cabeza**                             
+      249  
+      239 -*Lo que debe ser verdad*                                      
+      240 -                                                              
+      250  <div class="mt-3 space-y-2 text-sm text-gray-300">
+      251  
+      243 -- **Nickel schemas** → corrección estructural<br> en tiempo de
+          - definición                                                   
+      244 -- **ADR constraints** → "esto nunca puede violarse"           
+      245 -- **Config seals** → estados sellados, verificables con sha256
+      246 -- **Ontology invariants** → lo que no puede cambiar <br> sin u
+          -n nuevo ADR                                                   
+      247 -- **Hashes matemáticos** → prueba de qué fue sellado<br> y cuá
+          -ndo                                                           
+      252 +<div class="flex items-start gap-2">                          
+      253 +  <span class="text-orange-400 text-sm mt-1">→</span>         
+      254 +  <div>Por qué este servicio no puede tener más de 3 réplicas<
+          +/div>                                                         
+      255 +</div>                                                        
+      256  
+      257 +<div class="flex items-start gap-2">                          
+      258 +  <span class="text-orange-400 text-sm mt-1">→</span>         
+      259 +  <div>Qué pasó cuando intentamos escalar en enero y por qué f
+          +allamos</div>                                                 
+      260  </div>
+      261  
+      262 +<div class="flex items-start gap-2">                          
+      263 +  <span class="text-orange-400 text-sm mt-1">→</span>         
+      264 +  <div>La tensión entre el equipo de plataforma y el de produc
+          +to sobre quién controla el networking</div>                   
+      265  </div>
+      266  
+      253 -<div class="border-l-4 border-orange-500 pl-4">               
+      254 -                                                              
+      255 -**Yang — la capa operacional**                                
+      256 -                                                              
+      257 -*Cómo las cosas se mueven y cambian*                          
+      258 -                                                              
+      259 -<div class="-mt-2 space-y-2 text-sm text-gray-300">           
+      260 -                                                              
+      261 -- **Nu commands** → transformación estructurada de datos      
+      262 -- **Actors** (human/agent/CI) → mismo protocolo,<br> distintas
+          - capacidades                                                  
+      263 -- **Register flow** → captura cambios,<br> los enruta al artef
+          -acto correcto                                                 
+      264 -- **Mode definitions** → secuencias de operaciones<br> con ver
+          -ificación                                                     
+      265 -- **Pre-commit hooks** → sincronización forzada <br>en el mome
+          -nto del commit                                                
+      266 -                                                              
+      267 +<div class="flex items-start gap-2">                          
+      268 +  <span class="text-orange-400 text-sm mt-1">→</span>         
+      269 +  <div>Qué alternativa se rechazó y por qué</div>             
+      270  </div>
+      271  
+      272 +<div class="flex items-start gap-2">                          
+      273 +  <span class="text-orange-400 text-sm mt-1">→</span>         
+      274 +  <div>Qué invariante no puede violarse sin romper el acuerdo 
+          +con el equipo de seguridad</div>                              
+      275  </div>
+      276  
+      277  </div>
+      278  
+      279  <v-click>
+      280  
+      275 -<div class="mt-1 border border-gray-700 rounded p-4 bg-gray-90
+          -0 text-sm text-gray-300 font-mono leading-7">                 
+      276 -                                                              
+      277 -```text                                                       
+      278 -Yang sin Yin  = fluido pero caótico.  Cualquier cosa puede cam
+          -biar.       Nada es verificable.                              
+      279 -Yin  sin Yang = correcto pero inútil. Schemas perfectos sin op
+          -eraciones = documentación muerta.                             
+      280 -```                                                           
+      281 -                                                              
+      282 -<div class="text-orange-400 text-center mt-2 font-sans font-se
+          -mibold">El sistema vive en la coexistencia.</div>             
+      283 -                                                              
+      281 +<div class="mt-5 border-l-4 border-red-500 pl-4 text-sm text-g
+          +ray-400">                                                     
+      282 +  Lo que vive solo en tu cabeza <strong class="text-red-300">n
+          +o se puede validar</strong>.<br>                              
+      283 +  No puede consultarlo un agente. No puede consultarlo un comp
+          +añero nuevo.<br>                                              
+      284 +  No puedes consultarlo tú mismo en 6 meses.                  
+      285  </div>
+      286  
+      287  </v-click>
+      288  
+      288 -<Footer />                                                    
+      289 +</div>                                                        
+      290  
+      290 -<!--                                                          
+      291 -The Yin/Yang is not metaphor for aesthetics — it is the archit
+          -ectural duality.                                              
+      292 -Pure formal layer without operation is a museum. Pure operatio
+          -n without formal layer is controlled chaos.                   
+      293 -Hold the tension statement at the end. It is the thesis.      
+      294 --->                                                           
+      291 +<div>                                                         
+      292  
+      296 ----                                                           
+      297 -layout: center                                                
+      298 ----                                                           
+      293 +<v-click>                                                     
+      294  
+      300 -# Los ingredientes                                            
+      295 +**La trampa del README**                                      
+      296  
+      302 -<div class="mt-2 space-y-3">                                  
+      297 +<div class="mt-3 font-mono text-xs text-gray-400">            
+      298  
+      304 -<div class="border-l-4 border-blue-400 pl-4 py-2 bg-gray-900 r
+          -ounded-r">                                                    
+      305 -  <div class="text-blue-300 font-mono font-semibold text-sm">C
+          -APA DECLARATIVA &nbsp;·&nbsp; Nickel</div>                    
+      306 -  <div class="text-gray-400 text-sm mt-1"><code>.ontology/</co
+          -de> &nbsp;·&nbsp; <code>adrs/</code> &nbsp;·&nbsp; <code>refle
+          -ction/schemas/</code> &nbsp;·&nbsp; <code>reflection/configs/<
+          -/code></div>                                                  
+      307 -  <div class="text-gray-500 text-xs mt-1">Tipos fuertes, contr
+          -atos, enums. Falla en definición, no en runtime.</div>        
+      308 -</div>                                                        
+      299 +```markdown                                                   
+      300 +# Mi Proyecto                                                 
+      301  
+      310 -<div class="border-l-4 border-orange-400 pl-4 py-2 bg-gray-900
+          - rounded-r">                                                  
+      311 -  <div class="text-orange-300 font-mono font-semibold text-sm"
+          ->CAPA OPERACIONAL &nbsp;·&nbsp; Nushell</div>                 
+      312 -  <div class="text-gray-400 text-sm mt-1"><code>adr</code> &nb
+          -sp;·&nbsp; <code>register</code> &nbsp;·&nbsp; <code>config</c
+          -ode> &nbsp;·&nbsp; <code>backlog</code> &nbsp;·&nbsp; <code>fo
+          -rms</code> &nbsp;·&nbsp; <code>prereqs</code></div>           
+      313 -  <div class="text-gray-500 text-xs mt-1">Pipelines tipadas so
+          -bre datos estructurados. No streams de texto.</div>           
+      314 -</div>                                                        
+      302 +Este servicio hace X. Está desplegado en GCP.                 
+      303 +Usa Terraform. Tiene 3 entornos: dev, staging, prod.          
+      304  
+      316 -<div class="border-l-4 border-green-400 pl-4 py-2 bg-gray-900 
+          -rounded-r">                                                   
+      317 -  <div class="text-green-300 font-mono font-semibold text-sm">
+          -PUNTO DE ENTRADA &nbsp;·&nbsp; Bash → Nu</div>                
+      318 -  <div class="text-gray-400 text-sm mt-1"><code>stratum.sh</co
+          -de> &nbsp;·&nbsp; actor detection &nbsp;·&nbsp; locking &nbsp;
+          -·&nbsp; <code>NICKEL_IMPORT_PATH</code></div>                 
+      319 -  <div class="text-gray-500 text-xs mt-1">Un único entry point
+          -. Detecta actor, adquiere lock, despacha al módulo Nu correcto
+          -.</div>                                                       
+      320 -</div>                                                        
+      305 +## Arquitectura                                               
+      306 +Ver diagrama (desactualizado desde marzo).                    
+      307  
+      322 -<div class="border-l-4 border-purple-400 pl-4 py-2 bg-gray-900
+          - rounded-r">                                                  
+      323 -  <div class="text-purple-300 font-mono font-semibold text-sm"
+          ->GRAFO DE CONOCIMIENTO &nbsp;·&nbsp; Ontología + ADRs</div>   
+      324 -  <div class="text-gray-400 text-sm mt-1">nodes &nbsp;·&nbsp; 
+          -invariants &nbsp;·&nbsp; gates &nbsp;·&nbsp; dimensions &nbsp;
+          -·&nbsp; states</div>                                          
+      325 -  <div class="text-gray-500 text-xs mt-1">El sistema sabe qué 
+          -sabe. Actor-agnostic. Machine-queryable.</div>                
+      326 -</div>                                                        
+      308 +## Decisiones                                                 
+      309 +- Elegimos PostgreSQL porque MySQL no nos gustaba.            
+      310 +- Usamos gRPC internamente.                                   
+      311 +```                                                           
+      312  
+      328 -<div class="border-l-4 border-yellow-400 pl-4 py-2 bg-gray-900
+          - rounded-r">                                                  
+      329 -  <div class="text-yellow-300 font-mono font-semibold text-sm"
+          ->ESTADOS SELLADOS &nbsp;·&nbsp; Config + Historia</div>       
+      330 -  <div class="text-gray-400 text-sm mt-1">profiles &nbsp;·&nbs
+          -p; sha256 seals &nbsp;·&nbsp; audit trail &nbsp;·&nbsp; rollba
+          -ck</div>                                                      
+      331 -  <div class="text-gray-500 text-xs mt-1">Inmutabilidad verifi
+          -cable. Drift detection. Trazabilidad completa ADR/PR/bug.</div
+          ->                                                             
+      313  </div>
+      314  
+      334 -</div>                                                        
+      315 +<div class="mt-4 space-y-2 text-sm text-gray-400">            
+      316  
+      336 -<Footer />                                                    
+      337 -                                                              
+      338 -<!--                                                          
+      339 -This is the map, not the territory. Each layer will get its ow
+          -n slide.                                                      
+      340 -Point bottom-up: sealed states give you history; the graph giv
+          -es you meaning; the entry point gives you discipline.         
+      341 --->                                                           
+      342 -                                                              
+      343 ----                                                           
+      344 -layout: default                                               
+      345 ----                                                           
+      346 -                                                              
+      347 -# Nickel — configuración con tipos, sin runtime               
+      348 -                                                              
+      349 -<div class="grid grid-cols-2 gap-8 mt-4">                     
+      350 -                                                              
+      351 -<div>                                                         
+      352 -                                                              
+      353 -**Qué es Nickel**                                             
+      354 -                                                              
+      355 -<div class="text-sm text-gray-300 space-y-1 mt-2">            
+      356 -                                                              
+      357 -- Lenguaje de configuración con tipos fuertes (de Tweag)      
+      358 -- Puramente declarativo: sin efectos secundarios, sin loops, s
+          -in runtime                                                    
+      359 -- Superset de JSON: JSON válido es Nickel válido              
+      360 -- Imports: <small>`let s = import "schema.ncl" in`</small>    
+      361 -- Contratos: `field | TypeName` — verificado al merge         
+      362 -- Enums tipados: <small>`[| 'Proposed, 'Accepted, 'Superseded 
+          -|]`</small>                                                   
+      363 -- Registros abiertos/cerrados: <br> <small>`{ _: String }` vs 
+          -`{ field | String }`</small>                                  
+      364 -- `nickel export file.ncl` → JSON estable pipelines Nu        
+      365 -                                                              
+      317 +<div class="flex items-start gap-2">                          
+      318 +  <span class="text-red-400 text-sm">✗</span>                 
+      319 +  <div>No dice qué no puede cambiar</div>                     
+      320  </div>
+      321  
+      368 -<div class="mt-4 text-lg text-gray-500 border-l-2 border-orang
+          -e-500 pl-3 italic">                                           
+      369 -  El sistema de tipos <em>es</em> la documentación.           
+      322 +<div class="flex items-start gap-2">                          
+      323 +  <span class="text-red-400 text-sm">✗</span>                 
+      324 +  <div>No captura las alternativas rechazadas ni el por qué</d
+          +iv>                                                           
+      325  </div>
+      326  
+      327 +<div class="flex items-start gap-2">                          
+      328 +  <span class="text-red-400 text-sm">✗</span>                 
+      329 +  <div>No es consultable ni verificable — es prosa</div>      
+      330  </div>
+      331  
+      374 -<div>                                                         
+      375 -                                                              
+      376 -**Por qué no alternativas**                                   
+      377 -                                                              
+      378 -<div class="font-mono text-xs text-gray-400 -mt-4 space-y-1"> 
+      379 -                                                              
+      380 -```text                                                       
+      381 -YAML  → sin tipos, whitespace-sensitive,                      
+      382 -        coerciones implícitas                                 
+      383 -TOML  → sin tipos ≥ primitivos, sin imports                   
+      384 -JSON  → sin comentarios, sin imports, sin tipos               
+      385 -HCL   → efectos secundarios, no es config pura                
+      386 -CUE   → similar pero peor historia de imports                 
+      387 -KCL   → tipos + validación, pero vendedor                     
+      388 -        lenguaje opaco                                        
+      389 -```                                                           
+      390 -                                                              
+      332 +<div class="flex items-start gap-2">                          
+      333 +  <span class="text-red-400 text-sm">✗</span>                 
+      334 +  <div>No refleja el estado actual — refleja el estado de cuan
+          +do se escribió</div>                                          
+      335  </div>
+      336  
+      393 -<div class="mt-i2 text-lg text-gray-300 font-semibold">Por qué
+          - aquí específicamente</div>                                   
+      394 -                                                              
+      395 -<div class="text-xs text-gray-400 mt-1 space-y-1">            
+      396 -                                                              
+      397 -- ADR constraints necesitan typecheck, no validación manual   
+      398 -- Perfiles de config deben fallar en definición, no en runtime
+      399 -- `nickel export` produce JSON estable para consumo Nu        
+      400 -                                                              
+      337  </div>
+      338  
+      403 -<v-click>                                                     
+      404 -                                                              
+      405 -<div class="mt-1 font-mono text-xs">                          
+      406 -                                                              
+      407 -```bash                                                       
+      408 -# YAML: silently accepts garbage                              
+      409 -actor: developr   # typo, accepted                            
+      410 -# Nickel: rejected at typecheck                               
+      411 -{ actor | [| 'developer, 'agent, 'ci |] = "developr" }        
+      412 -#                       ^^^^^^^^^                             
+      413 -#                error: tag not in enum type                  
+      414 -```                                                           
+      415 -                                                              
+      416 -</div>                                                        
+      417 -                                                              
+      339  </v-click>
+      340  
+      341  </div>
+     ...
+      345  <Footer />
+      346  
+      347  <!--
+      427 -The before/after example is the punchline. Let it sit.        
+      428 -"The type system IS the documentation" — this is the core Nick
+          -el thesis. Not a feature, the philosophy.                     
+      348 +La pregunta clave que lanzar al público: ¿cuándo fue la última
+          + vez que actualizaste tu README?                              
+      349 +Un README es documentación. Una ontología es un modelo vivo de
+          +l sistema.                                                    
+      350  -->
+      351  
+      352  ---
+      432 -layout: default                                               
+      353 +layout: center                                                
+      354  ---
+      355  
+      435 -# Nushell — datos estructurados, no streams de texto          
+      356 +# La tesis                                                    
+      357  
+      437 -<div class="grid grid-cols-2 gap-8 mt-4">                     
+      358 +<div class="mt-8 text-center space-y-6">                      
+      359  
+      439 -<div>                                                         
+      360 +<div class="text-3xl font-semibold text-gray-100">            
+      361 +  La ontología no es para la IA.                              
+      362 +</div>                                                        
+      363  
+      441 -**Shell tradicional: texto opaco**                            
+      442 -                                                              
+      443 -<div class="mt-2 font-mono text-xs">                          
+      444 -                                                              
+      445 -```bash                                                       
+      446 -# Bash: parsear texto, error-prone                            
+      447 -grep 'severity' adrs/adr-001.json |\                          
+      448 -    python3 -c "import sys,json; ..."                         
+      449 -```                                                           
+      450 -                                                              
+      364 +<div class="text-3xl font-semibold text-orange-400">          
+      365 +  Es para ti.                                                 
+      366  </div>
+      367  
+      368  <v-click>
+      369  
+      455 -<div class="mt-4 text-sm font-semibold text-gray-200">Nushell:
+          - records nativos</div>                                        
+      456 -                                                              
+      457 -<div class="mt-2 font-mono text-xs">                          
+      458 -                                                              
+      459 -```nu                                                         
+      460 -nickel export adrs/adr-001.ncl                                
+      461 -  | from json                                                 
+      462 -  | get constraints                                           
+      463 -  | where severity == "Hard"                                  
+      464 -  | select id claim                                           
+      465 -```                                                           
+      466 -                                                              
+      370 +<div class="mt-8 text-lg text-gray-400 max-w-2xl mx-auto leadi
+          +ng-8">                                                        
+      371 +  Formalizar invariantes, tensiones y estado actual<br>       
+      372 +  te fuerza a saber qué es verdad <em>ahora mismo</em>.       
+      373  </div>
+      374  
+      469 -<div class="mt-3 font-mono text-xs text-gray-500">            
+      375 +</v-click>                                                    
+      376  
+      471 -```text                                                       
+      472 -╭───┬────────────┬──────────────────────────────╮             
+      473 -│ # │ id         │ claim                        │             
+      474 -├───┼────────────┼──────────────────────────────┤             
+      475 -│ 0 │ adr-001-c1 │ nickel export must succeed…  │             
+      476 -│ 1 │ adr-001-c2 │ schema is sole source of…    │             
+      477 -╰───┴────────────┴──────────────────────────────╯             
+      478 -```                                                           
+      377 +<v-click>                                                     
+      378  
+      379 +<div class="mt-6 text-lg text-gray-400 max-w-2xl mx-auto leadi
+          +ng-8">                                                        
+      380 +  La IA se beneficia como efecto secundario.<br>              
+      381 +  Opera sobre lo que <em>tú ya clarificaste</em>.             
+      382  </div>
+      383  
+      384  </v-click>
+      385  
+      484 -</div>                                                        
+      485 -                                                              
+      486 -<div>                                                         
+      487 -                                                              
+      386  <v-click>
+      387  
+      490 -**Nickel → Nu → Bash — tres capas, tres roles**               
+      491 -                                                              
+      492 -<div class="mt-3 space-y-3 text-sm">                          
+      493 -                                                              
+      494 -<div class="box-highlight border-l-4 border-blue-500 pl-3 py-2
+          -">                                                            
+      495 -  <div class="font-semibold text-blue-300">Nickel exporta JSON
+          - tipado</div>                                                 
+      496 -  <div class="text-gray-400 text-xs">Nunca strings mágicos. Co
+          -ntratos en la definición, no en runtime.</div>                
+      388 +<div class="mt-8 text-base text-gray-500 max-w-2xl mx-auto bor
+          +der-t border-gray-700 pt-6">                                  
+      389 +  El valor real: el proceso de <strong class="text-gray-300">e
+          +scribir la ontología</strong><br>                             
+      390 +  es donde te aclaras tú, no la IA.                           
+      391  </div>
+      392  
+      499 -<div class="box-highlight border-l-4 border-orange-400 pl-3 py
+          --2">                                                          
+      500 -  <div class="font-semibold text-orange-300">Nu consume estruc
+          -turas</div>                                                   
+      501 -  <div class="text-gray-400 text-xs">Tables, records, pipeline
+          -s. Filter, sort, merge — sin awk ni sed.</div>                
+      502 -</div>                                                        
+      503 -                                                              
+      504 -<div class="box-highlight border-l-4 border-gray-500 pl-3 py-2
+          -">                                                            
+      505 -  <div class="font-semibold text-gray-300">Bash orquesta proce
+          -sos</div>                                                     
+      506 -  <div class="text-gray-400 text-xs">Locking, signals, exec. S
+          -olo lo que bash hace de forma nativa.</div>                   
+      507 -</div>                                                        
+      508 -                                                              
+      509 -</div>                                                        
+      510 -                                                              
+      511 -<div class="mt-4 font-mono text-xs">                          
+      512 -                                                              
+      513 -```nu                                                         
+      514 -# Constraints Hard del sistema completo                       
+      515 -glob "adrs/adr-*.ncl"                                         
+      516 -| each { |f| nickel export $f | complete | get stdout }       
+      517 -| each { from json }                                          
+      518 -| where status == "Accepted"                                  
+      519 -| get constraints | flatten                                   
+      520 -| where severity == "Hard"                                    
+      521 -```                                                           
+      522 -                                                              
+      523 -</div>                                                        
+      524 -                                                              
+      393  </v-click>
+      394  
+      395  </div>
+      396  
+      529 -</div>                                                        
+      530 -                                                              
+      397  <Footer />
+      398  
+      399  <!--
+      534 -Nu is not "better bash". It is a different semantic layer.    
+      535 -Text pipelines are broken by design — structured pipelines are
+          - the fix.                                                     
+      536 -Nickel/Nu/Bash: each covers what the others cannot. The divisi
+          -on is deliberate.                                             
+      400 +Pausa larga después del primer par de líneas. Dejar que aterri
+          +ce.                                                           
+      401 +La gente espera que digas "la IA necesita contexto". La invers
+          +ión es el punto.                                              
+      402  -->
+      403  
+      404  ---
+      405  layout: default
+      406  ---
+      407  
+      543 -# DAGs y FSMs — grafos acíclicos y máquinas de estado         
+      408 +# Ontología ≠ lista de decisiones                             
+      409  
+      410  <div class="grid grid-cols-2 gap-8 mt-4">
+      411  
+      412  <div>
+      413  
+      549 -**FSM: ciclo de vida de un ADR**                              
+      414 +**Una lista de decisiones**                                   
+      415  
+      551 -<div class="mt-3 font-mono text-xs">                          
+      416 +<div class="mt-3 font-mono text-xs text-gray-400">            
+      417  
+      418  ```text
+      554 -  ┌──────────┐  adr accept  ┌──────────┐                      
+      555 -  │ Proposed │ ───────────▶ │ Accepted │                      
+      556 -  └──────────┘              └──────────┘                      
+      557 -                                  │ new ADR                   
+      558 -                                  │ supersedes                
+      559 -                                  ▼                           
+      560 -                            ┌────────────┐                    
+      561 -                            │ Superseded │                    
+      562 -                            └────────────┘                    
+      419 +- Elegimos Kubernetes                                         
+      420 +- Usamos Helm para despliegues                                
+      421 +- PostgreSQL como base de datos primaria                      
+      422 +- gRPC entre servicios internos                               
+      423 +- OAuth2 para autenticación externa                           
+      424  ```
+      425  
+      426  </div>
+      427  
+      567 -<div class="mt-3 text-sm text-gray-300">Transiciones typecheck
+          --enforced en Nickel:</div>                                    
+      568 -                                                              
+      569 -<div class="font-mono text-xs mt-1">                          
+      570 -                                                              
+      571 -```text                                                       
+      572 -status | [| 'Proposed, 'Accepted, 'Superseded |]              
+      573 -superseded_by | String | default = ""                         
+      574 -```                                                           
+      575 -                                                              
+      428 +<div class="mt-4 text-sm text-gray-500">                      
+      429 +Captura el <em>qué</em>. No captura el <em>por qué</em>,<br>  
+      430 +las alternativas rechazadas, ni las consecuencias.            
+      431  </div>
+      432  
+      578 -<v-click>                                                     
+      579 -                                                              
+      580 -<div class="mt-4 text-xs text-gray-400">                      
+      581 -Las transiciones inválidas no existen <br>                    
+      582 -— el tipo las elimina del espacio de posibilidades.<br>       
+      583 -Un ADR no puede volver a Proposed.<br>                        
+      584 -Superseded no puede tener constraints activas.                
+      433  </div>
+      434  
+      587 -</v-click>                                                    
+      588 -                                                              
+      589 -</div>                                                        
+      590 -                                                              
+      435  <div>
+      436  
+      593 -**DAG: ontología como grafo de dependencias**                 
+      437 +<v-click>                                                     
+      438  
+      595 -<div class="mt-3 font-mono text-xs text-gray-300">            
+      439 +**Una ontología**                                             
+      440  
+      597 -```text                                                       
+      598 -invariante: backend-agnostic-core                             
+      599 -  └─ tensión: nickel-as-canonical-schema                      
+      600 -       └─ gate: nickel-primacy-gate [activa]                  
+      601 -            └─ protege: nickel-integration-depth              
+      441 +<div class="mt-3 space-y-2 text-sm">                          
+      442  
+      603 -invariante: zero-external-runtime-core                        
+      604 -  └─ gate: core-dependency-gate [activa]                      
+      605 -       └─ protege: backend-agnostic-core                      
+      606 -```                                                           
+      443 +<div class="border-l-4 border-blue-400 pl-3 py-1 bg-gray-900 r
+          +ounded-r">                                                    
+      444 +  <div class="text-blue-300 font-mono text-xs font-semibold">I
+          +NVARIANTES</div>                                              
+      445 +  <div class="text-gray-400 text-xs">Lo que no puede cambiar s
+          +in decisión explícita y coordinación</div>                    
+      446 +</div>                                                        
+      447  
+      448 +<div class="border-l-4 border-orange-400 pl-3 py-1 bg-gray-900
+          + rounded-r">                                                  
+      449 +  <div class="text-orange-300 font-mono text-xs font-semibold"
+          +>TENSIONES ACTIVAS</div>                                      
+      450 +  <div class="text-gray-400 text-xs">Trade-offs que aún no se 
+          +han resuelto — viven en el sistema ahora mismo</div>          
+      451  </div>
+      452  
+      610 -<v-click>                                                     
+      453 +<div class="border-l-4 border-purple-400 pl-3 py-1 bg-gray-900
+          + rounded-r">                                                  
+      454 +  <div class="text-purple-300 font-mono text-xs font-semibold"
+          +>ESTADO ACTUAL → DESEADO</div>                                
+      455 +  <div class="text-gray-400 text-xs">Dimensiones de madurez. D
+          +ónde estás, dónde quieres estar, qué te bloquea</div>         
+      456 +</div>                                                        
+      457  
+      612 -<div class="mt-4 text-sm font-semibold text-gray-200">Por qué 
+          -importa la aciclicidad</div>                                  
+      458 +<div class="border-l-4 border-green-400 pl-3 py-1 bg-gray-900 
+          +rounded-r">                                                   
+      459 +  <div class="text-green-300 font-mono text-xs font-semibold">
+          +GATES ACTIVAS</div>                                           
+      460 +  <div class="text-gray-400 text-xs">Condiciones que deben cum
+          +plirse antes de cruzar una frontera arquitectural</div>       
+      461 +</div>                                                        
+      462  
+      614 -<div class="mt-2 space-y-1 text-sm text-gray-400">            
+      615 -<div>→ Un invariante no puede depender de sí mismo</div>      
+      616 -<div>→ Las gates protegen nodos hoja, no ciclos</div>         
+      617 -<div>→ El rollback es determinista: sin ambigüedad de orden</d
+          -iv>                                                           
+      463  </div>
+      464  
+      620 -<div class="mt-3 font-mono text-xs">                          
+      465 +</v-click>                                                    
+      466  
+      622 -```nu                                                         
+      623 -# Invariantes activos                                         
+      624 -nickel export .ontology/core.ncl                              
+      625 -| from json | get nodos                                       
+      626 -| where invariante == true                                    
+      627 -| select id descripcion                                       
+      628 -```                                                           
+      467 +<v-click>                                                     
+      468  
+      469 +<div class="mt-4 text-sm text-gray-300">                      
+      470 +  Es un grafo consultable. No prosa.<br>                      
+      471 +  <span class="text-gray-500 text-xs">La IA puede leerlo. Tu C
+          +I puede verificarlo. Tú puedes razonarlo.</span>              
+      472  </div>
+      473  
+      474  </v-click>
+     ...
+      480  <Footer />
+      481  
+      482  <!--
+      641 -State machines are architecture, not documentation.           
+      642 -The ADR FSM has 3 states and 2 transitions — that simplicity i
+          -s the power.                                                  
+      643 -The ontology DAG is the dependency graph of semantic constrain
+          -ts, not a diagram.                                            
+      483 +La diferencia no es el nivel de detalle — es la estructura.   
+      484 +Un grafo tiene propiedades que la prosa no tiene: es consultab
+          +le, verificable, traversable.                                 
+      485  -->
+      486  
+      487  ---
+      488  layout: default
+      489  ---
+      490  
+      650 -# La Ontología Operacional — el grafo operacional             
+      491 +# Ontoref — tres archivos, tres preguntas                     
+      492  
+      652 -<div class="grid grid-cols-2 gap-8 mt-4">                     
+      493 +<div class="grid grid-cols-3 gap-5 mt-6">                     
+      494  
+      654 -<div>                                                         
+      495 +<div class="border border-gray-700 rounded p-4 bg-gray-900">  
+      496  
+      656 -**`core.ncl` — lo que no puede cambiar**                      
+      497 +<div class="text-blue-300 font-mono font-semibold text-sm mb-3
+          +">core.ncl</div>                                              
+      498 +<div class="text-orange-400 text-xs mb-3 italic">"¿Qué ES este
+          + proyecto?"</div>                                             
+      499  
+      658 -<div class="mt-2 font-mono text-xs">                          
+      500 +<div class="font-mono text-xs text-gray-400">                 
+      501  
+      502  ```text
+      661 -# .ontology/core.ncl                                          
+      662 -{                                                             
+      663 -  nodos = [                                                   
+      664 -    {                                                         
+      665 -      id          = "backend-agnostic-core",                  
+      666 -      invariante  = true,                                     
+      667 -      descripcion = "business logic sin deps de UI",          
+      668 -      tensiones   = ["nickel-as-canonical-schema"],           
+      669 -    },                                                        
+      670 -  ],                                                          
+      671 -  practicas = [...],                                          
+      672 -}                                                             
+      503 +invariantes:                                                  
+      504 +  - backend-agnostic-core                                     
+      505 +  - zero-external-runtime                                     
+      506 +                                                              
+      507 +tensiones:                                                    
+      508 +  - formalization vs adoption-friction                        
+      509 +  - ontology vs reflection                                    
+      510 +                                                              
+      511 +prácticas:                                                    
+      512 +  - nickel-as-schema-layer                                    
+      513 +  - nushell-as-operator                                       
+      514  ```
+      515  
+      516  </div>
+      517  
+      677 -<div class="mt-3 text-sm text-gray-400">                      
+      678 -Tocar un nodo con <code class="text-orange-300">invariante = t
+          -rue</code> requiere un nuevo ADR. Sin excepción.              
+      518 +<div class="mt-3 text-xs text-gray-600">                      
+      519 +Inmutable sin ADR nuevo.                                      
+      520  </div>
+      521  
+      522  </div>
+      523  
+      683 -<div>                                                         
+      524 +<div class="border border-gray-700 rounded p-4 bg-gray-900">  
+      525  
+      685 -**`state.ncl` y `gate.ncl` — dónde estamos**                  
+      526 +<div class="text-orange-300 font-mono font-semibold text-sm mb
+          +-3">state.ncl</div>                                           
+      527 +<div class="text-orange-400 text-xs mb-3 italic">"¿Dónde ESTAM
+          +OS?"</div>                                                    
+      528  
+      687 -<v-click>                                                     
+      529 +<div class="font-mono text-xs text-gray-400">                 
+      530  
+      689 -<div class="mt-2 font-mono text-xs">                          
+      531 +```text                                                       
+      532 +dimensiones:                                                  
+      533 +  backend-maturity:                                           
+      534 +    actual: multi-stable                                      
+      535 +    deseado: all-production                                   
+      536 +    bloqueador: missing-nats-tests                            
+      537  
+      691 -```nu                                                         
+      692 -nickel export .ontology/state.ncl                             
+      693 -| from json | get dimensiones                                 
+      694 -| select id estado_actual estado_deseado                      
+      538 +  nickel-depth:                                               
+      539 +    actual: schema-input                                      
+      540 +    deseado: bidirectional                                    
+      541 +    catalizador: export pipeline                              
+      542  ```
+      543  
+      544  </div>
+      545  
+      699 -<div class="mt-2 font-mono text-xs text-gray-400">            
+      546 +<div class="mt-3 text-xs text-gray-600">                      
+      547 +Evoluciona con el proyecto.                                   
+      548 +</div>                                                        
+      549  
+      701 -```text                                                       
+      702 -backend-maturity         multi-stable  → all-production       
+      703 -nickel-integration-depth schema-input  → bidirectional        
+      704 -```                                                           
+      705 -                                                              
+      550  </div>
+      551  
+      708 -</v-click>                                                    
+      552 +<div class="border border-gray-700 rounded p-4 bg-gray-900">  
+      553  
+      710 -<v-click>                                                     
+      554 +<div class="text-green-300 font-mono font-semibold text-sm mb-
+          +3">gate.ncl</div>                                             
+      555 +<div class="text-orange-400 text-xs mb-3 italic">"¿Cuándo esta
+          +mos LISTOS?"</div>                                            
+      556  
+      712 -<div class="mt-4 font-mono text-xs">                          
+      557 +<div class="font-mono text-xs text-gray-400">                 
+      558  
+      714 -```nu                                                         
+      715 -nickel export .ontology/gate.ncl                              
+      716 -| from json | get membranas                                   
+      717 -| where activa == true                                        
+      718 -| select id permeabilidad protege                             
+      559 +```text                                                       
+      560 +gates:                                                        
+      561 +  core-dependency-gate:                                       
+      562 +    activa: true                                              
+      563 +    permeabilidad: Low                                        
+      564 +    protege:                                                  
+      565 +      - backend-agnostic-core                                 
+      566 +                                                              
+      567 +  nickel-primacy-gate:                                        
+      568 +    activa: true                                              
+      569 +    permeabilidad: Low                                        
+      570 +    protege:                                                  
+      571 +      - nickel-as-schema-layer                                
+      572  ```
+      573  
+      574  </div>
+      575  
+      723 -<div class="mt-2 font-mono text-xs text-gray-400">            
+      724 -                                                              
+      725 -```text                                                       
+      726 -core-dependency-gate  Low  [backend-agnostic-core, ...]       
+      727 -nickel-primacy-gate   Low  [nickel-as-canonical-schema]       
+      728 -```                                                           
+      729 -                                                              
+      576 +<div class="mt-3 text-xs text-gray-600">                      
+      577 +Condiciones de cruce.                                         
+      578  </div>
+      579  
+      732 -</v-click>                                                    
+      733 -                                                              
+      734 -<div class="mt-3 text-xs text-gray-500">                      
+      735 -La ontología no es un diagrama.<br>                           
+      736 -Es un grafo consultable que el sistema y los agentes leen.    
+      580  </div>
+      581  
+      582  </div>
+      583  
+      584 +<v-click>                                                     
+      585 +                                                              
+      586 +<div class="mt-5 font-mono text-xs text-center text-gray-500">
+      587 +  Nickel tipado · exportable como JSON · consultable desde Nu 
+          +· verificable en CI                                           
+      588  </div>
+      589  
+      590 +</v-click>                                                    
+      591 +                                                              
+      592  <Footer />
+      593  
+      594  <!--
+      746 -The .ontology/ directory separates three orthogonal concerns i
+          -n three files:                                                
+      747 -  core.ncl  — what the project IS: invariant axioms and struct
+          -ural tensions.                                                
+      748 -  state.ncl — where it IS vs where it wants to BE.            
+      749 -  gate.ncl  — when it is READY to cross a boundary.           
+      750 -                                                              
+      751 -reflection/ reads all three and answers self-knowledge queries
+          -.                                                             
+      752 -This separation allows an agent to understand the project with
+          -out reading code —                                            
+      753 -only by consulting the declarative graph.                     
+      595 +Tres preguntas, tres responsabilidades. La separación es delib
+          +erada.                                                        
+      596 +core.ncl responde al SER. state.ncl al DEVENIR. gate.ncl al CR
+          +UZAR.                                                         
+      597 +Cada uno puede cambiar a ritmos diferentes sin contaminar los 
+          +otros.                                                        
+      598  -->
+      599  
+      600  ---
+      601  layout: default
+      602  ---
+      603  
+      760 -# Reflection — modos, formas, módulos, perfiles               
+      604 +# Reflection — lo que se CONVIERTE                            
+      605  
+      762 -<div class="grid grid-cols-2 gap-8 -mt-5">                    
+      606 +<div class="grid grid-cols-2 gap-8 mt-2">                     
+      607  
+      608  <div>
+      609  
+      766 -**Modo: especificación ejecutable**                           
+      610 +**Ontología = lo que ES**                                     
+      611  
+      768 -<div class="mt-1 font-mono text-xs">                          
+      769 -                                                              
+      770 -```bash                                                       
+      771 -# reflection/modes/new_service.ncl                            
+      772 -{                                                             
+      773 -  id      = "new_service",                                    
+      774 -  trigger = "When adding a service",                          
+      775 -  steps   = [                                                 
+      776 -    {                                                         
+      777 -      id       = "create-adr",                                
+      778 -      actor    = "developer",                                 
+      779 -      action   = "Draft ADR for service boundary",            
+      780 -      cmd      = "stratum form new_adr",                      
+      781 -      on_error = { strategy = "abort" },                      
+      782 -    },                                                        
+      783 -    {                                                         
+      784 -      id         = "seal-config",                             
+      785 -      actor      = "developer",                               
+      786 -      action     = "Apply initial config seal",               
+      787 -      cmd        = "stratum config apply development",        
+      788 -      depends_on = [{ step = "create-adr" }],                 
+      789 -      on_error   = { strategy = "abort" },                    
+      790 -    },                                                        
+      791 -  ],                                                          
+      792 -}                                                             
+      793 -```                                                           
+      794 -                                                              
+      612 +<div class="mt-3 text-sm text-gray-400 space-y-2">            
+      613 +<div>→ Invariantes — verdad estructural</div>                 
+      614 +<div>→ Tensiones — trade-offs vivos</div>                     
+      615 +<div>→ Estado — dónde estamos ahora</div>                     
+      616  </div>
+      617  
+      797 -</div>                                                        
+      618 +<div class="mt-5 text-sm font-semibold text-gray-200">Reflecti
+          +on = lo que DEVIENE</div>                                     
+      619  
+      799 -<div>                                                         
+      800 -                                                              
+      801 -**Forma: entrada estructurada tipada**                        
+      802 -                                                              
+      803 -<div class="mt-2 font-mono text-xs">                          
+      804 -                                                              
+      805 -```bash                                                       
+      806 -# reflection/forms/register_change.ncl                        
+      807 -{                                                             
+      808 -  elements = [                                                
+      809 -    { type = "section_header", label = "Change" },            
+      810 -    {                                                         
+      811 -      type     = "text",                                      
+      812 -      id       = "title",                                     
+      813 -      label    = "Title",                                     
+      814 -      required = true,                                        
+      815 -    },                                                        
+      816 -    {                                                         
+      817 -      type    = "select",                                     
+      818 -      id      = "change_type",                                
+      819 -      options = ["feature","fix","refactor","config"],        
+      820 -    },                                                        
+      821 -  ]                                                           
+      822 -}                                                             
+      823 -```                                                           
+      824 -                                                              
+      620 +<div class="mt-3 text-sm text-gray-400 space-y-2">            
+      621 +<div>→ Modos operacionales — secuencias con DAG de dependencia
+          +s</div>                                                       
+      622 +<div>→ ADRs con constraints tipadas — decisiones vivas, no pro
+          +sa</div>                                                      
+      623 +<div>→ Backlog estructurado — tensiones → tareas</div>        
+      624 +<div>→ Historial de sesiones — quién hizo qué y cuándo</div>  
+      625  </div>
+      626  
+      627  <v-click>
+      628  
+      829 -<div class="mt-3 space-y-2 text-sm">                          
+      830 -                                                              
+      831 -<div class="box-highlight border-l-4 border-purple-500 pl-3 py
+          --1">                                                          
+      832 -  <div class="font-semibold text-purple-300">Perfil de config 
+          -= estado sellado</div>                                        
+      833 -  <div class="text-gray-400 text-xs"><code>sha256(nickel expor
+          -t profile.ncl)</code> — drift detection instantánea</div>     
+      629 +<div class="mt-5 border-l-4 border-orange-500 pl-4 text-sm tex
+          +t-gray-300">                                                  
+      630 +  La ontología captura el <em>ser</em>.<br>                   
+      631 +  La reflection captura el <em>operar sobre ese ser</em>.     
+      632  </div>
+      633  
+      836 -<div class="text-xs text-gray-400 mt-2">                      
+      837 -<code>stratum config verify development</code> → compara hash 
+          -actual con hash sellado                                       
+      838 -</div>                                                        
+      839 -                                                              
+      840 -</div>                                                        
+      841 -                                                              
+      634  </v-click>
+      635  
+      636  </div>
+      637  
+      846 -</div>                                                        
+      847 -                                                              
+      848 -<Footer />                                                    
+      849 -                                                              
+      850 -<!--                                                          
+      851 -Modes are not runbooks. They are typed, versioned, queryable o
+          -perational specifications.                                    
+      852 -Forms are the structured capture layer — TypeDialog renders th
+          -em across CLI/TUI/Web backends.                               
+      853 -Config profiles seal the observable state with a cryptographic
+          - fingerprint.                                                 
+      854 --->                                                           
+      855 -                                                              
+      856 ----                                                           
+      857 -layout: default                                               
+      858 ----                                                           
+      859 -                                                              
+      860 -# ADRs vivos                                                  
+      861 -## — de documento muerto a constraint machine                 
+      862 -                                                              
+      863 -<div class="grid grid-cols-2 gap-8 mt-1">                     
+      864 -                                                              
+      638  <div>
+      639  
+      867 -**Estructura Nickel de un ADR**                               
+      640 +<v-click>                                                     
+      641  
+      869 -<div class="-mt-2 font-mono text-xs">                         
+      642 +**Un modo como especificación ejecutable**                    
+      643  
+      644 +<div class="mt-3 font-mono text-xs">                          
+      645 +                                                              
+      646  ```bash
+      872 -# adrs/adr-001-nickel-as-canonical.ncl                        
+      873 -let s = import "adrs/schema.ncl" in                           
+      647 +# reflection/modes/new_service.ncl                            
+      648  {
+      875 -  id      = "adr-001",                                        
+      876 -  title   = "Nickel as First-Class Form Definition",          
+      877 -  status  = "Accepted",                                       
+      878 -  date    = "2025-12-01",                                     
+      879 -  constraints = [                                             
+      649 +  id      = "new_service",                                    
+      650 +  trigger = "When adding a new service boundary",             
+      651 +  steps   = [                                                 
+      652      {
+      881 -      id       = "adr-001-c1",                                
+      882 -      severity = "Hard",                                      
+      883 -      claim    =                                              
+      884 -        "nickel export must succeed before "                  
+      885 -         ++ "any user interaction",                           
+      653 +      id     = "draft-adr",                                   
+      654 +      actor  = "developer",                                   
+      655 +      action = "Draft ADR for service boundary decision",     
+      656      },
+      657 +    {                                                         
+      658 +      id         = "verify-invariants",                       
+      659 +      actor      = "agent",                                   
+      660 +      action     = "Check no invariant is violated",          
+      661 +      depends_on = [{ step = "draft-adr" }],                  
+      662 +    },                                                        
+      663 +    {                                                         
+      664 +      id         = "seal-config",                             
+      665 +      actor      = "developer",                               
+      666 +      action     = "Apply initial config seal",               
+      667 +      depends_on = [{ step = "verify-invariants" }],          
+      668 +    },                                                        
+      669    ],
+      888 -  superseded_by = "",                                         
+      889 -} | s.ADR                                                     
+      670 +}                                                             
+      671  ```
+      672  
+      673  </div>
+      674  
+      675 +<div class="mt-2 text-xs text-gray-500">                      
+      676 +No es un runbook. Es un grafo acíclico tipado.<br>            
+      677 +La IA lo ejecuta. El sistema lo valida.                       
+      678  </div>
+      679  
+      896 -<div>                                                         
+      897 -                                                              
+      898 -<p class="-mt-2 text-sm font-semibold text-gray-200"> Ciclo de
+          - vida y consulta </p>                                         
+      899 -                                                              
+      900 -<div class="-mt-2 font-mono text-xs text-gray-300">           
+      901 -                                                              
+      902 -```text                                                       
+      903 -Proposed ──────▶ Accepted ──────▶ Superseded                  
+      904 -  │                                   ▲                       
+      905 -  │            new ADR                │                       
+      906 -  └──────── superseded_by ────────────┘                       
+      907 -```                                                           
+      908 -                                                              
+      909 -</div>                                                        
+      910 -                                                              
+      911 -<v-click>                                                     
+      912 -                                                              
+      913 -<div class="mt-2 text-sm font-semibold text-gray-200">Constrai
+          -nts Hard activas del sistema</div>                            
+      914 -                                                              
+      915 -<div class="mt-2 font-mono text-xs">                          
+      916 -                                                              
+      917 -```nu                                                         
+      918 -glob "adrs/adr-*.ncl"                                         
+      919 -| each { |f|                                                  
+      920 -    nickel export $f | complete                               
+      921 -    | if $in.exit_code == 0 {                                 
+      922 -        $in.stdout | from json                                
+      923 -    } else { null }                                           
+      924 -  }                                                           
+      925 -| compact                                                     
+      926 -| where status == "Accepted"                                  
+      927 -| get constraints | flatten                                   
+      928 -| where severity == "Hard"                                    
+      929 -| select id claim                                             
+      930 -```                                                           
+      931 -                                                              
+      932 -</div>                                                        
+      933 -                                                              
+      680  </v-click>
+      681  
+      936 -<v-click>                                                     
+      937 -                                                              
+      938 -<div class="mt-3 text-xs text-gray-500">                      
+      939 -El conjunto activo de constraints es el contrato en vigor — no
+          - la prosa del ADR.                                            
+      940 -La prosa explica el porqué. Los constraints enforcan el qué.  
+      682  </div>
+      683  
+      943 -</v-click>                                                    
+      944 -                                                              
+      684  </div>
+      685  
+      947 -</div>                                                        
+      948 -                                                              
+      686  <Footer />
+      687  
+      688  <!--
+      952 -"Living ADRs" is not a metaphor.                              
+      953 -The machine can query them, validate against them, and enforce
+          - them.                                                        
+      954 -The prose explains why. The constraints enforce what.         
+      689 +La dualidad ontología/reflection es la tensión central de onto
+          +ref.                                                          
+      690 +No pueden fusionarse — tienen ritmos de cambio diferentes y au
+          +diencias diferentes.                                          
+      691 +La reflection sin ontología es procedimiento sin semántica. La
+          + ontología sin reflection es diagrama muerto.                 
+      692  -->
+      693  
+      694  ---
+      695  layout: default
+      696  ---
+      697  
+      961 -# Implementación — las partes y cómo encajan                  
+      698 +# La IA como actor, no como oráculo                           
+      699  
+      700  <div class="grid grid-cols-2 gap-8 mt-4">
+      701  
+      702  <div>
+      703  
+      967 -**Stack de archivos operativos**                              
+      704 +**Sin ontología formalizada**                                 
+      705  
+      969 -<div class="mt-2 font-mono text-xs text-gray-300">            
+      706 +<div class="mt-3 font-mono text-xs">                          
+      707  
+      708  ```bash
+      972 -stratum.sh                 ← entry point, locking, env        
+      973 -  └─ reflection/bin/stratum.nu  ← dispatcher                  
+      974 -       ├─ modules/adr.nu        ← ADR lifecycle               
+      975 -       ├─ modules/register.nu   ← CHANGELOG + ontology        
+      976 -       ├─ modules/config.nu     ← sealed profiles             
+      977 -       ├─ modules/backlog.nu    ← backlog items               
+      978 -       └─ modules/forms.nu      ← TypeDialog integrated       
+      709 +# Lo que el agente recibe                                     
+      710 +"Ayúdame a escalar este servicio"                             
+      711  
+      980 -.ontology/                 ← queryable truth                  
+      981 -  ├─ core.ncl              ← invariants                       
+      982 -  ├─ state.ncl             ← dimensions                       
+      983 -  └─ gate.ncl              ← active guards                    
+      712 +# Lo que el agente asume                                      
+      713 +- que puede añadir réplicas libremente                        
+      714 +- que el estado actual es el deseado                          
+      715 +- que no hay restricciones de org                             
+      716 +- que la decisión de diciembre es irrelevante                 
+      717  
+      985 -adrs/                      ← typed decisions                  
+      986 -reflection/configs/        ← sealed config history            
+      718 +# Lo que genera                                               
+      719 +resource "google_compute_instance" {                          
+      720 +  count = 5  # razonable, ¿no?                                
+      721 +}                                                             
+      722  ```
+      723  
+      724  </div>
+     ...
+       727  
+       728  <div>
+       729  
+       995 -**Flujo de escritura protegida**                             
+       996 -                                                             
+       730  <v-click>
+       731  
+       999 -<div class="mt-2 space-y-2 text-sm">                         
+       732 +**Con ontología formalizada**                                
+       733  
+      1001 -<div class="box-highlight border-l-4 border-gray-600 pl-3 py-
+           -1">                                                          
+      1002 -  <div class="font-semibold text-gray-200">stratum.sh</div>  
+      1003 -  <div class="text-gray-400 text-xs">detect actor → acquire l
+           -ock (mkdir) → delegate → release</div>                       
+      1004 -</div>                                                       
+       734 +<div class="mt-3 font-mono text-xs">                         
+       735  
+      1006 -<div class="box-highlight border-l-4 border-orange-500 pl-3 p
+           -y-1">                                                        
+      1007 -  <div class="font-semibold text-orange-300">Lock resources</
+           -div>                                                         
+      1008 -  <div class="text-gray-400 text-xs"><code>manifest</code> (c
+           -onfig apply/rollback) · <code>changelog</code> (register) · <
+           -code>backlog</code> (done/cancel)</div>                      
+      1009 -</div>                                                       
+       736 +```bash                                                      
+       737 +# Lo que el agente recibe (SessionStart hook)                
+       738 +INVARIANT: max-replicas-per-zone = 2                         
+       739 +TENSION: scaling vs consistency [unresolved]                 
+       740 +CONSTRAINT (adr-003): no horizontal scaling                  
+       741 +  without explicit capacity review                           
+       742 +STATE: backend-maturity = single-stable                      
+       743 +  → desired: multi-stable                                    
+       744 +  → blocker: capacity-review-pending                         
+       745  
+      1011 -<div class="box-highlight border-l-4 border-blue-500 pl-3 py-
+           -1">                                                          
+      1012 -  <div class="font-semibold text-blue-300">Timestamp IDs</div
+           ->                                                            
+      1013 -  <div class="text-gray-400 text-xs"><code>cfg-20260309T04520
+           -6-developer</code> — colisión imposible, sin TOCTOU</div>    
+      1014 -</div>                                                       
+       746 +# Lo que el agente puede hacer                               
+       747 +- Preguntar si procede dado el bloqueador                    
+       748 +- Generar la config dentro de los límites                    
+       749 +- Indicar que el ADR debe revisarse antes                    
+       750 +```                                                          
+       751  
+       752  </div>
+       753  
+     ...
+       755  
+       756  <v-click>
+       757  
+      1022 -<div class="mt-4 font-mono text-xs text-gray-400">           
+      1023 -                                                             
+      1024 -```bash                                                      
+      1025 -# Advisory lock: mkdir es POSIX-atómico                      
+      1026 -mkdir .stratumiops/locks/manifest.lock 2>/dev/null           
+      1027 -echo "$$:developer:$(date -u +%Y%m%dT%H%M%SZ)"               
+      1028 -    > .stratumiops/locks/manifest.lock/owner                 
+      1029 -```                                                          
+      1030 -                                                             
+       758 +<div class="mt-4 border-l-4 border-green-500 pl-4 text-sm tex
+           +t-gray-300">                                                 
+       759 +  El agente opera sobre lo que tú ya clarificaste.<br>       
+       760 +  <span class="text-gray-500 text-xs">Deja de inventarse cont
+           +exto porque tú dejaste de tenerlo implícito.</span>          
+       761  </div>
+       762  
+       763  </v-click>
+     ...
+       769  <Footer />
+       770  
+       771  <!--
+      1042 -Implementation is derived from architecture — not the other w
+           -ay around.                                                   
+      1043 -Every piece exists because a specific problem required it.   
+      1044 -The lock is 4 lines of bash because mkdir is atomic on POSIX.
+           - No daemon needed.                                           
+       772 +El SessionStart hook en ontoref inyecta el contexto ontológic
+           +o al inicio de cada sesión.                                  
+       773 +El agente no es más inteligente — tiene más información real 
+           +y menos que inventar.                                        
+       774 +La mejora no es de la IA. Es tuya: formalizaste algo que ante
+           +s vivía solo en tu cabeza.                                   
+       775  -->
+       776  
+       777  ---
+       778  layout: default
+       779  ---
+       780  
+      1051 -# Mecánica y flujo — casos de uso reales                     
+       781 +# ¿Quién mantiene esto actualizado?                          
+       782  
+      1053 -<div class="grid grid-cols-2 gap-8 mt-4">                    
+       783 +<div class="mt-4 text-lg text-gray-400 text-center">La pregun
+           +ta que siempre sale.</div>                                   
+       784  
+       785 +<div class="grid grid-cols-2 gap-8 mt-6">                    
+       786 +                                                             
+       787  <div>
+       788  
+      1057 -**Flujo: registrar un cambio**                               
+       789 +**El problema real con el README**                           
+       790  
+      1059 -<div class="mt-2 text-xs space-y-1">                         
+       791 +<div class="mt-3 space-y-2 text-sm text-gray-400">           
+       792  
+      1061 -<div class="flex items-start gap-2 mt-1">                    
+      1062 -  <span class="text-orange-400 font-mono font-bold">①</span> 
+      1063 -  <div><code>stratum register</code> → forma TypeDialog inter
+           -activa</div>                                                 
+      1064 -</div>                                                       
+      1065 -                                                             
+       793  <div class="flex items-start gap-2">
+      1067 -  <span class="text-orange-400 font-mono font-bold">②</span> 
+      1068 -  <div>Clasificar: tipo de cambio, impacto, ADR relacionado</
+           -div>                                                         
+       794 +  <span class="text-red-400">✗</span>                        
+       795 +  <div>Nadie tiene incentivo para actualizarlo — no está en e
+           +l path crítico</div>                                         
+       796  </div>
+       797  
+       798  <div class="flex items-start gap-2">
+      1072 -  <span class="text-orange-400 font-mono font-bold">③</span> 
+      1073 -  <div>Artefacto: entrada CHANGELOG + stub ADR si aplica</div
+           ->                                                            
+       799 +  <span class="text-red-400">✗</span>                        
+       800 +  <div>No hay señal de cuando está desactualizado</div>      
+       801  </div>
+       802  
+       803  <div class="flex items-start gap-2">
+      1077 -  <span class="text-orange-400 font-mono font-bold">④</span> 
+      1078 -  <div>Verificar: <code>nickel typecheck</code> debe pasar</d
+           -iv>                                                          
+       804 +  <span class="text-red-400">✗</span>                        
+       805 +  <div>Actualizar requiere fricción de escritura libre</div> 
+       806  </div>
+       807  
+       808  <div class="flex items-start gap-2">
+      1082 -  <span class="text-orange-400 font-mono font-bold">⑤</span> 
+      1083 -  <div>Sellar si afecta config: <code>stratum config apply</c
+           -ode></div>                                                   
+       809 +  <span class="text-red-400">✗</span>                        
+       810 +  <div>No hay verificación — puede divergir silenciosamente</
+           +div>                                                         
+       811  </div>
+       812  
+       813  </div>
+       814  
+      1088 -<v-click>                                                    
+      1089 -                                                             
+      1090 -<div class="mt-4 text-sm font-semibold text-gray-200">Detecci
+           -ón de drift</div>                                            
+      1091 -                                                             
+      1092 -<div class="mt-2 font-mono text-xs">                         
+      1093 -                                                             
+      1094 -```nu                                                        
+      1095 -stratum config apply production --adr adr-006 --pr 42        
+      1096 -# Meses después:                                             
+      1097 -stratum config verify production                             
+      1098 -# → ✓ verified  a3f7c12d8b4e9f01...                          
+      1099 -# → ✗ DRIFT DETECTED stored vs current hash                  
+      1100 -```                                                          
+      1101 -                                                             
+       815  </div>
+       816  
+      1104 -</v-click>                                                   
+      1105 -                                                             
+      1106 -</div>                                                       
+      1107 -                                                             
+       817  <div>
+       818  
+      1110 -**Rollback verificado**                                      
+      1111 -                                                             
+       819  <v-click>
+       820  
+      1114 -<div class="mt-2 font-mono text-xs">                         
+       821 +**El diseño de ontoref**                                     
+       822  
+      1116 -```nu                                                        
+      1117 -stratum config history production                            
+       823 +<div class="mt-3 space-y-2 text-sm">                         
+       824  
+      1119 -stratum config rollback production \                         
+      1120 -    cfg-20260301T120000-developer  \                         
+      1121 -    --adr adr-007 \                                          
+      1122 -    --note "revert breaking change"                          
+      1123 -```                                                          
+      1124 -                                                             
+       825 +<div class="border-l-4 border-blue-400 pl-3 py-1 bg-gray-900 
+           +rounded-r">                                                  
+       826 +  <div class="text-blue-300 font-semibold text-xs">Pre-commit
+           + como sincronización forzada</div>                           
+       827 +  <div class="text-gray-400 text-xs"><code>nickel typecheck</
+           +code> en cada commit — la capa declarativa nunca queda rota</
+           +div>                                                         
+       828  </div>
+       829  
+      1127 -</v-click>                                                   
+      1128 -                                                             
+      1129 -<v-click>                                                    
+      1130 -                                                             
+      1131 -<div class="mt-3 text-xs text-gray-400 space-y-1">           
+      1132 -<div>→ <code>snapshot_hash</code> verifica integridad antes d
+           -e restaurar</div>                                            
+      1133 -<div>→ Rollback genera nuevo sello — la historia es append-on
+           -ly</div>                                                     
+      1134 -<div>→ Trazabilidad completa: ADR + PR + bug en cada sello</d
+           -iv>                                                          
+       830 +<div class="border-l-4 border-orange-400 pl-3 py-1 bg-gray-90
+           +0 rounded-r">                                                
+       831 +  <div class="text-orange-300 font-semibold text-xs">Modos co
+           +mo procedimientos obligatorios</div>                         
+       832 +  <div class="text-gray-400 text-xs">Añadir un servicio ejecu
+           +ta el modo — el ADR se escribe como parte del proceso</div>  
+       833  </div>
+       834  
+      1137 -<div class="mt-4 font-mono text-xs text-gray-500">           
+      1138 -                                                             
+      1139 -```text                                                      
+      1140 -cfg-20260301T120000-developer  (original)                    
+      1141 -  └─ cfg-20260309T045206-agent  (change)                     
+      1142 -       └─ cfg-20260309T120000-developer  (rollback)          
+      1143 -```                                                          
+      1144 -                                                             
+       835 +<div class="border-l-4 border-purple-400 pl-3 py-1 bg-gray-90
+           +0 rounded-r">                                                
+       836 +  <div class="text-purple-300 font-semibold text-xs">Drift de
+           +tection</div>                                                
+       837 +  <div class="text-gray-400 text-xs"><code>sha256(nickel expo
+           +rt)</code> — sabes cuándo el estado sellado difiere del actua
+           +l</div>                                                      
+       838  </div>
+       839  
+      1147 -</v-click>                                                   
+      1148 -                                                             
+       840 +<div class="border-l-4 border-green-400 pl-3 py-1 bg-gray-900
+           + rounded-r">                                                 
+       841 +  <div class="text-green-300 font-semibold text-xs">ADRs como
+           + constraints, no prosa</div>                                 
+       842 +  <div class="text-gray-400 text-xs">El constraint es machine
+           +-readable — el CI puede verificarlo, no solo leerlo</div>    
+       843  </div>
+       844  
+       845  </div>
+       846  
+      1153 -<Footer />                                                   
+      1154 -                                                             
+      1155 -<!--                                                         
+      1156 -Walk through a real scenario. The audience needs to see the f
+           -low, not just the components.                                
+      1157 -Drift detection is the killer feature: no guessing what chang
+           -ed in production.                                            
+      1158 --->                                                          
+      1159 -                                                             
+      1160 ----                                                          
+      1161 -layout: default                                              
+      1162 ----                                                          
+      1163 -                                                             
+      1164 -# Integración en proyectos — onboarding al ecosistema        
+      1165 -                                                             
+      1166 -<div class="grid grid-cols-2 gap-8 mt-4">                    
+      1167 -                                                             
+      1168 -<div>                                                        
+      1169 -                                                             
+      1170 -**Lo que un proyecto necesita**                              
+      1171 -                                                             
+      1172 -<div class="mt-2 font-mono text-xs text-gray-300">           
+      1173 -                                                             
+      1174 -```bash                                                      
+      1175 -proyecto/                                                    
+      1176 -  .ontology/                                                 
+      1177 -    core.ncl       ← invariants del proyecto                 
+      1178 -    state.ncl      ← dimensiones de madurez                  
+      1179 -    gate.ncl       ← guards activos                          
+      1180 -  adrs/                                                      
+      1181 -    schema.ncl                                               
+      1182 -    reflection.ncl                                           
+      1183 -    adr-001-*.ncl  ← ADR fundacional                         
+      1184 -  reflection/                                                
+      1185 -    forms/         ← TypeDialog forms                        
+      1186 -    modes/         ← operational modes                       
+      1187 -    configs/       ← sealed profiles                         
+      1188 -    modules/       ← Nu modules                              
+      1189 -    bin/stratum.nu                                           
+      1190 -  stratum.sh                                                 
+      1191 -```                                                          
+      1192 -                                                             
+      1193 -</div>                                                       
+      1194 -                                                             
+      1195 -</div>                                                       
+      1196 -                                                             
+      1197 -<div>                                                        
+      1198 -                                                             
+      1199 -**Inicialización**                                           
+      1200 -                                                             
+      1201 -<v-click>                                                    
+      1202 -                                                             
+      1203 -<div class="mt-2 text-xs space-y-2">                         
+      1204 -                                                             
+      1205 -<div class="box-highlight border-l-4 border-green-500 pl-3 py
+           --1">                                                         
+      1206 -  <div class="font-semibold text-green-300">Comando de onboar
+           -ding</div>                                                   
+      1207 -  <div class="font-mono text-gray-400 text-xs mt-1"><code>/on
+           -board-project</code> → genera <code>.ontology/ <br></code> + 
+           -ADR fundacional</div>                                        
+      1208 -</div>                                                       
+      1209 -                                                             
+      1210 -<div class="box-highlight border-l-4 border-blue-500 pl-3 py-
+           -1">                                                          
+      1211 -  <div class="font-semibold text-blue-300">Primer ADR obligat
+           -orio</div>                                                   
+      1212 -  <div class="text-gray-400 text-xs">Define las invariantes —
+           - qué no puede cambiar sin ADR nuevo</div>                    
+      1213 -</div>                                                       
+      1214 -                                                             
+      1215 -<div class="box-highlight border-l-4 border-orange-400 pl-3 p
+           -y-1">                                                        
+      1216 -  <div class="font-semibold text-orange-300">Pre-commit como 
+           -sincronización</div>                                         
+      1217 -  <div class="text-gray-400 text-xs"><code>nickel typecheck</
+           -code> en cada commit<br> — la capa declarativa nunca queda ro
+           -ta</div>                                                     
+      1218 -</div>                                                       
+      1219 -                                                             
+      1220 -</div>                                                       
+      1221 -                                                             
+       847  </v-click>
+       848  
+       849  <v-click>
+       850  
+      1226 -<div class="mt-4 font-mono text-xs">                         
+      1227 -                                                             
+      1228 -```bash                                                      
+      1229 -./stratum.sh check                                           
+      1230 -# ✓ nushell 0.110.0                                          
+      1231 -# ✓ nickel 1.9.0                                             
+      1232 -# ✓ .ontology/core.ncl exportable                            
+      1233 -# ✓ adrs/adr-001 Accepted                                    
+      1234 -# ✗ no sealed config profiles yet                            
+      1235 -```                                                          
+      1236 -                                                             
+       851 +<div class="mt-4 text-xs text-gray-500 border-l-2 border-gray
+           +-700 pl-3">                                                  
+       852 +  La ontología se mantiene porque es el path.<br>            
+       853 +  No porque alguien sea disciplinado.                        
+       854  </div>
+       855  
+       856  </v-click>
+     ...
+       862  <Footer />
+       863  
+       864  <!--
+      1248 -Onboarding is a contract, not a tutorial.                    
+      1249 -The system tells the project what it needs via `stratum check
+           -`.                                                           
+      1250 -The foundational ADR is the project's semantic constitution. 
+       865 +Esta es la objeción más honesta del público. Abrazarla, no es
+           +quivarla.                                                    
+       866 +La respuesta no es "la gente debería ser más disciplinada".  
+       867 +La respuesta es: el sistema lo hace obligatorio, no opcional.
+       868  -->
+       869  
+       870  ---
+       871  layout: default
+       872  ---
+       873  
+      1257 -# Integración con Claude Code & Uso concurrente              
+       874 +# Para el debate                                             
+       875  
+      1259 -<div class="grid grid-cols-2 gap-8 mt-4">                    
+       876 +<div class="grid grid-cols-2 gap-8 mt-6">                    
+       877  
+       878  <div>
+       879  
+      1263 -**SessionStart: contexto automático**                        
+       880 +**Propuesta 1 — Ontología**                                  
+       881  
+      1265 -<div class="mt-2 text-xs text-gray-300">                     
+      1266 -<code>stratum-session-start.sh</code> se ejecuta al inicio de
+           - cada sesión:                                                
+       882 +<div class="mt-4 space-y-4">                                 
+       883 +                                                             
+       884 +<div class="border border-gray-700 rounded p-3 bg-gray-900"> 
+       885 +  <div class="text-orange-300 text-sm font-semibold">¿No es s
+           +uficiente con un buen README?</div>                          
+       886 +  <div class="text-gray-500 text-xs mt-1">¿Cuándo fue la últi
+           +ma vez que el tuyo estaba al día?</div>                      
+       887  </div>
+       888  
+      1269 -<div class="mt-2 font-mono text-xs">                         
+       889 +<div class="border border-gray-700 rounded p-3 bg-gray-900"> 
+       890 +  <div class="text-orange-300 text-sm font-semibold">¿Qué dif
+           +erencia una ontología de una lista de decisiones?</div>      
+       891 +  <div class="text-gray-500 text-xs mt-1">Estructura, consult
+           +abilidad, verificabilidad.</div>                             
+       892 +</div>                                                       
+       893  
+      1271 -```bash                                                      
+      1272 -=== STRATUM CONTEXT: typedialog ===                          
+       894 +<div class="border border-gray-700 rounded p-3 bg-gray-900"> 
+       895 +  <div class="text-orange-300 text-sm font-semibold">¿Quién m
+           +antiene esto cuando el proyecto cambia rápido?</div>         
+       896 +  <div class="text-gray-500 text-xs mt-1">El diseño lo hace o
+           +bligatorio, no opcional.</div>                               
+       897 +</div>                                                       
+       898  
+      1274 -ADRS                                                         
+      1275 -  adr-001 [Accepted] Nickel as First-Class Form...           
+      1276 -  adr-002 [Accepted] NCL element order from position         
+      1277 -                                                             
+      1278 -HARD CONSTRAINTS                                             
+      1279 -  adr-001-c1  nickel export must succeed before...           
+      1280 -  adr-002-c1  web/tui must render headers interleaved        
+      1281 -                                                             
+      1282 -STATE                                                        
+      1283 -  backend-maturity: multi-stable → all-production            
+      1284 -                                                             
+      1285 -ACTIVE GATES                                                 
+      1286 -  nickel-primacy-gate [Low] Protects: nickel-as-...          
+      1287 -                                                             
+      1288 -=== END STRATUM CONTEXT ===                                  
+      1289 -```                                                          
+      1290 -                                                             
+       899  </div>
+       900  
+       901  </div>
+       902  
+       903  <div>
+       904  
+      1297 -**Uso concurrente: humano + agente IA**                      
+       905 +**Propuesta 2 — Infra + IA**                                 
+       906  
+      1299 -<v-click>                                                    
+       907 +<div class="mt-4 space-y-4">                                 
+       908  
+      1301 -<div class="mt-4 space-y-4 text-sm">                         
+      1302 -                                                             
+      1303 -<div class="box-highlight border-l-4 border-blue-500 pl-3 py-
+           -1">                                                          
+      1304 -  <div class="font-semibold text-blue-300">IDs timestamp con 
+           -actor</div>                                                  
+      1305 -  <div class="font-mono text-gray-400 text-xs">cfg-20260309T0
+           -45206-developer<br>cfg-20260309T045207-agent</div>           
+      1306 -  <div class="text-gray-400 text-xs">Sin coordinador. Sin TOC
+           -TOU. Trazabilidad por actor.</div>                           
+       909 +<div class="border border-gray-700 rounded p-3 bg-gray-900"> 
+       910 +  <div class="text-orange-300 text-sm font-semibold">¿Es dife
+           +rente a copiar Stack Overflow?</div>                         
+       911 +  <div class="text-gray-500 text-xs mt-1">En SO buscabas. Con
+           + la IA, delegas también la búsqueda.</div>                   
+       912  </div>
+       913  
+      1309 -<div class="box-highlight border-l-4 border-orange-400 pl-3 p
+           -y-1">                                                        
+      1310 -  <div class="font-semibold text-orange-300">Advisory locks (
+           -mkdir)</div>                                                 
+      1311 -  <div class="text-gray-400 text-xs">Serializa escrituras en 
+           -manifest/changelog/backlog. Stale detection por PID.</div>   
+       914 +<div class="border border-gray-700 rounded p-3 bg-gray-900"> 
+       915 +  <div class="text-orange-300 text-sm font-semibold">¿Dónde e
+           +stá el mínimo de comprensión necesario?</div>                
+       916 +  <div class="text-gray-500 text-xs mt-1">Saber qué no puede 
+           +romperse. Sin eso, no tienes control.</div>                  
+       917  </div>
+       918  
+      1314 -<div class="box-highlight border-l-4 border-gray-600 pl-3 py-
+           -1">                                                          
+      1315 -  <div class="font-semibold text-gray-300">Actor-agnostic pro
+           -tocol</div>                                                  
+      1316 -  <div class="text-gray-400 text-xs">El sistema no distingue 
+           -humano de agente. El protocolo enforcea igual para ambos.</di
+           -v>                                                           
+       919 +<div class="border border-gray-700 rounded p-3 bg-gray-900"> 
+       920 +  <div class="text-orange-300 text-sm font-semibold">¿Las pla
+           +taformas internas resuelven esto o lo esconden?</div>        
+       921 +  <div class="text-gray-500 text-xs mt-1">Una plataforma sin 
+           +ontología solo mueve la ignorancia un nivel arriba.</div>    
+       922  </div>
+       923  
+       924  </div>
+       925  
+      1321 -</v-click>                                                   
+      1322 -                                                             
+      1323 -<v-click>                                                    
+      1324 -                                                             
+      1325 -<div class="mt-8 text-sm text-gray-500">                     
+      1326 -Worktrees aíslan el working-tree (branches diferentes).<br>  
+      1327 -Locking serializa escrituras sobre el mismo árbol.<br>       
+      1328 -Son complementarios, no alternativos.                        
+       926  </div>
+       927  
+      1331 -</v-click>                                                   
+      1332 -                                                             
+       928  </div>
+       929  
+      1335 -</div>                                                       
+      1336 -                                                             
+       930  <Footer />
+       931  
+       932  <!--
+      1340 -AI agents are first-class actors in this system. The protocol
+           - does not distinguish — it enforces.                         
+      1341 -The session hook means Claude starts informed, not blind.    
+      1342 -Concurrent use without a central coordinator is the scaling p
+           -roperty that makes this practical.                           
+       933 +Open space: las mejores respuestas vendrán del público.      
+       934 +Tu rol aquí es provocar, no resolver. Lanza las preguntas y d
+           +eja el silencio.                                             
+       935  -->
+       936  
+       937  ---
+     ...
+       942  
+       943  <div class="standalone-slide text-shadow-lg text-xl">
+       944  
+      1353 -# Ontology & Reflection                                      
+       945 +<div class="text-3xl font-semibold text-gray-100 mb-2">      
+       946 +  La IA puede declarar tu infra.                             
+       947 +</div>                                                       
+       948  
+      1355 -<div class="meters-final -mt-3 mb-7">                        
+      1356 -  <span class="text-green-400">✓ Typed</span>                
+      1357 -  <span class="text-green-400">✓ Verified</span>             
+      1358 -  <span class="text-green-400">✓ Sealed</span>               
+      1359 -  <span class="text-green-400">✓ Decided</span>              
+       949 +<div class="text-3xl font-semibold text-orange-400 mb-8">    
+       950 +  No puede entenderla por ti.                                
+       951  </div>
+       952  
+      1362 -<div class="text-2xl leading-11 text-gray-200">              
+      1363 -El código y el entendimiento del código deben converger.<br> 
+      1364 -Los <em>ADRs</em> y la <em>ontología</em> son el contrato ent
+           -re ambos.<br>                                                
+      1365 -La <em>reflection</em> es la operación sobre ese contrato.   
+       953 +<v-click>                                                    
+       954 +                                                             
+       955 +<div class="text-xl text-gray-300 leading-10 max-w-2xl mx-aut
+           +o">                                                          
+       956 +  La ontología es el modelo que te fuerza a saber qué es verd
+           +ad ahora.<br>                                                
+       957 +  La reflection es la operación sobre ese modelo.<br>        
+       958 +  La IA es el actor que opera <em>sobre lo que ya clarificast
+           +e</em>.                                                      
+       959  </div>
+       960  
+      1368 -<div class="text-4xl leading-12 text-gray-200 my-5">         
+      1369 -Gracias. ¿Preguntas?                                         
+       961 +</v-click>                                                   
+       962 +                                                             
+       963 +<v-click>                                                    
+       964 +                                                             
+       965 +<div class="mt-10 text-2xl text-gray-200">                   
+       966 +  Gracias. ¿Preguntas?                                       
+       967  </div>
+       968  
+      1372 -<small>jesusperez.pro · ontoref.dev </small>                 
+       969 +<small class="mt-4 block text-gray-500">jesusperez.pro · onto
+           +ref.dev</small>                                              
+       970  
+       971 +</v-click>                                                   
+       972 +                                                             
+       973  </div>
+       974  
+       975  <style>
+      1377 -.meters-final {                                              
+      1378 -  display: inline-flex;                                      
+      1379 -  gap: 2rem;                                                 
+      1380 -  font-size: 1.4rem;                                         
+      1381 -  font-family: monospace;                                    
+       976 +.standalone-slide {                                          
+       977 +  text-align: center;                                        
+       978  }
+       979  </style>
+       980  
+       981  <Footer />
+       982  
+       983  <!--
+      1388 -Direct eye contact. Slow delivery.                           
+      1389 -The three green checks are the promise: typed, verified, seal
+           -ed. If those three hold, the system holds.                   
+       984 +Cierre lento. Las tres líneas son la síntesis completa.      
+       985 +No terminar con "ontoref es la solución" — terminar con la in
+           +versión cognitiva.                                           
+       986 +El punto no es adoptar ontoref. El punto es tomar posesión de
+           + tu propio modelo.                                           
+       987  -->
+
+⏺ ★ Insight ─────────────────────────────────────
+  Las slides fusionan las dos propuestas en una narrativa única: propuesta 2
+  (infra + IA) establece el problema, propuesta 1 (ontología como
+  auto-clarificación) es la respuesta. El slide final invierte la expectativa —
+  no "ontoref es la herramienta", sino "la mejora es tuya, no de la IA". El
+  slide de debate mantiene ambas líneas separadas para que en el open space el
+  facilitador pueda elegir qué hilo tirar.
+  ─────────────────────────────────────────────────
+
+  11 slides en total siguiendo el patrón Slidev exacto del archivo original:
+
+  ┌─────┬──────────────────────────┬────────────────────────────────────────┐
+  │  #  │          Slide           │           Función narrativa            │
+  ├─────┼──────────────────────────┼────────────────────────────────────────┤
+  │ 1   │ Cover                    │ Pregunta provocadora — combina ambas   │
+  │     │                          │ propuestas                             │
+  ├─────┼──────────────────────────┼────────────────────────────────────────┤
+  │ 2   │ El problema no es la     │ IA hace forma correcta, no significado │
+  │     │ herramienta              │                                        │
+  ├─────┼──────────────────────────┼────────────────────────────────────────┤
+  │ 3   │ Las 2am                  │ Caso concreto de Terraform sin         │
+  │     │                          │ contexto                               │
+  ├─────┼──────────────────────────┼────────────────────────────────────────┤
+  │ 4   │ Tu proyecto vive en tu   │ Transición al argumento de ontología   │
+  │     │ cabeza                   │                                        │
+  ├─────┼──────────────────────────┼────────────────────────────────────────┤
+  │ 5   │ La tesis                 │ Inversión central — ontología es para  │
+  │     │                          │ ti                                     │
+  ├─────┼──────────────────────────┼────────────────────────────────────────┤
+  │ 6   │ Ontología ≠ lista de     │ Diferencia estructural vs prosa        │
+  │     │ decisiones               │                                        │
+  ├─────┼──────────────────────────┼────────────────────────────────────────┤
+  │ 7   │ Ontoref — tres archivos  │ core.ncl, state.ncl, gate.ncl como     │
+  │     │                          │ modelo concreto                        │
+  ├─────┼──────────────────────────┼────────────────────────────────────────┤
+  │ 8   │ Reflection — lo que se   │ Dualidad ontología/reflection          │
+  │     │ CONVIERTE                │                                        │
+  ├─────┼──────────────────────────┼────────────────────────────────────────┤
+  │ 9   │ La IA como actor, no     │ SessionStart hook — antes/después      │
+  │     │ oráculo                  │                                        │
+  ├─────┼──────────────────────────┼────────────────────────────────────────┤
+  │ 10  │ ¿Quién mantiene esto?    │ La objeción más honesta + respuesta de │
+  │     │                          │  diseño                                │
+  ├─────┼──────────────────────────┼────────────────────────────────────────┤
+  │ 11  │ Debate + Cierre          │ Preguntas abiertas organizadas por     │
+  │     │                          │ propuesta                              │
+  └─────┴──────────────────────────┴────────────────────────────────────────┘
+
+✻ Baked for 2m 58s
+
+※ recap: Created open space slides in slides.md covering ontology and reflection as ontoref base concepts, blending both proposals into one narrative arc. No next action pending — slides are complete and ready to preview. (disable recaps in /config)
+
+❯ en la ontología filosófca exiten los universales, ¿ son los invariantes un    
+  equivalente de los mismos ?                                                   
+
+⏺ No exactamente — la correspondencia es más fina que eso.
+
+  Los universales responden a: ¿qué comparten múltiples particulares que los 
+  hace del mismo tipo? Son propiedades abstractas que se instancian en varios
+  individuos: "rojez", "humanidad". El debate clásico (realismo platónico vs
+  nominalismo) gira en torno a si existen independientemente de sus instancias.
+
+  Los invariantes en ontoref son otra cosa: son propiedades esenciales de un 
+  singular — el proyecto como entidad. No preguntan qué comparten múltiples
+  proyectos; preguntan qué no puede cambiar en este proyecto sin que deje de ser
+   este proyecto.
+
+  La analogía filosófica más precisa no es el universal — es la distinción
+  aristotélica entre propiedades esenciales y accidentales. Un invariante dice:
+  si esto cambia, el proyecto ha cambiado de naturaleza, no solo de estado.
+  Kripke lo formalizó mejor: ciertas propiedades son rígidas — no contingentes
+  en ningún mundo posible dado un individuo.
+
+  Donde sí hay resonancia con los universales: ambos capturan lo que no varía a 
+  través del cambio — el universal a través de la instanciación, el invariante a
+   través de la evolución temporal. La familia semántica es la misma aunque el
+  eje difiera.
+
+  Y hay otra diferencia de fondo: los universales platónicos existen
+  independientemente de que alguien los declare. Los invariantes de ontoref son
+  hechos institucionales en el sentido de Searle — son verdad porque los
+  arquitectos del proyecto los declararon como tales y los formalizaron. Son
+  constructivistas, no metafísicamente realistas. Eso los hace más frágiles
+  (requieren mantenimiento activo) pero también más honestos sobre su
+  naturaleza.
+
+  Las tensiones de ontoref tienen otro pedigree — se acercan más a la dialéctica
+   hegeliana: trade-offs que no se resuelven, solo se gestionan.
+
+✻ Crunched for 33s
+
+❯ relaciona los elementos básicos de la filosófica con respecto a la que usa el 
+   model operacional en software y compara sus autores y orígenes               
+
+⏺ Esta es una genealogía real. La tabla primero, luego los matices que no caben
+  en ella.
+
+  ---
+  Tabla comparativa
+
+  Concepto filosófico: Universales
+  Autor / Origen: Platón (428-348 aC) — República, Parménides
+  Equivalente operacional: Tipos / Schemas
+  Implementación en ontoref: Contratos Nickel — String, [| 'Accepted, 'Proposed 
+    |]
+  ────────────────────────────────────────
+  Concepto filosófico: Esencia / Accidente
+  Autor / Origen: Aristóteles (384-322 aC) — Metafísica
+  Equivalente operacional: Invariante / Estado mutable
+  Implementación en ontoref: invariante = true en core.ncl vs dimensiones en
+    state.ncl
+  ────────────────────────────────────────
+  Concepto filosófico: Potencia / Acto
+  Autor / Origen: Aristóteles — Metafísica IX
+  Equivalente operacional: Estado actual → deseado + Gates
+  Implementación en ontoref: current_state / desired_state / gate.ncl
+  ────────────────────────────────────────
+  Concepto filosófico: Categorías
+  Autor / Origen: Aristóteles — Categorías
+  Equivalente operacional: Tipos de nodo ontológico
+  Implementación en ontoref: invariants, tensions, practices, dimensions, gates
+  ────────────────────────────────────────
+  Concepto filosófico: Sustancia primera / segunda
+  Autor / Origen: Aristóteles — Categorías
+  Equivalente operacional: Instancia / Tipo
+  Implementación en ontoref: Proyecto concreto vs schema de proyecto
+  ────────────────────────────────────────
+  Concepto filosófico: Identidad a través del tiempo
+  Autor / Origen: Locke (Ensayo, 1689), Parfit (Reasons, 1984)
+  Equivalente operacional: Versioning + ADR lifecycle
+  Implementación en ontoref: status = Superseded, superseded_by — identidad que
+    evoluciona formalmente
+  ────────────────────────────────────────
+  Concepto filosófico: Dialéctica
+  Autor / Origen: Hegel (1770-1831) — Fenomenología del Espíritu
+  Equivalente operacional: Tensiones activas
+  Implementación en ontoref: tensions en core.ncl — sin síntesis, gestionadas
+  ────────────────────────────────────────
+  Concepto filosófico: Logos como tensión de opuestos
+  Autor / Origen: Heráclito (535-475 aC)
+  Equivalente operacional: Tensión irreducible
+  Implementación en ontoref: La dualidad ontología/reflection como tensión
+    constitutiva
+  ────────────────────────────────────────
+  Concepto filosófico: Mundos posibles
+  Autor / Origen: Leibniz (1646-1716) + Kripke (Naming, 1980)
+  Equivalente operacional: FSM — transiciones permitidas
+  Implementación en ontoref: ADRs como restricciones de accesibilidad entre
+    estados
+  ────────────────────────────────────────
+  Concepto filosófico: Propiedades esenciales rígidas
+  Autor / Origen: Kripke — Naming and Necessity (1980)
+  Equivalente operacional: Invariantes como necesidad, no contingencia
+  Implementación en ontoref: Un invariante es verdad en todos los estados
+    posibles del proyecto
+  ────────────────────────────────────────
+  Concepto filosófico: Límites del lenguaje = límites del mundo
+  Autor / Origen: Wittgenstein — Tractatus (1921)
+  Equivalente operacional: Schema como frontera de lo expresable
+  Implementación en ontoref: Lo que no tipea en Nickel no puede existir en el
+    sistema
+  ────────────────────────────────────────
+  Concepto filosófico: Juegos de lenguaje
+  Autor / Origen: Wittgenstein — Investigaciones (1953)
+  Equivalente operacional: Modos operacionales
+  Implementación en ontoref: Un modo define las reglas del juego para un
+  contexto
+    operacional
+  ────────────────────────────────────────
+  Concepto filosófico: Verdad como correspondencia
+  Autor / Origen: Tarski (1933) — El concepto de verdad
+  Equivalente operacional: Typecheck / validación
+  Implementación en ontoref: nickel export como verificación de correspondencia
+    con el schema
+  ────────────────────────────────────────
+  Concepto filosófico: Hechos institucionales
+  Autor / Origen: Searle — La construcción de la realidad social (1995)
+  Equivalente operacional: ADRs como decisiones declaradas
+  Implementación en ontoref: Un ADR es verdad porque el equipo lo declaró y
+    formalizó — no metafísicamente
+  ────────────────────────────────────────
+  Concepto filosófico: Ontología formal vs regional
+  Autor / Origen: Husserl — Ideas I (1913)
+  Equivalente operacional: Schema (formal) vs ontología del proyecto (regional)
+  Implementación en ontoref: adrs/schema.ncl vs .ontology/core.ncl
+  ────────────────────────────────────────
+  Concepto filosófico: Proceso vs sustancia
+  Autor / Origen: Whitehead — Process and Reality (1929)
+  Equivalente operacional: Ontología (ser) vs Reflection (devenir)
+  Implementación en ontoref: La dualidad estructural del protocolo
+  ────────────────────────────────────────
+  Concepto filosófico: Realismo estructural
+  Autor / Origen: Ladyman & Ross — Every Thing Must Go (2007)
+  Equivalente operacional: Grafo como lo real, no las entidades
+  Implementación en ontoref: El DAG captura relaciones; los nodos son
+  secundarios
+  ────────────────────────────────────────
+  Concepto filosófico: Lógica de predicados
+  Autor / Origen: Frege (1879) — Begriffsschrift
+  Equivalente operacional: Queries sobre el grafo
+  Implementación en ontoref: where severity == "Hard" | select id claim
+
+  ---
+  Los mapeos no triviales
+
+  Aristóteles es el arquitecto base. La esencia/accidente es exactamente
+  invariante/estado-mutable. La potencia/acto es exactamente
+  current_state/desired_state con la gate como umbral de actualización. Que esto
+   haya tardado 2.300 años en aparecer en software de forma explícita es
+  indicativo de cuánto tiempo estuvo implícito.
+
+  Whitehead vs Aristóteles es la tensión interna del protocolo. Aristóteles ve
+  la realidad como sustancias con propiedades. Whitehead la ve como procesos: no
+   hay cosas, hay eventos. Ontoref vive en esa grieta — la ontología es
+  aristotélica (qué ES el proyecto), la reflection es whiteheadiana (cómo
+  DEVIENE). La decisión de mantenerlos como capas separadas es un reconocimiento
+   implícito de que ninguno basta solo.
+
+  Kripke resuelve algo que Aristóteles dejó ambiguo. Aristóteles distingue
+  esencial de accidental pero no formaliza la modalidad. Kripke dice: una
+  propiedad esencial es rígida — verdad en todos los mundos posibles accesibles
+  desde el individuo. Los invariantes de ontoref son exactamente esto: no pueden
+   ser falsos en ningún estado posible del proyecto sin que el proyecto deje de
+  ser ese proyecto. Los ADRs que los cambian no violan los invariantes — crean
+  un nuevo individuo.
+
+  Searle es el que más se ignora y más aplica. Un ADR no captura una verdad
+  metafísica sobre el proyecto — es un hecho institucional. Es verdad porque el
+  equipo colectivamente lo declaró y comprometió. Exactamente como "este papel
+  es dinero" o "este jugador está fuera de juego". Sin el acto declarativo, la
+  constraint no existe. Esto tiene una implicación práctica: si el equipo no
+  cree en el ADR, no hay constraint real — hay teatro.
+
+  Wittgenstein en dos fases distintas. El Tractatus da el fundamento del schema:
+   los límites de lo tipable son los límites de lo que el sistema puede
+  representar. Las Investigaciones dan el fundamento de los modos: el
+  significado de una operación es su uso dentro de un juego de lenguaje — el
+  modo new_service no describe, prescribe una forma de vida operacional dentro
+  del proyecto.
+
+  La ausencia notable: Heidegger. El ser-en-el-mundo como condición de
+  posibilidad de cualquier ontología explícita no aparece en ningún modelo de
+  software operacional. Lo que Heidegger llamaría el trasfondo pre-temático — el
+   contexto que no se puede formalizar porque es la condición de posibilidad de
+  la formalización — es exactamente lo que los READMEs intentan capturar y
+  fracasan. Ontoref no lo resuelve tampoco. Lo que sí hace es reducir la
+  superficie de ese trasfondo haciendo explícito lo que puede serlo.
+
+✻ Cooked for 1m 2s
+
+❯ genera un pdf con estas dos últimas preguntas en un PDF y añade info sobre    
+  quíen creo la ontología operacional, por qué y dónde se usa actualmente       
+
+⏺ Primero verifico qué herramientas de conversión tengo disponibles.
+
+⏺ Bash(which pandoc && pandoc --version | head -1; which wkhtmltopdf; which
+      weasyprint; which typst)
+  ⎿  /usr/local/bin/pandoc
+     pandoc 3.7.0.2
+     /opt/homebrew/bin/weasyprint
+     /opt/homebrew/bin/typst
+
+⏺ Tengo pandoc y typst. Usaré typst — produce mejor output para contenido
+  académico/filosófico. Escribo el documento y compilo.
+
+⏺ Write(ontologia-filosofia-software.typ)
+  ⎿  Wrote 360 lines to ontologia-filosofia-software.typ
+       2   title: "Ontología filosófica y modelo operacional en software"
+         ,
+       3   author: "Jesús Pérez Lorenzo",
+       4 )
+       5 
+       7   paper: "a4",
+       8   margin: (top: 2.5cm, bottom: 2.5cm, left: 2.8cm, right: 2.8cm)
+         ,
+       9   numbering: "1",
+      10   number-align: right,
+     … +536 lines (ctrl+o to expand)
+
+❯ escribe el pdf tambien !!! 
+
+⏺ Bash(cd /Users/Akasha/Development/ontoref/assets/presentation && typst compile
+       ontologia-filosofia-software.typ ontologia-filosofia-software.pdf 2>&1)
+  ⎿  (No output)
+
+  Listed 1 directory (ctrl+o to expand)
+
+⏺ PDF generado en assets/presentation/ontologia-filosofia-software.pdf (179K).
+
+  Contiene tres secciones:
+
+  1. Universales e invariantes — la correspondencia no es directa: los
+  invariantes son propiedades esenciales de un singular (Aristóteles → Kripke)
+  con estatus de hechos institucionales (Searle), no universales en sentido
+  platónico ni aristotélico.
+
+  2. Tabla comparativa completa — 17 filas con concepto, autor/obra/fecha,
+  equivalente operacional e implementación concreta en ontoref.
+
+  3. Genealogía de la ontología operacional en software:
+
+  ┌────────────────┬────────────────────────────────────────────┬───────────┐
+  │     Figura     │                Contribución                │    Año    │
+  ├────────────────┼────────────────────────────────────────────┼───────────┤
+  │ McCarthy       │ Representación explícita del conocimiento, │ 1958–1969 │
+  │                │  situational calculus                      │           │
+  ├────────────────┼────────────────────────────────────────────┼───────────┤
+  │ Minsky         │ Frames — primera estructura análoga a la   │ 1974      │
+  │                │ sustancia aristotélica                     │           │
+  ├────────────────┼────────────────────────────────────────────┼───────────┤
+  │ Lenat          │ Cyc — 600K conceptos, el proyecto más      │ 1984      │
+  │                │ ambicioso                                  │           │
+  ├────────────────┼────────────────────────────────────────────┼───────────┤
+  │ Gruber         │ Definición canónica: "especificación de    │ 1993      │
+  │ (Stanford)     │ una conceptualización"                     │           │
+  ├────────────────┼────────────────────────────────────────────┼───────────┤
+  │ Guarino        │ Ontología formal en sistemas de            │ 1990s     │
+  │ (ISTC-CNR)     │ información, FOIS                          │           │
+  ├────────────────┼────────────────────────────────────────────┼───────────┤
+  │ Berners-Lee et │ Web Semántica, RDF, OWL, SPARQL            │ 2001      │
+  │  al.           │                                            │           │
+  ├────────────────┼────────────────────────────────────────────┼───────────┤
+  │ Google         │ Knowledge Graph — ontología operacional a  │ 2012      │
+  │                │ escala de producción                       │           │
+  └────────────────┴────────────────────────────────────────────┴───────────┘
+
+  Uso actual: biomedicina (Gene Ontology, SNOMED CT, ICD-11), grafos
+  empresariales (Wikidata, schema.org), manufactura (ISO 15926, BIM/IFC),
+  agentes IA — y la frontera actual: ontologías del propio proyecto de software
+  como modelo auto-descriptivo consultable.
+
+  El .typ fuente queda en el mismo directorio para ediciones futuras.
+
+✻ Brewed for 3m 16s
+
+※ recap: Generated slides for the open space talk (slides.md) and a philosophical PDF comparing ontology concepts to ontoref's operational model. No pending actions unless you want to review or present either file. (disable recaps in /config)
+
+❯ Qué relación tiene la ontología la filosófica y la operacional con el         
+  taoismo, yin/yang ?                                                           
+
+⏺ La relación es más profunda que una metáfora. Hay correspondencias
+  estructurales genuinas, pero también diferencias que el Taoísmo ilumina mejor
+  que la tradición occidental.
+
+  ---
+  Lo que el Taoísmo ve que la ontología occidental no ve
+
+  La ontología occidental desde Parménides privilegia el ser sobre el no-ser, la
+   presencia sobre la ausencia, la identidad sobre la diferencia. Toda la
+  tradición aristotélica-escolástica-analítica trabaja desde ahí.
+
+  El Taoísmo invierte la pregunta. El Tao (道) es previo a la distinción
+  ser/no-ser. El primer capítulo del Tao Te Ching:
+
+  ▎ 道可道，非常道。名可名，非常名。
+  ▎ El Tao que puede nombrarse no es el Tao eterno. El nombre que puede 
+  ▎ pronunciarse no es el nombre eterno.
+
+  Lo que no puede formalizarse no es una limitación — es la condición de
+  posibilidad de todo lo que sí puede formalizarse. La ontología operacional que
+   puede expresarse completamente en un schema no captura la realidad
+  operacional completa. Heidegger llegó al mismo lugar por otro camino — el
+  trasfondo pre-temático que hace posible cualquier acto de comprensión
+  explícita.
+
+  ---
+  Yin-Yang — no oposición sino polaridad constitutiva
+
+  La confusión más común: yin y yang no son contrarios sino complementarios que 
+  se definen mutuamente. Propiedades clave:
+
+  - Cada polo contiene una semilla del opuesto — el punto pequeño dentro de cada
+   mitad del símbolo
+  - El exceso de un polo genera su contrario (極則反 — la extremidad se
+  invierte)
+  - La tensión entre ellos nunca se resuelve — solo se gestiona dinámicamente
+  - Uno no existe sin el otro — yin sin yang es vacío inerte, yang sin yin es
+  movimiento sin forma
+
+  Esto es estructuralmente distinto de la dialéctica hegeliana, que busca
+  síntesis — resolver la contradicción en una unidad superior. El yin-yang no
+  resuelve. La tensión es constitutiva, permanente.
+
+  ---
+  Correspondencias con la ontología filosófica
+
+  Concepto taoísta: Tao (道) — el principio previo a toda distinción
+  Autor / Texto: Laozi, Tao Te Ching (~6-4 aC)
+  Equivalente occidental más cercano: Ser de Heidegger — previo a los entes
+  Diferencia clave: El Tao es innombrable; el Ser heideggeriano puede pensarse
+    aunque no definirse
+  ────────────────────────────────────────
+  Concepto taoísta: Yin/Yang — dualidad que no se resuelve
+  Autor / Texto: Tradición Zhou (~1100 aC), sistematizada en I Ching
+  Equivalente occidental más cercano: Heráclito — logos como tensión de opuestos
+  Diferencia clave: Heráclito es el único occidental pre-moderno equivalente
+  ────────────────────────────────────────
+  Concepto taoísta: Wu wei (無為) — acción sin forzar
+  Autor / Texto: Laozi, Tao Te Ching
+  Equivalente occidental más cercano: Aristóteles — acto que fluye de la esencia
+  Diferencia clave: Wu wei no tiene equivalente exacto en Occidente
+  ────────────────────────────────────────
+  Concepto taoísta: Ji (機) — el momento exacto de transformación
+  Autor / Texto: Zhuangzi, Zhuangzi (~4 aC)
+  Equivalente occidental más cercano: Kripke — accesibilidad entre mundos
+    posibles
+  Diferencia clave: Ji es cualitativo y contextual; Kripke es lógico-formal
+  ────────────────────────────────────────
+  Concepto taoísta: Wu ji → Tai ji — de lo indiferenciado a la polaridad
+  Autor / Texto: Cosmología taoísta, Zhou Dunyi (1017-1073)
+  Equivalente occidental más cercano: Hegel — de lo abstracto a lo concreto
+  Diferencia clave: Hegel sintetiza; el taoísmo mantiene la polaridad
+  ────────────────────────────────────────
+  Concepto taoísta: Relatividad perspectival
+  Autor / Texto: Zhuangzi — la perspectiva de la rana del pozo
+  Equivalente occidental más cercano: Wittgenstein tardío — juegos de lenguaje
+  Diferencia clave: Zhuangzi va más lejos: toda ontología es perspectival
+
+  Heráclito es el occidental más cercano. Su logos como tensión dinámica de
+  opuestos — "el camino hacia arriba y hacia abajo son el mismo" — es
+  estructuralmente idéntico al yin-yang. Hegel lo notó. La diferencia: Heráclito
+   vive en la tensión; Hegel la usa como palanca hacia la síntesis.
+
+  ---
+  Correspondencias con la ontología operacional en software
+
+  Aquí es donde se pone interesante. La ontología OWL-estilo (Gruber, Web
+  Semántica) es fundamentalmente clasificatoria — jerarquías de tipos, herencia,
+   subsunción. Aristotélica en su hueso.
+
+  El modelo operacional de ontoref es estructuralmente más cercano al Taoísmo:
+
+  1. Yin/Yang ↔ Ontología / Reflection
+
+  La dualidad ya está en el diseño del protocolo:
+
+  - Ontología (core.ncl): lo que ES — invariantes, estructura, ser. Yang — la
+  capa formal, dura, definida.
+  - Reflection: lo que DEVIENE — operaciones, modos, movimiento. Yin — la capa
+  fluida, procesual.
+
+  Y como en el símbolo: cada una contiene una semilla de la otra. La ontología
+  describe estados futuros (desired_state) — contiene movimiento. La reflection
+  puede actualizar la ontología vía ADRs — contiene forma. Separarlas es
+  necesario porque tienen ritmos distintos; fundirlas sería colapsar el yin en
+  el yang.
+
+  2. Wu wei ↔ "No enforcement"
+
+  Uno de los invariantes del protocolo: coherencia voluntaria que emerge de la 
+  adopción justificada, sin mecanismo de enforcement.
+
+  Esto es wu wei aplicado a diseño de sistemas. No impones — creas las
+  condiciones para que el alineamiento emerja naturalmente. Un sistema de tipos
+  fuertes fuerza. Una ontología operacional orienta. La diferencia es
+  exactamente la diferencia entre imponer el Tao y fluir con él.
+
+  3. Tensiones ↔ Yin-Yang permanente
+
+  Las tensiones en un modelo operacional ("formalización vs. fricción de
+  adopción", "ontología vs. reflection") son explícitamente irresolubles. No se
+  busca síntesis — se gestionan.
+
+  Esto es taoísta, no hegeliano. Si eliminas la tensión entre formalización y
+  adopción — maximizando uno — el sistema muere. O es un schema perfecto que
+  nadie adopta, o es caos que todos adoptan pero nadie puede razonar sobre. La
+  tensión viva entre los polos es la que sostiene el sistema.
+
+  4. Gates ↔ Ji (機) — el momento de transformación
+
+  El concepto taoísta de ji es el punto de inflexión donde el potencial se
+  actualiza — no un estado sino el umbral entre estados. Dinámico, cualitativo,
+  contextual.
+
+  Las gates en ontoref son exactamente esto. No son checkboxes — son condiciones
+   que marcan el momento en que un estado puede transformarse en otro. La gate
+  activa no dice "puedes cruzar cuando X sea verdad" sino "este es el umbral que
+   separa lo que eres de lo que puedes devenir".
+
+  5. Wu ji → Tai ji ↔ Schema formal → Ontología regional
+
+  La cosmología taoísta describe el origen como Wu ji (無極) — el vacío
+  indiferenciado previo a cualquier distinción — que da lugar a Tai ji (太極) —
+  la polaridad primordial yin/yang.
+
+  Husserl hizo la misma distinción: ontología formal (estructuras lógicas puras,
+   previas a cualquier dominio) vs. ontología regional (conceptos de un dominio
+  específico). El schema.ncl es el Wu ji — la estructura vacía que puede
+  instanciarse en cualquier proyecto. El .ontology/core.ncl de un proyecto
+  concreto es el Tai ji — la polaridad específica de ese mundo.
+
+  ---
+  Lo que el Taoísmo revela que la tradición occidental pierde
+
+  La crítica taoísta a OWL y las ontologías clasificatorias:
+
+  Los sistemas como OWL intentan eliminar la ambigüedad mediante clasificación
+  precisa. El Tao Te Ching advierte: cuanto más preciso el nombre, más excluyes.
+   La realidad vive en el espacio entre las categorías, no en las categorías
+  mismas.
+
+  Esto explica por qué las ontologías biomédicas — las más maduras del campo —
+  sufren continuamente con casos límite, dependencia del contexto y la necesidad
+   de múltiples ejes de clasificación simultáneos. La vista taoísta diría:
+  estáis intentando resolver una tensión irresolvable. La precisión y la
+  cobertura son yin y yang — solo se gestionan.
+
+  Zhuangzi y el límite de toda ontología:
+
+  Zhuangzi va más lejos que Laozi. Toda distinción ontológica es perspectival —
+  la perspectiva de la rana en el pozo que no puede imaginar el mar. Esto no
+  lleva al nihilismo sino a la apropiada contextualidad: la ontología correcta
+  para un proyecto no es la más formalmente precisa sino la que el equipo puede
+  sostener y evolucionar.
+
+  Gruber llegó intuitivamente a lo mismo en su definición de 1993: una ontología
+   no es una captura de la realidad — es "una especificación de una
+  conceptualización". Un punto de vista compartido, no una verdad metafísica.
+  Eso es profundamente zhuangziano sin saberlo.
+
+  ---
+  La síntesis imposible — y por qué eso es correcto
+
+  No hay síntesis entre el modelo ontológico occidental (jerarquías de tipos,
+  ser definido, identidad estable) y el Taoísmo (proceso, polaridad
+  irresolvable, Tao innombrable). Y no debería haberla.
+
+  Lo que el Taoísmo aporta a la ontología operacional en software no es una
+  alternativa — es una corrección de actitud. La tentación del ingeniero es
+  resolver las tensiones, eliminar la ambigüedad, llevar el sistema a un estado
+  estable final. El Taoísmo dice: ese estado final no existe. El yin-yang vivo —
+   la tensión gestionada — es el sistema funcionando correctamente, no el
+  sistema esperando ser completado.
+
+  Un modelo operacional que no tiene tensiones activas no es un modelo maduro.
+  Es un modelo muerto.
+
+✻ Sautéed for 1m 42s
+❯ Y cómo se llega a los DAG y su función en la ontología operacional            
+
+⏺ La genealogía es larga y viene de tres tradiciones que convergen. Vale la pena
+   recorrerla completa porque cada una aporta algo distinto.
+
+  ---
+  1. El problema que fuerza el grafo
+
+  Antes del grafo hay un problema: ¿cómo representar conocimiento estructurado 
+  de forma que se pueda razonar sobre él?
+
+  La prosa no sirve — no es consultable ni verificable. Las listas planas
+  pierden las relaciones. Las jerarquías de árbol pierden los nodos con
+  múltiples padres. El grafo emerge como la estructura mínima que captura
+  relaciones arbitrarias entre conceptos.
+
+  Pero el grafo general tiene un problema: puede tener ciclos. Y los ciclos en
+  ontología son epistemológicamente problemáticos.
+
+  ---
+  2. Por qué la aciclicidad — el problema de la justificación circular
+
+  Esto viene de dos tradiciones independientes que llegan al mismo resultado:
+
+  Aristóteles (384-322 aC) — Analíticos Posteriores: una definición válida
+  procede de género más diferencia específica. El género es más general que lo
+  definido — la jerarquía es estrictamente descendente. Si A se define en
+  términos de B y B en términos de A, ninguno está definido. La justificación
+  circular es un defecto lógico, no una propiedad tolerable.
+
+  Russell (1901) — la paradoja del conjunto de todos los conjuntos que no se
+  contienen a sí mismos. La solución que Russell y Whitehead desarrollan en
+  Principia Mathematica (1910) es la teoría de tipos: los tipos forman una
+  jerarquía estricta — un tipo de nivel N solo puede contener o referirse a
+  tipos de nivel N-1 o inferior. Esto es estructuralmente un DAG. La aciclicidad
+   no es una restricción arbitraria — es la solución formal al colapso por
+  autorreferencia.
+
+  Tarski (1933) — verdad como correspondencia entre proposiciones y estructuras.
+   Las estructuras que usa son relaciones sobre dominios — grafos. La teoría de
+  modelos formaliza esto.
+
+  La convergencia: en lógica, ontología formal y teoría de tipos, la aciclicidad
+   es el requisito de que las definiciones sean bien fundadas. No puede haber
+  descenso infinito ni circularidad.
+
+  ---
+  3. La entrada en inteligencia artificial — redes semánticas
+
+  Collins & Quillian (1969) — "Retrieval time from semantic memory", Cognitive
+  Psychology.
+
+  Primera implementación computacional de un grafo de conocimiento. Nodos =
+  conceptos, aristas = relaciones (is-a, has-property). La jerarquía is-a es
+  inherentemente un DAG — si "perro es-un animal" y "animal es-un ser-vivo", la
+  subsunción es transitiva y acíclica.
+
+  Collins y Quillian usaron esto para modelar cómo los humanos recuperamos
+  información semántica. El experimento: verificar "¿un canario puede respirar?"
+   tarda más que "¿un canario puede cantar?" porque hay que recorrer más niveles
+   del DAG. La estructura del grafo predice tiempos de reacción — el
+  conocimiento humano tiene topología.
+
+  Minsky (1974) — frames. Ya discutido, pero la herencia entre frames crea un
+  DAG. Un frame puede tener múltiples padres (herencia múltiple) — eso descarta
+  el árbol y exige el grafo. Pero los ciclos de herencia son patológicos — un
+  frame no puede ser su propio ancestro.
+
+  KL-ONE (Brachman & Schmolze, 1985) — el primer sistema formal de Description
+  Logic. La relación de subsunción entre conceptos forma un DAG. Este trabajo es
+   el antecedente directo de OWL. La contribución clave: separar la taxonomía
+  (qué es-un qué) del esquema (qué propiedades tiene qué). Dos DAGs superpuestos
+   sobre el mismo dominio.
+
+  ---
+  4. La formalización de la causalidad como DAG — Pearl
+
+  Judea Pearl (1988, 2000) — Causality: Models, Reasoning, and Inference.
+
+  Pearl formalizó algo que hasta entonces era intuitivo: la causalidad es 
+  acíclica porque la causalidad es temporal. Si A causa B, B ocurre después de
+  A. Si B también causara A, tendríamos una contradicción temporal — B ocurre
+  antes de A y después de A simultáneamente.
+
+  Los Bayesian Networks de Pearl son DAGs donde las aristas representan
+  relaciones causales y los nodos son variables aleatorias. La estructura del
+  DAG codifica las independencias condicionales del sistema — qué variables son
+  relevantes para predecir qué otras.
+
+  Esto añade algo que las redes semánticas no tenían: dirección con significado 
+  causal, no solo jerárquico. La arista no solo dice "A depende de B" — dice "B
+  influye sobre A en esta dirección y no al revés".
+
+  Para la ontología operacional esto es crucial. Las dependencias entre
+  invariantes, tensiones, prácticas y gates no son solo jerárquicas — son
+  cuasi-causales. Un invariante justifica una tensión; una tensión motiva una
+  práctica; una práctica activa una gate. La dirección importa y no puede
+  invertirse.
+
+  ---
+  5. El DAG en ingeniería — Make, Git, build systems
+
+  Paralelo al trabajo académico, los DAGs aparecen en ingeniería por necesidad
+  práctica:
+
+  Make (Feldman, 1976) — el primer build system formal. Las reglas de
+  compilación forman un DAG: el ejecutable depende de los objetos, los objetos
+  dependen de los fuentes. Make recorre el DAG en orden topológico — compilar
+  solo lo que cambió, en el orden correcto. Un ciclo en las dependencias es un
+  error explícito.
+
+  PERT charts (1958, US Navy — programa Polaris) — planificación de proyectos
+  como DAG de tareas. Las aristas son dependencias temporales. El camino crítico
+   es el camino más largo en el DAG — determina la duración mínima del proyecto.
+   Primer uso masivo de DAGs en gestión.
+
+  Git (Torvalds, 2005) — el historial de commits es un DAG. Cada commit apunta a
+   sus padres; los merges crean nodos con múltiples padres; las branches
+  divergen. La propiedad clave: no puede existir un commit que sea su propio
+  ancestro. La historia es irreversible — el tiempo es un DAG.
+
+  ---
+  6. La función del DAG en ontología operacional
+
+  Cuando llega a la ontología operacional, el DAG hace cuatro cosas que ninguna
+  otra estructura hace juntas:
+
+  ① Traversal determinista para consultas de impacto
+
+  Si cambias un nodo, el DAG define exactamente qué nodos son alcanzables desde
+  él (descendientes) y qué nodos lo alcanzan (ancestros). Esto es análisis de
+  impacto: "si modifico este invariante, ¿qué tensiones, prácticas y gates se
+  ven afectadas?"
+
+  En un grafo cíclico esto es indecidible — el cierre transitivo de un nodo en
+  un grafo cíclico puede ser todo el grafo.
+
+  ② Orden topológico para ejecución
+
+  Los modos operacionales en ontoref son DAGs de pasos con depends_on. El
+  ordenamiento topológico del DAG da el orden de ejecución válido — sin
+  necesidad de un scheduler externo. Si el paso B depende del paso A, A siempre
+  precede a B en cualquier topological sort.
+
+  ③ Justificación bien fundada — sin circularidad epistémica
+
+  En el debate epistemológico clásico, el fundacionalismo dice que el
+  conocimiento descansa en fundamentos no circulares — como un árbol o DAG con
+  nodos raíz. El coherentismo dice que el conocimiento es una red de creencias
+  mutuamente sustentadas — puede ser cíclico.
+
+  La ontología operacional es fundacionalista por diseño. Los invariantes son
+  los nodos raíz — se justifican por decisión arquitectural explícita (ADR), no
+  por otros nodos del grafo. Las tensiones dependen de los invariantes; las
+  prácticas dependen de las tensiones; las gates dependen de las prácticas. La
+  justificación fluye en una dirección.
+
+  Un ciclo en ese grafo significaría: "este invariante existe porque esta
+  práctica existe, y esta práctica existe porque este invariante existe". Eso no
+   es arquitectura — es dogma circular.
+
+  ④ Rollback con orden definido
+
+  Si necesitas deshacer cambios, el DAG da el orden inverso. En un grafo
+  cíclico, no hay orden inverso bien definido.
+
+  ---
+  7. La tensión con el Taoísmo — DAG vs. ciclo
+
+  Aquí está el punto más interesante, conectando con la conversación anterior.
+
+  El DAG es yang en su estructura: direccional, jerárquico, asimétrico,
+  irreversible. Codifica la flecha del tiempo — las causas preceden a los
+  efectos, las definiciones preceden a lo definido.
+
+  El yin-yang taoísta es cíclico: yin genera yang, yang genera yin. Las
+  estaciones, el día y la noche, vida y muerte — todo retorna. El Tao es el
+  ciclo eterno.
+
+  Esto genera una tensión real en la ontología operacional:
+
+  DAG                          Yin-Yang
+  ─────────────────────────    ──────────────────────────
+  Irreversible                 Cíclico
+  Asimétrico                   Simétrico en transformación
+  El pasado no depende         El pasado se convierte
+    del futuro                   en futuro
+  Fundacionalista              Coherentista
+
+  La resolución que ontoref hace implícitamente — sin nombrarlo — es separar dos
+   temporalidades:
+
+  - Sincrónica (snapshot): el DAG es válido dentro de un momento dado. La
+  ontología en t₀ es un DAG acíclico.
+  - Diacrónica (evolución): la secuencia de snapshots puede tener patrones
+  cíclicos. Lo que era tensión puede convertirse en invariante; lo que era
+  práctica puede ser abandonado y redescubierto.
+
+  El DAG captura la estructura del ser en un momento. El yin-yang captura el
+  proceso de transformación a través del tiempo. Son complementarios, no
+  contradictorios — exactamente como ontología y reflection son complementarios.
+
+  Dicho de otra manera: el contenido es un DAG; la evolución es un ciclo. El I
+  Ching — el libro de los cambios — es un sistema para razonar sobre las
+  transformaciones de un estado a otro. Los 64 hexagramas son estados (nodos);
+  las transformaciones entre ellos son las aristas. El sistema completo es... un
+   grafo. No un DAG — un grafo con ciclos, porque la realidad vuelve.
+
+  ---
+  Síntesis
+
+  ┌───────────────────────┬─────────────────────────────────────────────────┐
+  │        Origen         │  Contribución al DAG en ontología operacional   │
+  ├───────────────────────┼─────────────────────────────────────────────────┤
+  │ Aristóteles (384 aC)  │ Definición bien fundada — jerarquía sin         │
+  │                       │ circularidad                                    │
+  ├───────────────────────┼─────────────────────────────────────────────────┤
+  │ Russell/Whitehead     │ Teoría de tipos — aciclicidad como solución al  │
+  │ (1910)                │ colapso por autorreferencia                     │
+  ├───────────────────────┼─────────────────────────────────────────────────┤
+  │ Collins & Quillian    │ Primera red semántica computacional — is-a como │
+  │ (1969)                │  DAG                                            │
+  ├───────────────────────┼─────────────────────────────────────────────────┤
+  │ Minsky (1974)         │ Frames con herencia múltiple — DAG sobre árbol  │
+  ├───────────────────────┼─────────────────────────────────────────────────┤
+  │ Brachman/Schmolze     │ Description Logic — subsunción formal como DAG  │
+  │ (1985)                │                                                 │
+  ├───────────────────────┼─────────────────────────────────────────────────┤
+  │ Pearl (1988)          │ Causalidad como DAG — dirección con significado │
+  │                       │  temporal                                       │
+  ├───────────────────────┼─────────────────────────────────────────────────┤
+  │ Feldman/Make (1976)   │ Dependencias de build como DAG — orden          │
+  │                       │ topológico práctico                             │
+  ├───────────────────────┼─────────────────────────────────────────────────┤
+  │ Torvalds/Git (2005)   │ Historia como DAG — irreversibilidad del tiempo │
+  │                       │  en VCS                                         │
+  └───────────────────────┴─────────────────────────────────────────────────┘
+
+  Lo que el DAG aporta que ninguna otra estructura da: razonamiento sobre 
+  dependencias con garantía de terminación y justificación no circular. Es el
+  mínimo estructural para que una ontología pueda responder "¿qué afecta a qué?"
+   sin entrar en bucle infinito.
+
+  Y la tensión con el Taoísmo no es un defecto — es la tensión constitutiva
+  entre estructura y proceso, entre el instante formalizados y el tiempo que
+  fluye. El DAG captura el yang del sistema. El ciclo de ADRs, versiones y
+  transformaciones captura el yin.
+
+✻ Brewed for 1m 56s
+
+❯ ¿ tal vez quieras incluir algo como material relevante y añadir alguna        
+  refencia y perpectiva ?                                                       
+
+⏺ Voy a leer el archivo typst actual y añadir las dos nuevas secciones más la
+  bibliografía.
+
+  Read 1 file (ctrl+o to expand)
+  ⎿  Interrupted · What should Claude do instead?
+
+❯ continua                                                                      
+
+⏺ Update(ontologia-filosofia-software.typ)
+  ⎿  Added 323 lines
+      358  La ontología operacional en software no es una importación rec
+           iente de la filosofía — es una convergencia que lleva setenta 
+           años ocurriendo, desde los frames de Minsky hasta los grafos d
+           e conocimiento de Google. Lo que ha cambiado con la proliferac
+           ión de agentes de IA es la urgencia: cuando un agente puede mo
+           dificar código, infraestructura o configuración de forma autón
+           oma, la ausencia de un modelo explícito del proyecto no es una
+            deuda técnica. Es una pérdida de control.
+      359  
+      360  La ontología no es la solución a ese problema. Es el nombre de
+           l problema que hay que resolver.
+      361 +                                                              
+      362 +#pagebreak()                                                  
+      363 +                                                              
+      364 +// ─── 4. TAOÍSMO Y ONTOLOGÍA OPERACIONAL ────────────────────
+          +────────────────                                              
+      365 +                                                              
+      366 += Taoísmo, Yin-Yang y ontología operacional                   
+      367 +                                                              
+      368 +== Lo que la tradición occidental no vio                      
+      369 +                                                              
+      370 +La ontología occidental desde Parménides privilegia el *ser* s
+          +obre el *no-ser*, la *presencia* sobre la *ausencia*, la *iden
+          +tidad* sobre la *diferencia*. Toda la tradición aristotélica-e
+          +scolástica-analítica trabaja desde ese punto de partida.      
+      371 +                                                              
+      372 +El Taoísmo invierte la pregunta. El Tao (道) es previo a la di
+          +stinción ser/no-ser. El primer capítulo del _Tao Te Ching_ (La
+          +ozi, ~siglo VI-IV aC):                                        
+      373 +                                                              
+      374 +#align(center)[                                               
+      375 +  #text(style: "italic")[                                     
+      376 +    道可道，非常道。名可名，非常名。\                         
+      377 +    El Tao que puede nombrarse no es el Tao eterno.\          
+      378 +    El nombre que puede pronunciarse no es el nombre eterno.  
+      379 +  ]                                                           
+      380 +]                                                             
+      381 +                                                              
+      382 +Lo que no puede formalizarse no es una limitación del modelo —
+          + es la condición de posibilidad de todo lo que sí puede formal
+          +izarse. La ontología operacional que puede expresarse completa
+          +mente en un schema no captura la realidad operacional completa
+          +. Heidegger llegó al mismo lugar por camino diferente: el *tra
+          +sfondo pre-temático* que hace posible cualquier acto de compre
+          +nsión explícita es precisamente lo que escapa a la formalizaci
+          +ón.                                                           
+      383 +                                                              
+      384 +== Yin-Yang — polaridad constitutiva, no oposición            
+      385 +                                                              
+      386 +La confusión más frecuente: yin y yang no son *contrarios* sin
+          +o *complementarios que se definen mutuamente*. Las propiedades
+          + estructurales del par son:                                   
+      387 +                                                              
+      388 +- Cada polo *contiene una semilla del opuesto* — representada 
+          +por el punto de color contrario en el símbolo                 
+      389 +- El exceso de un polo *genera su contrario* (極則反 — la extr
+          +emidad se invierte)                                           
+      390 +- La tensión entre ellos *nunca se resuelve* en síntesis — sol
+          +o se gestiona dinámicamente                                   
+      391 +- Uno *no existe sin el otro* — yin sin yang es vacío inerte, 
+          +yang sin yin es movimiento sin forma                          
+      392 +                                                              
+      393 +Esto es estructuralmente distinto de la dialéctica hegeliana, 
+          +que busca síntesis — resolver la contradicción en una unidad s
+          +uperior. El yin-yang no resuelve. La tensión es constitutiva, 
+          +permanente. Hegel usa la contradicción como *palanca*; el Taoí
+          +smo la habita como *condición*.                               
+      394 +                                                              
+      395 +Heráclito (535-475 aC) es el occidental más cercano: su logos 
+          +como tensión dinámica de opuestos — "el camino hacia arriba y 
+          +hacia abajo son el mismo" — es estructuralmente idéntico al yi
+          +n-yang. La diferencia: Heráclito vive en la tensión, Hegel la 
+          +usa para superarla.                                           
+      396 +                                                              
+      397 +== Correspondencias con el modelo operacional                 
+      398 +                                                              
+      399 +#table(                                                       
+      400 +  columns: (1.4fr, 1.3fr, 1.8fr, 1.8fr),                      
+      401 +  inset: 7pt,                                                 
+      402 +  fill: (col, row) => if row == 0 { rgb("#c04000") } else if c
+          +alc.odd(row) { rgb("#fdf6f0") } else { white },               
+      403 +  table.header(                                               
+      404 +    text(fill: white, weight: "bold")[Concepto taoísta],      
+      405 +    text(fill: white, weight: "bold")[Texto / Época],         
+      406 +    text(fill: white, weight: "bold")[Equivalente occidental m
+          +ás cercano],                                                  
+      407 +    text(fill: white, weight: "bold")[Correspondencia operacio
+          +nal],                                                         
+      408 +  ),                                                          
+      409 +  [*Tao* (道) — principio previo a toda distinción],          
+      410 +  [Laozi, _Tao Te Ching_ \ (~6-4 aC)],                        
+      411 +  [Ser de Heidegger — previo a los entes],                    
+      412 +  [Trasfondo pre-temático que ningún schema captura completame
+          +nte],                                                         
+      413 +                                                              
+      414 +  [*Yin / Yang* — polaridad que no se resuelve],              
+      415 +  [Tradición Zhou (~1100 aC) \ sistematizada en _I Ching_],   
+      416 +  [Heráclito — logos como tensión de opuestos],               
+      417 +  [Ontología (ser) / Reflection (devenir) — capas que se neces
+          +itan mutuamente],                                             
+      418 +                                                              
+      419 +  [*Wu wei* (無為) — acción sin forzar],                      
+      420 +  [Laozi, _Tao Te Ching_],                                    
+      421 +  [Sin equivalente exacto en Occidente],                      
+      422 +  ["No enforcement" — coherencia que emerge de adopción justif
+          +icada, no impuesta],                                          
+      423 +                                                              
+      424 +  [*Ji* (機) — el momento exacto de transformación],          
+      425 +  [Zhuangzi, _Zhuangzi_ \ (~4 aC)],                           
+      426 +  [Kripke — mundos posibles accesibles],                      
+      427 +  [Gates — umbral cualitativo entre estados, no condición bina
+          +ria],                                                         
+      428 +                                                              
+      429 +  [*Wu ji → Tai ji* — de lo indiferenciado a la polaridad],   
+      430 +  [Zhou Dunyi (1017-1073) \ _Taijitu shuo_],                  
+      431 +  [Husserl — ontología formal vs. regional],                  
+      432 +  [schema.ncl (vacío universal) → core.ncl (polaridad concreta
+          + del proyecto)],                                              
+      433 +                                                              
+      434 +  [*Relatividad perspectival*],                               
+      435 +  [Zhuangzi — la rana del pozo],                              
+      436 +  [Wittgenstein tardío — juegos de lenguaje],                 
+      437 +  [Toda ontología es una conceptualización, no una captura de 
+          +la realidad (Gruber)],                                        
+      438 +)                                                             
+      439 +                                                              
+      440 +== Las cuatro correspondencias profundas                      
+      441 +                                                              
+      442 +*Yin/Yang ↔ Ontología/Reflection.* La dualidad ya está en el d
+          +iseño del protocolo. La ontología (core.ncl) captura lo que ES
+          + — invariantes, estructura, ser; capa yang: formal, dura, defi
+          +nida. La reflection captura lo que DEVIENE — operaciones, modo
+          +s, movimiento; capa yin: fluida, procesual. Como en el símbolo
+          +: cada una contiene una semilla de la otra. La ontología descr
+          +ibe estados *futuros* (desired_state) — contiene movimiento. L
+          +a reflection puede *actualizar* la ontología vía ADRs — contie
+          +ne forma. Separarlas es necesario porque tienen ritmos de camb
+          +io distintos; fundirlas colapsaría el yin en el yang.         
+      443 +                                                              
+      444 +*Wu wei ↔ "No enforcement".* Uno de los invariantes del protoc
+          +olo: coherencia voluntaria que emerge de la adopción justifica
+          +da, sin mecanismo de enforcement. Esto es wu wei aplicado a di
+          +seño de sistemas. No impones — creas las condiciones para que 
+          +el alineamiento emerja naturalmente. Un sistema de tipos fuert
+          +es fuerza; una ontología operacional orienta. La diferencia es
+          + exactamente la diferencia entre imponer el Tao y fluir con él
+          +.                                                             
+      445 +                                                              
+      446 +*Tensiones ↔ Yin-Yang permanente.* Las tensiones en un modelo 
+          +operacional ("formalización vs. fricción de adopción", "ontolo
+          +gía vs. reflection") son explícitamente irresolubles. No se bu
+          +sca síntesis — se gestionan. Si eliminas la tensión entre form
+          +alización y adopción maximizando uno de los polos, el sistema 
+          +muere: o es un schema perfecto que nadie adopta, o es caos ope
+          +racional que todos adoptan pero nadie puede razonar sobre él. 
+      447 +                                                              
+      448 +*Gates ↔ Ji (機).* El concepto taoísta de ji es el punto de in
+          +flexión donde el potencial se actualiza — no un estado sino el
+          + umbral entre estados. Dinámico, cualitativo, contextual. Las 
+          +gates en el modelo operacional son exactamente esto: no checkb
+          +oxes — son condiciones que marcan el momento en que un estado 
+          +puede transformarse en otro.                                  
+      449 +                                                              
+      450 +== La crítica taoísta a la ontología clasificatoria           
+      451 +                                                              
+      452 +Los sistemas como OWL intentan eliminar la ambigüedad mediante
+          + clasificación precisa. El _Tao Te Ching_ advierte: cuanto más
+          + preciso el nombre, más excluyes. La realidad vive en el espac
+          +io *entre* las categorías, no en las categorías mismas.       
+      453 +                                                              
+      454 +Esto explica por qué las ontologías biomédicas — las más madur
+          +as del campo — sufren continuamente con casos límite, dependen
+          +cia del contexto y la necesidad de múltiples ejes de clasifica
+          +ción simultáneos. La vista taoísta diría: estáis intentando re
+          +solver una tensión constitutiva. La precisión y la cobertura s
+          +on yin y yang — no se resuelven, se equilibran.               
+      455 +                                                              
+      456 +Zhuangzi va más lejos que Laozi. Toda distinción ontológica es
+          + perspectival. Esto no lleva al nihilismo sino a la *apropiada
+          + contextualidad*: la ontología correcta para un proyecto no es
+          + la más formalmente precisa sino la que el equipo puede sosten
+          +er y evolucionar. Gruber llegó al mismo lugar en 1993 sin sabe
+          +rlo: una ontología es "una especificación de una conceptualiza
+          +ción" — un punto de vista compartido, no una verdad metafísica
+          +.                                                             
+      457 +                                                              
+      458 +#pagebreak()                                                  
+      459 +                                                              
+      460 +// ─── 5. DAGs ───────────────────────────────────────────────
+          +─────────────────                                             
+      461 +                                                              
+      462 += DAGs — genealogía y función en ontología operacional        
+      463 +                                                              
+      464 +== El problema que fuerza el grafo                            
+      465 +                                                              
+      466 +Antes del grafo hay un problema: ¿cómo representar conocimient
+          +o estructurado de forma que se pueda razonar sobre él? La pros
+          +a no es consultable ni verificable. Las listas planas pierden 
+          +las relaciones. Las jerarquías de árbol pierden los nodos con 
+          +múltiples padres. El grafo emerge como la estructura mínima qu
+          +e captura relaciones arbitrarias entre conceptos.             
+      467 +                                                              
+      468 +Pero el grafo general tiene un problema: puede tener ciclos. Y
+          + los ciclos en ontología son epistemológicamente problemáticos
+          +.                                                             
+      469 +                                                              
+      470 +== Por qué la aciclicidad — el problema de la justificación ci
+          +rcular                                                        
+      471 +                                                              
+      472 +Dos tradiciones independientes llegan al mismo resultado:     
+      473 +                                                              
+      474 +*Aristóteles* (_Analíticos Posteriores_): una definición válid
+          +a procede de género más diferencia específica. El género es má
+          +s general que lo definido — la jerarquía es estrictamente desc
+          +endente. Si A se define en términos de B y B en términos de A,
+          + ninguno está definido. La justificación circular es un defect
+          +o lógico.                                                     
+      475 +                                                              
+      476 +*Russell* (1901): la paradoja del conjunto de todos los conjun
+          +tos que no se contienen a sí mismos. La solución en _Principia
+          + Mathematica_ (1910, con Whitehead) es la *teoría de tipos*: l
+          +os tipos forman una jerarquía estricta — un tipo de nivel N so
+          +lo puede referirse a tipos de nivel N-1 o inferior. Esto es es
+          +tructuralmente un DAG. La aciclicidad no es una restricción ar
+          +bitraria — es la solución formal al colapso por autorreferenci
+          +a.                                                            
+      477 +                                                              
+      478 +La convergencia: en lógica, ontología formal y teoría de tipos
+          +, la aciclicidad es el requisito de que las definiciones sean 
+          +*bien fundadas*. No puede haber descenso infinito ni circulari
+          +dad.                                                          
+      479 +                                                              
+      480 +== Línea genealógica                                          
+      481 +                                                              
+      482 +#table(                                                       
+      483 +  columns: (0.8fr, 2fr, 2.5fr),                               
+      484 +  inset: 7pt,                                                 
+      485 +  fill: (col, row) => if row == 0 { rgb("#c04000") } else if c
+          +alc.odd(row) { rgb("#fdf6f0") } else { white },               
+      486 +  table.header(                                               
+      487 +    text(fill: white, weight: "bold")[Año],                   
+      488 +    text(fill: white, weight: "bold")[Autor / Hito],          
+      489 +    text(fill: white, weight: "bold")[Contribución al DAG en o
+          +ntología],                                                    
+      490 +  ),                                                          
+      491 +  [384-322 aC],                                               
+      492 +  [Aristóteles — _Analíticos Posteriores_],                   
+      493 +  [Definición bien fundada: jerarquía sin circularidad. Género
+          + → diferencia → especie.],                                    
+      494 +                                                              
+      495 +  [1879],                                                     
+      496 +  [Frege — _Begriffsschrift_],                                
+      497 +  [Lógica de predicados: relaciones formales entre conceptos. 
+          +Base para representar grafos de conocimiento.],               
+      498 +                                                              
+      499 +  [1901-1910],                                                
+      500 +  [Russell & Whitehead — _Principia Mathematica_],            
+      501 +  [Teoría de tipos como DAG: aciclicidad como solución al cola
+          +pso por autorreferencia.],                                    
+      502 +                                                              
+      503 +  [1933],                                                     
+      504 +  [Tarski — "El concepto de verdad"],                         
+      505 +  [Verdad como correspondencia entre proposiciones y estructur
+          +as (grafos). Base de la validación formal.],                  
+      506 +                                                              
+      507 +  [1958],                                                     
+      508 +  [US Navy — PERT charts (programa Polaris)],                 
+      509 +  [DAG para planificación de proyectos: camino crítico como re
+          +corrido del grafo de dependencias temporales.],               
+      510 +                                                              
+      511 +  [1969],                                                     
+      512 +  [Collins & Quillian — "Retrieval time from semantic memory"]
+          +,                                                             
+      513 +  [Primera red semántica computacional: is-a como DAG. La topo
+          +logía del grafo predice tiempos de reacción humanos.],        
+      514 +                                                              
+      515 +  [1974],                                                     
+      516 +  [Minsky — "A Framework for Representing Knowledge"],        
+      517 +  [Frames con herencia múltiple: DAG sobre árbol. Primer siste
+          +ma con nodos de múltiples padres.],                           
+      518 +                                                              
+      519 +  [1976],                                                     
+      520 +  [Feldman — Make],                                           
+      521 +  [Build system como DAG de dependencias. Orden topológico par
+          +a compilación incremental. Ciclos = error explícito.],        
+      522 +                                                              
+      523 +  [1985],                                                     
+      524 +  [Brachman & Schmolze — KL-ONE],                             
+      525 +  [Description Logic: subsunción formal como DAG. Antecedente 
+          +directo de OWL.],                                             
+      526 +                                                              
+      527 +  [1988],                                                     
+      528 +  [Pearl — _Probabilistic Reasoning in Intelligent Systems_], 
+      529 +  [Causalidad como DAG: la asimetría temporal hace la causalid
+          +ad inherentemente acíclica.],                                 
+      530 +                                                              
+      531 +  [2000],                                                     
+      532 +  [Pearl — _Causality_],                                      
+      533 +  [Do-calculus: intervención como operación sobre el DAG causa
+          +l. Separación entre correlación y causa.],                    
+      534 +                                                              
+      535 +  [2001],                                                     
+      536 +  [Berners-Lee et al. — Web Semántica],                       
+      537 +  [RDF/OWL: grafos de conocimiento como estándar industrial. D
+          +AG de subsunción en jerarquías de clases.],                   
+      538 +                                                              
+      539 +  [2005],                                                     
+      540 +  [Torvalds — Git],                                           
+      541 +  [Historia de commits como DAG: irreversibilidad del tiempo e
+          +n control de versiones. Merges = nodos con múltiples padres.],
+      542 +)                                                             
+      543 +                                                              
+      544 +== Las cuatro funciones del DAG en ontología operacional      
+      545 +                                                              
+      546 +*① Traversal determinista para análisis de impacto.* Si cambia
+          +s un nodo, el DAG define exactamente qué nodos son alcanzables
+          + desde él (descendientes) y qué nodos lo alcanzan (ancestros).
+          + Esto es análisis de impacto: "si modifico este invariante, ¿q
+          +ué tensiones, prácticas y gates se ven afectadas?" En un grafo
+          + cíclico esto es indecidible — el cierre transitivo de un nodo
+          + puede ser todo el grafo.                                     
+      547 +                                                              
+      548 +*② Orden topológico para ejecución.* Los modos operacionales s
+          +on DAGs de pasos con dependencias. El ordenamiento topológico 
+          +da el orden de ejecución válido sin necesidad de un scheduler 
+          +externo. Si el paso B depende del paso A, A siempre precede a 
+          +B en cualquier topological sort.                              
+      549 +                                                              
+      550 +*③ Justificación bien fundada — sin circularidad epistémica.* 
+          +El debate clásico en epistemología enfrenta fundacionalismo (e
+          +l conocimiento descansa en fundamentos no circulares — DAG con
+          + nodos raíz) contra coherentismo (red de creencias mutuamente 
+          +sustentadas — puede ser cíclica). La ontología operacional es 
+          +fundacionalista por diseño: los invariantes son nodos raíz jus
+          +tificados por ADR explícito, no por otros nodos del grafo. Un 
+          +ciclo significaría: "este invariante existe porque esta prácti
+          +ca existe, y esta práctica existe porque este invariante exist
+          +e" — dogma circular, no arquitectura.                         
+      551 +                                                              
+      552 +*④ Rollback con orden definido.* Deshacer cambios requiere un 
+          +orden inverso bien definido. En un grafo cíclico no existe ese
+          + orden. El DAG garantiza que el rollback siempre tiene una sec
+          +uencia válida.                                                
+      553 +                                                              
+      554 +== La tensión con el Taoísmo — DAG vs. ciclo                  
+      555 +                                                              
+      556 +El DAG es *yang* en su estructura: direccional, jerárquico, as
+          +imétrico, irreversible. Codifica la flecha del tiempo — las ca
+          +usas preceden a los efectos, las definiciones preceden a lo de
+          +finido.                                                       
+      557 +                                                              
+      558 +El yin-yang taoísta es *cíclico*: yin genera yang, yang genera
+          + yin. Las estaciones, el día y la noche, vida y muerte — todo 
+          +retorna.                                                      
+      559 +                                                              
+      560 +La resolución está en separar dos temporalidades:             
+      561 +                                                              
+      562 +- *Sincrónica* (snapshot): el DAG es válido dentro de un momen
+          +to dado. La ontología en $t_0$ es un grafo acíclico.          
+      563 +- *Diacrónica* (evolución): la secuencia de snapshots puede te
+          +ner patrones cíclicos. Lo que era tensión puede convertirse en
+          + invariante; lo que era práctica puede ser abandonado y redesc
+          +ubierto.                                                      
+      564 +                                                              
+      565 +El DAG captura la *estructura del ser en un momento*. El ciclo
+          + taoísta captura el *proceso de transformación a través del ti
+          +empo*. Son complementarios, no contradictorios — exactamente c
+          +omo ontología y reflection son complementarios en el protocolo
+          +. El *I Ching* — el libro de los cambios — es un sistema para 
+          +razonar sobre las transformaciones de un estado a otro: 64 hex
+          +agramas como estados (nodos), transformaciones entre ellos com
+          +o aristas. El sistema completo es un grafo. No un DAG — un gra
+          +fo con ciclos, porque la realidad vuelve.                     
+      566 +                                                              
+      567 +#pagebreak()                                                  
+      568 +                                                              
+      569 +// ─── 6. REFERENCIAS ────────────────────────────────────────
+          +─────────────────                                             
+      570 +                                                              
+      571 += Referencias y lecturas recomendadas                         
+      572 +                                                              
+      573 +== Filosofía clásica y ontología                              
+      574 +                                                              
+      575 +#set par(first-line-indent: 0em)                              
+      576 +                                                              
+      577 +*Aristóteles.* _Metafísica_. Traducción de Tomás Calvo Martíne
+          +z. Madrid: Gredos, 1994. [~330 aC] \                          
+      578 +#text(fill: rgb("#666666"), size: 9.5pt)[El libro fundamental 
+          +para la distinción esencia/accidente, potencia/acto y las cate
+          +gorías. El libro IX sobre potencia es la base del modelo estad
+          +o-actual/estado-deseado.]                                     
+      579 +                                                              
+      580 +*Aristóteles.* _Analíticos Posteriores_. Traducción de Miguel 
+          +Candel Sanmartín. Madrid: Gredos, 1988. [~350 aC] \           
+      581 +#text(fill: rgb("#666666"), size: 9.5pt)[Formaliza el requisit
+          +o de definiciones bien fundadas — antecedente directo de la ac
+          +iclicidad en DAGs.]                                           
+      582 +                                                              
+      583 +*Laozi.* _Tao Te Ching_. Traducción y comentarios de Stephen M
+          +itchell. Nueva York: Harper Perennial, 1992. [~siglo VI-IV aC]
+          + \                                                            
+      584 +#text(fill: rgb("#666666"), size: 9.5pt)[Texto fundacional del
+          + Taoísmo. Los primeros y últimos capítulos son los más relevan
+          +tes para la relación con ontología operacional.]              
+      585 +                                                              
+      586 +*Zhuangzi.* _Zhuangzi: The Complete Writings_. Traducción de B
+          +rook Ziporyn. Indianapolis: Hackett, 2020. [~siglo IV aC] \   
+      587 +#text(fill: rgb("#666666"), size: 9.5pt)[Especialmente los cap
+          +ítulos internos (1-7). La relatividad perspectival de Zhuangzi
+          + como crítica de toda ontología clasificatoria.]              
+      588 +                                                              
+      589 +*Frege, G.* _Begriffsschrift_. Halle: Nebert, 1879. \         
+      590 +#text(fill: rgb("#666666"), size: 9.5pt)[Primera formulación d
+          +e la lógica de predicados. El antecedente formal de toda repre
+          +sentación de conocimiento estructurado.]                      
+      591 +                                                              
+      592 +*Russell, B. & Whitehead, A.N.* _Principia Mathematica_. Cambr
+          +idge: Cambridge University Press, 1910-1913. \                
+      593 +#text(fill: rgb("#666666"), size: 9.5pt)[La teoría de tipos co
+          +mo DAG para evitar la autorreferencia circular. El capítulo in
+          +troductorio explica el principio del círculo vicioso.]        
+      594 +                                                              
+      595 +*Wittgenstein, L.* _Tractatus Logico-Philosophicus_. Londres: 
+          +Kegan Paul, 1922. \                                           
+      596 +#text(fill: rgb("#666666"), size: 9.5pt)["Los límites de mi le
+          +nguaje son los límites de mi mundo." El fundamento de por qué 
+          +el schema Nickel define la frontera de lo representable en el 
+          +sistema.]                                                     
+      597 +                                                              
+      598 +*Wittgenstein, L.* _Investigaciones filosóficas_. Oxford: Blac
+          +kwell, 1953. \                                                
+      599 +#text(fill: rgb("#666666"), size: 9.5pt)[Juegos de lenguaje co
+          +mo práctica. El significado como uso — los modos operacionales
+          + como juegos de lenguaje de un proyecto.]                     
+      600 +                                                              
+      601 +*Hegel, G.W.F.* _Fenomenología del Espíritu_. Traducción de Ma
+          +nuel Jiménez Redondo. Valencia: Pre-Textos, 2006. [1807] \    
+      602 +#text(fill: rgb("#666666"), size: 9.5pt)[Dialéctica tesis-antí
+          +tesis-síntesis. Contrástese con el yin-yang: la diferencia ent
+          +re síntesis y tensión gestionada permanentemente.]            
+      603 +                                                              
+      604 +*Husserl, E.* _Ideas relativas a una fenomenología pura y una 
+          +filosofía fenomenológica_. México: FCE, 1993. [1913] \        
+      605 +#text(fill: rgb("#666666"), size: 9.5pt)[La distinción ontolog
+          +ía formal/regional — estructura vacía universal vs. conceptos 
+          +de dominio específico.]                                       
+      606 +                                                              
+      607 +*Whitehead, A.N.* _Process and Reality_. Nueva York: Free Pres
+          +s, 1978. [1929] \                                             
+      608 +#text(fill: rgb("#666666"), size: 9.5pt)[La realidad como proc
+          +eso, no sustancia. El fundamento filosófico de la capa reflect
+          +ion como devenir distinto del ser ontológico.]                
+      609 +                                                              
+      610 +*Kripke, S.* _Naming and Necessity_. Cambridge: Harvard Univer
+          +sity Press, 1980. \                                           
+      611 +#text(fill: rgb("#666666"), size: 9.5pt)[Propiedades esenciale
+          +s rígidas y semántica de mundos posibles. El fundamento modal 
+          +de los invariantes operacionales.]                            
+      612 +                                                              
+      613 +*Searle, J.* _The Construction of Social Reality_. Nueva York:
+          + Free Press, 1995. \                                          
+      614 +#text(fill: rgb("#666666"), size: 9.5pt)[Hechos institucionale
+          +s como fundamento de las decisiones declaradas. El ADR como ac
+          +to de habla colectivo que crea realidad operacional.]         
+      615 +                                                              
+      616 +*Parfit, D.* _Reasons and Persons_. Oxford: Oxford University 
+          +Press, 1984. \                                                
+      617 +#text(fill: rgb("#666666"), size: 9.5pt)[Identidad a través de
+          +l tiempo. La pregunta "¿es el mismo proyecto?" cuando sus inva
+          +riantes cambian vía ADR.]                                     
+      618 +                                                              
+      619 +*Ladyman, J. & Ross, D.* _Every Thing Must Go: Metaphysics Nat
+          +uralized_. Oxford: Oxford University Press, 2007. \           
+      620 +#text(fill: rgb("#666666"), size: 9.5pt)[Realismo estructural:
+          + las relaciones son más reales que las entidades. El DAG como 
+          +estructura fundamental, los nodos como secundarios.]          
+      621 +                                                              
+      622 +*Quine, W.V.O.* "On What There Is." _Review of Metaphysics_, 2
+          +(5), 1948. \                                                  
+      623 +#text(fill: rgb("#666666"), size: 9.5pt)[El concepto de compro
+          +miso ontológico: toda teoría asume la existencia de ciertas en
+          +tidades. Todo schema asume una ontología aunque no lo declare.
+          +]                                                             
+      624 +                                                              
+      625 +*Tarski, A.* "The Concept of Truth in Formalized Languages." E
+          +n _Logic, Semantics, Metamathematics_. Oxford: Clarendon Press
+          +, 1956. [1933] \                                              
+      626 +#text(fill: rgb("#666666"), size: 9.5pt)[Verdad como correspon
+          +dencia. El fundamento de la validación formal — nickel typeche
+          +ck como verificación tarsquiana.]                             
+      627 +                                                              
+      628 +== Inteligencia artificial y representación del conocimiento  
+      629 +                                                              
+      630 +*Minsky, M.* "A Framework for Representing Knowledge." MIT AI 
+          +Lab Memo 306, 1974. \                                         
+      631 +#text(fill: rgb("#666666"), size: 9.5pt)[Los frames como prime
+          +ra estructura computacional análoga a la sustancia aristotélic
+          +a con propiedades heredables.]                                
+      632 +                                                              
+      633 +*Collins, A.M. & Quillian, M.R.* "Retrieval time from semantic
+          + memory." _Journal of Verbal Learning and Verbal Behavior_, 8(
+          +2), 240-247, 1969. \                                          
+      634 +#text(fill: rgb("#666666"), size: 9.5pt)[Primera red semántica
+          + computacional. Demuestra que la topología del grafo de conoci
+          +miento predice el comportamiento cognitivo.]                  
+      635 +                                                              
+      636 +*Brachman, R.J. & Schmolze, J.G.* "An overview of the KL-ONE k
+          +nowledge representation system." _Cognitive Science_, 9(2), 17
+          +1-216, 1985. \                                                
+      637 +#text(fill: rgb("#666666"), size: 9.5pt)[Description Logic: la
+          + subsunción como relación formal que genera el DAG de concepto
+          +s. Antecedente teórico de OWL.]                               
+      638 +                                                              
+      639 +*Gruber, T.* "A translation approach to portable ontology spec
+          +ifications." _Knowledge Acquisition_, 5(2), 199-220, 1993. \  
+      640 +#text(fill: rgb("#666666"), size: 9.5pt)[La definición canónic
+          +a: ontología como especificación de una conceptualización. Ref
+          +erencia obligatoria para cualquier trabajo en ontología de sof
+          +tware.]                                                       
+      641 +                                                              
+      642 +*Lenat, D.B. & Guha, R.V.* _Building Large Knowledge-Based Sys
+          +tems: Representation and Inference in the Cyc Project_. Readin
+          +g: Addison-Wesley, 1990. \                                    
+      643 +#text(fill: rgb("#666666"), size: 9.5pt)[El proyecto más ambic
+          +ioso de ontología operacional. Sus fracasos son tan informativ
+          +os como sus logros.]                                          
+      644 +                                                              
+      645 +*Pearl, J.* _Probabilistic Reasoning in Intelligent Systems_. 
+          +San Mateo: Morgan Kaufmann, 1988. \                           
+      646 +#text(fill: rgb("#666666"), size: 9.5pt)[Bayesian networks com
+          +o DAGs causales. La formalización de por qué la causalidad es 
+          +inherentemente acíclica.]                                     
+      647 +                                                              
+      648 +*Pearl, J.* _Causality: Models, Reasoning, and Inference_. Cam
+          +bridge: Cambridge University Press, 2000. \                   
+      649 +#text(fill: rgb("#666666"), size: 9.5pt)[El do-calculus. Disti
+          +nción entre correlación y causa como operaciones sobre el DAG.
+          + Segunda edición (2009) incluye el modelo causal completo.]   
+      650 +                                                              
+      651 +*Guarino, N. (ed.)* _Formal Ontology in Information Systems_. 
+          +Amsterdam: IOS Press, 1998. \                                 
+      652 +#text(fill: rgb("#666666"), size: 9.5pt)[Actas de FOIS'98. La 
+          +conferencia fundacional del campo. Especialmente el artículo i
+          +ntroductorio de Guarino sobre ontología formal vs. de dominio.
+          +]                                                             
+      653 +                                                              
+      654 +*Berners-Lee, T., Hendler, J. & Lassila, O.* "The Semantic Web
+          +." _Scientific American_, 284(5), 34-43, 2001. \              
+      655 +#text(fill: rgb("#666666"), size: 9.5pt)[El artículo que popul
+          +arizó las ontologías en ingeniería de software. La visión de l
+          +a web como grafo de conocimiento estructurado.]               
+      656 +                                                              
+      657 +== Ingeniería de software y praxis                            
+      658 +                                                              
+      659 +*Nygard, M.* "Documenting Architecture Decisions." Cognitect B
+          +log, 2011. Disponible en: cognitect.com/blog \                
+      660 +#text(fill: rgb("#666666"), size: 9.5pt)[La propuesta original
+          + de ADRs como documentación ligera de decisiones. El precursor
+          + directo de los ADRs con constraints tipadas.]                
+      661 +                                                              
+      662 +*Brown, S.* _Software Architecture for Developers_. Leanpub, 2
+          +014. \                                                        
+      663 +#text(fill: rgb("#666666"), size: 9.5pt)[El modelo C4 como alt
+          +ernativa a los diagramas UML. Introduce la idea de múltiples n
+          +iveles de abstracción para describir un sistema.]             
+      664 +                                                              
+      665 +*Hohpe, G. & Woolf, B.* _Enterprise Integration Patterns_. Bos
+          +ton: Addison-Wesley, 2003. \                                  
+      666 +#text(fill: rgb("#666666"), size: 9.5pt)[Patrones como vocabul
+          +ario compartido — la idea de que nombrar los patrones operacio
+          +nales es un acto ontológico.]                                 
+      667 +                                                              
+      668 +== Perspectivas adicionales                                   
+      669 +                                                              
+      670 +*Zhou Dunyi.* _Taijitu shuo_ (Explicación del Diagrama del Sup
+          +remo Último). [1017-1073] \                                   
+      671 +#text(fill: rgb("#666666"), size: 9.5pt)[La cosmología taoísta
+          + sistematizada: wu ji → tai ji → yin-yang. La transición de lo
+          + indiferenciado a la polaridad como modelo de emergencia estru
+          +ctural.]                                                      
+      672 +                                                              
+      673 +*Heidegger, M.* _Ser y Tiempo_. Traducción de Jorge Eduardo Ri
+          +vera. Santiago: Editorial Universitaria, 1997. [1927] \       
+      674 +#text(fill: rgb("#666666"), size: 9.5pt)[El trasfondo pre-temá
+          +tico que ninguna ontología captura completamente. La ausencia 
+          +notable en todos los modelos operacionales de software.]      
+      675 +                                                              
+      676 +*Heráclito.* _Fragmentos_. Edición de Rodolfo Mondolfo. Buenos
+          + Aires: Losada, 2007. [~500 aC] \                             
+      677 +#text(fill: rgb("#666666"), size: 9.5pt)[El único pensador pre
+          +socrático estructuralmente equivalente al yin-yang. El fragmen
+          +to DK22B60 — "el camino hacia arriba y hacia abajo son el mism
+          +o" — es el punto de partida.]                                 
+      678 +                                                              
+      679 +*Bateson, G.* _Steps to an Ecology of Mind_. Nueva York: Balla
+          +ntine, 1972. \                                                
+      680 +#text(fill: rgb("#666666"), size: 9.5pt)[La diferencia que hac
+          +e una diferencia. Epistemología sistémica que conecta Taoísmo,
+          + cibernética y representación del conocimiento. Perspectiva po
+          +co citada en CS pero fundamentalmente relevante.]             
+      681 +                                                              
+      682 +*Winograd, T. & Flores, F.* _Understanding Computers and Cogni
+          +tion_. Norwood: Ablex, 1986. \                                
+      683 +#text(fill: rgb("#666666"), size: 9.5pt)[La crítica más influy
+          +ente a la IA simbólica desde Heidegger. Argumenta que el trasf
+          +ondo de comprensión no puede formalizarse — límite de toda ont
+          +ología operacional.]                                          
+
+⏺ Bash(typst compile ontologia-filosofia-software.typ
+      ontologia-filosofia-software.pdf 2>&1)
+  ⎿  (No output)
+
+⏺ Bash(ls -lh /Users/Akasha/Development/ontoref/assets/presentation/ontologia-fi
+      losofia-software.pdf && typst query --field value
+      ontologia-filosofia-software.typ "<sy…)
+  ⎿  -rw-r--r--@ 1 jesusperezlorenzo  admin   321K Apr 24 22:46 /Users/Akas
+     ha/Development/ontoref/assets/presentation/ontologia-filosofia-softwar
+     e.pdf
+     … +2 lines (ctrl+o to expand)
+
+⏺ PDF actualizado — 20 páginas, 321K. Lo que se añadió:
+
+  Sección 4 — Taoísmo, Yin-Yang y ontología operacional
+  - Por qué el Taoísmo ve lo que la tradición occidental no ve (el trasfondo
+  pre-temático que ningún schema captura)
+  - La distinción estructural entre yin-yang y dialéctica hegeliana — polaridad
+  permanente vs. síntesis
+  - Tabla comparativa: Tao, Wu wei, Ji, Wu ji/Tai ji, relatividad perspectival ↔
+   equivalentes occidentales y operacionales
+  - Las cuatro correspondencias profundas con el modelo de ontoref
+  - La crítica taoísta a la ontología OWL-clasificatoria
+
+  Sección 5 — DAGs: genealogía y función
+  - El problema que fuerza el grafo y por qué la aciclicidad no es arbitraria
+  - Tabla genealógica: Aristóteles (384 aC) → Git (2005), con contribución de
+  cada autor al DAG en ontología
+  - Las cuatro funciones del DAG: traversal determinista, orden topológico,
+  justificación bien fundada, rollback ordenado
+  - La tensión DAG/ciclo taoísta — resolución por separación
+  sincrónico/diacrónico
+
+  Sección 6 — Referencias (29 entradas anotadas)
+  - Filosofía clásica: Aristóteles, Laozi, Zhuangzi, Frege, Russell,
+  Wittgenstein, Hegel, Husserl, Whitehead, Kripke, Searle, Parfit, Ladyman,
+  Quine, Tarski
+  - IA y representación del conocimiento: Minsky, Collins/Quillian,
+  Brachman/Schmolze, Gruber, Lenat, Pearl, Guarino, Berners-Lee
+  - Ingeniería de software: Nygard, Brown, Hohpe/Woolf
+  - Perspectivas adicionales: Zhou Dunyi, Heidegger, Heráclito, Bateson,
+  Winograd/Flores
+
+✻ Baked for 3m 4s
+
+※ recap: Generating a philosophical PDF on ontology, Taoism, DAGs and their software parallels. The expanded document with new sections and 29 annotated references compiled successfully to 20 pages at assets/presentation/ontologia-filosofia-software.pdf. (disable recaps in /config)
+
+❯ ¿ En todo esto tiene algún sentido la física cuántica y la computación        
+  cuántica ?                                                                    
+
+⏺ Sí, pero hay que separar dónde la conexión es genuina de dónde es metáfora
+  seductora.
+
+  ---
+  Donde la física cuántica es genuinamente relevante — ontología filosófica
+
+  El golpe más profundo: Aristóteles estaba equivocado en el nivel fundamental.
+
+  La ontología aristotélica asume que las sustancias tienen propiedades
+  definidas. La mecánica cuántica demuestra que, en el nivel fundamental, las
+  propiedades no son definidas hasta que se miden. Un electrón no tiene posición
+   — la posición emerge del acto de medición. Esto no es ignorancia del
+  observador sobre una propiedad preexistente — el teorema de Kochen-Specker
+  (1967) lo prueba formalmente: es imposible asignar valores definidos a todos
+  los observables cuánticos simultáneamente, incluso antes de la medición. La
+  indeterminación es ontológica, no epistémica.
+
+  El suelo sobre el que se construyeron 2.300 años de ontología occidental se
+  mueve.
+
+  ---
+  Las tres conexiones genuinas
+
+  1. Complementariedad de Bohr ↔ Yin-Yang y la dualidad Ontología/Reflection
+
+  Niels Bohr (1885-1962) formuló el principio de complementariedad: onda y
+  partícula son dos descripciones mutuamente excluyentes pero igualmente
+  necesarias del mismo fenómeno. No puedes observar ambas simultáneamente — el
+  aparato de medición determina qué aspecto revelas.
+
+  Esto es estructuralmente idéntico al yin-yang. Bohr lo sabía — adoptó el
+  símbolo taoísta como parte de su escudo de armas cuando fue ennoblecido en
+  1947. La inscripción: Contraria sunt complementa — los contrarios son
+  complementarios.
+
+  La dualidad ontología/reflection del modelo operacional tiene la misma
+  estructura: no puedes capturar el ser y el devenir en una única
+  representación. El "aparato de medición" — qué preguntas haces al sistema —
+  determina qué aspecto revelas. La capa que uses (core.ncl o los modos de
+  reflection) determina qué aspecto del proyecto ves. Ninguna es más verdadera —
+   son complementarias.
+
+  2. Contextualidad cuántica ↔ Ontología perspectival
+
+  El teorema de Kochen-Specker (1967) prueba algo más radical que la
+  incertidumbre de Heisenberg: el resultado de medir un observable cuántico
+  depende de qué otros observables se miden simultáneamente. Las propiedades son
+   contextuales — no existen independientemente del contexto de medición.
+
+  Esto conecta con tres tradiciones de nuestra discusión:
+
+  - Zhuangzi: toda distinción ontológica es perspectival — la rana del pozo no
+  puede imaginar el mar
+  - Wittgenstein tardío: el significado depende del juego de lenguaje — del
+  contexto de uso
+  - Gruber: una ontología es una conceptualización, no una captura de la
+  realidad
+
+  La contextualidad cuántica convierte esto de posición filosófica a hecho
+  físico. Las propiedades no están en las cosas — emergen de la interacción
+  entre la cosa y el sistema de observación.
+
+  Para la ontología operacional: el "estado" de un proyecto es siempre relativo
+  al actor y al contexto de consulta. Un agente IA y un desarrollador humano
+  consultando el mismo core.ncl están haciendo mediciones con aparatos
+  distintos. Lo que ven es complementario, no idéntico.
+
+  3. El problema del observador ↔ El trasfondo que no puede formalizarse
+
+  En mecánica cuántica, la medición colapsa la función de onda — el observador
+  es parte del sistema. No hay view from nowhere (Thomas Nagel, 1986). Todo
+  conocimiento es conocimiento desde una posición.
+
+  Esto es exactamente lo que Heidegger llamó el trasfondo pre-temático — la
+  condición de posibilidad de cualquier formalización que ella misma no puede
+  ser formalizada. Y lo que el Tao Te Ching afirma en su primer verso: el Tao
+  que puede nombrarse no es el Tao eterno.
+
+  La mecánica cuántica convierte esto en física: el acto de observar constituye
+  lo observado. La ontología operacional que crees que describes objetivamente
+  está siendo constituida por el acto de escribirla.
+
+  ---
+  La interpretación más relevante — Mecánica Cuántica Relacional (Rovelli, 1996)
+
+  Carlo Rovelli propuso en 1996 una interpretación de la mecánica cuántica
+  directamente pertinente a esta conversación: no existen estados absolutos —
+  solo estados relativos a otros sistemas. El estado de una partícula no es una
+  propiedad intrínseca — es una propiedad relacional respecto al sistema con el
+  que interactúa.
+
+  Esto es la ontología más radical posible: no hay propiedades, solo relaciones.
+   Ladyman y Ross (que ya aparecen en nuestra tabla) llegaron al mismo lugar
+  desde la filosofía de la física: realismo estructural — las estructuras
+  relacionales son lo real, las entidades son secundarias.
+
+  Para el modelo operacional: el "estado" de un proyecto (current_state en
+  state.ncl) no es una propiedad intrínseca del proyecto — es una propiedad
+  relacional respecto al actor que consulta, los objetivos que se persiguen, el
+  momento en el ciclo de vida. Dos actores con marcos de referencia distintos
+  pueden tener mediciones del estado coherentes pero distintas. No hay
+  contradicción — hay contextualidad relacional.
+
+  ---
+  La computación cuántica — honestidad sobre sus límites aquí
+
+  La computación cuántica tiene relevancia limitada para la ontología
+  operacional en software hoy. Las razones son prácticas, no filosóficas:
+
+  ┌────────────────┬────────────────────────────────────────────────────────┐
+  │    Aspecto     │                     Situación real                     │
+  ├────────────────┼────────────────────────────────────────────────────────┤
+  │ Tamaño de los  │ Los grafos de ontología de proyecto son pequeños       │
+  │ grafos         │ (decenas a centenas de nodos). La ventaja cuántica     │
+  │                │ aparece a escala masiva.                               │
+  ├────────────────┼────────────────────────────────────────────────────────┤
+  │ Algoritmos     │ Grover (búsqueda no estructurada: O(√N)) y Shor        │
+  │ relevantes     │ (factorización) no mapean a consultas típicas sobre    │
+  │                │ ontologías de proyecto.                                │
+  ├────────────────┼────────────────────────────────────────────────────────┤
+  │                │ Los sistemas cuánticos son extremadamente frágiles —   │
+  │ Decoherencia   │ cualquier interacción con el entorno colapsa la        │
+  │                │ superposición. El uso práctico requiere aislamiento    │
+  │                │ que contradice la operación continua.                  │
+  ├────────────────┼────────────────────────────────────────────────────────┤
+  │ Estado del     │ Los computadores cuánticos actuales (2026) tienen      │
+  │ arte           │ decenas a cientos de qubits con alta tasa de error.    │
+  │                │ Los problemas de ontología no son el caso de uso.      │
+  └────────────────┴────────────────────────────────────────────────────────┘
+
+  Donde sí hay investigación activa y filosóficamente interesante:
+
+  Redes semánticas cuánticas — representar relaciones entre conceptos con
+  amplitudes de probabilidad en lugar de valores booleanos. Un concepto no es o
+  no es subclase de otro — lo es con cierta amplitud. Esto captura algo que las
+  ontologías OWL pierden: el conocimiento humano no es binario, es gradual y
+  contextual.
+
+  Bayesian networks cuánticas — Pearl formalizó la causalidad como DAG con
+  probabilidades condicionales clásicas. La extensión cuántica reemplaza las
+  probabilidades por amplitudes complejas, permitiendo modelar correlaciones
+  causales sin explicación clásica. Relevante para sistemas donde las
+  dependencias entre componentes exhiben comportamiento emergente no reducible a
+   sus partes.
+
+  Quantum knowledge graphs — investigación emergente (2020-presente) sobre
+  representar grafos de conocimiento en espacios de Hilbert. La similitud entre
+  dos conceptos se convierte en producto interno de vectores de estado. Esto es
+  la base matemática de los embeddings en modelos de lenguaje — los LLMs ya
+  están haciendo algo cuántico-adjacent sin saberlo.
+
+  ---
+  La conexión más profunda — y menos obvia
+
+  Los Large Language Models operan en espacios vectoriales de alta dimensión
+  donde el significado es relacional y contextual. Esto no es metáfora cuántica
+  — es la misma matemática de espacios de Hilbert que usa la mecánica cuántica,
+  aplicada a representación semántica.
+
+  La ontología operacional (core.ncl, invariantes, tensiones) es, en este
+  contexto, una base en ese espacio — un conjunto de vectores ortogonales que
+  fijan el marco de referencia en el que el agente opera. Sin esa base, el
+  agente opera en un espacio de alta dimensión sin orientación — genera texto
+  coherente localmente pero sin anclaje semántico al proyecto concreto.
+
+  Inyectar el contexto ontológico al inicio de una sesión de agente es, en la
+  geometría de estos modelos, fijar el marco de referencia antes de que el
+  agente empiece a computar. Es, literalmente, una operación de cambio de base
+  en el espacio de representación.
+
+  ---
+  Síntesis honesta
+
+  Dominio: Física cuántica → ontología filosófica
+  Relevancia: Alta y genuina
+  Por qué: Kochen-Specker destruye el Aristóteles ingenuo; complementariedad
+    formaliza el yin-yang; Rovelli formaliza la ontología relacional
+  ────────────────────────────────────────
+  Dominio: Física cuántica → ontología operacional en software
+  Relevancia: Media, principalmente filosófica
+  Por qué: Contextualidad y relacionalidad son principios de diseño, no solo
+    metáforas
+  ────────────────────────────────────────
+  Dominio: Computación cuántica → ontología operacional hoy
+  Relevancia: Baja, prácticamente nula
+  Por qué: Escala incorrecta, algoritmos que no mapean, hardware inmaduro
+  ────────────────────────────────────────
+  Dominio: Matemáticas cuánticas → LLMs y agentes
+  Relevancia: Alta y directa
+  Por qué: Espacios de Hilbert, productos internos, bases ortogonales — es la
+    misma geometría
+
+  La física cuántica importa aquí no porque vayamos a correr ontologías en
+  computadores cuánticos — sino porque demuestra que las asunciones ontológicas
+  sobre las que construimos el pensamiento clásico (propiedades definidas,
+  estados absolutos, observador separado del sistema) son físicamente
+  incorrectas en el nivel fundamental.
+
+  Eso no invalida la ontología operacional — la valida en otro sentido: si ni la
+   realidad física tiene propiedades absolutas independientes del contexto de
+  observación, menos las tendrá un proyecto de software. La ontología
+  operacional es contextual, relacional y perspectival por diseño correcto, no
+  por limitación.
+
+✻ Sautéed for 1m 48s
+
+❯ por qué entonces es relevante la ontología operaciona y la reflection con los 
+   DAG en desarrollo de software e infraestructuras digitales                   
+
+⏺ La respuesta honesta requiere separar tres planos: qué problema existe, qué
+  aporta cada capa, y por qué ahora más que nunca.
+
+  ---
+  El problema real — no es técnico, es epistémico
+
+  El software moderno tiene un déficit estructural que no se resuelve con
+  mejores herramientas:
+
+  La complejidad ha superado la capacidad de comprensión individual. Un sistema
+  de microservicios en Kubernetes con Terraform, pipelines de CI/CD, múltiples
+  equipos y agentes de IA trabajando en paralelo no cabe en la cabeza de ninguna
+   persona. Nunca cabió — pero durante años pudimos fingir que sí.
+
+  Lo que resulta de eso no es ignorancia puntual — es deriva epistémica 
+  sistémica:
+
+  Lo que el sistema hace          ≠  Lo que el equipo cree que hace
+  Lo que la infra tiene desplegado ≠  Lo que Terraform declara
+  Lo que el ADR decidió           ≠  Lo que el código implementa
+  Lo que el agente asume          ≠  Lo que el contexto real requiere
+
+  Cada una de esas brechas es silenciosa, acumulativa, y solo se descubre cuando
+   falla en producción.
+
+  ---
+  Por qué la ontología operacional — específicamente
+
+  Una ontología operacional no documenta el sistema. Es el modelo del sistema 
+  que el propio sistema puede leer.
+
+  La diferencia no es de grado — es de naturaleza:
+
+  ┌──────────────────────────┬──────────────────────────────────────────────┐
+  │      Documentación       │            Ontología operacional             │
+  │       tradicional        │                                              │
+  ├──────────────────────────┼──────────────────────────────────────────────┤
+  │ Prosa — no consultable   │ Grafo tipado — consultable con queries       │
+  ├──────────────────────────┼──────────────────────────────────────────────┤
+  │ Estática — se            │ Verificada en cada commit (typecheck)        │
+  │ desactualiza             │                                              │
+  ├──────────────────────────┼──────────────────────────────────────────────┤
+  │ Describe lo que hubo     │ Formaliza lo que ES y lo que debe ser        │
+  ├──────────────────────────┼──────────────────────────────────────────────┤
+  │ Solo legible por humanos │ Legible por humanos, agentes, CI             │
+  ├──────────────────────────┼──────────────────────────────────────────────┤
+  │ Captura el qué           │ Captura el qué, el por qué, el no puede      │
+  │                          │ cambiar                                      │
+  └──────────────────────────┴──────────────────────────────────────────────┘
+
+  Lo que la ontología operacional hace en software e infra que nada más hace:
+
+  ① Formaliza lo que no puede romperse. Los invariantes no son comentarios — son
+   constraints tipadas que el sistema verifica. Cuando alguien o algo intenta
+  violarlos, falla antes del runtime, no después.
+
+  ② Captura las tensiones activas como ciudadanos de primera clase. "Coste vs.
+  disponibilidad", "velocidad de despliegue vs. estabilidad" — estos trade-offs
+  viven en la ontología, no en la cabeza de quien lleva más tiempo en el equipo.
+   Cuando ese alguien se va, los trade-offs permanecen.
+
+  ③ Separa el ser del devenir. Lo que el proyecto es (core.ncl) tiene ritmo de
+  cambio diferente a lo que el proyecto opera (reflection). Fusionarlos — como
+  hace cualquier README o wiki — produce documentación que siempre está
+  desactualizada en la mitad más importante.
+
+  ④ Da al agente de IA el contexto que no puede inferir del código. Un agente
+  puede leer todo el código de un sistema y no saber que hay un acuerdo con el
+  equipo de seguridad que impide escalar horizontalmente cierto servicio, o que
+  la decisión de usar gRPC interno fue consecuencia de una tensión específica
+  con latencia que no está en ningún archivo. La ontología operacional hace eso
+  explícito y machine-readable.
+
+  ---
+  Por qué los DAGs — específicamente
+
+  Los DAGs no son una elección estética. Son la única estructura que garantiza
+  cuatro propiedades simultáneamente:
+
+  ① Análisis de impacto decidible. Antes de cambiar un invariante, una práctica,
+   o un nodo de infra — recorre el DAG y sabes exactamente qué se ve afectado.
+  En prosa o grafo cíclico esto no es computable de forma general.
+
+  Concreto: en Terraform, el grafo de recursos es un DAG. terraform plan recorre
+   ese grafo para determinar qué debe crearse, modificarse o destruirse en qué
+  orden. Eso funciona porque el grafo es acíclico. Si hubiera ciclos, el plan
+  sería indecidible.
+
+  ② Orden topológico para ejecución fiable. Cualquier operación con dependencias
+   — despliegue, migración, rollback, onboarding — tiene un orden correcto. El
+  DAG lo da sin necesidad de un coordinador central. Este es el principio detrás
+   de Make (1976), Bazel, Gradle, y cualquier build system no trivial.
+
+  ③ Justificación sin circularidad. Si el invariante A existe porque la práctica
+   B lo requiere, y la práctica B existe porque el invariante A lo define —
+  tienes dogma, no arquitectura. El DAG hace esa circularidad visible e
+  inaceptable. Fuerza que cada nodo tenga justificación bien fundada.
+
+  ④ Rollback determinista. El inverso topológico del DAG es siempre computable.
+  Cualquier operación es reversible en orden definido. Sin DAG, el rollback es
+  un proceso manual, frágil y que depende de que alguien recuerde el orden
+  correcto.
+
+  ---
+  Por qué la reflection — específicamente
+
+  La ontología captura el ser. Pero los sistemas de software no solo son —
+  operan. La reflection es la capa que hace la ontología ejecutable:
+
+  Modos como DAGs operacionales. Un procedimiento sin DAG es una lista de pasos
+  que alguien puede ejecutar en cualquier orden, saltar, o ignorar. Un modo como
+   DAG tiene dependencias explícitas — el sistema puede verificar que se ejecutó
+   correctamente antes de avanzar. Esto es la diferencia entre un runbook y una
+  especificación ejecutable.
+
+  ADRs con constraints tipadas. Una decisión sin constraints es prosa. Una
+  decisión con constraints tipadas es un contrato que el CI puede verificar en
+  cada commit. La pregunta "¿estamos violando alguna decisión arquitectural?"
+  pasa de ser reflexión manual a ser una comprobación automática.
+
+  Historial como append-only. La reflection registra qué operaciones ocurrieron,
+   quién las ejecutó, cuándo, con qué contexto. Eso no es logging — es la
+  historia del devenir del proyecto, trazable y consultable. Cuando algo falla a
+   las 2am, la pregunta "¿qué cambió y quién lo autorizó?" tiene respuesta en el
+   grafo de history, no en Slack.
+
+  Detección de drift. La diferencia entre el estado sellado (lo que declaraste
+  que era verdad en t₀) y el estado actual (lo que realmente hay desplegado) es
+  medible si y solo si formalizaste el estado sellado. Sin eso, la pregunta "¿ha
+   derivado esto?" no tiene respuesta objetiva.
+
+  ---
+  Por qué ahora — la convergencia de tres presiones
+
+  Estos conceptos no son nuevos. Lo nuevo es que tres presiones convergen
+  simultáneamente y hacen la ausencia del modelo operacional costosa de una
+  forma que antes era tolerable:
+
+  Presión 1: Los sistemas son demasiado complejos para modelos mentales.
+  Un monolito de 2005 podía vivir en la cabeza de dos personas. Un sistema
+  cloud-native de 2026 con 40 microservicios, múltiples regiones, pipelines de
+  despliegue automatizados y SLOs por servicio no cabe en ninguna cabeza. El
+  modelo mental individual ha dejado de ser suficiente como mecanismo de
+  coordinación.
+
+  Presión 2: Los agentes de IA tienen acceso de escritura.
+  Hasta hace muy poco, una herramienta de IA que generaba código incorrecto
+  producía texto que un humano revisaba antes de ejecutar. Hoy los agentes
+  pueden abrir PRs, modificar configuración, ejecutar Terraform, y desplegar. El
+   ciclo humano-en-el-loop se está comprimiendo. Un agente que opera sin
+  contexto ontológico en ese entorno no es un asistente — es un actor autónomo
+  con comprensión incompleta del sistema que modifica.
+
+  La ontología operacional es el mecanismo por el cual el agente opera sobre lo 
+  que tú ya clarificaste en lugar de inventar el contexto que le falta.
+
+  Presión 3: Los equipos son multi-actor y distribuidos.
+  Humanos en zonas horarias distintas, agentes de IA con distintos contextos de
+  sesión, pipelines de CI ejecutando en paralelo — todos escribiendo sobre el
+  mismo sistema sin un coordinador central. Sin un protocolo compartido y
+  machine-readable para "qué puede cambiarse, bajo qué condiciones, con qué
+  justificación", la coordinación depende de que todo el mundo lea el mismo
+  Confluence actualizado. Eso nunca funcionó.
+
+  ---
+  El punto de convergencia
+
+  Lo que la ontología operacional + reflection + DAGs aportan en conjunto es
+  algo que ninguna de sus partes aporta sola:
+
+  Un sistema que sabe lo que es, puede verificar que sigue siendo lo que dice 
+  ser, y puede comunicarlo a cualquier actor — humano, agente, o CI — de forma 
+  estructurada y sin ambigüedad.
+
+  Eso no es documentación. Es un modelo operacional vivo.
+
+  La ausencia de ese modelo no es un problema de disciplina del equipo — es una
+  propiedad emergente de sistemas complejos operados sin representación
+  explícita. El drift, la amnesia de decisiones, la ilusión de control en
+  infraestructura declarativa — no son fallos de las personas. Son consecuencias
+   inevitables de operar un sistema complejo sin un modelo del propio sistema.
+
+  La ontología operacional es la respuesta a la pregunta: ¿qué estructura mínima
+   necesita un sistema para conocerse a sí mismo de forma que ese conocimiento 
+  pueda ser usado, verificado y compartido?
+
+✻ Cooked for 1m 27s
+
+※ recap: Explorando la genealogía filosófica y científica de la ontología operacional en software. El siguiente paso es generar el PDF actualizado con las secciones de Taoísmo, DAGs y referencias.
+
+❯ quiero incluir las últimas cuestione y respuestas en los docuemntos con las   
+  referecias que puedes completar al final                                      
+
+⏺ Voy a leer el final del documento para hacer la inserción correcta.
+
+  Read 1 file (ctrl+o to expand)
+
+⏺ Ahora hago la edición — inserto las dos nuevas secciones antes de referencias
+  y amplío la bibliografía al final.
+
+⏺ Update(ontologia-filosofia-software.typ)
+  ⎿  Added 213 lines, removed 1 line
+      566  
+      567  #pagebreak()
+      568  
+      569 -// ─── 6. REFERENCIAS ────────────────────────────────────────
+          -─────────────────                                             
+      569 +// ─── 6. FÍSICA CUÁNTICA ────────────────────────────────────
+          +─────────────────                                             
+      570  
+      571 += Física cuántica, computación cuántica y ontología operaciona
+          +l                                                             
+      572 +                                                              
+      573 +== El golpe más profundo — Aristóteles estaba equivocado en el
+          + nivel fundamental                                            
+      574 +                                                              
+      575 +La ontología aristotélica asume que las sustancias tienen prop
+          +iedades definidas. La mecánica cuántica demuestra que, en el n
+          +ivel fundamental, las propiedades no son definidas hasta que s
+          +e miden. Un electrón no *tiene* posición — la posición emerge 
+          +del acto de medición.                                         
+      576 +                                                              
+      577 +Esto no es ignorancia del observador sobre una propiedad preex
+          +istente. El *teorema de Kochen-Specker* (1967) lo prueba forma
+          +lmente: es imposible asignar valores definidos a todos los obs
+          +ervables cuánticos simultáneamente, incluso antes de la medici
+          +ón. La indeterminación es *ontológica*, no epistémica. El suel
+          +o sobre el que se construyeron 2.300 años de ontología occiden
+          +tal se mueve.                                                 
+      578 +                                                              
+      579 +== Las tres conexiones genuinas                               
+      580 +                                                              
+      581 +=== Complementariedad de Bohr ↔ Yin-Yang y la dualidad Ontolog
+          +ía/Reflection                                                 
+      582 +                                                              
+      583 +Niels Bohr (1885-1962) formuló el *principio de complementarie
+          +dad*: onda y partícula son dos descripciones mutuamente excluy
+          +entes pero igualmente necesarias del mismo fenómeno. No puedes
+          + observar ambas simultáneamente — el aparato de medición deter
+          +mina qué aspecto revelas.                                     
+      584 +                                                              
+      585 +Bohr lo conectó explícitamente con el Taoísmo: cuando fue enno
+          +blecido en 1947 adoptó el símbolo yin-yang como parte de su es
+          +cudo de armas con la inscripción _Contraria sunt complementa_ 
+          +— los contrarios son complementarios.                         
+      586 +                                                              
+      587 +La dualidad ontología/reflection del modelo operacional tiene 
+          +la misma estructura. El "aparato de medición" — qué preguntas 
+          +haces al sistema — determina qué aspecto revelas. La capa que 
+          +uses (core.ncl o los modos de reflection) determina qué ves. N
+          +inguna es más verdadera — son complementarias.                
+      588 +                                                              
+      589 +=== Contextualidad cuántica ↔ Ontología perspectival          
+      590 +                                                              
+      591 +El teorema de Kochen-Specker prueba algo más radical que la in
+          +certidumbre de Heisenberg: el resultado de medir un observable
+          + cuántico depende de *qué otros observables se miden simultáne
+          +amente*. Las propiedades son *contextuales* — no existen indep
+          +endientemente del contexto de medición.                       
+      592 +                                                              
+      593 +Esto convierte en hecho físico lo que tres tradiciones de esta
+          + discusión habían sostenido filosóficamente:                  
+      594 +                                                              
+      595 +- *Zhuangzi*: toda distinción ontológica es perspectival      
+      596 +- *Wittgenstein tardío*: el significado depende del juego de l
+          +enguaje                                                       
+      597 +- *Gruber*: una ontología es una especificación de una concept
+          +ualización, no una captura de la realidad                     
+      598 +                                                              
+      599 +Para la ontología operacional: el "estado" de un proyecto es s
+          +iempre relativo al actor y al contexto de consulta. Un agente 
+          +de IA y un desarrollador humano consultando el mismo core.ncl 
+          +están haciendo mediciones con aparatos distintos. Lo que ven e
+          +s complementario, no idéntico.                                
+      600 +                                                              
+      601 +=== El problema del observador ↔ El trasfondo que no puede for
+          +malizarse                                                     
+      602 +                                                              
+      603 +En mecánica cuántica, la medición colapsa la función de onda —
+          + el observador es parte del sistema. No hay _view from nowhere
+          +_ (Nagel, 1986). Todo conocimiento es conocimiento *desde* una
+          + posición.                                                    
+      604 +                                                              
+      605 +Esto es exactamente lo que Heidegger llamó el trasfondo pre-te
+          +mático — la condición de posibilidad de cualquier formalizació
+          +n que ella misma no puede ser formalizada. Y lo que el _Tao Te
+          + Ching_ afirma: el Tao que puede nombrarse no es el Tao eterno
+          +.                                                             
+      606 +                                                              
+      607 +La ontología operacional que crees que describes objetivamente
+          + está siendo constituida por el acto de escribirla. Esto no es
+          + un defecto del modelo — es un límite estructural de toda repr
+          +esentación.                                                   
+      608 +                                                              
+      609 +== La interpretación más relevante — Mecánica Cuántica Relacio
+          +nal                                                           
+      610 +                                                              
+      611 +Carlo Rovelli propuso en 1996 (_Relational Quantum Mechanics_,
+          + International Journal of Theoretical Physics) que no existen 
+          +estados absolutos — solo estados *relativos a otros sistemas*.
+          + El estado de una partícula no es una propiedad intrínseca — e
+          +s una propiedad relacional respecto al sistema con el que inte
+          +ractúa.                                                       
+      612 +                                                              
+      613 +Esto es la ontología más radical posible y conecta directament
+          +e con Ladyman y Ross (realismo estructural): no hay propiedade
+          +s, solo relaciones. Lo real son las estructuras relacionales, 
+          +no las entidades.                                             
+      614 +                                                              
+      615 +Para el modelo operacional: el "estado" de un proyecto (curren
+          +t_state en state.ncl) no es una propiedad intrínseca del proye
+          +cto — es relacional respecto al actor que consulta, los objeti
+          +vos perseguidos, el momento del ciclo de vida. Dos actores con
+          + marcos de referencia distintos pueden tener mediciones del es
+          +tado coherentes pero distintas. No hay contradicción — hay con
+          +textualidad relacional.                                       
+      616 +                                                              
+      617 +== La computación cuántica — honestidad sobre sus límites     
+      618 +                                                              
+      619 +#table(                                                       
+      620 +  columns: (1.5fr, 2fr, 2fr),                                 
+      621 +  inset: 7pt,                                                 
+      622 +  fill: (col, row) => if row == 0 { rgb("#c04000") } else if c
+          +alc.odd(row) { rgb("#fdf6f0") } else { white },               
+      623 +  table.header(                                               
+      624 +    text(fill: white, weight: "bold")[Aspecto],               
+      625 +    text(fill: white, weight: "bold")[Situación real (2026)], 
+      626 +    text(fill: white, weight: "bold")[Relevancia para ontologí
+          +a operacional],                                               
+      627 +  ),                                                          
+      628 +  [Tamaño de grafos],                                         
+      629 +  [La ventaja cuántica aparece a escala masiva],              
+      630 +  [Los grafos de ontología de proyecto son pequeños — sin vent
+          +aja cuántica],                                                
+      631 +                                                              
+      632 +  [Algoritmos relevantes],                                    
+      633 +  [Grover: O(√N) búsqueda. Shor: factorización polinómica],   
+      634 +  [No mapean a consultas típicas sobre ontologías de proyecto]
+          +,                                                             
+      635 +                                                              
+      636 +  [Decoherencia],                                             
+      637 +  [Cualquier interacción colapsa la superposición],           
+      638 +  [El uso práctico requiere aislamiento que contradice operaci
+          +ón continua],                                                 
+      639 +                                                              
+      640 +  [Redes semánticas cuánticas],                               
+      641 +  [Investigación activa — relaciones con amplitudes de probabi
+          +lidad],                                                       
+      642 +  [Captura gradualidad que OWL binario pierde — relevante a me
+          +diano plazo],                                                 
+      643 +                                                              
+      644 +  [Bayesian networks cuánticas],                              
+      645 +  [Extensión de Pearl con amplitudes complejas],              
+      646 +  [Modelar dependencias con comportamiento emergente no reduci
+          +ble],                                                         
+      647 +                                                              
+      648 +  [LLMs y espacios de Hilbert],                               
+      649 +  [Los embeddings son productos internos en espacio vectorial 
+          +de alta dimensión],                                           
+      650 +  [Directamente relevante: la misma matemática cuántica, aplic
+          +ada a semántica],                                             
+      651 +)                                                             
+      652 +                                                              
+      653 +== La conexión no obvia — LLMs como geometría cuántica aplicad
+          +a                                                             
+      654 +                                                              
+      655 +Los Large Language Models operan en espacios vectoriales de al
+          +ta dimensión donde el significado es relacional y contextual. 
+          +No es metáfora cuántica — es la misma matemática de espacios d
+          +e Hilbert que usa la mecánica cuántica, aplicada a representac
+          +ión semántica.                                                
+      656 +                                                              
+      657 +La ontología operacional (core.ncl, invariantes, tensiones) es
+          +, en este contexto, una *base* en ese espacio — un conjunto de
+          + vectores que fijan el marco de referencia en el que el agente
+          + opera. Inyectar el contexto ontológico al inicio de una sesió
+          +n es, en la geometría de estos modelos, fijar el marco de refe
+          +rencia antes de que el agente empiece a computar.             
+      658 +                                                              
+      659 +La física cuántica importa aquí no porque vayamos a correr ont
+          +ologías en computadores cuánticos — sino porque demuestra que 
+          +las asunciones ontológicas clásicas (propiedades definidas, es
+          +tados absolutos, observador separado del sistema) son físicame
+          +nte incorrectas en el nivel fundamental. La ontología operacio
+          +nal es contextual, relacional y perspectival *por diseño corre
+          +cto*, no por limitación.                                      
+      660 +                                                              
+      661 +#pagebreak()                                                  
+      662 +                                                              
+      663 +// ─── 7. RELEVANCIA EN SOFTWARE E INFRA ─────────────────────
+          +─────────────────                                             
+      664 +                                                              
+      665 += Por qué es relevante en software e infraestructuras digitale
+          +s                                                             
+      666 +                                                              
+      667 +== El problema real no es técnico — es epistémico             
+      668 +                                                              
+      669 +El software moderno tiene un déficit estructural que no se res
+          +uelve con mejores herramientas:                               
+      670 +                                                              
+      671 +#align(center)[                                               
+      672 +  #block(                                                     
+      673 +    fill: rgb("#1a1a1a"),                                     
+      674 +    radius: 4pt,                                              
+      675 +    inset: 12pt,                                              
+      676 +    width: 90%,                                               
+      677 +  )[                                                          
+      678 +    #set text(font: "New Computer Modern Mono", size: 9pt, fil
+          +l: rgb("#cccccc"))                                            
+      679 +    Lo que el sistema hace           ≠  Lo que el equipo cree 
+          +que hace \                                                    
+      680 +    Lo que la infra tiene desplegado ≠  Lo que Terraform decla
+          +ra \                                                          
+      681 +    Lo que el ADR decidió            ≠  Lo que el código imple
+          +menta \                                                       
+      682 +    Lo que el agente asume           ≠  Lo que el contexto rea
+          +l requiere                                                    
+      683 +  ]                                                           
+      684 +]                                                             
+      685 +                                                              
+      686 +Cada una de esas brechas es silenciosa, acumulativa, y solo se
+          + descubre cuando falla en producción. No son fallos de las per
+          +sonas — son consecuencias inevitables de operar un sistema com
+          +plejo sin un modelo del propio sistema.                       
+      687 +                                                              
+      688 +== Lo que la ontología operacional hace que nada más hace     
+      689 +                                                              
+      690 +#table(                                                       
+      691 +  columns: (1.5fr, 1.5fr),                                    
+      692 +  inset: 7pt,                                                 
+      693 +  fill: (col, row) => if row == 0 { rgb("#c04000") } else if c
+          +alc.odd(row) { rgb("#fdf6f0") } else { white },               
+      694 +  table.header(                                               
+      695 +    text(fill: white, weight: "bold")[Documentación tradiciona
+          +l],                                                           
+      696 +    text(fill: white, weight: "bold")[Ontología operacional], 
+      697 +  ),                                                          
+      698 +  [Prosa — no consultable],                                   
+      699 +  [Grafo tipado — consultable con queries estructuradas],     
+      700 +                                                              
+      701 +  [Estática — se desactualiza sin señal],                     
+      702 +  [Verificada en cada commit mediante typecheck],             
+      703 +                                                              
+      704 +  [Describe lo que hubo],                                     
+      705 +  [Formaliza lo que ES y lo que debe ser],                    
+      706 +                                                              
+      707 +  [Solo legible por humanos],                                 
+      708 +  [Legible por humanos, agentes de IA y pipelines de CI],     
+      709 +                                                              
+      710 +  [Captura el qué],                                           
+      711 +  [Captura el qué, el por qué, y lo que no puede cambiar],    
+      712 +                                                              
+      713 +  [Nadie la actualiza bajo presión],                          
+      714 +  [El proceso la actualiza — es el path, no documentación para
+          +lela],                                                        
+      715 +)                                                             
+      716 +                                                              
+      717 +*Formaliza lo que no puede romperse.* Los invariantes no son c
+          +omentarios — son constraints tipadas que el sistema verifica a
+          +ntes del runtime, no después.                                 
+      718 +                                                              
+      719 +*Captura las tensiones activas.* "Coste vs. disponibilidad", "
+          +velocidad de despliegue vs. estabilidad" viven en la ontología
+          +, no en la cabeza del miembro más veterano del equipo. Cuando 
+          +ese alguien se va, los trade-offs permanecen.                 
+      720 +                                                              
+      721 +*Da al agente de IA el contexto que no puede inferir del códig
+          +o.* Un agente puede leer todo el código de un sistema y no sab
+          +er que hay un acuerdo que impide escalar horizontalmente ciert
+          +o servicio, o que la decisión de usar gRPC interno fue consecu
+          +encia de una tensión específica de latencia que no está en nin
+          +gún archivo. La ontología operacional hace eso explícito y mac
+          +hine-readable.                                                
+      722 +                                                              
+      723 +== Por qué los DAGs — especificidad técnica                   
+      724 +                                                              
+      725 +Los DAGs no son una elección estética. Son la única estructura
+          + que garantiza cuatro propiedades simultáneamente en sistemas 
+          +operacionales:                                                
+      726 +                                                              
+      727 +*Análisis de impacto decidible.* Antes de cambiar un invariant
+          +e o un nodo de infra — recorre el DAG y sabes exactamente qué 
+          +se ve afectado. Concreto: en Terraform, el grafo de recursos e
+          +s un DAG. `terraform plan` recorre ese grafo para determinar q
+          +ué debe crearse, modificarse o destruirse en qué orden. Si hub
+          +iera ciclos, el plan sería indecidible.                       
+      728 +                                                              
+      729 +*Orden topológico para ejecución fiable.* Cualquier operación 
+          +con dependencias — despliegue, migración, rollback, onboarding
+          + — tiene un orden correcto. El DAG lo da sin necesidad de un c
+          +oordinador central. Este es el principio detrás de Make (1976)
+          +, Bazel, Gradle, y cualquier build system no trivial.         
+      730 +                                                              
+      731 +*Justificación sin circularidad.* Si el invariante A existe po
+          +rque la práctica B lo requiere, y la práctica B existe porque 
+          +el invariante A lo define — tienes dogma, no arquitectura. El 
+          +DAG hace esa circularidad visible e inaceptable.              
+      732 +                                                              
+      733 +*Rollback determinista.* El inverso topológico del DAG es siem
+          +pre computable. Sin DAG, el rollback es un proceso manual que 
+          +depende de que alguien recuerde el orden correcto.            
+      734 +                                                              
+      735 +== Por qué la reflection — la ontología sin operación es un mu
+          +seo                                                           
+      736 +                                                              
+      737 +La ontología captura el *ser*. Pero los sistemas de software n
+          +o solo son — *operan*. La reflection es la capa que hace la on
+          +tología ejecutable:                                           
+      738 +                                                              
+      739 +*Modos como DAGs operacionales.* Un procedimiento sin DAG es u
+          +na lista de pasos que alguien puede ejecutar en cualquier orde
+          +n, saltar, o ignorar. Un modo como DAG tiene dependencias expl
+          +ícitas — el sistema puede verificar que se ejecutó correctamen
+          +te antes de avanzar.                                          
+      740 +                                                              
+      741 +*ADRs con constraints tipadas.* Una decisión sin constraints e
+          +s prosa. Con constraints tipadas es un contrato que el CI pued
+          +e verificar en cada commit. La pregunta "¿estamos violando alg
+          +una decisión arquitectural?" pasa de ser reflexión manual a se
+          +r comprobación automática.                                    
+      742 +                                                              
+      743 +*Detección de drift.* La diferencia entre el estado sellado (l
+          +o que declaraste que era verdad) y el estado actual (lo que re
+          +almente hay desplegado) es medible si y solo si formalizaste e
+          +l estado sellado. Sin eso, la pregunta "¿ha derivado esto?" no
+          + tiene respuesta objetiva.                                    
+      744 +                                                              
+      745 +== Las tres presiones que hacen esto urgente ahora            
+      746 +                                                              
+      747 +Estos conceptos no son nuevos. Lo nuevo es que tres presiones 
+          +convergen simultáneamente haciendo que la ausencia del modelo 
+          +operacional sea costosa de una forma que antes era tolerable: 
+      748 +                                                              
+      749 +*Presión 1: Los sistemas superan la capacidad de comprensión i
+          +ndividual.* Un monolito de 2005 podía vivir en la cabeza de do
+          +s personas. Un sistema cloud-native de 2026 con decenas de mic
+          +roservicios, múltiples regiones, pipelines automatizados y SLO
+          +s por servicio no cabe en ninguna cabeza. El modelo mental ind
+          +ividual ha dejado de ser suficiente como mecanismo de coordina
+          +ción.                                                         
+      750 +                                                              
+      751 +*Presión 2: Los agentes de IA tienen acceso de escritura.* Has
+          +ta hace poco, una herramienta de IA que generaba código incorr
+          +ecto producía texto que un humano revisaba antes de ejecutar. 
+          +Hoy los agentes pueden abrir PRs, modificar configuración, eje
+          +cutar Terraform, y desplegar. El ciclo humano-en-el-loop se co
+          +mprime. Un agente sin contexto ontológico en ese entorno no es
+          + un asistente — es un actor autónomo con comprensión incomplet
+          +a del sistema que modifica.                                   
+      752 +                                                              
+      753 +*Presión 3: Los equipos son multi-actor y distribuidos.* Human
+          +os en zonas horarias distintas, agentes de IA con distintos co
+          +ntextos de sesión, pipelines de CI ejecutando en paralelo — to
+          +dos escribiendo sobre el mismo sistema sin coordinador central
+          +. Sin un protocolo compartido y machine-readable para "qué pue
+          +de cambiarse, bajo qué condiciones, con qué justificación", la
+          + coordinación depende de que todo el mundo lea el mismo Conflu
+          +ence actualizado. Eso nunca funcionó.                         
+      754 +                                                              
+      755 +== La síntesis                                                
+      756 +                                                              
+      757 +Lo que la ontología operacional, reflection y DAGs aportan en 
+          +conjunto es algo que ninguna de sus partes aporta sola:       
+      758 +                                                              
+      759 +#align(center)[                                               
+      760 +  #block(                                                     
+      761 +    stroke: rgb("#c04000") + 1pt,                             
+      762 +    radius: 4pt,                                              
+      763 +    inset: 14pt,                                              
+      764 +    width: 88%,                                               
+      765 +  )[                                                          
+      766 +    #set text(size: 11pt)                                     
+      767 +    Un sistema que sabe lo que es, puede verificar que sigue s
+          +iendo \                                                       
+      768 +    lo que dice ser, y puede comunicarlo a cualquier actor — h
+          +umano, \                                                      
+      769 +    agente o CI — de forma estructurada y sin ambigüedad.     
+      770 +  ]                                                           
+      771 +]                                                             
+      772 +                                                              
+      773 +Eso no es documentación. Es un modelo operacional vivo.       
+      774 +                                                              
+      775 +La ausencia de ese modelo no es un problema de disciplina del 
+          +equipo. Es una propiedad emergente de sistemas complejos opera
+          +dos sin representación explícita. El drift, la amnesia de deci
+          +siones, la ilusión de control en infraestructura declarativa —
+          + no son fallos de las personas. Son consecuencias inevitables.
+      776 +                                                              
+      777 +La ontología operacional es la respuesta a la pregunta: *¿qué 
+          +estructura mínima necesita un sistema para conocerse a sí mism
+          +o de forma que ese conocimiento pueda ser usado, verificado y 
+          +compartido?*                                                  
+      778 +                                                              
+      779 +#pagebreak()                                                  
+      780 +                                                              
+      781 +// ─── 8. REFERENCIAS ────────────────────────────────────────
+          +─────────────────                                             
+      782 +                                                              
+      783  = Referencias y lecturas recomendadas
+      784  
+      785  == Filosofía clásica y ontología
+
+⏺ Ahora añado las referencias de física cuántica e ingeniería de software al
+  final de la bibliografía.
+
+⏺ Update(ontologia-filosofia-software.typ)
+  ⎿  Added 68 lines
+      892  #text(fill: rgb("#666666"), size: 9.5pt)[La diferencia que hac
+           e una diferencia. Epistemología sistémica que conecta Taoísmo,
+            cibernética y representación del conocimiento. Perspectiva po
+           co citada en CS pero fundamentalmente relevante.]
+      893  
+      894  *Winograd, T. & Flores, F.* _Understanding Computers and Cogni
+           tion_. Norwood: Ablex, 1986. \
+      895 +#text(fill: rgb("#666666"), size: 9.5pt)[La crítica más influy
+          +ente a la IA simbólica desde Heidegger. Argumenta que el trasf
+          +ondo de comprensión no puede formalizarse — límite estructural
+          + de toda ontología operacional.]                              
+      896 +                                                              
+      897 +== Física cuántica y fundamentos                              
+      898 +                                                              
+      899 +*Bohr, N.* "Can Quantum-Mechanical Description of Physical Rea
+          +lity Be Considered Complete?" _Physical Review_, 48, 696-702, 
+          +1935. \                                                       
+      900 +#text(fill: rgb("#666666"), size: 9.5pt)[La respuesta de Bohr 
+          +al artículo EPR de Einstein-Podolsky-Rosen. Formula el princip
+          +io de complementariedad como posición filosófica irrenunciable
+          +.]                                                            
+      901 +                                                              
+      902 +*Bohr, N.* "Discussion with Einstein on Epistemological Proble
+          +ms in Atomic Physics." En Schilpp (ed.), _Albert Einstein: Phi
+          +losopher-Scientist_. Evanston: Library of Living Philosophers,
+          + 1949. \                                                      
+      903 +#text(fill: rgb("#666666"), size: 9.5pt)[El debate más importa
+          +nte de la física del siglo XX sobre interpretación cuántica. B
+          +ohr y Einstein representan dos ontologías irreconciliables.]  
+      904 +                                                              
+      905 +*Heisenberg, W.* _Physics and Philosophy: The Revolution in Mo
+          +dern Science_. Nueva York: Harper & Row, 1958. \              
+      906 +#text(fill: rgb("#666666"), size: 9.5pt)[La reflexión filosófi
+          +ca más accesible de uno de los fundadores. El capítulo sobre e
+          +l papel del lenguaje en física cuántica conecta directamente c
+          +on Wittgenstein.]                                             
+      907 +                                                              
+      908 +*Schrödinger, E.* "Die gegenwärtige Situation in der Quantenme
+          +chanik." _Naturwissenschaften_, 23(48-49-50), 1935. \         
+      909 +#text(fill: rgb("#666666"), size: 9.5pt)[El artículo del gato.
+          + Introduce el concepto de entrelazamiento (_Verschränkung_) y 
+          +demuestra que la superposición cuántica no puede interpretarse
+          + clásicamente.]                                               
+      910 +                                                              
+      911 +*Kochen, S. & Specker, E.P.* "The Problem of Hidden Variables 
+          +in Quantum Mechanics." _Journal of Mathematics and Mechanics_,
+          + 17(1), 59-87, 1967. \                                        
+      912 +#text(fill: rgb("#666666"), size: 9.5pt)[Prueba formal de que 
+          +es imposible asignar valores definidos a todos los observables
+          + cuánticos simultáneamente. La contextualidad como hecho matem
+          +ático, no filosófico.]                                        
+      913 +                                                              
+      914 +*Bell, J.S.* _Speakable and Unspeakable in Quantum Mechanics_.
+          + Cambridge: Cambridge University Press, 1987. \               
+      915 +#text(fill: rgb("#666666"), size: 9.5pt)[Colección de ensayos 
+          +del autor del teorema de Bell (1964). El capítulo "Against mea
+          +surement" es una crítica filosófica fundamental al vocabulario
+          + de la física cuántica.]                                      
+      916 +                                                              
+      917 +*Aspect, A., Grangier, P. & Roger, G.* "Experimental Realizati
+          +on of Einstein-Podolsky-Rosen-Bohm Gedankenexperiment." _Physi
+          +cal Review Letters_, 49(2), 91-94, 1982. \                    
+      918 +#text(fill: rgb("#666666"), size: 9.5pt)[La confirmación exper
+          +imental de que las desigualdades de Bell se violan. Prueba que
+          + la naturaleza es no-local — no hay variables ocultas locales.
+          +]                                                             
+      919 +                                                              
+      920 +*Everett, H.* "Relative State Formulation of Quantum Mechanics
+          +." _Reviews of Modern Physics_, 29(3), 454-462, 1957. \       
+      921 +#text(fill: rgb("#666666"), size: 9.5pt)[La interpretación de 
+          +muchos mundos. Relevante porque ofrece una ontología donde tod
+          +os los estados posibles son reales — el antípoda de la ontolog
+          +ía operacional que colapsa posibilidades en decisiones.]      
+      922 +                                                              
+      923 +*Rovelli, C.* "Relational Quantum Mechanics." _International J
+          +ournal of Theoretical Physics_, 35(8), 1637-1678, 1996. \     
+      924 +#text(fill: rgb("#666666"), size: 9.5pt)[No existen estados ab
+          +solutos — solo estados relativos a otros sistemas. La ontologí
+          +a más radical de la lista y la más relevante para entender por
+          + qué el estado de un proyecto es siempre relacional.]         
+      925 +                                                              
+      926 +*Rovelli, C.* _Helgoland_. Milano: Adelphi, 2020. \           
+      927 +#text(fill: rgb("#666666"), size: 9.5pt)[La versión divulgativ
+          +a de la mecánica cuántica relacional. Accesible y filosóficame
+          +nte riguroso. El capítulo sobre relaciones y el budismo es rel
+          +evante para la conexión con el Taoísmo.]                      
+      928 +                                                              
+      929 +*Nagel, T.* _The View from Nowhere_. Oxford: Oxford University
+          + Press, 1986. \                                               
+      930 +#text(fill: rgb("#666666"), size: 9.5pt)[El argumento de que t
+          +oda objetividad requiere un punto de vista. En tensión directa
+          + con el problema del observador cuántico y la ontología perspe
+          +ctival de Zhuangzi.]                                          
+      931 +                                                              
+      932 +*Nielsen, M.A. & Chuang, I.L.* _Quantum Computation and Quantu
+          +m Information_. Cambridge: Cambridge University Press, 2000. \
+      933 +#text(fill: rgb("#666666"), size: 9.5pt)[El texto estándar de 
+          +computación cuántica. Los capítulos 2-3 sobre el formalismo ma
+          +temático (espacios de Hilbert, operadores, entrelazamiento) so
+          +n la base para entender la conexión con LLMs.]                
+      934 +                                                              
+      935 +*Deutsch, D.* _The Fabric of Reality_. Londres: Allen Lane, 19
+          +97. \                                                         
+      936 +#text(fill: rgb("#666666"), size: 9.5pt)[La defensa más ambici
+          +osa de la interpretación de muchos mundos. El capítulo sobre e
+          +pistemología cuántica conecta con la cuestión de qué puede con
+          +ocerse sobre un sistema complejo.]                            
+      937 +                                                              
+      938 +== Ingeniería de software e infraestructura                   
+      939 +                                                              
+      940 +*Evans, E.* _Domain-Driven Design: Tackling Complexity in the 
+          +Heart of Software_. Boston: Addison-Wesley, 2003. \           
+      941 +#text(fill: rgb("#666666"), size: 9.5pt)[El lenguaje ubicuo co
+          +mo ontología operacional informal. El bounded context es estru
+          +cturalmente equivalente a una ontología regional de Husserl ap
+          +licada a dominios de negocio.]                                
+      942 +                                                              
+      943 +*Kleppmann, M.* _Designing Data-Intensive Applications_. Sebas
+          +topol: O'Reilly, 2017. \                                      
+      944 +#text(fill: rgb("#666666"), size: 9.5pt)[El capítulo sobre mod
+          +elos de datos es el más relevante: la elección de modelo (rela
+          +cional, documental, grafo) es una decisión ontológica — define
+          + qué relaciones pueden existir.]                              
+      945 +                                                              
+      946 +*Morris, K.* _Infrastructure as Code_. Sebastopol: O'Reilly, 2
+          +016. \                                                        
+      947 +#text(fill: rgb("#666666"), size: 9.5pt)[Formaliza la declarat
+          +ividad en infraestructura. El capítulo sobre principios hace e
+          +xplícita la distinción entre estado deseado y estado actual — 
+          +el núcleo del modelo state.ncl.]                              
+      948 +                                                              
+      949 +*Humble, J. & Farley, D.* _Continuous Delivery_. Boston: Addis
+          +on-Wesley, 2010. \                                            
+      950 +#text(fill: rgb("#666666"), size: 9.5pt)[El pipeline de despli
+          +egue como DAG de verificaciones. La idea de que cada etapa del
+          + pipeline es un gate que el sistema debe cruzar formalmente.] 
+      951 +                                                              
+      952 +*Kim, G., Humble, J., Debois, P. & Willis, J.* _The DevOps Han
+          +dbook_. Portland: IT Revolution Press, 2016. \                
+      953 +#text(fill: rgb("#666666"), size: 9.5pt)[El flujo de trabajo D
+          +evOps como protocolo operacional. La sección sobre feedback lo
+          +ops es estructuralmente equivalente al ciclo ontología/reflect
+          +ion.]                                                         
+      954 +                                                              
+      955 +*Dehghani, Z.* "How to Move Beyond a Monolithic Data Lake to a
+          + Distributed Data Mesh." Martin Fowler Blog, 2019. \          
+      956 +#text(fill: rgb("#666666"), size: 9.5pt)[Data mesh como ontolo
+          +gía operacional distribuida para datos. Cada dominio es respon
+          +sable de su propia ontología de datos — la aplicación más cerc
+          +ana al modelo ontoref en infraestructura de datos.]           
+      957 +                                                              
+      958 +*Fowler, M.* "Who Needs an Architect?" _IEEE Software_, 20(5),
+          + 11-13, 2003. \                                               
+      959 +#text(fill: rgb("#666666"), size: 9.5pt)[Define arquitectura c
+          +omo "las decisiones que son difíciles de cambiar" — equivalent
+          +e exacto de los invariantes en ontología operacional.]        
+      960 +                                                              
+      961 +*Richardson, C.* _Microservices Patterns_. Shelter Island: Man
+          +ning, 2018. \                                                 
+      962 +#text(fill: rgb("#666666"), size: 9.5pt)[El patrón Saga como D
+          +AG de transacciones distribuidas. Cada paso del saga tiene dep
+          +ends_on implícito — el DAG operacional aplicado a consistencia
+          + eventual.]                                                   
+      963  #text(fill: rgb("#666666"), size: 9.5pt)[La crítica más influy
+           ente a la IA simbólica desde Heidegger. Argumenta que el trasf
+           ondo de comprensión no puede formalizarse — límite de toda ont
+           ología operacional.]
+⏺ Bash(typst compile ontologia-filosofia-software.typ
+      ontologia-filosofia-software.pdf 2>&1)
+  ⎿  warning: unknown font family: new computer modern mono
+         ┌─ ontologia-filosofia-software.typ:678:20
+         │
+     … +3 lines (ctrl+o to expand)
+
+⏺ Update(ontologia-filosofia-software.typ)
+  ⎿  Added 1 line, removed 1 line
+      675      inset: 12pt,
+      676      width: 90%,
+      677    )[
+      678 -    #set text(font: "New Computer Modern Mono", size: 9pt, fil
+          -l: rgb("#cccccc"))                                            
+      678 +    #set text(font: "Courier New", size: 9pt, fill: rgb("#cccc
+          +cc"))                                                         
+      679      Lo que el sistema hace           ≠  Lo que el equipo cree 
+           que hace \
+      680      Lo que la infra tiene desplegado ≠  Lo que Terraform decla
+           ra \
+      681      Lo que el ADR decidió            ≠  Lo que el código imple
+           menta \
+
+⏺ Bash(typst compile ontologia-filosofia-software.typ
+      ontologia-filosofia-software.pdf 2>&1 && echo "OK" && ls -lh
+      ontologia-filosofia-software.pdf && echo "pages: $(p…)
+  ⎿  OK
+     -rw-r--r--@ 1 jesusperezlorenzo  admin   440K Apr 24 23:47 ontologia-f
+     ilosofia-software.pdf
+     pages: 28
+
+⏺ PDF actualizado — 28 páginas, 440K.
+
+  Lo que se añadió:
+
+  Sección 6 — Física cuántica y ontología operacional
+  - El teorema de Kochen-Specker como prueba formal de que Aristóteles falla en
+  el nivel fundamental
+  - Tres conexiones genuinas: complementariedad de Bohr ↔ yin-yang/dualidad
+  ontología-reflection; contextualidad cuántica ↔ perspectivismo (Zhuangzi,
+  Wittgenstein, Gruber); problema del observador ↔ trasfondo pre-temático
+  - Mecánica Cuántica Relacional de Rovelli — estados sin propiedades absolutas,
+   solo relacionales
+  - Tabla honesta sobre computación cuántica: dónde aplica y dónde no
+  - La conexión LLMs/espacios de Hilbert — la misma matemática cuántica aplicada
+   a semántica
+
+  Sección 7 — Relevancia en software e infraestructuras
+  - Las cuatro brechas epistémicas sistémicas (sistema vs. creencia del equipo,
+  infra vs. Terraform, ADR vs. código, agente vs. contexto real)
+  - Tabla documentación tradicional vs. ontología operacional
+  - Por qué los DAGs técnicamente: impacto decidible, orden topológico,
+  justificación sin circularidad, rollback determinista
+  - Por qué la reflection: modos ejecutables, ADR constraints en CI, drift
+  detection
+  - Las tres presiones convergentes: complejidad supera comprensión individual,
+  agentes con acceso de escritura, equipos multi-actor distribuidos
+
+  Referencias ampliadas — 14 nuevas entradas:
+  - Física cuántica: Bohr (×2), Heisenberg, Schrödinger, Kochen-Specker, Bell,
+  Aspect, Everett, Rovelli (×2), Nagel, Nielsen/Chuang, Deutsch
+  - Ingeniería: Evans (DDD), Kleppmann, Morris (IaC), Humble/Farley, Kim et al.
+  (DevOps), Dehghani (data mesh), Fowler, Richardson
+
+✻ Worked for 3m 43s
\ No newline at end of file
diff --git a/assets/presentation/docs/2026-04-25-000346-puedes-sacar-el-texto-de-este-contenido.txt b/assets/presentation/docs/2026-04-25-000346-puedes-sacar-el-texto-de-este-contenido.txt
new file mode 100644
index 0000000..8058ee2
--- /dev/null
+++ b/assets/presentation/docs/2026-04-25-000346-puedes-sacar-el-texto-de-este-contenido.txt
@@ -0,0 +1,905 @@
+ ▐▛███▜▌   Claude Code v2.1.119
+▝▜█████▛▘  Sonnet 4.6 · Claude Max
+  ▘▘ ▝▝    /Users/Akasha/Development/ontoref
+
+❯ puedes sacar el texto de este contenido ?                                     
+  <div class="available-content reader2-post-content"><div dir="auto"           
+  class="body markup"><div data-component-name="SubscribeWidget"                
+  class="subscribe-widget is-signed-up"><div class="pencraft pc-reset           
+  button-wrapper"><div class="pencraft pc-display-flex pc-justifyContent-center 
+   pc-reset"><button tabindex="0" type="button"                                 
+  data-href="https://jessicatalisman.substack.com/subscribe?utm_medium=web&amp; 
+  utm_source=subscribe-widget&amp;utm_content=194462271&amp;next=https%3A%2F%2F 
+  substack.com%2F%40jessicatalisman%2Fp-194462271" class="pencraft pc-reset     
+  pencraft buttonBase-GK1x3M buttonText-X0uSmG buttonStyle-r7yGCK               
+  priority_primary-RfbeYt size_md-gCDS3o"><span>Upgrade to                      
+  paid</span></button></div></div></div><p><span>In 1986, Hirotaka Takeuchi and 
+   Ikujiro Nonaka published “The New New Product Development Game” in the       
+  </span><em>Harvard Business Review</em><span>, describing how Honda, Canon,   
+  and Fuji-Xerox organized teams to build copiers and automobiles.</span><span  
+  data-state="closed" style="min-width: 0px;"><a                                
+  data-component-name="FootnoteAnchorToDOM" id="footnote-anchor-1"              
+  href="#footnote-1" target="_self" class="footnote-anchor">1</a></span><span>  
+  Jeff Sutherland read that article, combined it with lean manufacturing        
+  principles and object oriented programming, and in 1993 adapted it for        
+  software development at Easel Corporation. Ken Schwaber formalized the method 
+   at the OOPSLA conference in 1995.</span><span data-state="closed"            
+  style="min-width: 0px;"><a data-component-name="FootnoteAnchorToDOM"          
+  id="footnote-anchor-2" href="#footnote-2" target="_self"                      
+  class="footnote-anchor">2</a></span><span> By 2001, seventeen software        
+  developers gathered at a ski lodge in Snowbird, Utah, and signed the Agile    
+  Manifesto, declaring that they were “uncovering better ways of developing     
+  software.”</span><span data-state="closed" style="min-width: 0px;"><a         
+  data-component-name="FootnoteAnchorToDOM" id="footnote-anchor-3"              
+  href="#footnote-3" target="_self" class="footnote-anchor">3</a></span><span>  
+  Not knowledge or meaning. A manufacturing framework for                       
+  software.</span></p><div class="captioned-image-container"><figure><a         
+  target="_blank" href="https://substackcdn.com/image/fetch/$s_!c-hA!,f_auto,q_ 
+  auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws 
+  .com%2Fpublic%2Fimages%2F66254527-6290-4907-9186-3be299a56c91_664x861.jpeg"   
+  data-component-name="Image2ToDOM" class="image-link image2                    
+  is-viewable-img"><div class="image2-                                          
+  ──── (0 lines hidden) ─────────────────────────────────────────────────────── 
+  paddingBottom-0 pc-alignItems-center pc-reset"><a                             
+  href="https://jessicatalisman.substack.com/p/the-semantics-of-semantics"      
+  class="pencraft pc-reset align-center-y7ZD4w line-height-20-t4M0El            
+  font-text-qe4AeH size-13-hZTUKr weight-medium-fw81nC reset-IxiVJZ"><div       
+  class="pencraft pc-display-flex pc-gap-8 pc-alignItems-center pc-reset        
+  link-HREYZo"><span class="pencraft pc-reset color-accent-BVX_7M               
+  line-height-20-t4M0El font-text-qe4AeH size-14-MLPa7j weight-semibold-uqA4FV  
+  reset-IxiVJZ">Read full story</span><svg xmlns="http://www.w3.org/2000/svg"   
+  width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor"  
+  stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide 
+   lucide-arrow-right"><path d="M5 12h14"></path><path d="m12 5 7 7-7           
+  7"></path></svg></div></a></div></div></div><p></p></div></div></div><div     
+  class="pencraft pc-display-contents pc-reset dark-theme"><div class="pencraft 
+   pc-display-flex pc-flexDirection-column pc-gap-4 pc-padding-4                
+  pc-justifyContent-center pc-boxShadow-md pc-reset outline-detail-vcQLyr       
+  pc-borderRadius-md popover-Q5L48i  " style="top: 208.797px; left:             
+  557.645px;"><button tabindex="0" type="button" class="pencraft pc-reset       
+  pencraft button-r00Nrm buttonBase-GK1x3M"><svg                                
+  xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" 
+   fill="none" stroke="var(--color-utility-white)" stroke-width="2"             
+  stroke-linecap="round" stroke-linejoin="round" class="lucide                  
+  lucide-refresh-cw"><path d="M3 12a9 9 0 0 1 9-9 9.75 9.75 0 0 1 6.74 2.74L21  
+  8"></path><path d="M21 3v5h-5"></path><path d="M21 12a9 9 0 0 1-9 9 9.75 9.75 
+   0 0 1-6.74-2.74L3 16"></path><path d="M8 16H3v5"></path></svg><div           
+  class="pencraft pc-reset color-vibrance-primary-KHCdqV line-height-20-t4M0El  
+  font-text-qe4AeH size-13-hZTUKr weight-medium-fw81nC                          
+  reset-IxiVJZ">Restack</div></button><button tabindex="0" type="button"        
+  class="pencraft pc-reset pencraft button-r00Nrm buttonBase-GK1x3M"><svg       
+  xmlns="http://www.w3.org/2000/svg" width="20" height="20" viewBox="0 0 24 24" 
+   fill="none" stroke="var(--color-utility-white)" stroke-width="2"             
+  stroke-linecap="round" stroke-linejoin="round" class="lucide                  
+  lucide-share"><path d="M4 12v8a2 2 0 0 0 2 2h12a2 2 0 0 0                     
+  2-2v-8"></path><polyline points="16 6 12 2 8 6"></polyline><line x1="12"      
+  x2="12" y1="2" y2="15"></line></svg><div class="pencraft pc-reset             
+  color-vibrance-primary-KHCdqV line-height-20-t4M0El font-text-qe4AeH          
+  size-13-hZTUKr weight-medium-fw81nC                                           
+  reset-IxiVJZ">Share</div></button></div></div></div>                          
+
+⏺ Here's the plain text extracted from the HTML:
+
+  ---
+  In 1986, Hirotaka Takeuchi and Ikujiro Nonaka published "The New New Product
+  Development Game" in the Harvard Business Review, describing how Honda, Canon,
+   and Fuji-Xerox organized teams to build copiers and automobiles. Jeff
+  Sutherland read that article, combined it with lean manufacturing principles
+  and object oriented programming, and in 1993 adapted it for software
+  development at Easel Corporation. Ken Schwaber formalized the method at the
+  OOPSLA conference in 1995. By 2001, seventeen software developers gathered at
+  a ski lodge in Snowbird, Utah, and signed the Agile Manifesto, declaring that
+  they were "uncovering better ways of developing software." Not knowledge or
+  meaning. A manufacturing framework for software.
+
+  The Agile Manifesto's own language is explicit about its scope, yet the
+  industry has applied Agile methodology to virtually everything in a
+  technological environment — including knowledge work, semantic systems and now
+   artificial intelligence. And within the same breath, organizations and
+  experts regale the industry with talk of innovation and the emergence of a new
+   paradigm in ways of working. Yet those ways of working are locked inside of
+  Agile methodologies and Scrum, designed for the factory floor and the codebase
+   — not for the messy, recursive, deeply human work of making meaning out of
+  complexity.
+
+  Scrum Is a Manufacturing Schedule
+
+  Let's return to the Takeuchi and Nonaka paper, where the title—"The New New
+  Product Development Game" describes how product teams at Japanese firms
+  developed physical goods such as the FX-3500 copier and the Honda City car.
+  They draw on the rugby metaphor because the team "tries to go the distance as
+  a unit, passing the ball back and forth." The unit of delivery is a working
+  prototype of a material artifact. The feedback loop is a machine that works or
+   fails.
+
+  Sutherland's 2014 book, Scrum: The Art of Doing Twice the Work in Half the 
+  Time, reinforces this with military and manufacturing case studies such as the
+   F-86 fighter production and Toyota. Sutherland's book rebuilds the argument
+  with the same assumption—that the thing being built has discrete, inspectable,
+   testable parts that, when finished, are finished.
+
+  The 2020 Scrum Guide defines the Increment as "a concrete stepping stone
+  toward the Product Goal," one that is "additive to all prior Increments and
+  thoroughly verified, ensuring that all Increments work together." These words
+  are sensible if you are building a mobile application, with a login screen and
+   a settings page. But for knowledge systems, these words are destructive,
+  because the components of a semantic architecture are not additive in the way
+  Scrum dictates. A controlled vocabulary is not a subcomponent of a taxonomy
+  the way a login screen is a subcomponent of an app. A taxonomy without
+  disciplined, vocabulary control as its underpinnings does not function well,
+  if at all. You cannot ship half a taxonomy and call it an Increment.
+
+  You see, Scrum assumes tame problems that can be time-boxed, estimated with
+  planning poker, decomposed into user stories, tracked on a velocity chart and
+  declared done against a checklist. If only knowledge were this tame. Knowledge
+   organization is anything but tame. More than fifty-years-old, in 1973 Horst
+  Rittel and Melvin Webber warned that applying the engineering paradigm
+  uniformly to all problems was "bound to fail." And yet here we are, applying a
+   manufacturing schedule and methodologies to knowledge, expecting that
+  knowledge will fit into the confines of old ways of working — an artifact of
+  the Third Industrial Revolution imposed on the demands of the Fourth.
+
+  In 2016, Klaus Schwab wrote about this historical pattern. The First
+  Industrial Revolution, beginning around 1760, gave us steam and mechanical
+  production. The Second, in the late nineteenth century, gave us electricity
+  and the assembly line — the world that produced Toyota, lean manufacturing,
+  and the production logic that Scrum still carries as its core ethos. The
+  Third, beginning in the 1960s, gave us semiconductors, mainframe computing,
+  and eventually the internet — the world in which Agile was born, by software
+  developers, for software development. The Fourth Industrial Revolution, Schwab
+   argued, is fundamentally different, in kind, as it is defined by the fusion
+  of artificial intelligence, cyber-physical systems and knowledge technologies
+  that blur the boundaries between the physical, digital and biological worlds.
+
+  The problems of this era are not necessarily engineering problems with stable
+  requirements. We are wrestling with problems such as how to ground machine
+  intelligence in meaning, how to organize evolving knowledge and how to build
+  systems that can explain themselves. These are wicked problems that demand new
+   epistemologies and frameworks for problem solving. Agile may have been the
+  right methodology for the Third Industrial Revolution but we are no longer in
+  the Third Industrial Revolution. The frameworks have not caught up, and the
+  cost of that lag is exacerbated every time an organization forces Third
+  Industrial Revolution principles applied to Fourth Industrial Revolution work.
+
+  In other words, this is not a product problem or a platform problem—it is
+  learning to embrace new ways of working. This is an infrastructure problem
+  that is governed by epistemology—a fundamental philosophy pertaining to
+  knowledge.
+
+  Why Knowledge Cannot Be Sprinted
+
+  An ontology is, per Thomas Gruber's foundational definition, "a specification
+  of a conceptualization." A conceptualization is a whole, not parts of a whole.
+   Specifying half of it produces logical incoherence, not a minimally viable
+  product. For example, ANSI/NISO Z39.19, the standard for constructing
+  controlled vocabularies, requires resolving synonymy, homonymy and scope
+  before terms enter production, because any inconsistency propagates downstream
+   into every subsequent use. SKOS, the W3C standard for encoding knowledge
+  organization systems, requires that hierarchical relationships form coherent
+  structures with transitive closure and OWL 2 requires that class axioms be
+  satisfiable. These are preconditions, necessary for logical systems to reason
+  and express meaning.
+
+  In practice, I have watched organizations attempt Agile delivery of knowledge
+  systems in two patterns and neither ends well. In the first, the team holds
+  stand-ups, ceremonies and retrospectives. The taxonomist brings tickets and
+  adds fifteen terms to the product vocabulary. Tickets close and a velocity
+  chart goes up. Nothing is integrated, because integration is the hard part and
+   integration does not fit into two weeks so that goes to the backlog. Six
+  months later, the vocabulary is a collection of inconsistently scoped terms,
+  the ontology has orphan classes and the knowledge graph — if it exists —
+  cannot reliably answer queries. But the burndown chart is immaculate and
+  leadership is blindly satisfied because boxes were checked.
+
+  In the second pattern, the team borrows from other methodologies while still
+  calling it Agile. The taxonomist conducts multi-week domain analysis and
+  stakeholder interviews — that is ethnography. She does card sorting and
+  concept modeling — that is design research. She iterates with subject matter
+  experts — that is participatory design. She ships, eventually, a coherent
+  model. The sprint cadence was irrelevant to the work. Martin Fowler named this
+   phenomenon in his 2018 keynote: "faux-agile" — the Agile Industrial Complex.
+
+  Ron Jeffries, another original signatory of the Agile Manifesto, told
+  developers the same year, 2018, to abandon the word entirely. Dave Thomas
+  declared in 2014 that "Agile" had become "a magnet for anyone with points to
+  espouse, hours to bill or products to sell." When three of the seventeen
+  authors of a framework say it has been hollowed out, one has to wonder if
+  Agile is appropriate for all technology work and if it is time to revisit ways
+   of working.
+
+  Innovation Does Not Fit in a Backlog
+
+  The deeper issue is that Agile and Scrum were never designed to inspire or
+  foster innovation. They were designed to deliver known requirements with
+  increasing efficiency. The Scrum Guide's entire apparatus — product backlog,
+  sprint planning, Definition of Done — presupposes that the problem has been
+  formulated by a product owner who can write it in a user story. But knowledge
+  work, like design, is wicked. Rittel and Webber's first property of a wicked
+  problem is that "There is no definitive formulation." The second property
+  states that there is no stopping rule. And the fifth property illuminates that
+   every solution is a "one-shot operation" because there is no opportunity to
+  learn by trial-and-error without consequence.
+
+  Given the nature of wicked problems, it would seem that Agile is ill fitted
+  for wicked problems. And for that matter, ill fitted for the dynamic nature of
+   knowledge, where knowledge infrastructures are never one-shot operations,
+  necessitate stopping rules and without a doubt defies definitive formulations.
+
+  Therefore, knowledge organization in an enterprise domain is wicked in exactly
+   this sense. The problem of how to represent "customer" in a
+  telecommunications company's semantic model has no pre-existing correct
+  answer. Its formulation depends on who is doing what with the representation,
+  and that changes as you build. The scope of the vocabulary shifts as you
+  interview stakeholders. The taxonomy reorganizes as you encounter edge cases.
+  The ontology's axioms tighten or loosen as the reasoner surfaces
+  contradictions. Don't be mistaken—this does not bend to the Agile principle of
+   welcoming changing requirements, as the requirements are constituted through
+  the work. The problem is under construction, often evolving, during the
+  solving.
+
+  Organizations that manage all work through Scrum are, in effect, optimizing
+  for predictability in domains that demand exploration. The results resist
+  innovation and take on the appearance of productivity. Velocity charts rise
+  while the underlying intellectual architecture is flattened into bit sized
+  deployments, uncoupled from the continuum that is knowledge.
+
+  Design Thinking Belongs in the Swamp
+
+  Donald Schön drew the distinction that "In the varied topography of
+  professional practice, there is a high, hard ground where practitioners can
+  make effective use of research-based theory and technique, and there is a
+  swampy lowland where situations are confusing 'messes' incapable of technical
+  solution." Agile lives on the high ground while knowledge work lives in the
+  swamp. Schön's term for how competent practitioners navigate the swamp was
+  "reflection-in-action" — the iterative, tacit recalibration that happens while
+   the work is happening.
+
+  Design thinking, in the lineage of Herbert Simon, Richard Buchanan and Schön,
+  is the methodology of the swamp. It assumes the problem is under construction.
+   Its five phases — empathize, define, ideate, prototype, test — are not a
+  linear pipeline but a recursive, iterative process that expects the definition
+   of the problem to change as understanding deepens.
+
+  Research in the Handbook of Human-Centered Artificial Intelligence confirms
+  this alignment—integrating design thinking into AI development makes systems
+  more user-focused, iterative and impactful, emphasizing data governance,
+  transparency, explainability and bias mitigation. Carnegie Mellon University's
+   Tepper School of Business now teaches a dedicated program on design thinking
+  with AI, arguing that the combination addresses "complex, systemic problems"
+  that linear frameworks cannot manage successfully. IBM's enterprise design
+  thinking practice, developed over two decades, has been applied across client
+  engagements precisely it keeps human experience at the center, while
+  technology evolves underneath — design thinking is "timeless," while
+  technology is "timestamped."
+
+  A 2026 study in Scientific Reports examined the relationship between higher
+  order thinking, generative AI chatbot use and engineering creativity, finding
+  that higher order thinking had significantly greater importance for creative
+  outcomes than AI tool use alone — reinforcing that cognitive depth, not tool
+  velocity, drives innovation. Research in the Proceedings of the Design Society
+   further argues that generative AI is shifting the designer's role from
+  executor to paradigm-setter, making design thinking more essential as AI
+  handles production tasks while humans must guide framing, ethics and meaning.
+
+  To be clear, these frameworks are not different flavors of the same
+  methodology. Scrum and design thinking represent totally different
+  epistemologies. Scrum assumes the problem was handed to you while design
+  thinking assumes the problem must be found and wrestled with.
+
+  Is AI a Knowledge Tool?
+
+  This brings us to the question the industry avoids and I, for one, have been
+  itching to ask. Is AI a knowledge tool? How do you classify it?
+
+  A large language model generates plausible sequences of text. Returning to the
+   Bender, Gebru, McMillan-Major, and Mitchell paper mentioned earlier, the
+  authors defined a language model as a system for stitching together linguistic
+   forms according to probabilistic patterns, "without any reference to
+  meaning." Without reference to meaning, as LLMs out-of-the-box are absent of
+  external grounding. They optimize for plausibility, not truth and therefore as
+   a standalone, can be best categorized as an information-processing tool.
+
+  It becomes a knowledge tool only when grounded in a semantic layer —
+  controlled vocabularies, taxonomies, ontologies, knowledge graphs — that
+  provides reference and provenance. The industry's own corrective measures and
+  associated marketplace confirms this as fact. Retrieval-augmented generation,
+  knowledge graphs for LLMs and what recent discourse calls "context
+  engineering" are, underneath the packaging, the semantic infrastructure that
+  should have been built first. But it was bolted on after the fact to prevent
+  hallucination that was pretty much guaranteed, the moment the model was
+  deployed without one.
+
+  The RAG paper from 2020 called out this problem, identifying that parametric
+  memory is opaque and cannot be updated and that retrieval from structured
+  sources is how provenance is captured and encoded. But it seems the industry
+  has a memory problem or it was too early for most to hear the message.
+
+  Ironically, provenance is a core element of knowledge management and ontology
+  work, as provenance is a principled tenant of the discipline and problem
+  space. Without provenance, knowledge cannot exist. Organizations are only now
+  discovering they need it, because their generative systems cannot explain
+  themselves. Call it lineage, call it traces. But if you are building for
+  knowledge, call it provenance, as provenance includes traces, lineage,
+  citations, the temporal and the spatial.
+
+  So if generative AI is an information processing tool, I posture that for AI
+  to become a knowledge tool, there must be provenance and descriptive logic.
+  And to be frank, nobody sprints their way into provenance as it is architected
+   and sustained.
+
+  Knowledge Is a Spiral, Not a Sprint
+
+  The irony does not escape the moment. The same Nonaka who co-authored the 1986
+   paper that inspired Scrum, went on to develop the SECI model of knowledge
+  creation in 1995, describing knowledge as a spiral between tacit and explicit
+  forms, moving through socialization, externalization, combination and
+  internalization.
+
+  The spiral does not have an end, as is true with spirals, emblematic of life.
+  Each rotation creates new knowledge that becomes the substrate for the next
+  rotation. There is no Definition of Done because the definition of what is
+  being done is itself being constructed. Scrum took the team structure from
+  Nonaka's first paper and discarded his next decade of work because it did not
+  fit neatly with Scrum. The mechanisms of knowledge work, as Nonaka spent a
+  decade arguing, are metaphor, tacit skill, apprenticeship, shared space and
+  the patient conversion of hunches into models and models back into practice.
+
+  The Ontology Pipeline™ mirrors this spiral. You start with a controlled
+  vocabulary — a disciplined process involving agreements, as to what terms
+  mean. You mature it into a taxonomy by introducing hierarchy, structuring
+  terms into parent-child relations. Thereafter, the taxonomy is matured into a
+  thesaurus by adding associative relationships and synonymy, encoded in SKOS.
+  The thesaurus is baseline work from which an ontology is developed, adding
+  formal semantics using OWL and RDF. You architect a knowledge graph,
+  assembling components to include the knowledge assets developed through the
+  Ontology Pipeline™ iterative framework.
+
+  Each stage prepares the ground for the next and stages cannot be skipped or
+  bypassed. You cannot ship the knowledge graph without an ontology, and the
+  ontology must be logically consistent and coherent to make sense of data. A
+  knowledge graph without a vocabulary control risks becoming a confident liar
+  or failing to work with natural language. Because knowledge assets are
+  dimensional interconnected, knowledge is indeed a spiral, unable to be
+  contained by manufacturing logic.
+
+  The Pattern and the Cure
+
+  Organizations budget for platforms — Confluence, SharePoint, LLM
+  subscriptions, vector databases, a RAG vendor. But they do not budget for
+  descriptive logic and knowledge, perhaps because the problem is not solved by
+  a single tool and is seemingly too amorphous in an Agile, data centric world.
+  The controlled vocabulary is cut unless it delivers value for the database.
+  Often, the taxonomy is cut because it does not demo well and its value sits in
+   the plumbing, delivering its punch over time. The result is an ever-growing
+  pile of documents with a chatbot stapled to the front, confidently answering
+  questions in the voice of a system that has no reference for any of the terms
+  sprinkled throughout the documents.
+
+  Then the system hallucinates. It contradicts itself across sessions. It
+  surfaces outdated policy as current and surfaces private outlook messages from
+   deactivated employees. I have seen this IRL more times than I can count. And
+  the optics are worse when a vendor tool is responsible for revealing private
+  chats and email transactions. Usually the executive sponsor asks why the
+  investment has not worked, ignoring the evidence of faux pas such as the
+  surfacing of private communications. All in a days work, when burndown charts
+  and velocity set the precedence for ways of working and reward systems.
+
+  Executives ignore the fact that the system was never grounded. It was never
+  grounded because grounding requires the slow, expensive, disciplined work of
+  knowledge organization — work that does not fit into a sprint, does not demo
+  well and fails to produce a velocity chart. The organization chooses to bypass
+   investment in the descriptive knowledge apparatus because the lines are too
+  blurry to clearly define according to Agile speak. Because knowledge is
+  infrastructure, not a product or a platform. Knowledge is the gas in the tank,
+   the fuel for meaning, the material that justifies logical reasoning.
+
+  Design thinking gives that knowledge work a home. It is comfortable with
+  ambiguity. It treats problem discovery as an equal to problem solving. It
+  certainly does not demand a shippable increment every fourteen days. Instead,
+  design thinking demands that you understand what you are building before you
+  build it — and that you keep revising that understanding as the work proceeds.
+   Library and information science, with over a century of standards for
+  knowledge organization, provides the methods while design thinking provides
+  the epistemology. Together, they offer what Agile never could—a framework
+  suited to the nature of knowledge itself.
+
+  A two-week sprint to deliver what? If you are building knowledge systems, two
+  weeks will get you nothing coherent. Knowledge is not built fast and broken
+  into shape. It is built carefully, with provenance, governance and patience.
+  Knowledge is not a product to be manufactured and no—you cannot buy
+  off-the-shelf knowledge that perfectly matches a domain or organization.
+  Knowledge is nuanced, specific to each organization, domain and person.
+
+  In the spirit of VC-speak, knowledge is the moat, and that moat is not Agile.
+
+  ---
+  By Jessica Talisman — Intentional Arrangement (Substack)
+
+✻ Brewed for 2m 25s
+
+❯ puedes generar una traducción precisa y elegante del contenido al español     
+
+⏺ ---
+  En 1986, Hirotaka Takeuchi e Ikujiro Nonaka publicaron "The New New Product
+  Development Game" en Harvard Business Review, describiendo cómo Honda, Canon y
+   Fuji-Xerox organizaban sus equipos para fabricar copiadoras y automóviles.
+  Jeff Sutherland leyó ese artículo, lo combinó con los principios de
+  manufactura lean y la programación orientada a objetos, y en 1993 lo adaptó al
+   desarrollo de software en Easel Corporation. Ken Schwaber formalizó el método
+   en la conferencia OOPSLA de 1995. En 2001, diecisiete desarrolladores de
+  software se reunieron en un albergue de esquí en Snowbird, Utah, y firmaron el
+   Manifiesto Ágil, declarando que estaban "descubriendo mejores maneras de
+  desarrollar software." No conocimiento ni significado. Un marco de manufactura
+   aplicado al software.
+
+  El propio lenguaje del Manifiesto Ágil es explícito sobre su alcance, y sin
+  embargo la industria ha aplicado la metodología ágil a prácticamente todo lo
+  que existe en un entorno tecnológico — incluyendo el trabajo del conocimiento,
+   los sistemas semánticos y ahora la inteligencia artificial. Y en el mismo
+  aliento, organizaciones y expertos colman la industria con discursos sobre
+  innovación y el surgimiento de un nuevo paradigma en las formas de trabajo.
+  Sin embargo, esas formas de trabajo permanecen encerradas dentro de las
+  metodologías ágiles y Scrum, diseñadas para la línea de producción y la base
+  de código — no para el trabajo profundamente humano, recursivo y desordenado
+  de construir significado a partir de la complejidad.
+
+  Scrum es un calendario de manufactura
+
+  Volvamos al artículo de Takeuchi y Nonaka, cuyo título — "The New New Product
+  Development Game" — describe cómo los equipos de producto en empresas
+  japonesas desarrollaban bienes físicos como la copiadora FX-3500 y el Honda
+  City. Los autores recurren a la metáfora del rugby porque el equipo "intenta
+  recorrer la distancia como una unidad, pasándose el balón de un lado al otro."
+   La unidad de entrega es un prototipo funcional de un artefacto material. El
+  ciclo de retroalimentación es una máquina que funciona o falla.
+
+  El libro de Sutherland de 2014, Scrum: The Art of Doing Twice the Work in Half
+   the Time, refuerza esto con casos de estudio militares y de manufactura como
+  la producción del caza F-86 y Toyota. El libro reconstruye el argumento sobre
+  la misma premisa: que aquello que se construye tiene partes discretas,
+  inspeccionables y verificables que, una vez terminadas, están terminadas.
+
+  La Guía Scrum de 2020 define el Incremento como "un peldaño concreto hacia el
+  Objetivo del Producto", uno que es "aditivo respecto a todos los Incrementos
+  anteriores y rigurosamente verificado, asegurando que todos los Incrementos
+  funcionen en conjunto." Estas palabras tienen sentido si se está construyendo
+  una aplicación móvil, con una pantalla de inicio de sesión y una página de
+  configuración. Pero para los sistemas de conocimiento, estas palabras son
+  destructivas, porque los componentes de una arquitectura semántica no son
+  aditivos de la manera en que Scrum prescribe. Un vocabulario controlado no es
+  un subcomponente de una taxonomía del mismo modo en que una pantalla de inicio
+   de sesión lo es de una aplicación. Una taxonomía sin un control de
+  vocabulario disciplinado como fundamento no funciona bien, o directamente no
+  funciona. No se puede entregar media taxonomía y llamarla un Incremento.
+
+  Scrum asume problemas dóciles que pueden enmarcarse en el tiempo, estimarse
+  con planning poker, descomponerse en historias de usuario, rastrearse en un
+  gráfico de velocidad y declararse resueltos frente a una lista de
+  verificación. Si tan solo el conocimiento fuera así de dócil. La organización
+  del conocimiento está muy lejos de serlo. Con más de cincuenta años de
+  antigüedad, en 1973 Horst Rittel y Melvin Webber advirtieron que aplicar el
+  paradigma de la ingeniería de manera uniforme a todos los problemas estaba
+  "condenado al fracaso." Y sin embargo aquí estamos, aplicando un calendario de
+   manufactura y sus metodologías al conocimiento, esperando que el conocimiento
+   quepa en los confines de viejas formas de trabajar — un artefacto de la
+  Tercera Revolución Industrial impuesto sobre las exigencias de la Cuarta.
+
+  En otras palabras, esto no es un problema de producto ni de plataforma — es el
+   desafío de aprender a abrazar nuevas formas de trabajar. Es un problema de
+  infraestructura gobernado por la epistemología: una filosofía fundamental
+  sobre el conocimiento.
+
+  Por qué el conocimiento no puede ejecutarse en sprints
+
+  Una ontología es, según la definición fundacional de Thomas Gruber, "una
+  especificación de una conceptualización." Una conceptualización es un todo, no
+   partes de un todo. Especificar la mitad produce incoherencia lógica, no un
+  producto mínimamente viable. Por ejemplo, la norma ANSI/NISO Z39.19 para la
+  construcción de vocabularios controlados exige resolver la sinonimia, la
+  homonimia y el alcance antes de que los términos entren en producción, porque
+  cualquier inconsistencia se propaga hacia abajo en cada uso posterior. SKOS,
+  el estándar del W3C para codificar sistemas de organización del conocimiento,
+  requiere que las relaciones jerárquicas formen estructuras coherentes con
+  cierre transitivo, y OWL 2 exige que los axiomas de clase sean satisfacibles.
+  Estas son precondiciones necesarias para que los sistemas lógicos razonen y
+  expresen significado.
+
+  En la práctica, he observado organizaciones intentando entregar sistemas de
+  conocimiento de forma ágil en dos patrones, y ninguno termina bien. En el
+  primero, el equipo celebra stand-ups, ceremonias y retrospectivas. El
+  taxonomista trae tickets y añade quince términos al vocabulario del producto.
+  Los tickets se cierran y el gráfico de velocidad sube. Nada se integra, porque
+   la integración es la parte difícil y no cabe en dos semanas, de modo que va
+  al backlog. Seis meses después, el vocabulario es una colección de términos
+  con alcances inconsistentes, la ontología tiene clases huérfanas y el grafo de
+   conocimiento — si existe — no puede responder consultas de manera confiable.
+  Pero el burndown chart es impecable y el liderazgo está ciegamente satisfecho
+  porque las casillas fueron marcadas.
+
+  En el segundo patrón, el equipo toma prestado de otras metodologías sin dejar
+  de llamarlo ágil. La taxonomista conduce análisis de dominio y entrevistas con
+   partes interesadas durante varias semanas — eso es etnografía. Realiza card
+  sorting y modelado conceptual — eso es investigación en diseño. Itera con
+  expertos en la materia — eso es diseño participativo. Finalmente entrega un
+  modelo coherente. El ritmo del sprint fue irrelevante para el trabajo. Martin
+  Fowler denominó este fenómeno en su conferencia magistral de 2018:
+  "faux-agile" — el Complejo Industrial Ágil.
+
+  Ron Jeffries, otro signatario original del Manifiesto Ágil, les dijo a los
+  desarrolladores ese mismo año, 2018, que abandonaran la palabra por completo.
+  Dave Thomas declaró en 2014 que "Ágil" se había convertido en "un imán para
+  cualquiera con puntos que exponer, horas que facturar o productos que vender."
+   Cuando tres de los diecisiete autores de un marco dicen que ha sido vaciado
+  de contenido, uno tiene que preguntarse si Agile es apropiado para todo el
+  trabajo tecnológico y si ha llegado el momento de revisar las formas de
+  trabajar.
+
+  La innovación no cabe en un backlog
+
+  El problema más profundo es que Agile y Scrum nunca fueron diseñados para
+  inspirar o fomentar la innovación. Fueron diseñados para entregar requisitos
+  conocidos con eficiencia creciente. Todo el aparato de la Guía Scrum — product
+   backlog, sprint planning, Definición de Hecho — presupone que el problema ha
+  sido formulado por un product owner capaz de escribirlo en una historia de
+  usuario. Pero el trabajo del conocimiento, como el diseño, es rebelde. La
+  primera propiedad de un problema retorcido según Rittel y Webber es que "No
+  existe una formulación definitiva." La segunda propiedad establece que no hay
+  regla de parada. Y la quinta propiedad ilumina que cada solución es una
+  "operación de un solo disparo" porque no hay oportunidad de aprender por
+  ensayo y error sin consecuencias.
+
+  Dada la naturaleza de los problemas retorcidos, parecería que Agile está mal
+  equipado para enfrentarlos. Y por la misma razón, mal equipado para la
+  naturaleza dinámica del conocimiento, donde las infraestructuras de
+  conocimiento nunca son operaciones de un solo disparo, requieren reglas de
+  parada y desafían sin lugar a dudas las formulaciones definitivas.
+
+  Por lo tanto, la organización del conocimiento en un dominio empresarial es
+  retorcida exactamente en este sentido. El problema de cómo representar
+  "cliente" en el modelo semántico de una empresa de telecomunicaciones no tiene
+   una respuesta correcta preexistente. Su formulación depende de quién hace qué
+   con esa representación, y eso cambia a medida que se construye. El alcance
+  del vocabulario se desplaza conforme se entrevista a las partes interesadas.
+  La taxonomía se reorganiza al encontrar casos límite. Los axiomas de la
+  ontología se tensan o aflojan cuando el razonador descubre contradicciones. No
+   hay que equivocarse — esto no se pliega al principio ágil de bienvenida a los
+   requisitos cambiantes, porque los requisitos se constituyen a través del
+  trabajo. El problema está en construcción, evolucionando frecuentemente,
+  durante su resolución.
+
+  Las organizaciones que gestionan todo el trabajo a través de Scrum están, en
+  efecto, optimizando para la previsibilidad en dominios que exigen exploración.
+   Los resultados inhiben la innovación y adoptan la apariencia de
+  productividad. Los gráficos de velocidad suben mientras la arquitectura
+  intelectual subyacente se aplana en entregas fragmentadas, desconectadas del
+  continuo que es el conocimiento.
+
+  El pensamiento de diseño pertenece al pantano
+
+  Donald Schön trazó la distinción de que "En la variada topografía de la
+  práctica profesional, hay una meseta alta y firme donde los profesionales
+  pueden hacer uso efectivo de la teoría y la técnica basadas en la
+  investigación, y hay una tierras bajas pantanosas donde las situaciones son
+  'enredos' confusos incapaces de solución técnica." Agile vive en la meseta
+  mientras que el trabajo del conocimiento vive en el pantano. El término de
+  Schön para describir cómo los profesionales competentes navegan el pantano fue
+   "reflexión-en-acción" — la recalibración iterativa y tácita que ocurre
+  mientras el trabajo transcurre.
+
+  El pensamiento de diseño, en el linaje de Herbert Simon, Richard Buchanan y
+  Schön, es la metodología del pantano. Asume que el problema está en
+  construcción. Sus cinco fases — empatizar, definir, idear, prototipar, probar
+  — no son un flujo lineal sino un proceso recursivo e iterativo que espera que
+  la definición del problema cambie a medida que la comprensión se profundiza.
+
+  La investigación en el Handbook of Human-Centered Artificial Intelligence
+  confirma esta alineación — integrar el pensamiento de diseño en el desarrollo
+  de IA hace que los sistemas sean más centrados en el usuario, iterativos e
+  impactantes, enfatizando la gobernanza de datos, la transparencia, la
+  explicabilidad y la mitigación del sesgo. La Tepper School of Business de
+  Carnegie Mellon University imparte ahora un programa dedicado a pensamiento de
+   diseño con IA, argumentando que la combinación aborda "problemas complejos y
+  sistémicos" que los marcos lineales no pueden gestionar con éxito. La práctica
+   de diseño empresarial de IBM, desarrollada a lo largo de dos décadas, se ha
+  aplicado en compromisos con clientes precisamente porque mantiene la
+  experiencia humana en el centro mientras la tecnología evoluciona por debajo —
+   el pensamiento de diseño es "atemporal", mientras que la tecnología está
+  "fechada".
+
+  Un estudio de 2026 en Scientific Reports examinó la relación entre el
+  pensamiento de orden superior, el uso de chatbots de IA generativa y la
+  creatividad en ingeniería, encontrando que el pensamiento de orden superior
+  tenía una importancia significativamente mayor para los resultados creativos
+  que el uso de herramientas de IA por sí solo — reforzando que la profundidad
+  cognitiva, no la velocidad de las herramientas, impulsa la innovación. La
+  investigación en las Proceedings of the Design Society argumenta además que la
+   IA generativa está desplazando el rol del diseñador de ejecutor a
+  configurador de paradigmas, haciendo que el pensamiento de diseño sea más
+  esencial a medida que la IA gestiona las tareas de producción mientras los
+  humanos deben guiar el encuadre, la ética y el significado.
+
+  Para ser claros, estos marcos no son distintos sabores de la misma
+  metodología. Scrum y el pensamiento de diseño representan epistemologías
+  completamente diferentes. Scrum asume que el problema te fue entregado,
+  mientras que el pensamiento de diseño asume que el problema debe encontrarse y
+   lidiarse con él.
+
+  ¿Es la IA una herramienta de conocimiento?
+
+  Esto nos lleva a la pregunta que la industria evita y que yo, por mi parte,
+  llevo tiempo deseando formular. ¿Es la IA una herramienta de conocimiento?
+  ¿Cómo se clasifica?
+
+  Un modelo de lenguaje de gran escala genera secuencias plausibles de texto.
+  Volviendo al artículo de Bender, Gebru, McMillan-Major y Mitchell mencionado
+  anteriormente, los autores definieron un modelo de lenguaje como un sistema
+  para ensamblar formas lingüísticas según patrones probabilísticos, "sin
+  ninguna referencia al significado." Sin referencia al significado, dado que
+  los LLMs sin configuración adicional carecen de anclaje externo. Optimizan
+  para la plausibilidad, no para la verdad, y por lo tanto, como sistemas
+  autónomos, pueden clasificarse mejor como herramientas de procesamiento de
+  información.
+
+  Se convierte en una herramienta de conocimiento únicamente cuando se ancla en
+  una capa semántica — vocabularios controlados, taxonomías, ontologías, grafos
+  de conocimiento — que proporciona referencia y procedencia. Las propias
+  medidas correctivas de la industria y el mercado asociado confirman esto como
+  un hecho. La generación aumentada por recuperación, los grafos de conocimiento
+   para LLMs y lo que el discurso reciente denomina "ingeniería de contexto"
+  son, bajo el empaquetado, la infraestructura semántica que debería haberse
+  construido primero. Pero fue añadida a posteriori para prevenir alucinaciones
+  que estaban prácticamente garantizadas desde el momento en que el modelo fue
+  desplegado sin ella.
+
+  El artículo sobre RAG de 2020 señaló este problema, identificando que la
+  memoria paramétrica es opaca y no puede actualizarse, y que la recuperación
+  desde fuentes estructuradas es el modo en que la procedencia se captura y
+  codifica. Pero parece que la industria tiene un problema de memoria, o el
+  mensaje llegó demasiado pronto para que la mayoría pudiera escucharlo.
+
+  Irónicamente, la procedencia es un elemento central de la gestión del
+  conocimiento y el trabajo ontológico, pues constituye un principio fundacional
+   de la disciplina y su espacio de problemas. Sin procedencia, el conocimiento
+  no puede existir. Las organizaciones están descubriendo ahora que la
+  necesitan, porque sus sistemas generativos no pueden explicarse a sí mismos.
+  Llámalo linaje, llámalo trazabilidad. Pero si construyes para el conocimiento,
+   llámalo procedencia, porque la procedencia abarca trazas, linaje, citaciones,
+   lo temporal y lo espacial.
+
+  De modo que si la IA generativa es una herramienta de procesamiento de
+  información, sostengo que para que la IA se convierta en una herramienta de
+  conocimiento debe haber procedencia y lógica descriptiva. Y para ser francos,
+  nadie llega a la procedencia haciendo sprints — esta se arquitecta y se
+  sostiene.
+
+  El conocimiento es una espiral, no un sprint
+
+  La ironía no pasa desapercibida en este momento. El mismo Nonaka que
+  coescribió el artículo de 1986 que inspiró Scrum, desarrolló el modelo SECI de
+   creación de conocimiento en 1995, describiendo el conocimiento como una
+  espiral entre formas tácitas y explícitas, que avanza a través de la
+  socialización, la externalización, la combinación y la internalización.
+
+  La espiral no tiene fin, como es propio de las espirales, emblema de la vida.
+  Cada rotación crea nuevo conocimiento que se convierte en el sustrato de la
+  siguiente. No existe una Definición de Hecho porque la definición de lo que se
+   está haciendo está siendo construida en sí misma. Scrum tomó la estructura de
+   equipo del primer artículo de Nonaka y descartó su siguiente década de
+  trabajo porque no encajaba limpiamente en Scrum. Los mecanismos del trabajo
+  del conocimiento, como Nonaka argumentó durante una década, son la metáfora,
+  la habilidad tácita, el aprendizaje por proximidad, el espacio compartido y la
+   paciente conversión de intuiciones en modelos y de modelos de vuelta en
+  práctica.
+
+  El Ontology Pipeline™ refleja esta espiral. Se comienza con un vocabulario
+  controlado — un proceso disciplinado que implica acuerdos sobre lo que
+  significan los términos. Se madura hacia una taxonomía introduciendo jerarquía
+   y estructurando los términos en relaciones padre-hijo. Posteriormente, la
+  taxonomía se madura hacia un tesauro añadiendo relaciones asociativas y
+  sinonimia, codificadas en SKOS. El tesauro es el trabajo base a partir del
+  cual se desarrolla una ontología, incorporando semántica formal mediante OWL y
+   RDF. Se arquitecta un grafo de conocimiento, ensamblando componentes que
+  incluyen los activos de conocimiento desarrollados a través del marco
+  iterativo del Ontology Pipeline™.
+
+  Cada etapa prepara el terreno para la siguiente y las etapas no pueden
+  saltarse ni eludirse. No se puede entregar el grafo de conocimiento sin una
+  ontología, y la ontología debe ser lógicamente consistente y coherente para
+  dotar de sentido a los datos. Un grafo de conocimiento sin control de
+  vocabulario arriesga convertirse en un mentiroso convincente o en fallar al
+  trabajar con lenguaje natural. Porque los activos de conocimiento son
+  dimensionalmente interdependientes, el conocimiento es efectivamente una
+  espiral, incapaz de ser contenida por la lógica de la manufactura.
+
+  El patrón y el remedio
+
+  Las organizaciones presupuestan para plataformas — Confluence, SharePoint,
+  suscripciones a LLMs, bases de datos vectoriales, un proveedor de RAG. Pero no
+   presupuestan para la lógica descriptiva y el conocimiento, quizás porque el
+  problema no se resuelve con una sola herramienta y resulta demasiado amorfo en
+   un mundo centrado en los datos y en Agile. El vocabulario controlado se
+  recorta a menos que genere valor inmediato para la base de datos. Con
+  frecuencia, la taxonomía se recorta porque no hace una buena demostración y su
+   valor reside en la fontanería, entregando su impacto con el tiempo. El
+  resultado es una pila siempre creciente de documentos con un chatbot pegado en
+   la entrada, respondiendo preguntas con confianza en la voz de un sistema que
+  no tiene referencia para ninguno de los términos dispersos a lo largo de los
+  documentos.
+
+  Entonces el sistema alucina. Se contradice a sí mismo entre sesiones. Presenta
+   como vigente una política desactualizada y muestra mensajes privados de
+  Outlook de empleados cuyas cuentas fueron desactivadas. He visto esto en la
+  vida real más veces de las que puedo contar. Y la percepción es peor cuando
+  una herramienta de proveedor es responsable de revelar chats y transacciones
+  de correo privadas. Habitualmente, el patrocinador ejecutivo pregunta por qué
+  la inversión no ha funcionado, ignorando la evidencia de los errores, como la
+  exposición de comunicaciones privadas. Trabajo de un día cualquiera, cuando
+  los burndown charts y la velocidad marcan el precedente de las formas de
+  trabajar y los sistemas de recompensa.
+
+  Los ejecutivos ignoran el hecho de que el sistema nunca fue anclado en el
+  conocimiento. Nunca lo fue porque anclar requiere el trabajo lento, costoso y
+  disciplinado de la organización del conocimiento — trabajo que no cabe en un
+  sprint, que no hace una buena demostración y que no genera un gráfico de
+  velocidad. La organización elige eludir la inversión en el aparato de
+  conocimiento descriptivo porque las líneas son demasiado difusas para
+  definirse claramente según el vocabulario de Agile. Porque el conocimiento es
+  infraestructura, no un producto ni una plataforma. El conocimiento es la
+  gasolina del depósito, el combustible del significado, el material que
+  justifica el razonamiento lógico.
+
+  El pensamiento de diseño le da al trabajo del conocimiento un hogar. Es cómodo
+   con la ambigüedad. Trata el descubrimiento del problema como un igual a la
+  resolución del problema. Desde luego no exige un incremento entregable cada
+  catorce días. En cambio, el pensamiento de diseño exige que entiendas lo que
+  estás construyendo antes de construirlo — y que sigas revisando esa
+  comprensión a medida que el trabajo avanza. La biblioteconomía y la ciencia de
+   la información, con más de un siglo de estándares para la organización del
+  conocimiento, proporcionan los métodos mientras el pensamiento de diseño
+  proporciona la epistemología. Juntos ofrecen lo que Agile nunca pudo — un
+  marco adecuado a la naturaleza del conocimiento mismo.
+
+  ¿Un sprint de dos semanas para entregar qué? Si estás construyendo sistemas de
+   conocimiento, dos semanas no producirán nada coherente. El conocimiento no se
+   construye rápido y se moldea a la fuerza. Se construye con cuidado, con
+  procedencia, gobernanza y paciencia. El conocimiento no es un producto que
+  pueda manufacturarse y no — no se puede comprar conocimiento estándar que
+  encaje perfectamente con un dominio u organización. El conocimiento es
+  matizado, específico a cada organización, dominio y persona.
+
+  En el espíritu del argot del capital de riesgo: el conocimiento es el foso, y
+  ese foso no es Ágil.
+
+  ---
+  Por Jessica Talisman — Intentional Arrangement (Substack)
+
+✻ Baked for 2m 38s
+
+❯ Con el título A Two-Week Sprint for Knowledge                                 
+  From Agile to Design Thinking                                                 
+  Jessica Talisman, MLS                                                         
+  Apr 21, 2026                                                                  
+  y su traducción al espanõl                                                    
+  genera documentos en pdf y md en inglés y en español                          
+
+⏺ Primero verifico si pandoc está disponible para generar los PDFs.
+
+⏺ Bash(which pandoc && pandoc --version | head -1)
+  ⎿  /usr/local/bin/pandoc
+     pandoc 3.7.0.2
+
+⏺ Bien. Creo los cuatro archivos en /tmp/talisman-article/.
+
+⏺ Bash(mkdir -p /tmp/talisman-article)
+  ⎿  Done
+
+⏺ Write(/tmp/talisman-article/a-two-week-sprint-for-knowledge.md)
+  ⎿  Wrote 101 lines to
+     ../../../../tmp/talisman-article/a-two-week-sprint-for-knowledge.md
+      2 ## From Agile to Design Thinkinge
+      4 **Jessica Talisman, MLS**
+      6 *Apr 21, 2026*
+      7 ---
+      9 In 1986, Hirotaka Takeuchi and Ikujiro Nonaka published "The New 
+        escribing how Honda, Canon, and Fuji-Xerox organized teams to bui
+        ld copiers and automobiles. Jeff Sutherland read that article, co
+        programming, and in 1993 adapted it for software development at E
+        A conference in 1995. By 2001, seventeen software developers gath
+        ered at a ski lodge in Snowbird, Utah, and signed the A
+     … +364 lines (ctrl+o to expand)
+
+❯ usa typst ? 
+
+⏺ Write(/tmp/talisman-article/un-sprint-de-dos-semanas-para-el-conocimiento.md)
+  ⎿  Wrote 101 lines to ../../../../tmp/talisman-article/un-sprint-de-dos-s
+     emanas-para-el-conocimiento.md
+      2 ## De Agile al Pensamiento de Diseñoonocimiento
+      4 **Jessica Talisman, MLS**
+      6 *21 de abril de 2026*
+      7 ---
+      9 En 1986, Hirotaka Takeuchi e Ikujiro Nonaka publicaron "The New N
+        biendo cómo Honda, Canon y Fuji-Xerox organizaban sus equipos par
+        a fabricar copiadoras y automóviles. Jeff Sutherland leyó ese art
+        gramación orientada a objetos, y en 1993 lo adaptó al desarrollo 
+        o en la conferencia OOPSLA de 1995. En 2001, diecisiete desarroll
+        adores de software se reuniero
+     … +407 lines (ctrl+o to expand)
+
+⏺ El usuario preguntó sobre Typst. Verifico si está disponible.
+
+⏺ Bash(which typst && typst --version)
+  ⎿  /opt/homebrew/bin/typst
+     typst 0.14.2 (unknown hash)
+
+⏺ Typst disponible. Genera PDFs mucho más elegantes que pandoc. Creo los fuentes
+   .typ y compilo.
+
+⏺ Write(/tmp/talisman-article/en.typ)
+  ⎿  Wrote 160 lines to ../../../../tmp/talisman-article/en.typ
+       1 #set document(
+       2   title: "A Two-Week Sprint for Knowledge",
+       3   author: "Jessica Talisman, MLS",
+       4 )
+       5 
+       6 #set page(
+       7   paper: "a4",
+       8   margin: (top: 2.8cm, bottom: 3cm, left: 3cm, right: 3cm),
+       9   footer: context [
+      10     #set text(size: 9pt, fill: luma(140))
+     … +433 lines (ctrl+o to expand)
+
+⏺ Write(/tmp/talisman-article/es.typ)
+  ⎿  Wrote 160 lines to ../../../../tmp/talisman-article/es.typ
+       1 #set document(
+       2   title: "Un Sprint de Dos Semanas para el Conocimiento",
+       3   author: "Jessica Talisman, MLS",
+       4 )
+       5 
+       6 #set page(
+       7   paper: "a4",
+       8   margin: (top: 2.8cm, bottom: 3cm, left: 3cm, right: 3cm),
+       9   footer: context [
+      10     #set text(size: 9pt, fill: luma(140))
+     … +475 lines (ctrl+o to expand)
+
+⏺ Ahora compilo los dos PDFs en paralelo.
+
+⏺ Bash(typst compile /tmp/talisman-article/en.typ
+      /tmp/talisman-article/a-two-week-sprint-for-knowledge.pdf 2>&1)
+  ⎿  warning: unknown font family: linux libertine
+        ┌─ ../../../../private/tmp/talisman-article/en.typ:17:8
+        │
+     … +2 lines (ctrl+o to expand)
+
+⏺ Bash(typst compile /tmp/talisman-article/es.typ
+      /tmp/talisman-article/un-sprint-de-dos-semanas-para-el-conocimiento.pdf
+      2>&1)
+  ⎿  warning: unknown font family: linux libertine
+        ┌─ ../../../../private/tmp/talisman-article/es.typ:17:8
+        │
+     … +2 lines (ctrl+o to expand)
+
+⏺ Solo warnings de fallback de fuente, no errores. Verifico los archivos
+  generados.
+
+  Listed 1 directory (ctrl+o to expand)
+⏺ Los 4 archivos están en /tmp/talisman-article/:
+
+  ┌───────────────────────────────────────────────────┬────────┐
+  │                      Archivo                      │ Tamaño │
+  ├───────────────────────────────────────────────────┼────────┤
+  │ a-two-week-sprint-for-knowledge.md                │ 20 KB  │
+  ├───────────────────────────────────────────────────┼────────┤
+  │ a-two-week-sprint-for-knowledge.pdf               │ 101 KB │
+  ├───────────────────────────────────────────────────┼────────┤
+  │ un-sprint-de-dos-semanas-para-el-conocimiento.md  │ 23 KB  │
+  ├───────────────────────────────────────────────────┼────────┤
+  │ un-sprint-de-dos-semanas-para-el-conocimiento.pdf │ 108 KB │
+  └───────────────────────────────────────────────────┴────────┘
+
+  Sí, usé Typst para los PDFs — produce tipografía profesional con
+  justificación, guiones automáticos por idioma (lang: "en" / lang: "es"),
+  paginación, y un bloque de autor limpio. Georgia como fuente serif principal.
+  Los fuentes .typ también están ahí si quieres ajustar el estilo.
+
+✻ Cogitated for 11m 45s
\ No newline at end of file
diff --git a/assets/presentation/docs/INDEX.md b/assets/presentation/docs/INDEX.md
new file mode 100644
index 0000000..c696986
--- /dev/null
+++ b/assets/presentation/docs/INDEX.md
@@ -0,0 +1,69 @@
+# Presentation Docs — Index
+
+Material de referencia para posts, slides y presentaciones de ontoref.
+
+---
+
+## Análisis y posicionamiento ontoref
+
+| Archivo | Contenido |
+|---|---|
+| [ontoref-yin-yang-diferenciacion.md](ontoref-yin-yang-diferenciacion.md) | Por qué Talisman no resuelve el lado operacional. Ontología sin Reflection = buenas intenciones. |
+| [ontoref-dags-diferenciacion.md](ontoref-dags-diferenciacion.md) | Por qué los DAGs de ontoref son distintos a CI/CD, compiladores, runbooks. DAGs como conocimiento vs. ejecución. |
+| [ontoref-scope-proyecto-infra-personal.md](ontoref-scope-proyecto-infra-personal.md) | Scope completo: proyecto + infraestructura + personal. Stack soberano: jj, Radicle, NATS, CLI/UI/MCP/GraphQL. |
+
+---
+
+## Posts — Serie ontoref (EN + ES)
+
+Publicados en jesusperez.pro/blog. Estado: borradores (`published: false`).
+
+| Post | EN | ES |
+|---|---|---|
+| AI is a knowledge tool — who keeps it alive? | [en](posts/en/ai-knowledge-tool-who-keeps-it-alive.md) | [es](posts/es/la-ia-herramienta-conocimiento-quien-mantiene-vivo.md) |
+| DAGs are everywhere — none know what they are | [en](posts/en/dags-everywhere-none-know-what-they-are.md) | [es](posts/es/dags-en-todos-lados-ninguno-sabe-lo-que-es.md) |
+| One protocol, three subjects | [en](posts/en/one-protocol-three-subjects.md) | [es](posts/es/un-protocolo-tres-sujetos.md) |
+| Your ontology should live with your code | [en](posts/en/your-ontology-should-live-with-your-code.md) | [es](posts/es/tu-ontologia-deberia-vivir-con-tu-codigo.md) |
+
+---
+
+## Fuentes externas — Talisman (KGC 2026)
+
+| Archivo | Contenido |
+|---|---|
+| [Steal_This_Deck.pdf](Steal_This_Deck.pdf) | Deck completo. "Stop Betting, Start Building", KGC 2026, Jessica Talisman. |
+| [talisman-steal-this-deck-extractions.md](talisman-steal-this-deck-extractions.md) | Extracciones del deck: citas, datos citable, mapeos a ontoref, nota de atribución. |
+| [talisman-article/](talisman-article/) | Artículo de Talisman "A Two-Week Sprint for Knowledge" en EN/ES, PDF y Typst. |
+
+---
+
+## Fuentes externas — José Celano
+
+| Archivo | Contenido |
+|---|---|
+| [JoseCelano_2026-6-08_Ontologías_y_DAGs.pdf](JoseCelano_2026-6-08_Ontologías_y_DAGs.pdf) | Ontologías y DAGs — presentación José Celano, junio 2026. |
+
+---
+
+## Material previo — ontología como filosofía del software
+
+| Archivo | Contenido |
+|---|---|
+| [ontologia-filosofia-software.md](ontologia-filosofia-software.md) | Documento principal: ontología como fundamento filosófico del software. |
+| [ontologia-filosofia-software.pdf](ontologia-filosofia-software.pdf) | Versión PDF. |
+| [ontologia-filosofia-software.typ](ontologia-filosofia-software.typ) | Fuente Typst. |
+| [ontologia_motor.md](ontologia_motor.md) | Ontología como motor — concepto central. |
+| [ontologia_motor_resumen.md](ontologia_motor_resumen.md) | Resumen ejecutivo del motor ontológico. |
+
+---
+
+## Material previo — sprints y contexto
+
+| Archivo | Contenido |
+|---|---|
+| [2-week-sprint-knowledge_done.md](2-week-sprint-knowledge_done.md) | Sprint de dos semanas para conocimiento — completado. |
+| [about_plantir.md](about_plantir.md) | Contexto sobre Plantir como referencia de posicionamiento. |
+| [2026-02-17-notas_voz.md](2026-02-17-notas_voz.md) | Notas de voz, 17 Feb 2026. |
+| [2026-02-27-notas_voz_rust.md](2026-02-27-notas_voz_rust.md) | Notas de voz sobre Rust, 27 Feb 2026. |
+| [2026-04-24-234910-me-planteo-estas-dos-propuestas-para-un-open-spac.txt](2026-04-24-234910-me-planteo-estas-dos-propuestas-para-un-open-spac.txt) | Propuestas para open space, 24 Abr 2026. |
+| [2026-04-25-000346-puedes-sacar-el-texto-de-este-contenido.txt](2026-04-25-000346-puedes-sacar-el-texto-de-este-contenido.txt) | Extracción de contenido, 25 Abr 2026. |
diff --git a/assets/presentation/docs/JoseCelano_2026-6-08_Ontologías_y_DAGs.pdf b/assets/presentation/docs/JoseCelano_2026-6-08_Ontologías_y_DAGs.pdf
new file mode 100644
index 0000000..d3ff958
Binary files /dev/null and b/assets/presentation/docs/JoseCelano_2026-6-08_Ontologías_y_DAGs.pdf differ
diff --git a/assets/presentation/docs/Steal_This_Deck.pdf b/assets/presentation/docs/Steal_This_Deck.pdf
new file mode 100644
index 0000000..e9fd1e4
Binary files /dev/null and b/assets/presentation/docs/Steal_This_Deck.pdf differ
diff --git a/assets/presentation/docs/about_plantir.md b/assets/presentation/docs/about_plantir.md
new file mode 100644
index 0000000..d772b22
--- /dev/null
+++ b/assets/presentation/docs/about_plantir.md
@@ -0,0 +1,28 @@
+resume esto https://axiussdc.substack.com/p/palantir-just-said-the-quiet-part
+
+# Resumen: "Palantir acaba de decir la parte que se callaba"
+
+**Tesis central:** El artículo, publicado por Axius-SDC en abril de 2026, sostiene que Palantir admitió públicamente algo que la comunidad de arquitectura de datos lleva años discutiendo: **la IA no es el producto, la ontología sí lo es**. La IA es solo el motor de ejecución.
+
+**Lo que Palantir acierta**
+
+Citando un artículo de Sainath Palla sobre el AI-FDE (Foundry Development Environment) de Palantir, el autor reconoce que la ontología no es documentación, sino una capa operacional por la que fluyen datos, workflows y decisiones. El AI-FDE permite que la IA participe en construir el sistema (no solo asistir dentro de él), desplazando el cuello de botella de la implementación al juicio humano: qué construir y cómo debe comportarse.
+
+**Donde Palantir falla**
+
+El problema es estructural: la ontología vive *dentro* de la plataforma de Palantir y está gobernada centralmente. A medida que se añaden casos de uso, las intersecciones se multiplican y la carga de gobernanza crece más rápido que el ahorro en implementación. Las únicas personas capaces de manejar esa complejidad son los Forward Deployed Engineers (FDEs) de Palantir, lo que crea una dependencia del proveedor que ninguna IA puede eliminar. Estás "alquilando gobernanza", no comprando infraestructura.
+
+**La alternativa propuesta: gobernanza distribuida**
+
+En lugar de una ontología centralizada, el autor propone vincular las restricciones a los datos mismos a nivel de componente. La metáfora cambia: ya no construyes una red ferroviaria con intersecciones complejas, sino un *estándar de ancho de vía* donde cada segmento lleva su propia especificación. Modelos de workflow, atestación, cadenas de procedencia y roles viajan con el payload, no se gestionan desde una plataforma central. Una IA operando sobre este sustrato valida contra el modelo en tiempo de ejecución, sin necesidad de una ontología centralizada.
+
+**Tres preguntas para hacerle a tu account manager de Palantir el lunes:**
+1. ¿Cuántos FDEs requiere tu despliegue, y qué pasa si reduces ese número a la mitad?
+2. Cuando el AI-FDE escala un nuevo caso de uso, ¿quién valida que no desestabilice los workflows existentes?
+3. ¿Puedes exportar la ontología y desplegarla en tu propia infraestructura sin dependencia de Palantir?
+
+**Conclusión**
+
+Palantir no es el enemigo: es el "rompehielos" que está educando al Fortune 500 sobre la IA basada en ontologías. La pregunta no es *si* la arquitectura debe estar guiada por una ontología, sino si esa ontología debe ser propietaria, centralizada y gestionada por ingenieros del proveedor, o abierta, determinista y mantenida por practicantes propios. El autor promociona el programa SDC Practitioner (Apache 2.0, lanzamiento mayo 2026) como esta alternativa abierta.
+
+En una frase: **Palantir demostró la tesis; según el autor, no deberías tener que pagarla en alquiler.**
diff --git a/assets/presentation/docs/ontologia-filosofia-software.md b/assets/presentation/docs/ontologia-filosofia-software.md
new file mode 100644
index 0000000..d78f46d
--- /dev/null
+++ b/assets/presentation/docs/ontologia-filosofia-software.md
@@ -0,0 +1,544 @@
+# Ontología filosófica y modelo operacional en software
+
+**Genealogía conceptual, autores y uso contemporáneo**
+
+*Jesús Pérez Lorenzo · OntoRef · 2026*
+
+---
+
+## Contenidos
+
+1. [¿Son los invariantes el equivalente a los universales filosóficos?](#1-son-los-invariantes-el-equivalente-a-los-universales-filosóficos)
+2. [Genealogía filosófica del modelo operacional](#2-genealogía-filosófica-del-modelo-operacional)
+3. [La ontología operacional en software: origen, autores y uso actual](#3-la-ontología-operacional-en-software-origen-autores-y-uso-actual)
+4. [Taoísmo, Yin-Yang y ontología operacional](#4-taoísmo-yin-yang-y-ontología-operacional)
+5. [DAGs — genealogía y función en ontología operacional](#5-dags--genealogía-y-función-en-ontología-operacional)
+6. [Física cuántica, computación cuántica y ontología operacional](#6-física-cuántica-computación-cuántica-y-ontología-operacional)
+7. [Por qué es relevante en software e infraestructuras digitales](#7-por-qué-es-relevante-en-software-e-infraestructuras-digitales)
+8. [Referencias y lecturas recomendadas](#8-referencias-y-lecturas-recomendadas)
+
+---
+
+## 1. ¿Son los invariantes el equivalente a los universales filosóficos?
+
+La pregunta es legítima pero la correspondencia es más fina de lo que parece a primera vista.
+
+### Los universales en ontología filosófica
+
+Los *universales* responden a una pregunta específica: ¿qué comparten múltiples particulares que los hace del mismo tipo? Son propiedades, relaciones o cualidades que pueden instanciarse en varios individuos distintos — "rojez", "humanidad", "ser-mayor-que". El debate clásico gira en torno a su estatus ontológico:
+
+- **Realismo platónico**: los universales existen independientemente de sus instancias — las Formas habitan un reino separado del mundo sensible.
+- **Realismo moderado** (Aristóteles): los universales existen *en* las cosas, no aparte de ellas. No hay humanidad sin seres humanos concretos.
+- **Nominalismo** (Ockham, siglo XIV): solo existen los particulares. Los universales son nombres (*nomina*), convenciones lingüísticas sin referente ontológico propio.
+- **Conceptualismo** (posición media): los universales existen en la mente como conceptos, no en el mundo ni como meros nombres.
+
+### Los invariantes como propiedades esenciales de un singular
+
+Los invariantes en un modelo operacional de software no son universales en ninguno de los sentidos anteriores. No preguntan qué comparten múltiples proyectos — preguntan *qué no puede cambiar en este proyecto concreto sin que deje de ser este proyecto*.
+
+La analogía filosófica precisa es la distinción aristotélica entre **propiedades esenciales y accidentales**: la esencia es aquello sin lo cual una cosa no sería lo que es; el accidente es lo que puede cambiar sin alterar su naturaleza. Un invariante operacional declara: "esta propiedad es esencial para la identidad de este proyecto."
+
+Kripke formalizó esto de manera más rigurosa en *Naming and Necessity* (1980): ciertas propiedades son **rígidas** — verdaderas en todos los mundos posibles accesibles al individuo. El agua es necesariamente H₂O aunque alguien la llame "phlogiston". Un invariante operacional tiene este carácter modal: no puede ser falso en ningún estado posible del proyecto sin que el proyecto haya dejado de ser ese proyecto. Los ADRs que lo modifican no *violan* el invariante — constituyen un nuevo individuo.
+
+### Dónde sí hay resonancia con los universales
+
+La familia semántica es compartida: ambos capturan *lo que no varía*. El universal no varía a través de la instanciación. El invariante no varía a través de la evolución temporal del proyecto. El eje difiere — instanciación vs. tiempo — pero el gesto filosófico es análogo.
+
+### El estatus ontológico: hechos institucionales
+
+Los universales platónicos existen independientemente de que alguien los declare. Los invariantes operacionales son **hechos institucionales** en el sentido de John Searle (*La construcción de la realidad social*, 1995): son verdad porque un colectivo los declaró y formalizó. Como el dinero, los contratos o las reglas del ajedrez — su existencia depende de actos declarativos sostenidos por una práctica social.
+
+Un invariante que el equipo no reconoce como tal no es un invariante real; es teatro.
+
+---
+
+## 2. Genealogía filosófica del modelo operacional
+
+| Concepto filosófico | Autor / Origen | Equivalente operacional | Implementación |
+|---|---|---|---|
+| **Universales** | Platón (428–348 aC) — *República*, *Parménides* | Tipos / Schemas | Contratos Nickel: `String`, `[| 'Accepted, 'Proposed |]` |
+| **Esencia / Accidente** | Aristóteles (384–322 aC) — *Metafísica* | Invariante / Estado mutable | `invariante = true` en `core.ncl` vs. dimensiones en `state.ncl` |
+| **Potencia / Acto** | Aristóteles — *Metafísica* IX | Estado actual → deseado + Gates | `current_state` / `desired_state` / `gate.ncl` |
+| **Categorías** | Aristóteles — *Categorías* | Tipos de nodo ontológico | `invariants`, `tensions`, `practices`, `dimensions`, `gates` |
+| **Sustancia primera / segunda** | Aristóteles — *Categorías* | Instancia / Tipo | Proyecto concreto vs. schema de proyecto |
+| **Identidad a través del tiempo** | Locke (1689), Parfit (1984) | Versioning + ADR lifecycle | `status = Superseded`, `superseded_by` |
+| **Dialéctica** | Hegel (1807) — *Fenomenología del Espíritu* | Tensiones activas | `tensions` en `core.ncl` — sin síntesis, gestionadas |
+| **Logos como tensión de opuestos** | Heráclito (535–475 aC) | Tensión constitutiva | Dualidad ontología/reflection como tensión de diseño |
+| **Mundos posibles** | Leibniz (1710), Kripke (1980) | FSM — transiciones permitidas | ADRs como restricciones de accesibilidad entre estados |
+| **Propiedades rígidas** | Kripke — *Naming and Necessity* (1980) | Invariantes como necesidad | Un invariante es verdad en todos los estados posibles |
+| **Límites del lenguaje = límites del mundo** | Wittgenstein — *Tractatus* (1921) | Schema como frontera de lo expresable | Lo que no tipea en Nickel no puede existir en el sistema |
+| **Juegos de lenguaje** | Wittgenstein — *Investigaciones* (1953) | Modos operacionales | Un modo define las reglas del juego para un contexto |
+| **Verdad como correspondencia** | Tarski (1933) | Typecheck / validación formal | `nickel export` como verificación de correspondencia |
+| **Hechos institucionales** | Searle (1995) | ADRs como decisiones declaradas | Una constraint existe porque el equipo la declaró |
+| **Ontología formal vs. regional** | Husserl (1913) — *Ideas I* | Schema formal vs. ontología del proyecto | `adrs/schema.ncl` vs. `.ontology/core.ncl` |
+| **Filosofía del proceso** | Whitehead (1929) — *Process and Reality* | Ontología (ser) vs. Reflection (devenir) | La dualidad estructural del protocolo |
+| **Realismo estructural** | Ladyman & Ross (2007) | Grafo como lo real | El DAG captura relaciones; los nodos son secundarios |
+| **Lógica de predicados** | Frege (1879) — *Begriffsschrift* | Queries sobre el grafo | `where severity == "Hard" \| select id claim` |
+
+### Matices que la tabla no captura
+
+**Aristóteles es el arquitecto base.** La esencia/accidente es exactamente invariante/estado-mutable. La potencia/acto es exactamente `current_state`/`desired_state` con la gate como umbral de actualización.
+
+**Whitehead contra Aristóteles es la tensión interna del modelo.** Aristóteles ve la realidad como sustancias con propiedades. Whitehead la ve como procesos. Un modelo operacional de software vive en esa grieta: la ontología es aristotélica (qué *es* el proyecto), la capa de reflection es whiteheadiana (cómo *deviene*).
+
+**Kripke resuelve lo que Aristóteles dejó ambiguo.** Aristóteles distingue esencial de accidental pero no formaliza la modalidad. Kripke lo hace: una propiedad esencial es rígida — verdadera en todos los mundos posibles accesibles desde el individuo.
+
+**La ausencia notable: Heidegger.** El *ser-en-el-mundo* como condición de posibilidad de cualquier ontología explícita no aparece en ningún modelo operacional de software. Lo que Heidegger llamaría el trasfondo pre-temático es exactamente lo que los READMEs intentan capturar y fracasan. Un modelo operacional no lo resuelve; reduce la superficie de ese trasfondo haciendo explícito lo que puede serlo.
+
+---
+
+## 3. La ontología operacional en software: origen, autores y uso actual
+
+### El problema que había que resolver
+
+A finales de los años 1970, los sistemas de inteligencia artificial enfrentaban el *cuello de botella del conocimiento*: los sistemas expertos eran eficaces en dominios acotados pero no podían compartir conocimiento entre sí. Cada sistema construía su representación del mundo de forma ad hoc, incompatible con el resto.
+
+### Los fundadores
+
+**John McCarthy — el precursor (1958–1980s)**
+Fundador del término "inteligencia artificial" y creador de LISP, fue el primero en proponer que los sistemas de IA necesitaban representaciones explícitas del mundo para razonar sobre él. Su *situational calculus* (1969, con Patrick Hayes) — un lenguaje formal para describir el mundo y sus cambios — es el antecedente directo de la ontología operacional. McCarthy adoptó el término "ontología" de la filosofía para referirse a los compromisos sobre qué tipos de cosas existen en el dominio modelado.
+
+**Marvin Minsky — frames y estructuras de conocimiento (1974)**
+En *A Framework for Representing Knowledge* (MIT AI Lab, 1974), Minsky propuso los *frames* como estructuras para organizar el conocimiento: cada frame representa un concepto con slots (atributos) y valores por defecto. Es la primera implementación computacional de algo cercano a la sustancia aristotélica con propiedades.
+
+**Tom Gruber — la definición canónica (1993)**
+Formuló la definición que se convirtió en estándar: *"An ontology is a specification of a conceptualization"* (*Knowledge Acquisition*, 1993). La definición es deliberadamente minimalista: una ontología no captura la realidad — captura un *punto de vista compartido* sobre un dominio. Esto la sitúa en el territorio de los hechos institucionales de Searle.
+
+**Nicola Guarino — ontología formal en sistemas de información (1990s)**
+Desarrolló la distinción entre *ontología de alto nivel* (conceptos universales como tiempo, espacio, proceso) y *ontología de dominio* (conceptos específicos de un campo). Fundó la conferencia *Formal Ontology in Information Systems* (FOIS, 1998).
+
+**Doug Lenat — Cyc: el proyecto más ambicioso (1984–2017)**
+Inició el proyecto *Cyc* en MCC con el objetivo de codificar todo el conocimiento de sentido común humano. Alcanzó más de 600.000 conceptos y 5 millones de aserciones. Demostró tanto el potencial como los límites del enfoque: la escala del conocimiento implícito humano es prácticamente inabarcable mediante codificación manual.
+
+**Tim Berners-Lee, James Hendler y Ora Lassila — la Web Semántica (2001)**
+Propusieron extender la Web con significado semántico mediante RDF, OWL y SPARQL — el ecosistema que popularizó el uso de ontologías en ingeniería de software más allá de la IA.
+
+### Línea temporal
+
+| Año | Hito |
+|---|---|
+| 1958 | McCarthy propone representación explícita del conocimiento para IA |
+| 1969 | McCarthy & Hayes — situational calculus |
+| 1974 | Minsky — frames: primera estructura análoga a la sustancia aristotélica |
+| 1984 | Lenat inicia Cyc |
+| 1993 | Gruber — definición canónica |
+| 1998 | Primera conferencia FOIS |
+| 2001 | Berners-Lee et al. — Web Semántica: RDF, OWL, SPARQL |
+| 2004 | OWL se convierte en recomendación W3C |
+| 2012 | Google Knowledge Graph — ontología operacional en producción |
+| 2020s | LLMs entrenados sobre grafos ontológicos; ontologías como contexto para agentes |
+
+### Uso actual
+
+**Biomedicina:** Gene Ontology (GO), SNOMED CT, ICD-11, Human Phenotype Ontology.
+
+**Grafos de conocimiento empresariales:** Google Knowledge Graph (500.000M de hechos), Wikidata (100M de elementos), schema.org.
+
+**Industria y manufactura:** ISO 15926 (petróleo y gas), IFC/BIM (construcción), ontologías Industry 4.0.
+
+**Inteligencia artificial:** ConceptNet, entrenamiento de LLMs (Wikidata, DBpedia), contexto estructurado para agentes autónomos.
+
+**Software de proyecto — la frontera actual:** El uso de ontologías operacionales *dentro* de proyectos de software para describir el propio proyecto es el desarrollo más reciente. Los precursores son los ADRs (Nygard, 2011) y los modelos C4 (Brown). La propuesta de *ontoref* pertenece a esta frontera: el proyecto *es* el dominio, y la ontología *es* el modelo operacional de ese dominio.
+
+---
+
+## 4. Taoísmo, Yin-Yang y ontología operacional
+
+### Lo que la tradición occidental no vio
+
+La ontología occidental desde Parménides privilegia el *ser* sobre el *no-ser*, la *presencia* sobre la *ausencia*, la *identidad* sobre la *diferencia*. El Taoísmo invierte la pregunta. El Tao (道) es previo a la distinción ser/no-ser:
+
+> 道可道，非常道。名可名，非常名。  
+> *El Tao que puede nombrarse no es el Tao eterno. El nombre que puede pronunciarse no es el nombre eterno.*  
+> — Laozi, *Tao Te Ching*, cap. 1 (~siglo VI-IV aC)
+
+Lo que no puede formalizarse no es una limitación del modelo — es la condición de posibilidad de todo lo que sí puede formalizarse.
+
+### Yin-Yang — polaridad constitutiva, no oposición
+
+El yin y el yang no son *contrarios* sino *complementarios que se definen mutuamente*:
+
+- Cada polo **contiene una semilla del opuesto**
+- El exceso de un polo **genera su contrario** (極則反)
+- La tensión entre ellos **nunca se resuelve** — solo se gestiona
+- Uno **no existe sin el otro**
+
+Esto es estructuralmente distinto de la dialéctica hegeliana, que busca síntesis. El yin-yang no resuelve — la tensión es constitutiva, permanente. Heráclito es el occidental más cercano: "el camino hacia arriba y hacia abajo son el mismo."
+
+### Tabla de correspondencias
+
+| Concepto taoísta | Texto / Época | Equivalente occidental | Correspondencia operacional |
+|---|---|---|---|
+| **Tao** (道) | Laozi, *Tao Te Ching* (~6-4 aC) | Ser de Heidegger | Trasfondo pre-temático que ningún schema captura |
+| **Yin / Yang** | Tradición Zhou (~1100 aC), *I Ching* | Heráclito — logos como tensión | Ontología (ser) / Reflection (devenir) |
+| **Wu wei** (無為) | Laozi, *Tao Te Ching* | Sin equivalente exacto | "No enforcement" — coherencia sin imposición |
+| **Ji** (機) — momento de transformación | Zhuangzi (~4 aC) | Kripke — mundos posibles | Gates — umbral cualitativo entre estados |
+| **Wu ji → Tai ji** | Zhou Dunyi (1017-1073), *Taijitu shuo* | Husserl — formal vs. regional | schema.ncl (vacío) → core.ncl (polaridad concreta) |
+| **Relatividad perspectival** | Zhuangzi — la rana del pozo | Wittgenstein tardío | Toda ontología es una conceptualización (Gruber) |
+
+### Las cuatro correspondencias profundas
+
+**Yin/Yang ↔ Ontología/Reflection.** La ontología captura lo que ES — capa yang: formal, dura, definida. La reflection captura lo que DEVIENE — capa yin: fluida, procesual. Como en el símbolo: cada una contiene una semilla de la otra.
+
+**Wu wei ↔ "No enforcement".** Coherencia voluntaria que emerge de la adopción justificada, sin mecanismo de enforcement. No impones — creas las condiciones para que el alineamiento emerja naturalmente.
+
+**Tensiones ↔ Yin-Yang permanente.** Las tensiones operacionales son explícitamente irresolubles. Si eliminas la tensión entre formalización y adopción maximizando un polo, el sistema muere.
+
+**Gates ↔ Ji (機).** El punto de inflexión donde el potencial se actualiza. No un estado sino el umbral entre estados. Dinámico, cualitativo, contextual.
+
+### La crítica taoísta a la ontología clasificatoria
+
+Los sistemas OWL intentan eliminar la ambigüedad mediante clasificación precisa. El *Tao Te Ching* advierte: cuanto más preciso el nombre, más excluyes. Esto explica por qué las ontologías biomédicas sufren continuamente con casos límite y dependencia del contexto. La precisión y la cobertura son yin y yang — no se resuelven, se equilibran.
+
+Zhuangzi lo lleva más lejos: la ontología correcta para un proyecto no es la más formalmente precisa sino la que el equipo puede sostener y evolucionar. Gruber llegó al mismo lugar en 1993: una ontología es "una especificación de una conceptualización" — un punto de vista compartido, no una verdad metafísica.
+
+---
+
+## 5. DAGs — genealogía y función en ontología operacional
+
+### El problema que fuerza el grafo
+
+El grafo emerge como la estructura mínima que captura relaciones arbitrarias entre conceptos. Pero el grafo general puede tener ciclos — y los ciclos en ontología son epistemológicamente problemáticos.
+
+### Por qué la aciclicidad
+
+Dos tradiciones independientes llegan al mismo resultado:
+
+**Aristóteles** (*Analíticos Posteriores*): una definición válida procede de género más diferencia específica. La jerarquía es estrictamente descendente. Si A se define en términos de B y B en términos de A, ninguno está definido.
+
+**Russell** (1901): la paradoja del conjunto de todos los conjuntos que no se contienen a sí mismos. La solución en *Principia Mathematica* (1910) es la *teoría de tipos*: los tipos forman una jerarquía estricta — un tipo de nivel N solo puede referirse a tipos de nivel N-1 o inferior. Esto es estructuralmente un DAG. La aciclicidad es la solución formal al colapso por autorreferencia.
+
+### Línea genealógica
+
+| Año | Autor / Hito | Contribución |
+|---|---|---|
+| 384-322 aC | Aristóteles — *Analíticos Posteriores* | Definición bien fundada: jerarquía sin circularidad |
+| 1879 | Frege — *Begriffsschrift* | Lógica de predicados: relaciones formales entre conceptos |
+| 1901-1910 | Russell & Whitehead — *Principia Mathematica* | Teoría de tipos como DAG: aciclicidad contra autorreferencia |
+| 1933 | Tarski — "El concepto de verdad" | Verdad como correspondencia entre proposiciones y estructuras |
+| 1958 | US Navy — PERT charts (programa Polaris) | DAG para planificación: camino crítico como recorrido del grafo |
+| 1969 | Collins & Quillian — redes semánticas | Primera red semántica computacional: is-a como DAG |
+| 1974 | Minsky — frames | Herencia múltiple: DAG sobre árbol |
+| 1976 | Feldman — Make | Build system como DAG: orden topológico, ciclos = error |
+| 1985 | Brachman & Schmolze — KL-ONE | Description Logic: subsunción formal como DAG |
+| 1988 | Pearl — Bayesian networks | Causalidad como DAG: asimetría temporal inherentemente acíclica |
+| 2001 | Berners-Lee et al. — Web Semántica | RDF/OWL: grafos de conocimiento como estándar industrial |
+| 2005 | Torvalds — Git | Historia de commits como DAG: irreversibilidad del tiempo en VCS |
+
+### Las cuatro funciones del DAG en ontología operacional
+
+**① Análisis de impacto decidible.** Antes de cambiar un nodo, recorre el DAG y sabes exactamente qué se ve afectado. En un grafo cíclico esto es indecidible. Concreto: `terraform plan` recorre el DAG de recursos para determinar orden de creación/modificación/destrucción.
+
+**② Orden topológico para ejecución fiable.** Cualquier operación con dependencias tiene un orden correcto que el DAG da sin coordinador central. Principio detrás de Make, Bazel, Gradle, y cualquier build system no trivial.
+
+**③ Justificación sin circularidad epistémica.** La ontología operacional es fundacionalista: los invariantes son nodos raíz justificados por ADR explícito. Un ciclo significaría dogma circular, no arquitectura.
+
+**④ Rollback determinista.** El inverso topológico del DAG es siempre computable. Sin DAG, el rollback depende de que alguien recuerde el orden correcto.
+
+### La tensión con el Taoísmo — DAG vs. ciclo
+
+El DAG es *yang*: direccional, jerárquico, asimétrico, irreversible. El yin-yang taoísta es *cíclico*: yin genera yang, yang genera yin.
+
+La resolución está en separar dos temporalidades:
+
+- **Sincrónica** (snapshot): el DAG es válido dentro de un momento dado.
+- **Diacrónica** (evolución): la secuencia de snapshots puede tener patrones cíclicos.
+
+El DAG captura la *estructura del ser en un momento*. El ciclo taoísta captura el *proceso de transformación a través del tiempo*. El *I Ching* — el libro de los cambios — es exactamente esto: 64 hexagramas como estados (nodos), transformaciones entre ellos como aristas. Un grafo con ciclos, porque la realidad vuelve.
+
+---
+
+## 6. Física cuántica, computación cuántica y ontología operacional
+
+### El golpe más profundo — Aristóteles estaba equivocado en el nivel fundamental
+
+La mecánica cuántica demuestra que, en el nivel fundamental, las propiedades no son definidas hasta que se miden. Un electrón no *tiene* posición — la posición emerge del acto de medición.
+
+El **teorema de Kochen-Specker** (1967) lo prueba formalmente: es imposible asignar valores definidos a todos los observables cuánticos simultáneamente, incluso antes de la medición. La indeterminación es *ontológica*, no epistémica. El suelo sobre el que se construyeron 2.300 años de ontología occidental se mueve.
+
+### Las tres conexiones genuinas
+
+**Complementariedad de Bohr ↔ Yin-Yang y la dualidad Ontología/Reflection**
+
+Niels Bohr (1885-1962) formuló el *principio de complementariedad*: onda y partícula son dos descripciones mutuamente excluyentes pero igualmente necesarias. No puedes observar ambas simultáneamente — el aparato de medición determina qué aspecto revelas.
+
+Bohr lo conectó explícitamente con el Taoísmo: cuando fue ennoblecido en 1947 adoptó el símbolo yin-yang como parte de su escudo de armas con la inscripción *Contraria sunt complementa*.
+
+La dualidad ontología/reflection tiene la misma estructura: la capa que uses (core.ncl o los modos de reflection) determina qué aspecto del proyecto ves. Ninguna es más verdadera — son complementarias.
+
+**Contextualidad cuántica ↔ Ontología perspectival**
+
+El teorema de Kochen-Specker prueba que el resultado de medir un observable depende de *qué otros observables se miden simultáneamente*. Las propiedades son *contextuales*. Esto convierte en hecho físico lo que tres tradiciones habían sostenido filosóficamente: Zhuangzi (perspectivismo), Wittgenstein tardío (juegos de lenguaje), Gruber (conceptualización).
+
+Para la ontología operacional: el "estado" de un proyecto es siempre relativo al actor y al contexto de consulta.
+
+**El problema del observador ↔ El trasfondo que no puede formalizarse**
+
+En mecánica cuántica, la medición colapsa la función de onda — el observador es parte del sistema. No hay *view from nowhere* (Nagel, 1986). Esto es exactamente el trasfondo pre-temático de Heidegger — la condición de posibilidad de cualquier formalización que ella misma no puede ser formalizada.
+
+### La interpretación más relevante — Mecánica Cuántica Relacional (Rovelli, 1996)
+
+Carlo Rovelli propuso que no existen estados absolutos — solo estados *relativos a otros sistemas*. El estado de una partícula es una propiedad relacional, no intrínseca. Esto conecta directamente con Ladyman y Ross (realismo estructural) y con la ontología perspectival.
+
+Para el modelo operacional: el `current_state` en `state.ncl` no es una propiedad intrínseca del proyecto — es relacional respecto al actor que consulta, los objetivos perseguidos, el momento del ciclo de vida.
+
+### La computación cuántica — honestidad sobre sus límites
+
+| Aspecto | Situación real | Relevancia para ontología operacional |
+|---|---|---|
+| Tamaño de grafos | La ventaja cuántica aparece a escala masiva | Los grafos de proyecto son pequeños — sin ventaja |
+| Algoritmos relevantes | Grover O(√N), Shor factorización | No mapean a consultas típicas de ontología |
+| Decoherencia | Cualquier interacción colapsa la superposición | Incompatible con operación continua |
+| Redes semánticas cuánticas | Investigación activa — relaciones con amplitudes | Captura gradualidad que OWL binario pierde |
+| LLMs y espacios de Hilbert | Embeddings como productos internos | **Directamente relevante** — misma matemática |
+
+### La conexión no obvia — LLMs como geometría cuántica aplicada
+
+Los Large Language Models operan en espacios vectoriales de alta dimensión donde el significado es relacional y contextual — la misma matemática de espacios de Hilbert que usa la mecánica cuántica, aplicada a representación semántica.
+
+La ontología operacional (core.ncl, invariantes, tensiones) es, en este contexto, una *base* en ese espacio — un conjunto de vectores que fijan el marco de referencia en el que el agente opera. Inyectar el contexto ontológico al inicio de una sesión es fijar ese marco de referencia antes de que el agente empiece a computar.
+
+La física cuántica importa aquí no porque vayamos a correr ontologías en computadores cuánticos — sino porque demuestra que las asunciones ontológicas clásicas son físicamente incorrectas en el nivel fundamental. La ontología operacional es contextual, relacional y perspectival *por diseño correcto*, no por limitación.
+
+---
+
+## 7. Por qué es relevante en software e infraestructuras digitales
+
+### El problema real no es técnico — es epistémico
+
+```
+Lo que el sistema hace           ≠  Lo que el equipo cree que hace
+Lo que la infra tiene desplegado ≠  Lo que Terraform declara
+Lo que el ADR decidió            ≠  Lo que el código implementa
+Lo que el agente asume           ≠  Lo que el contexto real requiere
+```
+
+Cada una de esas brechas es silenciosa, acumulativa, y solo se descubre cuando falla en producción. No son fallos de las personas — son consecuencias inevitables de operar un sistema complejo sin un modelo del propio sistema.
+
+### Lo que la ontología operacional hace que nada más hace
+
+| Documentación tradicional | Ontología operacional |
+|---|---|
+| Prosa — no consultable | Grafo tipado — consultable con queries |
+| Estática — se desactualiza sin señal | Verificada en cada commit mediante typecheck |
+| Describe lo que hubo | Formaliza lo que ES y lo que debe ser |
+| Solo legible por humanos | Legible por humanos, agentes de IA y CI |
+| Captura el qué | Captura el qué, el por qué, y lo que no puede cambiar |
+| Nadie la actualiza bajo presión | El proceso la actualiza — es el path, no documentación paralela |
+
+**Formaliza lo que no puede romperse.** Los invariantes no son comentarios — son constraints tipadas que el sistema verifica antes del runtime.
+
+**Captura las tensiones activas.** "Coste vs. disponibilidad", "velocidad vs. estabilidad" viven en la ontología, no en la cabeza del miembro más veterano. Cuando ese alguien se va, los trade-offs permanecen.
+
+**Da al agente de IA el contexto que no puede inferir del código.** Un agente puede leer todo el código de un sistema y no saber que hay un acuerdo que impide escalar horizontalmente cierto servicio. La ontología operacional hace eso explícito y machine-readable.
+
+### Por qué los DAGs — especificidad técnica
+
+**Análisis de impacto decidible.** Antes de cambiar un invariante o un nodo de infra — recorre el DAG y sabes exactamente qué se ve afectado. Terraform hace exactamente esto: `terraform plan` recorre el DAG de recursos para determinar el orden de aplicación.
+
+**Orden topológico para ejecución fiable.** Cualquier operación con dependencias tiene un orden correcto sin necesidad de un coordinador central. El principio detrás de Make (1976), Bazel, Gradle.
+
+**Justificación sin circularidad.** "Este invariante existe porque esta práctica lo requiere, y esta práctica existe porque este invariante lo define" es dogma circular, no arquitectura. El DAG hace esa circularidad visible e inaceptable.
+
+**Rollback determinista.** El inverso topológico del DAG es siempre computable. Sin DAG, el rollback depende de memoria humana.
+
+### Por qué la reflection — la ontología sin operación es un museo
+
+**Modos como DAGs operacionales.** Un procedimiento sin DAG puede ejecutarse en cualquier orden, saltarse, o ignorarse. Un modo como DAG tiene dependencias explícitas — el sistema verifica que cada paso se ejecutó antes de avanzar.
+
+**ADRs con constraints tipadas.** Una decisión sin constraints es prosa. Con constraints tipadas es un contrato que el CI puede verificar en cada commit.
+
+**Detección de drift.** La diferencia entre el estado sellado y el estado actual es medible si y solo si formalizaste el estado sellado. Sin eso, "¿ha derivado esto?" no tiene respuesta objetiva.
+
+### Las tres presiones convergentes
+
+**Presión 1: Los sistemas superan la comprensión individual.** Un monolito de 2005 podía vivir en la cabeza de dos personas. Un sistema cloud-native con decenas de microservicios, múltiples regiones y SLOs por servicio no cabe en ninguna cabeza.
+
+**Presión 2: Los agentes de IA tienen acceso de escritura.** Hoy los agentes pueden abrir PRs, modificar configuración, ejecutar Terraform, y desplegar. Un agente sin contexto ontológico no es un asistente — es un actor autónomo con comprensión incompleta del sistema que modifica.
+
+**Presión 3: Los equipos son multi-actor y distribuidos.** Humanos en zonas horarias distintas, agentes de IA con distintos contextos de sesión, pipelines de CI en paralelo — todos escribiendo sobre el mismo sistema sin coordinador central.
+
+### La síntesis
+
+> Un sistema que sabe lo que es, puede verificar que sigue siendo lo que dice ser, y puede comunicarlo a cualquier actor — humano, agente o CI — de forma estructurada y sin ambigüedad.
+
+Eso no es documentación. Es un modelo operacional vivo.
+
+La ontología operacional es la respuesta a: *¿qué estructura mínima necesita un sistema para conocerse a sí mismo de forma que ese conocimiento pueda ser usado, verificado y compartido?*
+
+---
+
+## 8. Referencias y lecturas recomendadas
+
+### Filosofía clásica y ontología
+
+**Aristóteles.** *Metafísica*. Traducción de Tomás Calvo Martínez. Madrid: Gredos, 1994. [~330 aC]
+> El libro fundamental para la distinción esencia/accidente, potencia/acto y las categorías. El libro IX sobre potencia es la base del modelo estado-actual/estado-deseado.
+
+**Aristóteles.** *Analíticos Posteriores*. Traducción de Miguel Candel Sanmartín. Madrid: Gredos, 1988. [~350 aC]
+> Formaliza el requisito de definiciones bien fundadas — antecedente directo de la aciclicidad en DAGs.
+
+**Laozi.** *Tao Te Ching*. Traducción de Stephen Mitchell. Nueva York: Harper Perennial, 1992. [~siglo VI-IV aC]
+> Texto fundacional del Taoísmo. Los primeros y últimos capítulos son los más relevantes para la relación con ontología operacional.
+
+**Zhuangzi.** *Zhuangzi: The Complete Writings*. Traducción de Brook Ziporyn. Indianapolis: Hackett, 2020. [~siglo IV aC]
+> Especialmente los capítulos internos (1-7). La relatividad perspectival como crítica de toda ontología clasificatoria.
+
+**Frege, G.** *Begriffsschrift*. Halle: Nebert, 1879.
+> Primera formulación de la lógica de predicados. El antecedente formal de toda representación de conocimiento estructurado.
+
+**Russell, B. & Whitehead, A.N.** *Principia Mathematica*. Cambridge: Cambridge University Press, 1910-1913.
+> La teoría de tipos como DAG para evitar la autorreferencia circular. El principio del círculo vicioso como fundamento de la aciclicidad.
+
+**Wittgenstein, L.** *Tractatus Logico-Philosophicus*. Londres: Kegan Paul, 1922.
+> "Los límites de mi lenguaje son los límites de mi mundo." El schema Nickel define la frontera de lo representable.
+
+**Wittgenstein, L.** *Investigaciones filosóficas*. Oxford: Blackwell, 1953.
+> Juegos de lenguaje como práctica. Los modos operacionales como juegos de lenguaje de un proyecto.
+
+**Hegel, G.W.F.** *Fenomenología del Espíritu*. Traducción de Manuel Jiménez Redondo. Valencia: Pre-Textos, 2006. [1807]
+> Dialéctica tesis-antítesis-síntesis. Contrástese con el yin-yang: síntesis vs. tensión gestionada permanentemente.
+
+**Husserl, E.** *Ideas relativas a una fenomenología pura*. México: FCE, 1993. [1913]
+> La distinción ontología formal/regional — estructura vacía universal vs. conceptos de dominio específico.
+
+**Whitehead, A.N.** *Process and Reality*. Nueva York: Free Press, 1978. [1929]
+> La realidad como proceso, no sustancia. El fundamento filosófico de la capa reflection como devenir distinto del ser ontológico.
+
+**Kripke, S.** *Naming and Necessity*. Cambridge: Harvard University Press, 1980.
+> Propiedades esenciales rígidas y semántica de mundos posibles. El fundamento modal de los invariantes operacionales.
+
+**Searle, J.** *The Construction of Social Reality*. Nueva York: Free Press, 1995.
+> Hechos institucionales como fundamento de las decisiones declaradas. El ADR como acto de habla colectivo que crea realidad operacional.
+
+**Parfit, D.** *Reasons and Persons*. Oxford: Oxford University Press, 1984.
+> Identidad a través del tiempo. La pregunta "¿es el mismo proyecto?" cuando sus invariantes cambian vía ADR.
+
+**Ladyman, J. & Ross, D.** *Every Thing Must Go: Metaphysics Naturalized*. Oxford: Oxford University Press, 2007.
+> Realismo estructural: las relaciones son más reales que las entidades. El DAG como estructura fundamental, los nodos como secundarios.
+
+**Quine, W.V.O.** "On What There Is." *Review of Metaphysics*, 2(5), 1948.
+> El concepto de compromiso ontológico: toda teoría asume la existencia de ciertas entidades. Todo schema asume una ontología aunque no lo declare.
+
+**Tarski, A.** "The Concept of Truth in Formalized Languages." En *Logic, Semantics, Metamathematics*. Oxford: Clarendon Press, 1956. [1933]
+> Verdad como correspondencia. El fundamento de la validación formal — nickel typecheck como verificación tarsquiana.
+
+**Zhou Dunyi.** *Taijitu shuo* (Explicación del Diagrama del Supremo Último). [1017-1073]
+> La cosmología taoísta sistematizada: wu ji → tai ji → yin-yang. La transición de lo indiferenciado a la polaridad como modelo de emergencia estructural.
+
+**Heidegger, M.** *Ser y Tiempo*. Traducción de Jorge Eduardo Rivera. Santiago: Editorial Universitaria, 1997. [1927]
+> El trasfondo pre-temático que ninguna ontología captura completamente. La ausencia notable en todos los modelos operacionales de software.
+
+**Heráclito.** *Fragmentos*. Edición de Rodolfo Mondolfo. Buenos Aires: Losada, 2007. [~500 aC]
+> El fragmento DK22B60 — "el camino hacia arriba y hacia abajo son el mismo" — es el equivalente presocrático más cercano al yin-yang.
+
+**Bateson, G.** *Steps to an Ecology of Mind*. Nueva York: Ballantine, 1972.
+> La diferencia que hace una diferencia. Epistemología sistémica que conecta Taoísmo, cibernética y representación del conocimiento.
+
+**Winograd, T. & Flores, F.** *Understanding Computers and Cognition*. Norwood: Ablex, 1986.
+> La crítica más influyente a la IA simbólica desde Heidegger. El trasfondo de comprensión no puede formalizarse — límite estructural de toda ontología operacional.
+
+### Inteligencia artificial y representación del conocimiento
+
+**Minsky, M.** "A Framework for Representing Knowledge." MIT AI Lab Memo 306, 1974.
+> Los frames como primera estructura computacional análoga a la sustancia aristotélica con propiedades heredables.
+
+**Collins, A.M. & Quillian, M.R.** "Retrieval time from semantic memory." *Journal of Verbal Learning and Verbal Behavior*, 8(2), 240-247, 1969.
+> Primera red semántica computacional. La topología del grafo predice el comportamiento cognitivo.
+
+**Brachman, R.J. & Schmolze, J.G.** "An overview of the KL-ONE knowledge representation system." *Cognitive Science*, 9(2), 171-216, 1985.
+> Description Logic: la subsunción como relación formal que genera el DAG de conceptos. Antecedente teórico de OWL.
+
+**Gruber, T.** "A translation approach to portable ontology specifications." *Knowledge Acquisition*, 5(2), 199-220, 1993.
+> La definición canónica: ontología como especificación de una conceptualización. Referencia obligatoria para cualquier trabajo en ontología de software.
+
+**Lenat, D.B. & Guha, R.V.** *Building Large Knowledge-Based Systems: Representation and Inference in the Cyc Project*. Reading: Addison-Wesley, 1990.
+> El proyecto más ambicioso de ontología operacional. Sus fracasos son tan informativos como sus logros.
+
+**Pearl, J.** *Probabilistic Reasoning in Intelligent Systems*. San Mateo: Morgan Kaufmann, 1988.
+> Bayesian networks como DAGs causales. La formalización de por qué la causalidad es inherentemente acíclica.
+
+**Pearl, J.** *Causality: Models, Reasoning, and Inference*. Cambridge: Cambridge University Press, 2000.
+> El do-calculus. Distinción entre correlación y causa como operaciones sobre el DAG causal.
+
+**Guarino, N. (ed.)** *Formal Ontology in Information Systems*. Amsterdam: IOS Press, 1998.
+> Actas de FOIS'98. La conferencia fundacional del campo. El artículo introductorio de Guarino sobre ontología formal vs. de dominio es referencia esencial.
+
+**Berners-Lee, T., Hendler, J. & Lassila, O.** "The Semantic Web." *Scientific American*, 284(5), 34-43, 2001.
+> El artículo que popularizó las ontologías en ingeniería de software. La visión de la web como grafo de conocimiento estructurado.
+
+### Física cuántica y fundamentos
+
+**Bohr, N.** "Can Quantum-Mechanical Description of Physical Reality Be Considered Complete?" *Physical Review*, 48, 696-702, 1935.
+> La respuesta de Bohr al artículo EPR. Formula el principio de complementariedad como posición filosófica irrenunciable.
+
+**Bohr, N.** "Discussion with Einstein on Epistemological Problems in Atomic Physics." En Schilpp (ed.), *Albert Einstein: Philosopher-Scientist*. Evanston: Library of Living Philosophers, 1949.
+> El debate más importante de la física del siglo XX sobre interpretación cuántica.
+
+**Heisenberg, W.** *Physics and Philosophy: The Revolution in Modern Science*. Nueva York: Harper & Row, 1958.
+> El capítulo sobre el papel del lenguaje en física cuántica conecta directamente con Wittgenstein.
+
+**Schrödinger, E.** "Die gegenwärtige Situation in der Quantenmechanik." *Naturwissenschaften*, 23(48-49-50), 1935.
+> El artículo del gato. Introduce el entrelazamiento (*Verschränkung*) y demuestra que la superposición no puede interpretarse clásicamente.
+
+**Kochen, S. & Specker, E.P.** "The Problem of Hidden Variables in Quantum Mechanics." *Journal of Mathematics and Mechanics*, 17(1), 59-87, 1967.
+> Prueba formal de que es imposible asignar valores definidos a todos los observables cuánticos simultáneamente. La contextualidad como hecho matemático.
+
+**Bell, J.S.** *Speakable and Unspeakable in Quantum Mechanics*. Cambridge: Cambridge University Press, 1987.
+> El capítulo "Against measurement" es una crítica filosófica fundamental al vocabulario de la física cuántica.
+
+**Aspect, A., Grangier, P. & Roger, G.** "Experimental Realization of Einstein-Podolsky-Rosen-Bohm Gedankenexperiment." *Physical Review Letters*, 49(2), 91-94, 1982.
+> La confirmación experimental de que las desigualdades de Bell se violan. La naturaleza es no-local.
+
+**Everett, H.** "Relative State Formulation of Quantum Mechanics." *Reviews of Modern Physics*, 29(3), 454-462, 1957.
+> La interpretación de muchos mundos. Ontología donde todos los estados posibles son reales — el antípoda de la ontología operacional que colapsa posibilidades en decisiones.
+
+**Rovelli, C.** "Relational Quantum Mechanics." *International Journal of Theoretical Physics*, 35(8), 1637-1678, 1996.
+> No existen estados absolutos — solo estados relativos a otros sistemas. La ontología más radical de la lista.
+
+**Rovelli, C.** *Helgoland*. Milano: Adelphi, 2020.
+> La versión divulgativa de la mecánica cuántica relacional. El capítulo sobre relaciones y el budismo es relevante para la conexión con el Taoísmo.
+
+**Nagel, T.** *The View from Nowhere*. Oxford: Oxford University Press, 1986.
+> El argumento de que toda objetividad requiere un punto de vista. En tensión directa con el problema del observador cuántico.
+
+**Nielsen, M.A. & Chuang, I.L.** *Quantum Computation and Quantum Information*. Cambridge: Cambridge University Press, 2000.
+> El texto estándar de computación cuántica. Los capítulos 2-3 sobre espacios de Hilbert son la base para entender la conexión con LLMs.
+
+**Deutsch, D.** *The Fabric of Reality*. Londres: Allen Lane, 1997.
+> La defensa más ambiciosa de la interpretación de muchos mundos. El capítulo sobre epistemología cuántica.
+
+### Ingeniería de software e infraestructura
+
+**Nygard, M.** "Documenting Architecture Decisions." Cognitect Blog, 2011.
+> La propuesta original de ADRs. El precursor directo de los ADRs con constraints tipadas.
+
+**Evans, E.** *Domain-Driven Design: Tackling Complexity in the Heart of Software*. Boston: Addison-Wesley, 2003.
+> El lenguaje ubicuo como ontología operacional informal. El bounded context como ontología regional de Husserl aplicada a dominios de negocio.
+
+**Kleppmann, M.** *Designing Data-Intensive Applications*. Sebastopol: O'Reilly, 2017.
+> El capítulo sobre modelos de datos: la elección de modelo (relacional, documental, grafo) es una decisión ontológica.
+
+**Morris, K.** *Infrastructure as Code*. Sebastopol: O'Reilly, 2016.
+> Formaliza la declaratividad en infraestructura. La distinción entre estado deseado y estado actual — el núcleo del modelo `state.ncl`.
+
+**Humble, J. & Farley, D.** *Continuous Delivery*. Boston: Addison-Wesley, 2010.
+> El pipeline de despliegue como DAG de verificaciones. Cada etapa es un gate que el sistema debe cruzar formalmente.
+
+**Kim, G., Humble, J., Debois, P. & Willis, J.** *The DevOps Handbook*. Portland: IT Revolution Press, 2016.
+> El flujo de trabajo DevOps como protocolo operacional. Los feedback loops como ciclo ontología/reflection.
+
+**Dehghani, Z.** "How to Move Beyond a Monolithic Data Lake to a Distributed Data Mesh." Martin Fowler Blog, 2019.
+> Data mesh como ontología operacional distribuida: cada dominio es responsable de su propia ontología de datos.
+
+**Fowler, M.** "Who Needs an Architect?" *IEEE Software*, 20(5), 11-13, 2003.
+> Define arquitectura como "las decisiones que son difíciles de cambiar" — equivalente exacto de los invariantes en ontología operacional.
+
+**Richardson, C.** *Microservices Patterns*. Shelter Island: Manning, 2018.
+> El patrón Saga como DAG de transacciones distribuidas. Cada paso tiene `depends_on` implícito — el DAG operacional aplicado a consistencia eventual.
+
+**Brown, S.** *Software Architecture for Developers*. Leanpub, 2014.
+> El modelo C4 como alternativa a UML. Múltiples niveles de abstracción para describir un sistema.
+
+**Hohpe, G. & Woolf, B.** *Enterprise Integration Patterns*. Boston: Addison-Wesley, 2003.
+> Patrones como vocabulario compartido — nombrar los patrones operacionales es un acto ontológico.
diff --git a/assets/presentation/docs/ontologia-filosofia-software.pdf b/assets/presentation/docs/ontologia-filosofia-software.pdf
new file mode 100644
index 0000000..cccafa2
--- /dev/null
+++ b/assets/presentation/docs/ontologia-filosofia-software.pdf
@@ -0,0 +1,19472 @@
+%PDF-1.7
+%����
+
+1 0 obj
+<<
+  /Type /Pages
+  /Count 28
+  /Kids [1385 0 R 1409 0 R 1411 0 R 1413 0 R 1415 0 R 1417 0 R 1419 0 R 1421 0 R 1423 0 R 1425 0 R 1427 0 R 1429 0 R 1431 0 R 1433 0 R 1435 0 R 1437 0 R 1439 0 R 1441 0 R 1443 0 R 1445 0 R 1447 0 R 1449 0 R 1451 0 R 1453 0 R 1455 0 R 1457 0 R 1459 0 R 1461 0 R]
+>>
+endobj
+
+2 0 obj
+<<
+  /Type /Outlines
+  /First 3 0 R
+  /Last 55 0 R
+  /Count 8
+>>
+endobj
+
+3 0 obj
+<<
+  /Parent 2 0 R
+  /Next 8 0 R
+  /First 4 0 R
+  /Last 7 0 R
+  /Count -4
+  /Title <FEFF0031002E002000BF0053006F006E0020006C006F007300200069006E00760061007200690061006E00740065007300200065006C0020006500710075006900760061006C0065006E00740065002000610020006C006F007300200075006E006900760065007200730061006C00650073002000660069006C006F007300F3006600690063006F0073003F>
+  /Dest 1307 0 R
+>>
+endobj
+
+4 0 obj
+<<
+  /Parent 3 0 R
+  /Next 5 0 R
+  /Title <FEFF0031002E0031002E0020004C006F007300200075006E006900760065007200730061006C0065007300200065006E0020006F006E0074006F006C006F006700ED0061002000660069006C006F007300F30066006900630061>
+  /Dest 1303 0 R
+>>
+endobj
+
+5 0 obj
+<<
+  /Parent 3 0 R
+  /Next 6 0 R
+  /Prev 4 0 R
+  /Title (1.2. Los invariantes como propiedades esenciales de un singular)
+  /Dest 1304 0 R
+>>
+endobj
+
+6 0 obj
+<<
+  /Parent 3 0 R
+  /Next 7 0 R
+  /Prev 5 0 R
+  /Title <FEFF0031002E0033002E0020004400F3006E006400650020007300ED00200068006100790020007200650073006F006E0061006E00630069006100200063006F006E0020006C006F007300200075006E006900760065007200730061006C00650073>
+  /Dest 1305 0 R
+>>
+endobj
+
+7 0 obj
+<<
+  /Parent 3 0 R
+  /Prev 6 0 R
+  /Title <FEFF0031002E0034002E00200045006C002000650073007400610074007500730020006F006E0074006F006C00F3006700690063006F003A00200068006500630068006F007300200069006E0073007400690074007500630069006F006E0061006C00650073>
+  /Dest 1306 0 R
+>>
+endobj
+
+8 0 obj
+<<
+  /Parent 2 0 R
+  /Next 10 0 R
+  /Prev 3 0 R
+  /First 9 0 R
+  /Last 9 0 R
+  /Count -1
+  /Title <FEFF0032002E002000470065006E00650061006C006F006700ED0061002000660069006C006F007300F30066006900630061002000640065006C0020006D006F00640065006C006F0020006F007000650072006100630069006F006E0061006C>
+  /Dest 1309 0 R
+>>
+endobj
+
+9 0 obj
+<<
+  /Parent 8 0 R
+  /Title (2.1. Matices que la tabla no captura)
+  /Dest 1308 0 R
+>>
+endobj
+
+10 0 obj
+<<
+  /Parent 2 0 R
+  /Next 27 0 R
+  /Prev 8 0 R
+  /First 11 0 R
+  /Last 26 0 R
+  /Count -5
+  /Title <FEFF0033002E0020004C00610020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C00200065006E00200073006F006600740077006100720065003A0020006F0072006900670065006E002C0020006100750074006F00720065007300200079002000750073006F002000610063007400750061006C>
+  /Dest 1326 0 R
+>>
+endobj
+
+11 0 obj
+<<
+  /Parent 10 0 R
+  /Next 12 0 R
+  /Title <FEFF0033002E0031002E00200045006C002000700072006F0062006C0065006D00610020007100750065002000680061006200ED006100200071007500650020007200650073006F006C007600650072>
+  /Dest 1310 0 R
+>>
+endobj
+
+12 0 obj
+<<
+  /Parent 10 0 R
+  /Next 19 0 R
+  /Prev 11 0 R
+  /First 13 0 R
+  /Last 18 0 R
+  /Count -6
+  /Title (3.2. Los fundadores)
+  /Dest 1317 0 R
+>>
+endobj
+
+13 0 obj
+<<
+  /Parent 12 0 R
+  /Next 14 0 R
+  /Title <FEFF0033002E0032002E0031002E0020004A006F0068006E0020004D00630043006100720074006800790020201400200065006C00200070007200650063007500720073006F00720020002800310039003500382013003100390038003000730029>
+  /Dest 1311 0 R
+>>
+endobj
+
+14 0 obj
+<<
+  /Parent 12 0 R
+  /Next 15 0 R
+  /Prev 13 0 R
+  /Title <FEFF0033002E0032002E0032002E0020004D0061007200760069006E0020004D0069006E0073006B0079002020140020006600720061006D006500730020007900200065007300740072007500630074007500720061007300200064006500200063006F006E006F00630069006D00690065006E0074006F0020002800310039003700340029>
+  /Dest 1312 0 R
+>>
+endobj
+
+15 0 obj
+<<
+  /Parent 12 0 R
+  /Next 16 0 R
+  /Prev 14 0 R
+  /Title <FEFF0033002E0032002E0033002E00200054006F006D0020004700720075006200650072002020140020006C006100200064006500660069006E00690063006900F3006E002000630061006E00F3006E0069006300610020002800310039003900330029>
+  /Dest 1313 0 R
+>>
+endobj
+
+16 0 obj
+<<
+  /Parent 12 0 R
+  /Next 17 0 R
+  /Prev 15 0 R
+  /Title <FEFF0033002E0032002E0034002E0020004E00690063006F006C0061002000470075006100720069006E006F002020140020006F006E0074006F006C006F006700ED006100200066006F0072006D0061006C00200065006E002000730069007300740065006D0061007300200064006500200069006E0066006F0072006D00610063006900F3006E00200028003100390039003000730029>
+  /Dest 1314 0 R
+>>
+endobj
+
+17 0 obj
+<<
+  /Parent 12 0 R
+  /Next 18 0 R
+  /Prev 16 0 R
+  /Title <FEFF0033002E0032002E0035002E00200044006F007500670020004C0065006E00610074002020140020004300790063003A00200065006C002000700072006F0079006500630074006F0020006D00E1007300200061006D0062006900630069006F0073006F002000280031003900380034201300320030003100370029>
+  /Dest 1315 0 R
+>>
+endobj
+
+18 0 obj
+<<
+  /Parent 12 0 R
+  /Prev 17 0 R
+  /Title <FEFF0033002E0032002E0036002E002000540069006D0020004200650072006E006500720073002D004C00650065002C0020004A0061006D00650073002000480065006E0064006C00650072002000790020004F007200610020004C0061007300730069006C0061002020140020006C00610020005700650062002000530065006D00E1006E00740069006300610020002800320030003000310029>
+  /Dest 1316 0 R
+>>
+endobj
+
+19 0 obj
+<<
+  /Parent 10 0 R
+  /Next 20 0 R
+  /Prev 12 0 R
+  /Title <FEFF0033002E0033002E0020004C00ED006E00650061002000740065006D0070006F00720061006C00200063006F006E00640065006E0073006100640061>
+  /Dest 1318 0 R
+>>
+endobj
+
+20 0 obj
+<<
+  /Parent 10 0 R
+  /Next 26 0 R
+  /Prev 19 0 R
+  /First 21 0 R
+  /Last 25 0 R
+  /Count -5
+  /Title (3.4. Uso actual)
+  /Dest 1324 0 R
+>>
+endobj
+
+21 0 obj
+<<
+  /Parent 20 0 R
+  /Next 22 0 R
+  /Title <FEFF0033002E0034002E0031002E002000420069006F006D00650064006900630069006E00610020201400200065006C00200064006F006D0069006E0069006F0020006D00E100730020006D0061006400750072006F>
+  /Dest 1319 0 R
+>>
+endobj
+
+22 0 obj
+<<
+  /Parent 20 0 R
+  /Next 23 0 R
+  /Prev 21 0 R
+  /Title (3.4.2. Grafos de conocimiento empresariales)
+  /Dest 1320 0 R
+>>
+endobj
+
+23 0 obj
+<<
+  /Parent 20 0 R
+  /Next 24 0 R
+  /Prev 22 0 R
+  /Title (3.4.3. Industria y manufactura)
+  /Dest 1321 0 R
+>>
+endobj
+
+24 0 obj
+<<
+  /Parent 20 0 R
+  /Next 25 0 R
+  /Prev 23 0 R
+  /Title (3.4.4. Inteligencia artificial y agentes)
+  /Dest 1322 0 R
+>>
+endobj
+
+25 0 obj
+<<
+  /Parent 20 0 R
+  /Prev 24 0 R
+  /Title <FEFF0033002E0034002E0035002E00200053006F006600740077006100720065002000640065002000700072006F0079006500630074006F002020140020006C0061002000660072006F006E0074006500720061002000610063007400750061006C>
+  /Dest 1323 0 R
+>>
+endobj
+
+26 0 obj
+<<
+  /Parent 10 0 R
+  /Prev 20 0 R
+  /Title <FEFF0033002E0035002E00200043006F006E0063006C00750073006900F3006E>
+  /Dest 1325 0 R
+>>
+endobj
+
+27 0 obj
+<<
+  /Parent 2 0 R
+  /Next 33 0 R
+  /Prev 10 0 R
+  /First 28 0 R
+  /Last 32 0 R
+  /Count -5
+  /Title <FEFF0034002E002000540061006F00ED0073006D006F002C002000590069006E002D00590061006E0067002000790020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C>
+  /Dest 1332 0 R
+>>
+endobj
+
+28 0 obj
+<<
+  /Parent 27 0 R
+  /Next 29 0 R
+  /Title <FEFF0034002E0031002E0020004C006F00200071007500650020006C00610020007400720061006400690063006900F3006E0020006F00630063006900640065006E00740061006C0020006E006F002000760069006F>
+  /Dest 1327 0 R
+>>
+endobj
+
+29 0 obj
+<<
+  /Parent 27 0 R
+  /Next 30 0 R
+  /Prev 28 0 R
+  /Title <FEFF0034002E0032002E002000590069006E002D00590061006E00670020201400200070006F006C00610072006900640061006400200063006F006E007300740069007400750074006900760061002C0020006E006F0020006F0070006F007300690063006900F3006E>
+  /Dest 1328 0 R
+>>
+endobj
+
+30 0 obj
+<<
+  /Parent 27 0 R
+  /Next 31 0 R
+  /Prev 29 0 R
+  /Title (4.3. Correspondencias con el modelo operacional)
+  /Dest 1329 0 R
+>>
+endobj
+
+31 0 obj
+<<
+  /Parent 27 0 R
+  /Next 32 0 R
+  /Prev 30 0 R
+  /Title (4.4. Las cuatro correspondencias profundas)
+  /Dest 1330 0 R
+>>
+endobj
+
+32 0 obj
+<<
+  /Parent 27 0 R
+  /Prev 31 0 R
+  /Title <FEFF0034002E0035002E0020004C006100200063007200ED0074006900630061002000740061006F00ED007300740061002000610020006C00610020006F006E0074006F006C006F006700ED006100200063006C00610073006900660069006300610074006F007200690061>
+  /Dest 1331 0 R
+>>
+endobj
+
+33 0 obj
+<<
+  /Parent 2 0 R
+  /Next 39 0 R
+  /Prev 27 0 R
+  /First 34 0 R
+  /Last 38 0 R
+  /Count -5
+  /Title <FEFF0035002E0020004400410047007300202014002000670065006E00650061006C006F006700ED006100200079002000660075006E0063006900F3006E00200065006E0020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C>
+  /Dest 1338 0 R
+>>
+endobj
+
+34 0 obj
+<<
+  /Parent 33 0 R
+  /Next 35 0 R
+  /Title (5.1. El problema que fuerza el grafo)
+  /Dest 1333 0 R
+>>
+endobj
+
+35 0 obj
+<<
+  /Parent 33 0 R
+  /Next 36 0 R
+  /Prev 34 0 R
+  /Title <FEFF0035002E0032002E00200050006F007200200071007500E90020006C006100200061006300690063006C0069006300690064006100640020201400200065006C002000700072006F0062006C0065006D00610020006400650020006C00610020006A007500730074006900660069006300610063006900F3006E002000630069007200630075006C00610072>
+  /Dest 1334 0 R
+>>
+endobj
+
+36 0 obj
+<<
+  /Parent 33 0 R
+  /Next 37 0 R
+  /Prev 35 0 R
+  /Title <FEFF0035002E0033002E0020004C00ED006E00650061002000670065006E00650061006C00F30067006900630061>
+  /Dest 1335 0 R
+>>
+endobj
+
+37 0 obj
+<<
+  /Parent 33 0 R
+  /Next 38 0 R
+  /Prev 36 0 R
+  /Title <FEFF0035002E0034002E0020004C00610073002000630075006100740072006F002000660075006E00630069006F006E00650073002000640065006C002000440041004700200065006E0020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C>
+  /Dest 1336 0 R
+>>
+endobj
+
+38 0 obj
+<<
+  /Parent 33 0 R
+  /Prev 37 0 R
+  /Title <FEFF0035002E0035002E0020004C0061002000740065006E0073006900F3006E00200063006F006E00200065006C002000540061006F00ED0073006D006F002020140020004400410047002000760073002E0020006300690063006C006F>
+  /Dest 1337 0 R
+>>
+endobj
+
+39 0 obj
+<<
+  /Parent 2 0 R
+  /Next 48 0 R
+  /Prev 33 0 R
+  /First 40 0 R
+  /Last 47 0 R
+  /Count -5
+  /Title <FEFF0036002E0020004600ED007300690063006100200063007500E1006E0074006900630061002C00200063006F006D00700075007400610063006900F3006E00200063007500E1006E0074006900630061002000790020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C>
+  /Dest 1347 0 R
+>>
+endobj
+
+40 0 obj
+<<
+  /Parent 39 0 R
+  /Next 41 0 R
+  /Title <FEFF0036002E0031002E00200045006C00200067006F006C007000650020006D00E10073002000700072006F00660075006E0064006F0020201400200041007200690073007400F300740065006C006500730020006500730074006100620061002000650071007500690076006F006300610064006F00200065006E00200065006C0020006E006900760065006C002000660075006E00640061006D0065006E00740061006C>
+  /Dest 1339 0 R
+>>
+endobj
+
+41 0 obj
+<<
+  /Parent 39 0 R
+  /Next 45 0 R
+  /Prev 40 0 R
+  /First 42 0 R
+  /Last 44 0 R
+  /Count -3
+  /Title (6.2. Las tres conexiones genuinas)
+  /Dest 1343 0 R
+>>
+endobj
+
+42 0 obj
+<<
+  /Parent 41 0 R
+  /Next 43 0 R
+  /Title <FEFF0036002E0032002E0031002E00200043006F006D0070006C0065006D0065006E0074006100720069006500640061006400200064006500200042006F0068007200202194002000590069006E002D00590061006E0067002000790020006C00610020006400750061006C00690064006100640020004F006E0074006F006C006F006700ED0061002F005200650066006C0065006300740069006F006E>
+  /Dest 1340 0 R
+>>
+endobj
+
+43 0 obj
+<<
+  /Parent 41 0 R
+  /Next 44 0 R
+  /Prev 42 0 R
+  /Title <FEFF0036002E0032002E0032002E00200043006F006E007400650078007400750061006C006900640061006400200063007500E1006E0074006900630061002021940020004F006E0074006F006C006F006700ED0061002000700065007200730070006500630074006900760061006C>
+  /Dest 1341 0 R
+>>
+endobj
+
+44 0 obj
+<<
+  /Parent 41 0 R
+  /Prev 43 0 R
+  /Title <FEFF0036002E0032002E0033002E00200045006C002000700072006F0062006C0065006D0061002000640065006C0020006F0062007300650072007600610064006F00720020219400200045006C002000740072006100730066006F006E0064006F00200071007500650020006E006F00200070007500650064006500200066006F0072006D0061006C0069007A0061007200730065>
+  /Dest 1342 0 R
+>>
+endobj
+
+45 0 obj
+<<
+  /Parent 39 0 R
+  /Next 46 0 R
+  /Prev 41 0 R
+  /Title <FEFF0036002E0033002E0020004C006100200069006E007400650072007000720065007400610063006900F3006E0020006D00E10073002000720065006C006500760061006E00740065002020140020004D0065006300E1006E00690063006100200043007500E1006E0074006900630061002000520065006C006100630069006F006E0061006C>
+  /Dest 1344 0 R
+>>
+endobj
+
+46 0 obj
+<<
+  /Parent 39 0 R
+  /Next 47 0 R
+  /Prev 45 0 R
+  /Title <FEFF0036002E0034002E0020004C006100200063006F006D00700075007400610063006900F3006E00200063007500E1006E00740069006300610020201400200068006F006E006500730074006900640061006400200073006F00620072006500200073007500730020006C00ED006D0069007400650073>
+  /Dest 1345 0 R
+>>
+endobj
+
+47 0 obj
+<<
+  /Parent 39 0 R
+  /Prev 46 0 R
+  /Title <FEFF0036002E0035002E0020004C006100200063006F006E00650078006900F3006E0020006E006F0020006F0062007600690061002020140020004C004C004D007300200063006F006D006F002000670065006F006D00650074007200ED006100200063007500E1006E0074006900630061002000610070006C00690063006100640061>
+  /Dest 1346 0 R
+>>
+endobj
+
+48 0 obj
+<<
+  /Parent 2 0 R
+  /Next 55 0 R
+  /Prev 39 0 R
+  /First 49 0 R
+  /Last 54 0 R
+  /Count -6
+  /Title <FEFF0037002E00200050006F007200200071007500E9002000650073002000720065006C006500760061006E0074006500200065006E00200073006F0066007400770061007200650020006500200069006E006600720061006500730074007200750063007400750072006100730020006400690067006900740061006C00650073>
+  /Dest 1354 0 R
+>>
+endobj
+
+49 0 obj
+<<
+  /Parent 48 0 R
+  /Next 50 0 R
+  /Title <FEFF0037002E0031002E00200045006C002000700072006F0062006C0065006D00610020007200650061006C0020006E006F0020006500730020007400E90063006E00690063006F0020201400200065007300200065007000690073007400E9006D00690063006F>
+  /Dest 1348 0 R
+>>
+endobj
+
+50 0 obj
+<<
+  /Parent 48 0 R
+  /Next 51 0 R
+  /Prev 49 0 R
+  /Title <FEFF0037002E0032002E0020004C006F00200071007500650020006C00610020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C0020006800610063006500200071007500650020006E0061006400610020006D00E1007300200068006100630065>
+  /Dest 1349 0 R
+>>
+endobj
+
+51 0 obj
+<<
+  /Parent 48 0 R
+  /Next 52 0 R
+  /Prev 50 0 R
+  /Title <FEFF0037002E0033002E00200050006F007200200071007500E90020006C006F00730020004400410047007300202014002000650073007000650063006900660069006300690064006100640020007400E90063006E006900630061>
+  /Dest 1350 0 R
+>>
+endobj
+
+52 0 obj
+<<
+  /Parent 48 0 R
+  /Next 53 0 R
+  /Prev 51 0 R
+  /Title <FEFF0037002E0034002E00200050006F007200200071007500E90020006C00610020007200650066006C0065006300740069006F006E002020140020006C00610020006F006E0074006F006C006F006700ED0061002000730069006E0020006F00700065007200610063006900F3006E00200065007300200075006E0020006D007500730065006F>
+  /Dest 1351 0 R
+>>
+endobj
+
+53 0 obj
+<<
+  /Parent 48 0 R
+  /Next 54 0 R
+  /Prev 52 0 R
+  /Title (7.5. Las tres presiones que hacen esto urgente ahora)
+  /Dest 1352 0 R
+>>
+endobj
+
+54 0 obj
+<<
+  /Parent 48 0 R
+  /Prev 53 0 R
+  /Title <FEFF0037002E0036002E0020004C00610020007300ED006E00740065007300690073>
+  /Dest 1353 0 R
+>>
+endobj
+
+55 0 obj
+<<
+  /Parent 2 0 R
+  /Prev 48 0 R
+  /First 56 0 R
+  /Last 61 0 R
+  /Count -6
+  /Title (8. Referencias y lecturas recomendadas)
+  /Dest 1361 0 R
+>>
+endobj
+
+56 0 obj
+<<
+  /Parent 55 0 R
+  /Next 57 0 R
+  /Title <FEFF0038002E0031002E002000460069006C006F0073006F006600ED006100200063006C00E10073006900630061002000790020006F006E0074006F006C006F006700ED0061>
+  /Dest 1355 0 R
+>>
+endobj
+
+57 0 obj
+<<
+  /Parent 55 0 R
+  /Next 58 0 R
+  /Prev 56 0 R
+  /Title <FEFF0038002E0032002E00200049006E00740065006C006900670065006E0063006900610020006100720074006900660069006300690061006C0020007900200072006500700072006500730065006E007400610063006900F3006E002000640065006C00200063006F006E006F00630069006D00690065006E0074006F>
+  /Dest 1356 0 R
+>>
+endobj
+
+58 0 obj
+<<
+  /Parent 55 0 R
+  /Next 59 0 R
+  /Prev 57 0 R
+  /Title <FEFF0038002E0033002E00200049006E00670065006E00690065007200ED006100200064006500200073006F006600740077006100720065002000790020007000720061007800690073>
+  /Dest 1357 0 R
+>>
+endobj
+
+59 0 obj
+<<
+  /Parent 55 0 R
+  /Next 60 0 R
+  /Prev 58 0 R
+  /Title (8.4. Perspectivas adicionales)
+  /Dest 1358 0 R
+>>
+endobj
+
+60 0 obj
+<<
+  /Parent 55 0 R
+  /Next 61 0 R
+  /Prev 59 0 R
+  /Title <FEFF0038002E0035002E0020004600ED007300690063006100200063007500E1006E007400690063006100200079002000660075006E00640061006D0065006E0074006F0073>
+  /Dest 1359 0 R
+>>
+endobj
+
+61 0 obj
+<<
+  /Parent 55 0 R
+  /Prev 60 0 R
+  /Title <FEFF0038002E0036002E00200049006E00670065006E00690065007200ED006100200064006500200073006F0066007400770061007200650020006500200069006E0066007200610065007300740072007500630074007500720061>
+  /Dest 1360 0 R
+>>
+endobj
+
+62 0 obj
+<<
+  /Nums [0 1217 0 R 1 1218 0 R 2 1219 0 R 3 1220 0 R 4 1221 0 R 5 1222 0 R 6 1223 0 R 7 1224 0 R 8 1225 0 R 9 1226 0 R 10 1227 0 R 11 1228 0 R 12 1229 0 R 13 1230 0 R 14 1231 0 R 15 1232 0 R 16 1233 0 R 17 1234 0 R 18 1235 0 R 19 1236 0 R 20 1237 0 R 21 1238 0 R 22 1239 0 R 23 1240 0 R 24 1241 0 R 25 1242 0 R 26 1243 0 R 27 1244 0 R]
+>>
+endobj
+
+63 0 obj
+<<
+  /Type /StructTreeRoot
+  /RoleMap <<
+    /Datetime /Span
+    /Terms /Part
+    /Title /P
+    /Strong /Span
+    /Em /Span
+  >>
+  /K [92 0 R]
+  /ParentTree <<
+    /Nums [0 1210 0 R 1 1206 0 R 2 1202 0 R 3 1198 0 R 4 1194 0 R 5 1189 0 R 6 1185 0 R 7 1180 0 R 8 1176 0 R 9 1172 0 R 10 1168 0 R 11 1164 0 R 12 1160 0 R 13 1155 0 R 14 1151 0 R 15 1147 0 R 16 1143 0 R 17 1139 0 R 18 1135 0 R 19 1130 0 R 20 1126 0 R 21 1122 0 R 22 1118 0 R 23 64 0 R 24 1114 0 R 25 1110 0 R 26 1105 0 R 27 1101 0 R 28 1097 0 R 29 1093 0 R 30 1089 0 R 31 1085 0 R 32 1080 0 R 33 1076 0 R 34 1072 0 R 35 1068 0 R 36 1064 0 R 37 1060 0 R 38 1056 0 R 39 1051 0 R 40 1047 0 R 41 1043 0 R 42 1039 0 R 43 1035 0 R 44 1031 0 R 45 1027 0 R 46 65 0 R 47 66 0 R 48 67 0 R 49 68 0 R 50 69 0 R 51 70 0 R 52 71 0 R 53 72 0 R 54 73 0 R 55 74 0 R 56 75 0 R 57 76 0 R 58 77 0 R 59 78 0 R 60 79 0 R 61 80 0 R 62 81 0 R 63 82 0 R 64 83 0 R 65 84 0 R 66 85 0 R 67 86 0 R 68 87 0 R 69 88 0 R 70 89 0 R 71 90 0 R 72 91 0 R]
+  >>
+  /IDTree <<
+    /Names [(U1x0y0) 978 0 R (U1x1y0) 977 0 R (U1x2y0) 976 0 R (U1x3y0) 975 0 R (U2x0y0) 767 0 R (U2x1y0) 766 0 R (U3x0y0) 594 0 R (U3x1y0) 593 0 R (U3x2y0) 592 0 R (U3x3y0) 591 0 R (U4x0y0) 511 0 R (U4x1y0) 510 0 R (U4x2y0) 509 0 R (U5x0y0) 371 0 R (U5x1y0) 370 0 R (U5x2y0) 369 0 R (U6x0y0) 328 0 R (U6x1y0) 327 0 R]
+  >>
+  /ParentTreeNextKey 73
+>>
+endobj
+
+64 0 obj
+[1216 0 R 1216 0 R 1215 0 R 1214 0 R 1213 0 R 1212 0 R 1211 0 R 1210 0 R 1210 0 R 1207 0 R 1206 0 R 1206 0 R 1203 0 R 1202 0 R 1202 0 R 1199 0 R 1198 0 R 1198 0 R 1195 0 R 1194 0 R 1194 0 R 1190 0 R 1189 0 R 1189 0 R 1186 0 R 1185 0 R 1185 0 R 1181 0 R 1180 0 R 1180 0 R 1177 0 R 1176 0 R 1176 0 R 1173 0 R 1172 0 R 1172 0 R 1169 0 R 1168 0 R 1168 0 R 1165 0 R 1164 0 R 1164 0 R 1161 0 R 1160 0 R 1160 0 R 1156 0 R 1155 0 R 1155 0 R 1152 0 R 1151 0 R 1151 0 R 1148 0 R 1147 0 R 1147 0 R 1144 0 R 1143 0 R 1143 0 R 1140 0 R 1139 0 R 1139 0 R 1136 0 R 1135 0 R 1135 0 R 1131 0 R 1130 0 R 1130 0 R 1127 0 R 1126 0 R 1126 0 R 1123 0 R 1122 0 R 1122 0 R 1119 0 R 1118 0 R 1118 0 R]
+endobj
+
+65 0 obj
+[1115 0 R 1114 0 R 1114 0 R 1111 0 R 1110 0 R 1110 0 R 1106 0 R 1105 0 R 1105 0 R 1102 0 R 1101 0 R 1101 0 R 1101 0 R 1098 0 R 1097 0 R 1097 0 R 1094 0 R 1093 0 R 1093 0 R 1090 0 R 1089 0 R 1089 0 R 1086 0 R 1085 0 R 1085 0 R 1081 0 R 1080 0 R 1080 0 R 1077 0 R 1076 0 R 1076 0 R 1073 0 R 1072 0 R 1072 0 R 1069 0 R 1068 0 R 1068 0 R 1065 0 R 1064 0 R 1064 0 R 1061 0 R 1060 0 R 1060 0 R 1057 0 R 1056 0 R 1056 0 R 1052 0 R 1051 0 R 1051 0 R 1048 0 R 1047 0 R 1047 0 R 1044 0 R 1043 0 R 1043 0 R 1040 0 R 1039 0 R 1039 0 R 1036 0 R 1035 0 R 1035 0 R 1032 0 R 1031 0 R 1031 0 R 1028 0 R 1027 0 R 1027 0 R]
+endobj
+
+66 0 obj
+[1022 0 R 1022 0 R 1022 0 R 1022 0 R 1021 0 R 1021 0 R 1020 0 R 1020 0 R 1017 0 R 1019 0 R 1017 0 R 1018 0 R 1018 0 R 1017 0 R 1017 0 R 1017 0 R 1017 0 R 1016 0 R 1015 0 R 1014 0 R 1014 0 R 1012 0 R 1011 0 R 1009 0 R 1010 0 R 1009 0 R 1009 0 R 1007 0 R 1006 0 R 1004 0 R 1004 0 R 1005 0 R 1004 0 R 1002 0 R 1001 0 R 1000 0 R 1000 0 R 997 0 R 997 0 R 995 0 R 995 0 R 996 0 R 996 0 R 995 0 R 993 0 R 994 0 R 994 0 R 993 0 R 993 0 R 993 0 R 989 0 R 992 0 R 989 0 R 989 0 R 991 0 R 989 0 R 989 0 R 989 0 R 989 0 R 989 0 R 989 0 R 989 0 R 990 0 R 989 0 R 989 0 R 988 0 R 988 0 R 986 0 R 987 0 R 986 0 R 986 0 R 986 0 R 986 0 R]
+endobj
+
+67 0 obj
+[985 0 R 985 0 R 982 0 R 982 0 R 982 0 R 982 0 R 984 0 R 982 0 R 983 0 R 983 0 R 982 0 R 982 0 R 982 0 R 982 0 R 981 0 R 981 0 R]
+endobj
+
+68 0 obj
+[980 0 R 980 0 R 979 0 R 979 0 R 978 0 R 977 0 R 976 0 R 976 0 R 976 0 R 975 0 R 972 0 R 968 0 R 970 0 R 968 0 R 969 0 R 969 0 R 969 0 R 967 0 R 964 0 R 966 0 R 964 0 R 965 0 R 965 0 R 962 0 R 959 0 R 959 0 R 960 0 R 958 0 R 958 0 R 958 0 R 957 0 R 954 0 R 956 0 R 954 0 R 954 0 R 955 0 R 952 0 R 949 0 R 950 0 R 949 0 R 948 0 R 948 0 R 947 0 R 944 0 R 946 0 R 944 0 R 945 0 R 942 0 R 941 0 R 941 0 R 941 0 R 939 0 R 940 0 R 938 0 R 937 0 R 932 0 R 936 0 R 932 0 R 935 0 R 932 0 R 934 0 R 932 0 R 933 0 R 930 0 R 930 0 R 927 0 R 928 0 R 926 0 R 925 0 R 925 0 R 923 0 R 923 0 R 919 0 R 921 0 R 919 0 R 920 0 R 918 0 R 918 0 R 918 0 R 917 0 R 915 0 R 916 0 R 915 0 R 915 0 R 915 0 R 915 0 R 915 0 R 913 0 R 910 0 R 911 0 R 911 0 R 909 0 R 908 0 R 906 0 R 907 0 R 906 0 R 906 0 R 906 0 R 904 0 R 904 0 R 902 0 R 902 0 R 901 0 R 901 0 R 901 0 R 900 0 R 900 0 R 900 0 R 900 0 R 898 0 R 894 0 R 896 0 R 896 0 R 896 0 R 894 0 R 895 0 R 895 0 R 895 0 R 893 0 R 893 0 R 893 0 R 892 0 R 892 0 R 892 0 R 892 0 R 890 0 R 890 0 R 887 0 R 888 0 R 888 0 R 888 0 R 886 0 R 886 0 R 886 0 R 885 0 R 885 0 R 885 0 R 885 0 R 883 0 R 883 0 R 880 0 R 881 0 R 881 0 R 879 0 R 879 0 R 878 0 R 878 0 R 878 0 R 878 0 R 876 0 R 873 0 R 874 0 R 874 0 R 874 0 R 872 0 R 871 0 R 871 0 R 871 0 R 871 0 R]
+endobj
+
+69 0 obj
+[869 0 R 869 0 R 869 0 R 866 0 R 867 0 R 867 0 R 867 0 R 865 0 R 865 0 R 864 0 R 863 0 R 863 0 R 863 0 R 863 0 R 863 0 R 861 0 R 858 0 R 859 0 R 859 0 R 857 0 R 857 0 R 856 0 R 856 0 R 856 0 R 856 0 R 854 0 R 854 0 R 851 0 R 852 0 R 850 0 R 850 0 R 850 0 R 849 0 R 847 0 R 847 0 R 848 0 R 847 0 R 847 0 R 845 0 R 842 0 R 843 0 R 841 0 R 841 0 R 841 0 R 840 0 R 840 0 R 840 0 R 838 0 R 835 0 R 835 0 R 836 0 R 836 0 R 834 0 R 834 0 R 833 0 R 833 0 R 833 0 R 833 0 R 833 0 R 831 0 R 828 0 R 829 0 R 827 0 R 826 0 R 826 0 R 821 0 R 821 0 R 820 0 R 817 0 R 817 0 R 819 0 R 817 0 R 818 0 R 817 0 R 817 0 R 817 0 R 816 0 R 813 0 R 813 0 R 813 0 R 813 0 R 815 0 R 813 0 R 814 0 R 813 0 R 813 0 R 812 0 R 811 0 R 811 0 R 811 0 R 811 0 R 811 0 R 810 0 R 808 0 R 809 0 R 808 0 R 808 0 R 808 0 R 808 0 R 808 0 R 808 0 R]
+endobj
+
+70 0 obj
+[807 0 R 807 0 R 807 0 R 806 0 R 806 0 R 804 0 R 804 0 R 805 0 R 804 0 R 804 0 R 804 0 R 804 0 R 803 0 R 803 0 R 802 0 R 802 0 R 801 0 R 801 0 R 798 0 R 798 0 R 800 0 R 800 0 R 798 0 R 799 0 R 798 0 R 798 0 R 798 0 R 797 0 R 797 0 R 797 0 R 796 0 R 796 0 R 793 0 R 795 0 R 793 0 R 793 0 R 794 0 R 793 0 R 793 0 R 793 0 R 793 0 R 792 0 R 792 0 R 791 0 R 791 0 R 788 0 R 788 0 R 790 0 R 788 0 R 789 0 R 789 0 R 788 0 R 786 0 R 786 0 R 787 0 R 786 0 R 786 0 R]
+endobj
+
+71 0 obj
+[785 0 R 785 0 R 784 0 R 784 0 R 781 0 R 781 0 R 783 0 R 781 0 R 781 0 R 782 0 R 781 0 R 777 0 R 780 0 R 777 0 R 777 0 R 777 0 R 779 0 R 777 0 R 777 0 R 778 0 R 777 0 R 777 0 R 776 0 R 776 0 R 774 0 R 775 0 R 774 0 R 774 0 R 774 0 R 774 0 R 773 0 R 773 0 R 773 0 R 772 0 R 772 0 R 772 0 R 770 0 R 771 0 R 770 0 R 770 0 R 770 0 R 769 0 R 769 0 R 769 0 R 769 0 R 768 0 R 768 0 R 767 0 R 766 0 R 763 0 R 762 0 R 760 0 R 759 0 R 757 0 R 756 0 R 756 0 R 754 0 R 753 0 R 751 0 R 750 0 R 750 0 R]
+endobj
+
+72 0 obj
+[748 0 R 747 0 R 747 0 R 747 0 R 745 0 R 744 0 R 742 0 R 741 0 R 739 0 R 738 0 R 736 0 R 735 0 R 733 0 R 732 0 R 732 0 R 730 0 R 729 0 R 729 0 R 725 0 R 725 0 R 724 0 R 724 0 R 723 0 R 722 0 R 721 0 R 720 0 R 720 0 R 720 0 R 718 0 R 717 0 R 716 0 R 716 0 R 714 0 R 713 0 R 712 0 R 712 0 R 710 0 R 709 0 R 708 0 R 708 0 R 705 0 R 705 0 R 704 0 R 703 0 R 702 0 R 702 0 R 700 0 R 699 0 R 698 0 R 698 0 R 698 0 R 698 0 R 696 0 R 695 0 R 694 0 R 694 0 R 694 0 R 691 0 R 691 0 R 690 0 R 689 0 R 688 0 R 688 0 R 686 0 R 685 0 R 684 0 R 684 0 R]
+endobj
+
+73 0 obj
+[682 0 R 681 0 R 680 0 R 680 0 R 677 0 R 677 0 R 676 0 R 675 0 R 674 0 R 674 0 R 672 0 R 671 0 R 670 0 R 670 0 R 668 0 R 667 0 R 666 0 R 666 0 R 666 0 R 663 0 R 663 0 R 661 0 R 662 0 R 661 0 R 661 0 R 661 0 R 660 0 R 660 0 R 660 0 R 656 0 R 659 0 R 656 0 R 656 0 R 656 0 R 658 0 R 656 0 R 656 0 R 657 0 R 656 0 R 655 0 R 654 0 R 653 0 R 652 0 R 650 0 R 649 0 R 648 0 R 646 0 R 645 0 R 644 0 R 642 0 R 641 0 R 640 0 R 640 0 R 637 0 R 637 0 R 636 0 R 636 0 R 636 0 R 636 0 R 636 0 R 636 0 R 635 0 R 635 0 R]
+endobj
+
+74 0 obj
+[634 0 R 634 0 R 633 0 R 633 0 R 626 0 R 632 0 R 626 0 R 631 0 R 626 0 R 630 0 R 626 0 R 626 0 R 629 0 R 626 0 R 628 0 R 626 0 R 627 0 R 626 0 R 626 0 R 626 0 R 624 0 R 624 0 R 624 0 R 624 0 R 624 0 R 625 0 R 624 0 R 623 0 R 623 0 R 623 0 R 621 0 R 621 0 R 621 0 R 621 0 R 621 0 R 622 0 R 621 0 R 621 0 R 620 0 R 620 0 R 617 0 R 619 0 R 617 0 R 618 0 R 618 0 R 617 0 R 616 0 R 614 0 R 615 0 R 614 0 R 614 0 R 612 0 R 610 0 R 611 0 R 610 0 R 610 0 R 610 0 R 610 0 R 608 0 R 606 0 R 607 0 R 606 0 R 604 0 R 602 0 R 603 0 R 602 0 R 602 0 R 597 0 R 597 0 R 597 0 R 597 0 R 599 0 R 597 0 R 598 0 R 597 0 R 596 0 R 596 0 R 596 0 R 595 0 R 595 0 R 594 0 R 593 0 R 592 0 R 592 0 R 591 0 R 591 0 R 591 0 R 588 0 R 587 0 R 587 0 R 587 0 R 587 0 R 587 0 R 587 0 R 587 0 R 585 0 R 586 0 R 586 0 R 585 0 R 584 0 R 584 0 R 583 0 R 583 0 R 583 0 R 583 0 R]
+endobj
+
+75 0 obj
+[581 0 R 580 0 R 580 0 R 580 0 R 580 0 R 580 0 R 578 0 R 578 0 R 578 0 R 579 0 R 579 0 R 577 0 R 577 0 R 577 0 R 576 0 R 576 0 R 576 0 R 574 0 R 573 0 R 573 0 R 573 0 R 573 0 R 571 0 R 572 0 R 572 0 R 570 0 R 570 0 R 569 0 R 569 0 R 569 0 R 569 0 R 569 0 R 567 0 R 566 0 R 566 0 R 566 0 R 566 0 R 566 0 R 566 0 R 566 0 R 566 0 R 564 0 R 565 0 R 564 0 R 563 0 R 563 0 R 562 0 R 562 0 R 562 0 R 560 0 R 559 0 R 559 0 R 559 0 R 557 0 R 557 0 R 558 0 R 556 0 R 556 0 R 555 0 R 555 0 R 555 0 R 555 0 R 553 0 R 553 0 R 553 0 R 551 0 R 551 0 R 550 0 R 550 0 R 550 0 R 549 0 R 549 0 R 549 0 R 549 0 R 549 0 R 545 0 R 545 0 R 544 0 R 544 0 R 544 0 R 541 0 R 541 0 R 541 0 R 541 0 R 541 0 R 543 0 R 541 0 R 541 0 R 542 0 R 541 0 R 541 0 R 540 0 R 540 0 R 540 0 R 539 0 R 539 0 R 539 0 R 539 0 R 539 0 R 539 0 R 538 0 R 538 0 R 538 0 R 537 0 R 537 0 R 537 0 R 537 0 R 537 0 R 536 0 R 536 0 R 536 0 R 536 0 R 536 0 R 535 0 R 535 0 R 535 0 R 535 0 R]
+endobj
+
+76 0 obj
+[534 0 R 534 0 R 531 0 R 533 0 R 531 0 R 531 0 R 532 0 R 531 0 R 530 0 R 530 0 R 530 0 R 530 0 R 528 0 R 528 0 R 529 0 R 528 0 R 528 0 R 528 0 R 528 0 R]
+endobj
+
+77 0 obj
+[527 0 R 527 0 R 526 0 R 526 0 R 525 0 R 525 0 R 525 0 R 525 0 R 524 0 R 524 0 R 523 0 R 523 0 R 522 0 R 521 0 R 519 0 R 520 0 R 519 0 R 519 0 R 519 0 R 519 0 R 519 0 R 518 0 R 515 0 R 515 0 R 517 0 R 515 0 R 516 0 R 516 0 R 515 0 R 515 0 R 515 0 R 513 0 R 513 0 R 514 0 R 513 0 R 513 0 R 513 0 R 512 0 R 512 0 R 511 0 R 510 0 R 509 0 R 506 0 R 504 0 R 505 0 R 503 0 R 503 0 R 503 0 R 501 0 R 499 0 R 500 0 R 498 0 R 498 0 R 498 0 R 498 0 R 496 0 R 494 0 R 495 0 R 495 0 R 493 0 R 493 0 R 493 0 R 491 0 R 490 0 R 489 0 R 489 0 R 489 0 R 489 0 R 487 0 R 486 0 R 486 0 R 485 0 R 485 0 R 485 0 R]
+endobj
+
+78 0 obj
+[483 0 R 482 0 R 482 0 R 481 0 R 481 0 R 481 0 R 479 0 R 478 0 R 478 0 R 478 0 R 477 0 R 477 0 R 475 0 R 474 0 R 473 0 R 473 0 R 473 0 R 471 0 R 470 0 R 469 0 R 469 0 R 467 0 R 465 0 R 466 0 R 466 0 R 466 0 R 464 0 R 464 0 R 462 0 R 460 0 R 461 0 R 459 0 R 459 0 R 459 0 R 457 0 R 456 0 R 455 0 R 455 0 R 455 0 R 455 0 R 453 0 R 452 0 R 451 0 R 451 0 R 451 0 R 447 0 R 447 0 R 446 0 R 446 0 R 445 0 R 445 0 R 445 0 R 445 0 R 445 0 R 444 0 R 444 0 R 443 0 R 443 0 R 443 0 R 443 0 R 442 0 R 442 0 R 441 0 R 441 0 R 441 0 R 441 0 R 441 0 R 441 0 R 441 0 R 440 0 R 440 0 R 439 0 R 439 0 R 439 0 R]
+endobj
+
+79 0 obj
+[438 0 R 438 0 R 436 0 R 437 0 R 436 0 R 436 0 R 434 0 R 435 0 R 434 0 R 434 0 R 433 0 R 432 0 R 431 0 R 429 0 R 430 0 R 430 0 R 429 0 R 427 0 R 426 0 R 425 0 R 425 0 R 425 0 R 419 0 R 422 0 R 419 0 R 421 0 R 419 0 R 419 0 R 419 0 R 420 0 R 420 0 R 419 0 R 419 0 R 419 0 R 419 0 R]
+endobj
+
+80 0 obj
+[418 0 R 418 0 R 418 0 R 418 0 R 417 0 R 417 0 R 417 0 R 415 0 R 415 0 R 415 0 R 416 0 R 415 0 R 412 0 R 414 0 R 414 0 R 412 0 R 412 0 R 412 0 R 412 0 R 413 0 R 412 0 R 412 0 R 411 0 R 411 0 R 410 0 R 410 0 R 410 0 R 410 0 R 410 0 R 410 0 R 408 0 R 409 0 R 408 0 R 408 0 R 408 0 R 408 0 R 406 0 R 406 0 R 407 0 R 407 0 R 406 0 R 405 0 R 405 0 R 405 0 R 405 0 R 404 0 R 404 0 R 404 0 R 404 0 R 401 0 R 401 0 R 403 0 R 403 0 R 401 0 R 402 0 R 401 0 R 401 0 R 400 0 R 400 0 R 399 0 R 398 0 R 397 0 R 395 0 R 394 0 R 393 0 R]
+endobj
+
+81 0 obj
+[391 0 R 390 0 R 389 0 R 389 0 R 386 0 R 386 0 R 386 0 R 386 0 R 385 0 R 385 0 R 385 0 R 385 0 R 382 0 R 382 0 R 384 0 R 382 0 R 383 0 R 382 0 R 380 0 R 380 0 R 381 0 R 380 0 R 379 0 R 379 0 R 379 0 R 378 0 R 378 0 R 375 0 R 377 0 R 375 0 R 375 0 R 376 0 R 376 0 R 375 0 R 375 0 R 374 0 R 374 0 R 374 0 R 373 0 R 373 0 R 373 0 R 373 0 R 373 0 R 372 0 R 372 0 R 371 0 R 370 0 R 369 0 R 369 0 R 369 0 R 366 0 R 365 0 R 365 0 R 364 0 R 364 0 R 362 0 R 361 0 R 361 0 R 361 0 R 360 0 R 360 0 R 358 0 R 357 0 R 357 0 R 357 0 R 356 0 R 356 0 R]
+endobj
+
+82 0 obj
+[354 0 R 354 0 R 354 0 R 353 0 R 353 0 R 352 0 R 352 0 R 352 0 R 352 0 R 350 0 R 350 0 R 350 0 R 349 0 R 349 0 R 348 0 R 348 0 R 348 0 R 346 0 R 345 0 R 345 0 R 345 0 R 345 0 R 344 0 R 344 0 R 344 0 R 344 0 R 340 0 R 340 0 R 339 0 R 339 0 R 339 0 R 337 0 R 338 0 R 337 0 R 337 0 R 337 0 R 337 0 R 335 0 R 335 0 R 335 0 R 335 0 R 336 0 R 335 0 R]
+endobj
+
+83 0 obj
+[334 0 R 334 0 R 333 0 R 333 0 R 332 0 R 331 0 R 331 0 R 331 0 R 331 0 R 330 0 R 330 0 R 330 0 R 329 0 R 329 0 R 328 0 R 327 0 R 324 0 R 323 0 R 323 0 R 323 0 R 321 0 R 320 0 R 318 0 R 317 0 R 315 0 R 314 0 R 312 0 R 311 0 R 309 0 R 308 0 R 308 0 R 308 0 R 304 0 R 303 0 R 303 0 R 302 0 R 301 0 R 301 0 R 301 0 R 300 0 R 299 0 R 299 0 R 299 0 R 299 0 R 299 0 R 298 0 R 298 0 R 297 0 R 297 0 R 296 0 R 294 0 R 294 0 R]
+endobj
+
+84 0 obj
+[294 0 R 295 0 R 294 0 R 294 0 R 293 0 R 292 0 R 292 0 R 292 0 R 292 0 R 291 0 R 290 0 R 290 0 R 290 0 R 289 0 R 288 0 R 288 0 R 288 0 R 287 0 R 287 0 R 284 0 R 286 0 R 284 0 R 285 0 R 284 0 R 284 0 R 283 0 R 282 0 R 282 0 R 282 0 R 282 0 R 281 0 R 280 0 R 280 0 R 280 0 R 280 0 R 279 0 R 278 0 R 278 0 R 278 0 R 278 0 R 277 0 R 277 0 R 276 0 R 276 0 R 276 0 R 275 0 R 274 0 R 274 0 R 274 0 R 274 0 R 274 0 R 273 0 R 272 0 R 272 0 R 272 0 R]
+endobj
+
+85 0 obj
+[272 0 R 272 0 R 272 0 R 271 0 R 270 0 R 270 0 R 270 0 R 270 0 R 270 0 R 270 0 R 270 0 R 269 0 R 269 0 R 268 0 R 268 0 R 267 0 R 267 0 R 267 0 R 266 0 R 265 0 R 265 0 R 265 0 R 265 0 R 263 0 R 264 0 R 264 0 R 264 0 R]
+endobj
+
+86 0 obj
+[262 0 R 262 0 R 261 0 R 261 0 R 260 0 R 258 0 R 259 0 R 258 0 R 258 0 R 258 0 R 258 0 R 257 0 R 255 0 R 256 0 R 255 0 R 255 0 R 255 0 R 254 0 R 252 0 R 253 0 R 252 0 R 252 0 R 252 0 R 252 0 R 251 0 R 249 0 R 250 0 R 249 0 R 249 0 R 249 0 R 249 0 R 248 0 R 246 0 R 247 0 R 246 0 R 246 0 R 246 0 R 246 0 R 245 0 R 243 0 R 244 0 R 243 0 R 243 0 R 243 0 R 243 0 R 243 0 R 242 0 R 240 0 R 241 0 R 240 0 R 240 0 R 240 0 R 239 0 R 237 0 R 238 0 R 237 0 R 237 0 R 237 0 R 236 0 R 234 0 R 235 0 R 234 0 R 234 0 R 234 0 R 234 0 R 234 0 R 233 0 R 231 0 R 232 0 R 231 0 R 231 0 R 231 0 R 230 0 R 228 0 R 229 0 R 228 0 R 228 0 R 228 0 R]
+endobj
+
+87 0 obj
+[227 0 R 225 0 R 226 0 R 225 0 R 225 0 R 225 0 R 224 0 R 222 0 R 223 0 R 222 0 R 222 0 R 222 0 R 221 0 R 219 0 R 220 0 R 219 0 R 219 0 R 219 0 R 218 0 R 216 0 R 217 0 R 216 0 R 216 0 R 216 0 R 216 0 R 215 0 R 213 0 R 214 0 R 213 0 R 213 0 R 213 0 R 212 0 R 210 0 R 211 0 R 211 0 R 210 0 R 210 0 R 210 0 R 209 0 R 209 0 R 208 0 R 207 0 R 207 0 R 207 0 R 206 0 R 204 0 R 205 0 R 205 0 R 204 0 R 204 0 R 204 0 R 203 0 R 201 0 R 201 0 R 201 0 R 202 0 R 201 0 R 201 0 R 201 0 R 200 0 R 198 0 R 199 0 R 199 0 R 198 0 R 198 0 R 198 0 R 197 0 R 195 0 R 196 0 R 196 0 R 195 0 R 195 0 R 194 0 R 192 0 R 193 0 R 192 0 R 192 0 R]
+endobj
+
+88 0 obj
+[192 0 R 192 0 R 191 0 R 189 0 R 190 0 R 189 0 R 189 0 R 189 0 R 189 0 R 188 0 R 186 0 R 187 0 R 186 0 R 186 0 R 186 0 R 185 0 R 183 0 R 184 0 R 183 0 R 183 0 R 183 0 R 183 0 R 182 0 R 182 0 R 181 0 R 180 0 R 180 0 R 180 0 R 180 0 R 179 0 R 177 0 R 178 0 R 177 0 R 177 0 R 177 0 R 176 0 R 174 0 R 175 0 R 174 0 R 174 0 R 174 0 R 173 0 R 173 0 R 172 0 R 170 0 R 171 0 R 170 0 R 170 0 R 170 0 R 169 0 R 167 0 R 168 0 R 167 0 R 167 0 R 167 0 R 167 0 R 166 0 R 164 0 R 165 0 R 164 0 R 164 0 R 164 0 R 163 0 R 161 0 R 162 0 R 161 0 R 161 0 R 161 0 R 161 0 R 160 0 R 158 0 R 159 0 R 158 0 R 158 0 R]
+endobj
+
+89 0 obj
+[158 0 R 158 0 R 157 0 R 157 0 R 156 0 R 154 0 R 154 0 R 155 0 R 154 0 R 154 0 R 154 0 R 154 0 R 153 0 R 151 0 R 151 0 R 152 0 R 151 0 R 151 0 R 151 0 R 151 0 R 151 0 R 150 0 R 148 0 R 149 0 R 148 0 R 148 0 R 148 0 R 148 0 R 147 0 R 144 0 R 146 0 R 146 0 R 146 0 R 144 0 R 144 0 R 145 0 R 144 0 R 144 0 R 143 0 R 141 0 R 142 0 R 141 0 R 141 0 R 141 0 R 141 0 R 140 0 R 138 0 R 139 0 R 138 0 R 138 0 R 138 0 R 138 0 R 137 0 R 135 0 R 135 0 R 136 0 R 135 0 R 135 0 R 135 0 R 135 0 R 134 0 R 132 0 R 133 0 R 133 0 R 132 0 R 132 0 R 132 0 R 131 0 R 129 0 R 130 0 R 130 0 R 130 0 R 129 0 R 129 0 R 129 0 R]
+endobj
+
+90 0 obj
+[128 0 R 126 0 R 127 0 R 126 0 R 126 0 R 126 0 R 125 0 R 123 0 R 124 0 R 123 0 R 123 0 R 123 0 R 122 0 R 120 0 R 121 0 R 120 0 R 120 0 R 120 0 R 120 0 R 119 0 R 117 0 R 118 0 R 117 0 R 117 0 R 117 0 R 116 0 R 116 0 R 115 0 R 113 0 R 114 0 R 113 0 R 113 0 R 113 0 R 113 0 R 113 0 R 112 0 R 110 0 R 111 0 R 110 0 R 110 0 R 110 0 R 109 0 R 107 0 R 108 0 R 107 0 R 107 0 R 107 0 R 106 0 R 104 0 R 105 0 R 104 0 R 104 0 R 104 0 R 103 0 R 101 0 R 102 0 R 101 0 R 101 0 R 101 0 R 101 0 R 100 0 R 99 0 R 99 0 R 99 0 R 99 0 R 98 0 R 96 0 R 97 0 R 96 0 R 96 0 R 96 0 R 95 0 R 93 0 R 94 0 R 93 0 R 93 0 R 93 0 R 93 0 R 93 0 R]
+endobj
+
+91 0 obj
+[93 0 R 93 0 R]
+endobj
+
+92 0 obj
+<<
+  /Type /StructElem
+  /S /Document
+  /P 63 0 R
+  /K [1216 0 R 1215 0 R 1214 0 R 1213 0 R 1212 0 R 1023 0 R 1022 0 R 1021 0 R 1020 0 R 1017 0 R 998 0 R 997 0 R 995 0 R 993 0 R 989 0 R 988 0 R 986 0 R 985 0 R 982 0 R 981 0 R 980 0 R 979 0 R 822 0 R 821 0 R 817 0 R 813 0 R 811 0 R 808 0 R 807 0 R 806 0 R 804 0 R 803 0 R 802 0 R 801 0 R 798 0 R 797 0 R 796 0 R 793 0 R 792 0 R 791 0 R 788 0 R 786 0 R 785 0 R 784 0 R 781 0 R 777 0 R 776 0 R 774 0 R 773 0 R 772 0 R 770 0 R 769 0 R 768 0 R 726 0 R 725 0 R 724 0 R 723 0 R 706 0 R 705 0 R 692 0 R 691 0 R 678 0 R 677 0 R 664 0 R 663 0 R 661 0 R 660 0 R 656 0 R 655 0 R 638 0 R 637 0 R 636 0 R 635 0 R 634 0 R 633 0 R 626 0 R 624 0 R 623 0 R 621 0 R 620 0 R 617 0 R 600 0 R 597 0 R 596 0 R 595 0 R 546 0 R 545 0 R 541 0 R 539 0 R 537 0 R 535 0 R 534 0 R 531 0 R 530 0 R 528 0 R 527 0 R 526 0 R 525 0 R 524 0 R 523 0 R 522 0 R 519 0 R 515 0 R 513 0 R 512 0 R 448 0 R 447 0 R 445 0 R 443 0 R 441 0 R 439 0 R 438 0 R 436 0 R 434 0 R 433 0 R 423 0 R 419 0 R 418 0 R 417 0 R 415 0 R 412 0 R 411 0 R 410 0 R 408 0 R 406 0 R 405 0 R 404 0 R 401 0 R 400 0 R 387 0 R 386 0 R 385 0 R 382 0 R 380 0 R 379 0 R 378 0 R 375 0 R 374 0 R 373 0 R 372 0 R 341 0 R 340 0 R 339 0 R 337 0 R 335 0 R 334 0 R 333 0 R 332 0 R 331 0 R 330 0 R 329 0 R 305 0 R 303 0 R 301 0 R 299 0 R 298 0 R 297 0 R 294 0 R 292 0 R 290 0 R 288 0 R 287 0 R 284 0 R 282 0 R 280 0 R 278 0 R 277 0 R 276 0 R 274 0 R 272 0 R 270 0 R 269 0 R 268 0 R 267 0 R 266 0 R 265 0 R 263 0 R 262 0 R 261 0 R 258 0 R 255 0 R 252 0 R 249 0 R 246 0 R 243 0 R 240 0 R 237 0 R 234 0 R 231 0 R 228 0 R 225 0 R 222 0 R 219 0 R 216 0 R 213 0 R 210 0 R 209 0 R 207 0 R 204 0 R 201 0 R 198 0 R 195 0 R 192 0 R 189 0 R 186 0 R 183 0 R 182 0 R 180 0 R 177 0 R 174 0 R 173 0 R 170 0 R 167 0 R 164 0 R 161 0 R 158 0 R 157 0 R 154 0 R 151 0 R 148 0 R 144 0 R 141 0 R 138 0 R 135 0 R 132 0 R 129 0 R 126 0 R 123 0 R 120 0 R 117 0 R 116 0 R 113 0 R 110 0 R 107 0 R 104 0 R 101 0 R 99 0 R 96 0 R 93 0 R]
+>>
+endobj
+
+93 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [95 0 R 72 94 0 R 74 75 76 77 78 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 1461 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 1461 0 R
+  >>]
+  /Pg 1459 0 R
+>>
+endobj
+
+94 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 93 0 R
+  /K [73]
+  /Pg 1459 0 R
+>>
+endobj
+
+95 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 93 0 R
+  /K [71]
+  /Pg 1459 0 R
+>>
+endobj
+
+96 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [98 0 R 66 97 0 R 68 69 70]
+  /Pg 1459 0 R
+>>
+endobj
+
+97 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 96 0 R
+  /K [67]
+  /Pg 1459 0 R
+>>
+endobj
+
+98 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 96 0 R
+  /K [65]
+  /Pg 1459 0 R
+>>
+endobj
+
+99 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [100 0 R 61 62 63 64]
+  /Pg 1459 0 R
+>>
+endobj
+
+100 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 99 0 R
+  /K [60]
+  /Pg 1459 0 R
+>>
+endobj
+
+101 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [103 0 R 54 102 0 R 56 57 58 59]
+  /Pg 1459 0 R
+>>
+endobj
+
+102 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 101 0 R
+  /K [55]
+  /Pg 1459 0 R
+>>
+endobj
+
+103 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 101 0 R
+  /K [53]
+  /Pg 1459 0 R
+>>
+endobj
+
+104 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [106 0 R 48 105 0 R 50 51 52]
+  /Pg 1459 0 R
+>>
+endobj
+
+105 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 104 0 R
+  /K [49]
+  /Pg 1459 0 R
+>>
+endobj
+
+106 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 104 0 R
+  /K [47]
+  /Pg 1459 0 R
+>>
+endobj
+
+107 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [109 0 R 42 108 0 R 44 45 46]
+  /Pg 1459 0 R
+>>
+endobj
+
+108 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 107 0 R
+  /K [43]
+  /Pg 1459 0 R
+>>
+endobj
+
+109 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 107 0 R
+  /K [41]
+  /Pg 1459 0 R
+>>
+endobj
+
+110 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [112 0 R 36 111 0 R 38 39 40]
+  /Pg 1459 0 R
+>>
+endobj
+
+111 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 110 0 R
+  /K [37]
+  /Pg 1459 0 R
+>>
+endobj
+
+112 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 110 0 R
+  /K [35]
+  /Pg 1459 0 R
+>>
+endobj
+
+113 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [115 0 R 28 114 0 R 30 31 32 33 34]
+  /Pg 1459 0 R
+>>
+endobj
+
+114 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 113 0 R
+  /K [29]
+  /Pg 1459 0 R
+>>
+endobj
+
+115 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 113 0 R
+  /K [27]
+  /Pg 1459 0 R
+>>
+endobj
+
+116 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF0049006E00670065006E00690065007200ED006100200064006500200073006F0066007400770061007200650020006500200069006E0066007200610065007300740072007500630074007500720061>
+  /K [25 26]
+  /Pg 1459 0 R
+>>
+endobj
+
+117 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [119 0 R 20 118 0 R 22 23 24]
+  /Pg 1459 0 R
+>>
+endobj
+
+118 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 117 0 R
+  /K [21]
+  /Pg 1459 0 R
+>>
+endobj
+
+119 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 117 0 R
+  /K [19]
+  /Pg 1459 0 R
+>>
+endobj
+
+120 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [122 0 R 13 121 0 R 15 16 17 18]
+  /Pg 1459 0 R
+>>
+endobj
+
+121 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 120 0 R
+  /K [14]
+  /Pg 1459 0 R
+>>
+endobj
+
+122 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 120 0 R
+  /K [12]
+  /Pg 1459 0 R
+>>
+endobj
+
+123 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [125 0 R 7 124 0 R 9 10 11]
+  /Pg 1459 0 R
+>>
+endobj
+
+124 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 123 0 R
+  /K [8]
+  /Pg 1459 0 R
+>>
+endobj
+
+125 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 123 0 R
+  /K [6]
+  /Pg 1459 0 R
+>>
+endobj
+
+126 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [128 0 R 1 127 0 R 3 4 5]
+  /Pg 1459 0 R
+>>
+endobj
+
+127 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 126 0 R
+  /K [2]
+  /Pg 1459 0 R
+>>
+endobj
+
+128 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 126 0 R
+  /K [0]
+  /Pg 1459 0 R
+>>
+endobj
+
+129 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [131 0 R 68 130 0 R 72 73 74]
+  /Pg 1457 0 R
+>>
+endobj
+
+130 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 129 0 R
+  /K [69 70 71]
+  /Pg 1457 0 R
+>>
+endobj
+
+131 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 129 0 R
+  /K [67]
+  /Pg 1457 0 R
+>>
+endobj
+
+132 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [134 0 R 61 133 0 R 64 65 66]
+  /Pg 1457 0 R
+>>
+endobj
+
+133 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 132 0 R
+  /K [62 63]
+  /Pg 1457 0 R
+>>
+endobj
+
+134 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 132 0 R
+  /K [60]
+  /Pg 1457 0 R
+>>
+endobj
+
+135 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [137 0 R 53 54 136 0 R 56 57 58 59]
+  /Pg 1457 0 R
+>>
+endobj
+
+136 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 135 0 R
+  /K [55]
+  /Pg 1457 0 R
+>>
+endobj
+
+137 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 135 0 R
+  /K [52]
+  /Pg 1457 0 R
+>>
+endobj
+
+138 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [140 0 R 46 139 0 R 48 49 50 51]
+  /Pg 1457 0 R
+>>
+endobj
+
+139 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 138 0 R
+  /K [47]
+  /Pg 1457 0 R
+>>
+endobj
+
+140 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 138 0 R
+  /K [45]
+  /Pg 1457 0 R
+>>
+endobj
+
+141 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [143 0 R 39 142 0 R 41 42 43 44]
+  /Pg 1457 0 R
+>>
+endobj
+
+142 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 141 0 R
+  /K [40]
+  /Pg 1457 0 R
+>>
+endobj
+
+143 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 141 0 R
+  /K [38]
+  /Pg 1457 0 R
+>>
+endobj
+
+144 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [147 0 R 29 146 0 R 33 34 145 0 R 36 37]
+  /Pg 1457 0 R
+>>
+endobj
+
+145 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 144 0 R
+  /K [35]
+  /Pg 1457 0 R
+>>
+endobj
+
+146 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 144 0 R
+  /K [30 31 32]
+  /Pg 1457 0 R
+>>
+endobj
+
+147 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 144 0 R
+  /K [28]
+  /Pg 1457 0 R
+>>
+endobj
+
+148 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [150 0 R 22 149 0 R 24 25 26 27]
+  /Pg 1457 0 R
+>>
+endobj
+
+149 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 148 0 R
+  /K [23]
+  /Pg 1457 0 R
+>>
+endobj
+
+150 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 148 0 R
+  /K [21]
+  /Pg 1457 0 R
+>>
+endobj
+
+151 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [153 0 R 13 14 152 0 R 16 17 18 19 20]
+  /Pg 1457 0 R
+>>
+endobj
+
+152 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 151 0 R
+  /K [15]
+  /Pg 1457 0 R
+>>
+endobj
+
+153 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 151 0 R
+  /K [12]
+  /Pg 1457 0 R
+>>
+endobj
+
+154 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [156 0 R 5 6 155 0 R 8 9 10 11]
+  /Pg 1457 0 R
+>>
+endobj
+
+155 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 154 0 R
+  /K [7]
+  /Pg 1457 0 R
+>>
+endobj
+
+156 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 154 0 R
+  /K [4]
+  /Pg 1457 0 R
+>>
+endobj
+
+157 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004600ED007300690063006100200063007500E1006E007400690063006100200079002000660075006E00640061006D0065006E0074006F0073>
+  /K [2 3]
+  /Pg 1457 0 R
+>>
+endobj
+
+158 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [160 0 R 70 159 0 R 72 73 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 1457 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 1457 0 R
+  >>]
+  /Pg 1455 0 R
+>>
+endobj
+
+159 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 158 0 R
+  /K [71]
+  /Pg 1455 0 R
+>>
+endobj
+
+160 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 158 0 R
+  /K [69]
+  /Pg 1455 0 R
+>>
+endobj
+
+161 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [163 0 R 63 162 0 R 65 66 67 68]
+  /Pg 1455 0 R
+>>
+endobj
+
+162 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 161 0 R
+  /K [64]
+  /Pg 1455 0 R
+>>
+endobj
+
+163 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 161 0 R
+  /K [62]
+  /Pg 1455 0 R
+>>
+endobj
+
+164 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [166 0 R 57 165 0 R 59 60 61]
+  /Pg 1455 0 R
+>>
+endobj
+
+165 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 164 0 R
+  /K [58]
+  /Pg 1455 0 R
+>>
+endobj
+
+166 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 164 0 R
+  /K [56]
+  /Pg 1455 0 R
+>>
+endobj
+
+167 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [169 0 R 50 168 0 R 52 53 54 55]
+  /Pg 1455 0 R
+>>
+endobj
+
+168 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 167 0 R
+  /K [51]
+  /Pg 1455 0 R
+>>
+endobj
+
+169 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 167 0 R
+  /K [49]
+  /Pg 1455 0 R
+>>
+endobj
+
+170 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [172 0 R 44 171 0 R 46 47 48]
+  /Pg 1455 0 R
+>>
+endobj
+
+171 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 170 0 R
+  /K [45]
+  /Pg 1455 0 R
+>>
+endobj
+
+172 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 170 0 R
+  /K [43]
+  /Pg 1455 0 R
+>>
+endobj
+
+173 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T (Perspectivas adicionales)
+  /K [41 42]
+  /Pg 1455 0 R
+>>
+endobj
+
+174 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [176 0 R 36 175 0 R 38 39 40]
+  /Pg 1455 0 R
+>>
+endobj
+
+175 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 174 0 R
+  /K [37]
+  /Pg 1455 0 R
+>>
+endobj
+
+176 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 174 0 R
+  /K [35]
+  /Pg 1455 0 R
+>>
+endobj
+
+177 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [179 0 R 30 178 0 R 32 33 34]
+  /Pg 1455 0 R
+>>
+endobj
+
+178 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 177 0 R
+  /K [31]
+  /Pg 1455 0 R
+>>
+endobj
+
+179 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 177 0 R
+  /K [29]
+  /Pg 1455 0 R
+>>
+endobj
+
+180 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [181 0 R 25 26 27 28]
+  /Pg 1455 0 R
+>>
+endobj
+
+181 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 180 0 R
+  /K [24]
+  /Pg 1455 0 R
+>>
+endobj
+
+182 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF0049006E00670065006E00690065007200ED006100200064006500200073006F006600740077006100720065002000790020007000720061007800690073>
+  /K [22 23]
+  /Pg 1455 0 R
+>>
+endobj
+
+183 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [185 0 R 16 184 0 R 18 19 20 21]
+  /Pg 1455 0 R
+>>
+endobj
+
+184 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 183 0 R
+  /K [17]
+  /Pg 1455 0 R
+>>
+endobj
+
+185 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 183 0 R
+  /K [15]
+  /Pg 1455 0 R
+>>
+endobj
+
+186 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [188 0 R 10 187 0 R 12 13 14]
+  /Pg 1455 0 R
+>>
+endobj
+
+187 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 186 0 R
+  /K [11]
+  /Pg 1455 0 R
+>>
+endobj
+
+188 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 186 0 R
+  /K [9]
+  /Pg 1455 0 R
+>>
+endobj
+
+189 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [191 0 R 3 190 0 R 5 6 7 8]
+  /Pg 1455 0 R
+>>
+endobj
+
+190 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 189 0 R
+  /K [4]
+  /Pg 1455 0 R
+>>
+endobj
+
+191 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 189 0 R
+  /K [2]
+  /Pg 1455 0 R
+>>
+endobj
+
+192 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [194 0 R 73 193 0 R 75 76 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 1455 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 1455 0 R
+  >>]
+  /Pg 1453 0 R
+>>
+endobj
+
+193 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 192 0 R
+  /K [74]
+  /Pg 1453 0 R
+>>
+endobj
+
+194 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 192 0 R
+  /K [72]
+  /Pg 1453 0 R
+>>
+endobj
+
+195 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [197 0 R 67 196 0 R 70 71]
+  /Pg 1453 0 R
+>>
+endobj
+
+196 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 195 0 R
+  /K [68 69]
+  /Pg 1453 0 R
+>>
+endobj
+
+197 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 195 0 R
+  /K [66]
+  /Pg 1453 0 R
+>>
+endobj
+
+198 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [200 0 R 60 199 0 R 63 64 65]
+  /Pg 1453 0 R
+>>
+endobj
+
+199 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 198 0 R
+  /K [61 62]
+  /Pg 1453 0 R
+>>
+endobj
+
+200 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 198 0 R
+  /K [59]
+  /Pg 1453 0 R
+>>
+endobj
+
+201 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [203 0 R 52 53 54 202 0 R 56 57 58]
+  /Pg 1453 0 R
+>>
+endobj
+
+202 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 201 0 R
+  /K [55]
+  /Pg 1453 0 R
+>>
+endobj
+
+203 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 201 0 R
+  /K [51]
+  /Pg 1453 0 R
+>>
+endobj
+
+204 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [206 0 R 45 205 0 R 48 49 50]
+  /Pg 1453 0 R
+>>
+endobj
+
+205 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 204 0 R
+  /K [46 47]
+  /Pg 1453 0 R
+>>
+endobj
+
+206 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 204 0 R
+  /K [44]
+  /Pg 1453 0 R
+>>
+endobj
+
+207 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [208 0 R 41 42 43]
+  /Pg 1453 0 R
+>>
+endobj
+
+208 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 207 0 R
+  /K [40]
+  /Pg 1453 0 R
+>>
+endobj
+
+209 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF0049006E00740065006C006900670065006E0063006900610020006100720074006900660069006300690061006C0020007900200072006500700072006500730065006E007400610063006900F3006E002000640065006C00200063006F006E006F00630069006D00690065006E0074006F>
+  /K [38 39]
+  /Pg 1453 0 R
+>>
+endobj
+
+210 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [212 0 R 32 211 0 R 35 36 37]
+  /Pg 1453 0 R
+>>
+endobj
+
+211 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 210 0 R
+  /K [33 34]
+  /Pg 1453 0 R
+>>
+endobj
+
+212 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 210 0 R
+  /K [31]
+  /Pg 1453 0 R
+>>
+endobj
+
+213 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [215 0 R 26 214 0 R 28 29 30]
+  /Pg 1453 0 R
+>>
+endobj
+
+214 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 213 0 R
+  /K [27]
+  /Pg 1453 0 R
+>>
+endobj
+
+215 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 213 0 R
+  /K [25]
+  /Pg 1453 0 R
+>>
+endobj
+
+216 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [218 0 R 19 217 0 R 21 22 23 24]
+  /Pg 1453 0 R
+>>
+endobj
+
+217 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 216 0 R
+  /K [20]
+  /Pg 1453 0 R
+>>
+endobj
+
+218 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 216 0 R
+  /K [18]
+  /Pg 1453 0 R
+>>
+endobj
+
+219 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [221 0 R 13 220 0 R 15 16 17]
+  /Pg 1453 0 R
+>>
+endobj
+
+220 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 219 0 R
+  /K [14]
+  /Pg 1453 0 R
+>>
+endobj
+
+221 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 219 0 R
+  /K [12]
+  /Pg 1453 0 R
+>>
+endobj
+
+222 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [224 0 R 7 223 0 R 9 10 11]
+  /Pg 1453 0 R
+>>
+endobj
+
+223 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 222 0 R
+  /K [8]
+  /Pg 1453 0 R
+>>
+endobj
+
+224 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 222 0 R
+  /K [6]
+  /Pg 1453 0 R
+>>
+endobj
+
+225 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [227 0 R 1 226 0 R 3 4 5]
+  /Pg 1453 0 R
+>>
+endobj
+
+226 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 225 0 R
+  /K [2]
+  /Pg 1453 0 R
+>>
+endobj
+
+227 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 225 0 R
+  /K [0]
+  /Pg 1453 0 R
+>>
+endobj
+
+228 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [230 0 R 73 229 0 R 75 76 77]
+  /Pg 1451 0 R
+>>
+endobj
+
+229 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 228 0 R
+  /K [74]
+  /Pg 1451 0 R
+>>
+endobj
+
+230 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 228 0 R
+  /K [72]
+  /Pg 1451 0 R
+>>
+endobj
+
+231 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [233 0 R 67 232 0 R 69 70 71]
+  /Pg 1451 0 R
+>>
+endobj
+
+232 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 231 0 R
+  /K [68]
+  /Pg 1451 0 R
+>>
+endobj
+
+233 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 231 0 R
+  /K [66]
+  /Pg 1451 0 R
+>>
+endobj
+
+234 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [236 0 R 59 235 0 R 61 62 63 64 65]
+  /Pg 1451 0 R
+>>
+endobj
+
+235 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 234 0 R
+  /K [60]
+  /Pg 1451 0 R
+>>
+endobj
+
+236 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 234 0 R
+  /K [58]
+  /Pg 1451 0 R
+>>
+endobj
+
+237 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [239 0 R 53 238 0 R 55 56 57]
+  /Pg 1451 0 R
+>>
+endobj
+
+238 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 237 0 R
+  /K [54]
+  /Pg 1451 0 R
+>>
+endobj
+
+239 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 237 0 R
+  /K [52]
+  /Pg 1451 0 R
+>>
+endobj
+
+240 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [242 0 R 47 241 0 R 49 50 51]
+  /Pg 1451 0 R
+>>
+endobj
+
+241 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 240 0 R
+  /K [48]
+  /Pg 1451 0 R
+>>
+endobj
+
+242 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 240 0 R
+  /K [46]
+  /Pg 1451 0 R
+>>
+endobj
+
+243 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [245 0 R 39 244 0 R 41 42 43 44 45]
+  /Pg 1451 0 R
+>>
+endobj
+
+244 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 243 0 R
+  /K [40]
+  /Pg 1451 0 R
+>>
+endobj
+
+245 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 243 0 R
+  /K [38]
+  /Pg 1451 0 R
+>>
+endobj
+
+246 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [248 0 R 32 247 0 R 34 35 36 37]
+  /Pg 1451 0 R
+>>
+endobj
+
+247 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 246 0 R
+  /K [33]
+  /Pg 1451 0 R
+>>
+endobj
+
+248 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 246 0 R
+  /K [31]
+  /Pg 1451 0 R
+>>
+endobj
+
+249 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [251 0 R 25 250 0 R 27 28 29 30]
+  /Pg 1451 0 R
+>>
+endobj
+
+250 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 249 0 R
+  /K [26]
+  /Pg 1451 0 R
+>>
+endobj
+
+251 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 249 0 R
+  /K [24]
+  /Pg 1451 0 R
+>>
+endobj
+
+252 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [254 0 R 18 253 0 R 20 21 22 23]
+  /Pg 1451 0 R
+>>
+endobj
+
+253 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 252 0 R
+  /K [19]
+  /Pg 1451 0 R
+>>
+endobj
+
+254 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 252 0 R
+  /K [17]
+  /Pg 1451 0 R
+>>
+endobj
+
+255 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [257 0 R 12 256 0 R 14 15 16]
+  /Pg 1451 0 R
+>>
+endobj
+
+256 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 255 0 R
+  /K [13]
+  /Pg 1451 0 R
+>>
+endobj
+
+257 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 255 0 R
+  /K [11]
+  /Pg 1451 0 R
+>>
+endobj
+
+258 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [260 0 R 5 259 0 R 7 8 9 10]
+  /Pg 1451 0 R
+>>
+endobj
+
+259 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 258 0 R
+  /K [6]
+  /Pg 1451 0 R
+>>
+endobj
+
+260 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 258 0 R
+  /K [4]
+  /Pg 1451 0 R
+>>
+endobj
+
+261 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF00460069006C006F0073006F006600ED006100200063006C00E10073006900630061002000790020006F006E0074006F006C006F006700ED0061>
+  /K [2 3]
+  /Pg 1451 0 R
+>>
+endobj
+
+262 0 obj
+<<
+  /Type /StructElem
+  /S /H1
+  /P 92 0 R
+  /T (Referencias y lecturas recomendadas)
+  /K [0 1]
+  /Pg 1451 0 R
+>>
+endobj
+
+263 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [23 264 0 R]
+  /Pg 1449 0 R
+>>
+endobj
+
+264 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 263 0 R
+  /K [24 25 26]
+  /Pg 1449 0 R
+>>
+endobj
+
+265 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [19 20 21 22]
+  /Pg 1449 0 R
+>>
+endobj
+
+266 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [18]
+  /Pg 1449 0 R
+>>
+endobj
+
+267 0 obj
+<<
+  /Type /StructElem
+  /S /Span
+  /P 92 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [15 16 17]
+  /Pg 1449 0 R
+>>
+endobj
+
+268 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [13 14]
+  /Pg 1449 0 R
+>>
+endobj
+
+269 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004C00610020007300ED006E00740065007300690073>
+  /K [11 12]
+  /Pg 1449 0 R
+>>
+endobj
+
+270 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [271 0 R 4 5 6 7 8 9 10]
+  /Pg 1449 0 R
+>>
+endobj
+
+271 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 270 0 R
+  /K [3]
+  /Pg 1449 0 R
+>>
+endobj
+
+272 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [273 0 R 52 53 54 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 1449 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 1449 0 R
+  >> <<
+    /Type /MCR
+    /MCID 2
+    /Pg 1449 0 R
+  >>]
+  /Pg 1447 0 R
+>>
+endobj
+
+273 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 272 0 R
+  /K [51]
+  /Pg 1447 0 R
+>>
+endobj
+
+274 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [275 0 R 46 47 48 49 50]
+  /Pg 1447 0 R
+>>
+endobj
+
+275 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 274 0 R
+  /K [45]
+  /Pg 1447 0 R
+>>
+endobj
+
+276 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [42 43 44]
+  /Pg 1447 0 R
+>>
+endobj
+
+277 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T (Las tres presiones que hacen esto urgente ahora)
+  /K [40 41]
+  /Pg 1447 0 R
+>>
+endobj
+
+278 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [279 0 R 36 37 38 39]
+  /Pg 1447 0 R
+>>
+endobj
+
+279 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 278 0 R
+  /K [35]
+  /Pg 1447 0 R
+>>
+endobj
+
+280 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [281 0 R 31 32 33 34]
+  /Pg 1447 0 R
+>>
+endobj
+
+281 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 280 0 R
+  /K [30]
+  /Pg 1447 0 R
+>>
+endobj
+
+282 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [283 0 R 26 27 28 29]
+  /Pg 1447 0 R
+>>
+endobj
+
+283 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 282 0 R
+  /K [25]
+  /Pg 1447 0 R
+>>
+endobj
+
+284 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [19 286 0 R 21 285 0 R 23 24]
+  /Pg 1447 0 R
+>>
+endobj
+
+285 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 284 0 R
+  /K [22]
+  /Pg 1447 0 R
+>>
+endobj
+
+286 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 284 0 R
+  /K [20]
+  /Pg 1447 0 R
+>>
+endobj
+
+287 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF0050006F007200200071007500E90020006C00610020007200650066006C0065006300740069006F006E002020140020006C00610020006F006E0074006F006C006F006700ED0061002000730069006E0020006F00700065007200610063006900F3006E00200065007300200075006E0020006D007500730065006F>
+  /K [17 18]
+  /Pg 1447 0 R
+>>
+endobj
+
+288 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [289 0 R 14 15 16]
+  /Pg 1447 0 R
+>>
+endobj
+
+289 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 288 0 R
+  /K [13]
+  /Pg 1447 0 R
+>>
+endobj
+
+290 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [291 0 R 10 11 12]
+  /Pg 1447 0 R
+>>
+endobj
+
+291 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 290 0 R
+  /K [9]
+  /Pg 1447 0 R
+>>
+endobj
+
+292 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [293 0 R 5 6 7 8]
+  /Pg 1447 0 R
+>>
+endobj
+
+293 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 292 0 R
+  /K [4]
+  /Pg 1447 0 R
+>>
+endobj
+
+294 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [296 0 R 50 51 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 1447 0 R
+  >> 295 0 R <<
+    /Type /MCR
+    /MCID 2
+    /Pg 1447 0 R
+  >> <<
+    /Type /MCR
+    /MCID 3
+    /Pg 1447 0 R
+  >>]
+  /Pg 1445 0 R
+>>
+endobj
+
+295 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 294 0 R
+  /Lang (en)
+  /K [1]
+  /Pg 1447 0 R
+>>
+endobj
+
+296 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 294 0 R
+  /K [49]
+  /Pg 1445 0 R
+>>
+endobj
+
+297 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [47 48]
+  /Pg 1445 0 R
+>>
+endobj
+
+298 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF0050006F007200200071007500E90020006C006F00730020004400410047007300202014002000650073007000650063006900660069006300690064006100640020007400E90063006E006900630061>
+  /K [45 46]
+  /Pg 1445 0 R
+>>
+endobj
+
+299 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [300 0 R 40 41 42 43 44]
+  /Pg 1445 0 R
+>>
+endobj
+
+300 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 299 0 R
+  /K [39]
+  /Pg 1445 0 R
+>>
+endobj
+
+301 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [302 0 R 36 37 38]
+  /Pg 1445 0 R
+>>
+endobj
+
+302 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 301 0 R
+  /K [35]
+  /Pg 1445 0 R
+>>
+endobj
+
+303 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [304 0 R 33 34]
+  /Pg 1445 0 R
+>>
+endobj
+
+304 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 303 0 R
+  /K [32]
+  /Pg 1445 0 R
+>>
+endobj
+
+305 0 obj
+<<
+  /Type /StructElem
+  /S /Table
+  /P 92 0 R
+  /A [<<
+    /O /Layout
+    /BorderColor [0 0 0]
+    /BorderThickness 1
+  >>]
+  /K [325 0 R 306 0 R]
+>>
+endobj
+
+306 0 obj
+<<
+  /Type /StructElem
+  /S /TBody
+  /P 305 0 R
+  /K [322 0 R 319 0 R 316 0 R 313 0 R 310 0 R 307 0 R]
+>>
+endobj
+
+307 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 306 0 R
+  /K [309 0 R 308 0 R]
+>>
+endobj
+
+308 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 307 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U6x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [29 30 31]
+  /Pg 1445 0 R
+>>
+endobj
+
+309 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 307 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U6x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [28]
+  /Pg 1445 0 R
+>>
+endobj
+
+310 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 306 0 R
+  /K [312 0 R 311 0 R]
+>>
+endobj
+
+311 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 310 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U6x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [27]
+  /Pg 1445 0 R
+>>
+endobj
+
+312 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 310 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U6x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [26]
+  /Pg 1445 0 R
+>>
+endobj
+
+313 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 306 0 R
+  /K [315 0 R 314 0 R]
+>>
+endobj
+
+314 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 313 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U6x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [25]
+  /Pg 1445 0 R
+>>
+endobj
+
+315 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 313 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U6x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [24]
+  /Pg 1445 0 R
+>>
+endobj
+
+316 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 306 0 R
+  /K [318 0 R 317 0 R]
+>>
+endobj
+
+317 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 316 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U6x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [23]
+  /Pg 1445 0 R
+>>
+endobj
+
+318 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 316 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U6x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [22]
+  /Pg 1445 0 R
+>>
+endobj
+
+319 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 306 0 R
+  /K [321 0 R 320 0 R]
+>>
+endobj
+
+320 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 319 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U6x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [21]
+  /Pg 1445 0 R
+>>
+endobj
+
+321 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 319 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U6x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [20]
+  /Pg 1445 0 R
+>>
+endobj
+
+322 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 306 0 R
+  /K [324 0 R 323 0 R]
+>>
+endobj
+
+323 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 322 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U6x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [17 18 19]
+  /Pg 1445 0 R
+>>
+endobj
+
+324 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 322 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U6x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [16]
+  /Pg 1445 0 R
+>>
+endobj
+
+325 0 obj
+<<
+  /Type /StructElem
+  /S /THead
+  /P 305 0 R
+  /K [326 0 R]
+>>
+endobj
+
+326 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 325 0 R
+  /K [328 0 R 327 0 R]
+>>
+endobj
+
+327 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 326 0 R
+  /ID (U6x1y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [15]
+  /Pg 1445 0 R
+>>
+endobj
+
+328 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 326 0 R
+  /ID (U6x0y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [14]
+  /Pg 1445 0 R
+>>
+endobj
+
+329 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004C006F00200071007500650020006C00610020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C0020006800610063006500200071007500650020006E0061006400610020006D00E1007300200068006100630065>
+  /K [12 13]
+  /Pg 1445 0 R
+>>
+endobj
+
+330 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [9 10 11]
+  /Pg 1445 0 R
+>>
+endobj
+
+331 0 obj
+<<
+  /Type /StructElem
+  /S /Span
+  /P 92 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [5 6 7 8]
+  /Pg 1445 0 R
+>>
+endobj
+
+332 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [4]
+  /Pg 1445 0 R
+>>
+endobj
+
+333 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF0045006C002000700072006F0062006C0065006D00610020007200650061006C0020006E006F0020006500730020007400E90063006E00690063006F0020201400200065007300200065007000690073007400E9006D00690063006F>
+  /K [2 3]
+  /Pg 1445 0 R
+>>
+endobj
+
+334 0 obj
+<<
+  /Type /StructElem
+  /S /H1
+  /P 92 0 R
+  /T <FEFF0050006F007200200071007500E9002000650073002000720065006C006500760061006E0074006500200065006E00200073006F0066007400770061007200650020006500200069006E006600720061006500730074007200750063007400750072006100730020006400690067006900740061006C00650073>
+  /K [0 1]
+  /Pg 1445 0 R
+>>
+endobj
+
+335 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [37 38 39 40 336 0 R 42]
+  /Pg 1443 0 R
+>>
+endobj
+
+336 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 335 0 R
+  /K [41]
+  /Pg 1443 0 R
+>>
+endobj
+
+337 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [31 338 0 R 33 34 35 36]
+  /Pg 1443 0 R
+>>
+endobj
+
+338 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 337 0 R
+  /K [32]
+  /Pg 1443 0 R
+>>
+endobj
+
+339 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [28 29 30]
+  /Pg 1443 0 R
+>>
+endobj
+
+340 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004C006100200063006F006E00650078006900F3006E0020006E006F0020006F0062007600690061002020140020004C004C004D007300200063006F006D006F002000670065006F006D00650074007200ED006100200063007500E1006E0074006900630061002000610070006C00690063006100640061>
+  /K [26 27]
+  /Pg 1443 0 R
+>>
+endobj
+
+341 0 obj
+<<
+  /Type /StructElem
+  /S /Table
+  /P 92 0 R
+  /A [<<
+    /O /Layout
+    /BorderColor [0 0 0]
+    /BorderThickness 1
+  >>]
+  /K [367 0 R 342 0 R]
+>>
+endobj
+
+342 0 obj
+<<
+  /Type /StructElem
+  /S /TBody
+  /P 341 0 R
+  /K [363 0 R 359 0 R 355 0 R 351 0 R 347 0 R 343 0 R]
+>>
+endobj
+
+343 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 342 0 R
+  /K [346 0 R 345 0 R 344 0 R]
+>>
+endobj
+
+344 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 343 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [22 23 24 25]
+  /Pg 1443 0 R
+>>
+endobj
+
+345 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 343 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [18 19 20 21]
+  /Pg 1443 0 R
+>>
+endobj
+
+346 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 343 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [17]
+  /Pg 1443 0 R
+>>
+endobj
+
+347 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 342 0 R
+  /K [350 0 R 349 0 R 348 0 R]
+>>
+endobj
+
+348 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 347 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [14 15 16]
+  /Pg 1443 0 R
+>>
+endobj
+
+349 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 347 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [12 13]
+  /Pg 1443 0 R
+>>
+endobj
+
+350 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 347 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [9 10 11]
+  /Pg 1443 0 R
+>>
+endobj
+
+351 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 342 0 R
+  /K [354 0 R 353 0 R 352 0 R]
+>>
+endobj
+
+352 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 351 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [5 6 7 8]
+  /Pg 1443 0 R
+>>
+endobj
+
+353 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 351 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [3 4]
+  /Pg 1443 0 R
+>>
+endobj
+
+354 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 351 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [0 1 2]
+  /Pg 1443 0 R
+>>
+endobj
+
+355 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 342 0 R
+  /K [358 0 R 357 0 R 356 0 R]
+>>
+endobj
+
+356 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 355 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [65 66]
+  /Pg 1441 0 R
+>>
+endobj
+
+357 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 355 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [62 63 64]
+  /Pg 1441 0 R
+>>
+endobj
+
+358 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 355 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [61]
+  /Pg 1441 0 R
+>>
+endobj
+
+359 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 342 0 R
+  /K [362 0 R 361 0 R 360 0 R]
+>>
+endobj
+
+360 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 359 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [59 60]
+  /Pg 1441 0 R
+>>
+endobj
+
+361 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 359 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [56 57 58]
+  /Pg 1441 0 R
+>>
+endobj
+
+362 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 359 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [55]
+  /Pg 1441 0 R
+>>
+endobj
+
+363 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 342 0 R
+  /K [366 0 R 365 0 R 364 0 R]
+>>
+endobj
+
+364 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 363 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [53 54]
+  /Pg 1441 0 R
+>>
+endobj
+
+365 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 363 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [51 52]
+  /Pg 1441 0 R
+>>
+endobj
+
+366 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 363 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U5x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [50]
+  /Pg 1441 0 R
+>>
+endobj
+
+367 0 obj
+<<
+  /Type /StructElem
+  /S /THead
+  /P 341 0 R
+  /K [368 0 R]
+>>
+endobj
+
+368 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 367 0 R
+  /K [371 0 R 370 0 R 369 0 R]
+>>
+endobj
+
+369 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 368 0 R
+  /ID (U5x2y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [47 48 49]
+  /Pg 1441 0 R
+>>
+endobj
+
+370 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 368 0 R
+  /ID (U5x1y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [46]
+  /Pg 1441 0 R
+>>
+endobj
+
+371 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 368 0 R
+  /ID (U5x0y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [45]
+  /Pg 1441 0 R
+>>
+endobj
+
+372 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004C006100200063006F006D00700075007400610063006900F3006E00200063007500E1006E00740069006300610020201400200068006F006E006500730074006900640061006400200073006F00620072006500200073007500730020006C00ED006D0069007400650073>
+  /K [43 44]
+  /Pg 1441 0 R
+>>
+endobj
+
+373 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [38 39 40 41 42]
+  /Pg 1441 0 R
+>>
+endobj
+
+374 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [35 36 37]
+  /Pg 1441 0 R
+>>
+endobj
+
+375 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [27 377 0 R 29 30 376 0 R 33 34]
+  /Pg 1441 0 R
+>>
+endobj
+
+376 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 375 0 R
+  /K [31 32]
+  /Pg 1441 0 R
+>>
+endobj
+
+377 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 375 0 R
+  /K [28]
+  /Pg 1441 0 R
+>>
+endobj
+
+378 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004C006100200069006E007400650072007000720065007400610063006900F3006E0020006D00E10073002000720065006C006500760061006E00740065002020140020004D0065006300E1006E00690063006100200043007500E1006E0074006900630061002000520065006C006100630069006F006E0061006C>
+  /K [25 26]
+  /Pg 1441 0 R
+>>
+endobj
+
+379 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [22 23 24]
+  /Pg 1441 0 R
+>>
+endobj
+
+380 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [18 19 381 0 R 21]
+  /Pg 1441 0 R
+>>
+endobj
+
+381 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 380 0 R
+  /K [20]
+  /Pg 1441 0 R
+>>
+endobj
+
+382 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [12 13 384 0 R 15 383 0 R 17]
+  /Pg 1441 0 R
+>>
+endobj
+
+383 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 382 0 R
+  /K [16]
+  /Pg 1441 0 R
+>>
+endobj
+
+384 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 382 0 R
+  /K [14]
+  /Pg 1441 0 R
+>>
+endobj
+
+385 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T <FEFF0045006C002000700072006F0062006C0065006D0061002000640065006C0020006F0062007300650072007600610064006F00720020219400200045006C002000740072006100730066006F006E0064006F00200071007500650020006E006F00200070007500650064006500200066006F0072006D0061006C0069007A0061007200730065>
+  /K [8 9 10 11]
+  /Pg 1441 0 R
+>>
+endobj
+
+386 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [4 5 6 7]
+  /Pg 1441 0 R
+>>
+endobj
+
+387 0 obj
+<<
+  /Type /StructElem
+  /S /L
+  /P 92 0 R
+  /A [<<
+    /O /List
+    /ListNumbering /Circle
+  >>]
+  /K [396 0 R 392 0 R 388 0 R]
+>>
+endobj
+
+388 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 387 0 R
+  /K [391 0 R 389 0 R]
+>>
+endobj
+
+389 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 388 0 R
+  /K [390 0 R 2 3]
+  /Pg 1441 0 R
+>>
+endobj
+
+390 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 389 0 R
+  /K [1]
+  /Pg 1441 0 R
+>>
+endobj
+
+391 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 388 0 R
+  /K [0]
+  /Pg 1441 0 R
+>>
+endobj
+
+392 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 387 0 R
+  /K [395 0 R 393 0 R]
+>>
+endobj
+
+393 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 392 0 R
+  /K [394 0 R 64]
+  /Pg 1439 0 R
+>>
+endobj
+
+394 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 393 0 R
+  /K [63]
+  /Pg 1439 0 R
+>>
+endobj
+
+395 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 392 0 R
+  /K [62]
+  /Pg 1439 0 R
+>>
+endobj
+
+396 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 387 0 R
+  /K [399 0 R 397 0 R]
+>>
+endobj
+
+397 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 396 0 R
+  /K [398 0 R 61]
+  /Pg 1439 0 R
+>>
+endobj
+
+398 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 397 0 R
+  /K [60]
+  /Pg 1439 0 R
+>>
+endobj
+
+399 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 396 0 R
+  /K [59]
+  /Pg 1439 0 R
+>>
+endobj
+
+400 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [57 58]
+  /Pg 1439 0 R
+>>
+endobj
+
+401 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [49 50 403 0 R 53 402 0 R 55 56]
+  /Pg 1439 0 R
+>>
+endobj
+
+402 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 401 0 R
+  /K [54]
+  /Pg 1439 0 R
+>>
+endobj
+
+403 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 401 0 R
+  /K [51 52]
+  /Pg 1439 0 R
+>>
+endobj
+
+404 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T <FEFF0043006F006E007400650078007400750061006C006900640061006400200063007500E1006E0074006900630061002021940020004F006E0074006F006C006F006700ED0061002000700065007200730070006500630074006900760061006C>
+  /K [45 46 47 48]
+  /Pg 1439 0 R
+>>
+endobj
+
+405 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [41 42 43 44]
+  /Pg 1439 0 R
+>>
+endobj
+
+406 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [36 37 407 0 R 40]
+  /Pg 1439 0 R
+>>
+endobj
+
+407 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 406 0 R
+  /K [38 39]
+  /Pg 1439 0 R
+>>
+endobj
+
+408 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [30 409 0 R 32 33 34 35]
+  /Pg 1439 0 R
+>>
+endobj
+
+409 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 408 0 R
+  /K [31]
+  /Pg 1439 0 R
+>>
+endobj
+
+410 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T <FEFF0043006F006D0070006C0065006D0065006E0074006100720069006500640061006400200064006500200042006F0068007200202194002000590069006E002D00590061006E0067002000790020006C00610020006400750061006C00690064006100640020004F006E0074006F006C006F006700ED0061002F005200650066006C0065006300740069006F006E>
+  /K [24 25 26 27 28 29]
+  /Pg 1439 0 R
+>>
+endobj
+
+411 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T (Las tres conexiones genuinas)
+  /K [22 23]
+  /Pg 1439 0 R
+>>
+endobj
+
+412 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [12 414 0 R 15 16 17 18 413 0 R 20 21]
+  /Pg 1439 0 R
+>>
+endobj
+
+413 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 412 0 R
+  /K [19]
+  /Pg 1439 0 R
+>>
+endobj
+
+414 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 412 0 R
+  /K [13 14]
+  /Pg 1439 0 R
+>>
+endobj
+
+415 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [7 8 9 416 0 R 11]
+  /Pg 1439 0 R
+>>
+endobj
+
+416 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 415 0 R
+  /K [10]
+  /Pg 1439 0 R
+>>
+endobj
+
+417 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF0045006C00200067006F006C007000650020006D00E10073002000700072006F00660075006E0064006F0020201400200041007200690073007400F300740065006C006500730020006500730074006100620061002000650071007500690076006F006300610064006F00200065006E00200065006C0020006E006900760065006C002000660075006E00640061006D0065006E00740061006C>
+  /K [4 5 6]
+  /Pg 1439 0 R
+>>
+endobj
+
+418 0 obj
+<<
+  /Type /StructElem
+  /S /H1
+  /P 92 0 R
+  /T <FEFF004600ED007300690063006100200063007500E1006E0074006900630061002C00200063006F006D00700075007400610063006900F3006E00200063007500E1006E0074006900630061002000790020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C>
+  /K [0 1 2 3]
+  /Pg 1439 0 R
+>>
+endobj
+
+419 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [22 422 0 R 24 421 0 R 26 27 28 420 0 R 31 32 33 34]
+  /Pg 1437 0 R
+>>
+endobj
+
+420 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 419 0 R
+  /K [29 30]
+  /Pg 1437 0 R
+>>
+endobj
+
+421 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 419 0 R
+  /K [25]
+  /Pg 1437 0 R
+>>
+endobj
+
+422 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 419 0 R
+  /K [23]
+  /Pg 1437 0 R
+>>
+endobj
+
+423 0 obj
+<<
+  /Type /StructElem
+  /S /L
+  /P 92 0 R
+  /A [<<
+    /O /List
+    /ListNumbering /Circle
+  >>]
+  /K [428 0 R 424 0 R]
+>>
+endobj
+
+424 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 423 0 R
+  /K [427 0 R 425 0 R]
+>>
+endobj
+
+425 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 424 0 R
+  /K [426 0 R 19 20 21]
+  /Pg 1437 0 R
+>>
+endobj
+
+426 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 425 0 R
+  /K [18]
+  /Pg 1437 0 R
+>>
+endobj
+
+427 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 424 0 R
+  /K [17]
+  /Pg 1437 0 R
+>>
+endobj
+
+428 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 423 0 R
+  /K [432 0 R 429 0 R]
+>>
+endobj
+
+429 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 428 0 R
+  /K [431 0 R 13 430 0 R 16]
+  /Pg 1437 0 R
+>>
+endobj
+
+430 0 obj
+<<
+  /Type /StructElem
+  /S /Formula
+  /P 429 0 R
+  /K [14 15]
+  /Pg 1437 0 R
+>>
+endobj
+
+431 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 429 0 R
+  /K [12]
+  /Pg 1437 0 R
+>>
+endobj
+
+432 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 428 0 R
+  /K [11]
+  /Pg 1437 0 R
+>>
+endobj
+
+433 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [10]
+  /Pg 1437 0 R
+>>
+endobj
+
+434 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [6 435 0 R 8 9]
+  /Pg 1437 0 R
+>>
+endobj
+
+435 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 434 0 R
+  /K [7]
+  /Pg 1437 0 R
+>>
+endobj
+
+436 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [2 437 0 R 4 5]
+  /Pg 1437 0 R
+>>
+endobj
+
+437 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 436 0 R
+  /K [3]
+  /Pg 1437 0 R
+>>
+endobj
+
+438 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004C0061002000740065006E0073006900F3006E00200063006F006E00200065006C002000540061006F00ED0073006D006F002020140020004400410047002000760073002E0020006300690063006C006F>
+  /K [0 1]
+  /Pg 1437 0 R
+>>
+endobj
+
+439 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [440 0 R 71 72 73]
+  /Pg 1435 0 R
+>>
+endobj
+
+440 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 439 0 R
+  /K [69 70]
+  /Pg 1435 0 R
+>>
+endobj
+
+441 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [442 0 R 62 63 64 65 66 67 68]
+  /Pg 1435 0 R
+>>
+endobj
+
+442 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 441 0 R
+  /K [60 61]
+  /Pg 1435 0 R
+>>
+endobj
+
+443 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [444 0 R 56 57 58 59]
+  /Pg 1435 0 R
+>>
+endobj
+
+444 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 443 0 R
+  /K [54 55]
+  /Pg 1435 0 R
+>>
+endobj
+
+445 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [446 0 R 49 50 51 52 53]
+  /Pg 1435 0 R
+>>
+endobj
+
+446 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 445 0 R
+  /K [47 48]
+  /Pg 1435 0 R
+>>
+endobj
+
+447 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004C00610073002000630075006100740072006F002000660075006E00630069006F006E00650073002000640065006C002000440041004700200065006E0020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C>
+  /K [45 46]
+  /Pg 1435 0 R
+>>
+endobj
+
+448 0 obj
+<<
+  /Type /StructElem
+  /S /Table
+  /P 92 0 R
+  /A [<<
+    /O /Layout
+    /BorderColor [0 0 0]
+    /BorderThickness 1
+  >>]
+  /K [507 0 R 449 0 R]
+>>
+endobj
+
+449 0 obj
+<<
+  /Type /StructElem
+  /S /TBody
+  /P 448 0 R
+  /K [502 0 R 497 0 R 492 0 R 488 0 R 484 0 R 480 0 R 476 0 R 472 0 R 468 0 R 463 0 R 458 0 R 454 0 R 450 0 R]
+>>
+endobj
+
+450 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [453 0 R 452 0 R 451 0 R]
+>>
+endobj
+
+451 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 450 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [42 43 44]
+  /Pg 1435 0 R
+>>
+endobj
+
+452 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 450 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [41]
+  /Pg 1435 0 R
+>>
+endobj
+
+453 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 450 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [40]
+  /Pg 1435 0 R
+>>
+endobj
+
+454 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [457 0 R 456 0 R 455 0 R]
+>>
+endobj
+
+455 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 454 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [36 37 38 39]
+  /Pg 1435 0 R
+>>
+endobj
+
+456 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 454 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [35]
+  /Pg 1435 0 R
+>>
+endobj
+
+457 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 454 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [34]
+  /Pg 1435 0 R
+>>
+endobj
+
+458 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [462 0 R 460 0 R 459 0 R]
+>>
+endobj
+
+459 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 458 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [31 32 33]
+  /Pg 1435 0 R
+>>
+endobj
+
+460 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 458 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [29 461 0 R]
+  /Pg 1435 0 R
+>>
+endobj
+
+461 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 460 0 R
+  /K [30]
+  /Pg 1435 0 R
+>>
+endobj
+
+462 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 458 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [28]
+  /Pg 1435 0 R
+>>
+endobj
+
+463 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [467 0 R 465 0 R 464 0 R]
+>>
+endobj
+
+464 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 463 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [26 27]
+  /Pg 1435 0 R
+>>
+endobj
+
+465 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 463 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [22 466 0 R]
+  /Pg 1435 0 R
+>>
+endobj
+
+466 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 465 0 R
+  /K [23 24 25]
+  /Pg 1435 0 R
+>>
+endobj
+
+467 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 463 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [21]
+  /Pg 1435 0 R
+>>
+endobj
+
+468 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [471 0 R 470 0 R 469 0 R]
+>>
+endobj
+
+469 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 468 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [19 20]
+  /Pg 1435 0 R
+>>
+endobj
+
+470 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 468 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [18]
+  /Pg 1435 0 R
+>>
+endobj
+
+471 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 468 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [17]
+  /Pg 1435 0 R
+>>
+endobj
+
+472 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [475 0 R 474 0 R 473 0 R]
+>>
+endobj
+
+473 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 472 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [14 15 16]
+  /Pg 1435 0 R
+>>
+endobj
+
+474 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 472 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [13]
+  /Pg 1435 0 R
+>>
+endobj
+
+475 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 472 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [12]
+  /Pg 1435 0 R
+>>
+endobj
+
+476 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [479 0 R 478 0 R 477 0 R]
+>>
+endobj
+
+477 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 476 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [10 11]
+  /Pg 1435 0 R
+>>
+endobj
+
+478 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 476 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [7 8 9]
+  /Pg 1435 0 R
+>>
+endobj
+
+479 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 476 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [6]
+  /Pg 1435 0 R
+>>
+endobj
+
+480 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [483 0 R 482 0 R 481 0 R]
+>>
+endobj
+
+481 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 480 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [3 4 5]
+  /Pg 1435 0 R
+>>
+endobj
+
+482 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 480 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [1 2]
+  /Pg 1435 0 R
+>>
+endobj
+
+483 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 480 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [0]
+  /Pg 1435 0 R
+>>
+endobj
+
+484 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [487 0 R 486 0 R 485 0 R]
+>>
+endobj
+
+485 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 484 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [71 72 73]
+  /Pg 1433 0 R
+>>
+endobj
+
+486 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 484 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [69 70]
+  /Pg 1433 0 R
+>>
+endobj
+
+487 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 484 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [68]
+  /Pg 1433 0 R
+>>
+endobj
+
+488 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [491 0 R 490 0 R 489 0 R]
+>>
+endobj
+
+489 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 488 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [64 65 66 67]
+  /Pg 1433 0 R
+>>
+endobj
+
+490 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 488 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [63]
+  /Pg 1433 0 R
+>>
+endobj
+
+491 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 488 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [62]
+  /Pg 1433 0 R
+>>
+endobj
+
+492 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [496 0 R 494 0 R 493 0 R]
+>>
+endobj
+
+493 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 492 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [59 60 61]
+  /Pg 1433 0 R
+>>
+endobj
+
+494 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 492 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [56 495 0 R]
+  /Pg 1433 0 R
+>>
+endobj
+
+495 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 494 0 R
+  /K [57 58]
+  /Pg 1433 0 R
+>>
+endobj
+
+496 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 492 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [55]
+  /Pg 1433 0 R
+>>
+endobj
+
+497 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [501 0 R 499 0 R 498 0 R]
+>>
+endobj
+
+498 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 497 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [51 52 53 54]
+  /Pg 1433 0 R
+>>
+endobj
+
+499 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 497 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [49 500 0 R]
+  /Pg 1433 0 R
+>>
+endobj
+
+500 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 499 0 R
+  /K [50]
+  /Pg 1433 0 R
+>>
+endobj
+
+501 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 497 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [48]
+  /Pg 1433 0 R
+>>
+endobj
+
+502 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 449 0 R
+  /K [506 0 R 504 0 R 503 0 R]
+>>
+endobj
+
+503 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 502 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [45 46 47]
+  /Pg 1433 0 R
+>>
+endobj
+
+504 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 502 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [43 505 0 R]
+  /Pg 1433 0 R
+>>
+endobj
+
+505 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 504 0 R
+  /K [44]
+  /Pg 1433 0 R
+>>
+endobj
+
+506 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 502 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U4x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [42]
+  /Pg 1433 0 R
+>>
+endobj
+
+507 0 obj
+<<
+  /Type /StructElem
+  /S /THead
+  /P 448 0 R
+  /K [508 0 R]
+>>
+endobj
+
+508 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 507 0 R
+  /K [511 0 R 510 0 R 509 0 R]
+>>
+endobj
+
+509 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 508 0 R
+  /ID (U4x2y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [41]
+  /Pg 1433 0 R
+>>
+endobj
+
+510 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 508 0 R
+  /ID (U4x1y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [40]
+  /Pg 1433 0 R
+>>
+endobj
+
+511 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 508 0 R
+  /ID (U4x0y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [39]
+  /Pg 1433 0 R
+>>
+endobj
+
+512 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004C00ED006E00650061002000670065006E00650061006C00F30067006900630061>
+  /K [37 38]
+  /Pg 1433 0 R
+>>
+endobj
+
+513 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [31 32 514 0 R 34 35 36]
+  /Pg 1433 0 R
+>>
+endobj
+
+514 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 513 0 R
+  /K [33]
+  /Pg 1433 0 R
+>>
+endobj
+
+515 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [518 0 R 22 23 517 0 R 25 516 0 R 28 29 30]
+  /Pg 1433 0 R
+>>
+endobj
+
+516 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 515 0 R
+  /K [26 27]
+  /Pg 1433 0 R
+>>
+endobj
+
+517 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 515 0 R
+  /K [24]
+  /Pg 1433 0 R
+>>
+endobj
+
+518 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 515 0 R
+  /K [21]
+  /Pg 1433 0 R
+>>
+endobj
+
+519 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [521 0 R 14 520 0 R 16 17 18 19 20]
+  /Pg 1433 0 R
+>>
+endobj
+
+520 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 519 0 R
+  /K [15]
+  /Pg 1433 0 R
+>>
+endobj
+
+521 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 519 0 R
+  /K [13]
+  /Pg 1433 0 R
+>>
+endobj
+
+522 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [12]
+  /Pg 1433 0 R
+>>
+endobj
+
+523 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF0050006F007200200071007500E90020006C006100200061006300690063006C0069006300690064006100640020201400200065006C002000700072006F0062006C0065006D00610020006400650020006C00610020006A007500730074006900660069006300610063006900F3006E002000630069007200630075006C00610072>
+  /K [10 11]
+  /Pg 1433 0 R
+>>
+endobj
+
+524 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [8 9]
+  /Pg 1433 0 R
+>>
+endobj
+
+525 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [4 5 6 7]
+  /Pg 1433 0 R
+>>
+endobj
+
+526 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T (El problema que fuerza el grafo)
+  /K [2 3]
+  /Pg 1433 0 R
+>>
+endobj
+
+527 0 obj
+<<
+  /Type /StructElem
+  /S /H1
+  /P 92 0 R
+  /T <FEFF004400410047007300202014002000670065006E00650061006C006F006700ED006100200079002000660075006E0063006900F3006E00200065006E0020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C>
+  /K [0 1]
+  /Pg 1433 0 R
+>>
+endobj
+
+528 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [12 13 529 0 R 15 16 17 18]
+  /Pg 1431 0 R
+>>
+endobj
+
+529 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 528 0 R
+  /K [14]
+  /Pg 1431 0 R
+>>
+endobj
+
+530 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [8 9 10 11]
+  /Pg 1431 0 R
+>>
+endobj
+
+531 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [2 533 0 R 4 5 532 0 R 7]
+  /Pg 1431 0 R
+>>
+endobj
+
+532 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 531 0 R
+  /K [6]
+  /Pg 1431 0 R
+>>
+endobj
+
+533 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 531 0 R
+  /K [3]
+  /Pg 1431 0 R
+>>
+endobj
+
+534 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004C006100200063007200ED0074006900630061002000740061006F00ED007300740061002000610020006C00610020006F006E0074006F006C006F006700ED006100200063006C00610073006900660069006300610074006F007200690061>
+  /K [0 1]
+  /Pg 1431 0 R
+>>
+endobj
+
+535 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [536 0 R 113 114 115 116]
+  /Pg 1429 0 R
+>>
+endobj
+
+536 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 535 0 R
+  /K [108 109 110 111 112]
+  /Pg 1429 0 R
+>>
+endobj
+
+537 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [538 0 R 103 104 105 106 107]
+  /Pg 1429 0 R
+>>
+endobj
+
+538 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 537 0 R
+  /K [100 101 102]
+  /Pg 1429 0 R
+>>
+endobj
+
+539 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [540 0 R 94 95 96 97 98 99]
+  /Pg 1429 0 R
+>>
+endobj
+
+540 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 539 0 R
+  /K [91 92 93]
+  /Pg 1429 0 R
+>>
+endobj
+
+541 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [544 0 R 80 81 82 83 84 543 0 R 86 87 542 0 R 89 90]
+  /Pg 1429 0 R
+>>
+endobj
+
+542 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 541 0 R
+  /K [88]
+  /Pg 1429 0 R
+>>
+endobj
+
+543 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 541 0 R
+  /K [85]
+  /Pg 1429 0 R
+>>
+endobj
+
+544 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 541 0 R
+  /K [77 78 79]
+  /Pg 1429 0 R
+>>
+endobj
+
+545 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T (Las cuatro correspondencias profundas)
+  /K [75 76]
+  /Pg 1429 0 R
+>>
+endobj
+
+546 0 obj
+<<
+  /Type /StructElem
+  /S /Table
+  /P 92 0 R
+  /A [<<
+    /O /Layout
+    /BorderColor [0 0 0]
+    /BorderThickness 1
+  >>]
+  /K [589 0 R 547 0 R]
+>>
+endobj
+
+547 0 obj
+<<
+  /Type /StructElem
+  /S /TBody
+  /P 546 0 R
+  /K [582 0 R 575 0 R 568 0 R 561 0 R 554 0 R 548 0 R]
+>>
+endobj
+
+548 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 547 0 R
+  /K [552 0 R 551 0 R 550 0 R 549 0 R]
+>>
+endobj
+
+549 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 548 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [70 71 72 73 74]
+  /Pg 1429 0 R
+>>
+endobj
+
+550 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 548 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [67 68 69]
+  /Pg 1429 0 R
+>>
+endobj
+
+551 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 548 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [65 66]
+  /Pg 1429 0 R
+>>
+endobj
+
+552 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 548 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [553 0 R]
+>>
+endobj
+
+553 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 552 0 R
+  /K [62 63 64]
+  /Pg 1429 0 R
+>>
+endobj
+
+554 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 547 0 R
+  /K [559 0 R 557 0 R 556 0 R 555 0 R]
+>>
+endobj
+
+555 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 554 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [58 59 60 61]
+  /Pg 1429 0 R
+>>
+endobj
+
+556 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 554 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [56 57]
+  /Pg 1429 0 R
+>>
+endobj
+
+557 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 554 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [53 54 558 0 R]
+  /Pg 1429 0 R
+>>
+endobj
+
+558 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 557 0 R
+  /K [55]
+  /Pg 1429 0 R
+>>
+endobj
+
+559 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 554 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [560 0 R 50 51 52]
+  /Pg 1429 0 R
+>>
+endobj
+
+560 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 559 0 R
+  /K [49]
+  /Pg 1429 0 R
+>>
+endobj
+
+561 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 547 0 R
+  /K [566 0 R 564 0 R 563 0 R 562 0 R]
+>>
+endobj
+
+562 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 561 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [46 47 48]
+  /Pg 1429 0 R
+>>
+endobj
+
+563 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 561 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [44 45]
+  /Pg 1429 0 R
+>>
+endobj
+
+564 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 561 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [41 565 0 R 43]
+  /Pg 1429 0 R
+>>
+endobj
+
+565 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 564 0 R
+  /K [42]
+  /Pg 1429 0 R
+>>
+endobj
+
+566 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 561 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [567 0 R 33 34 35 36 37 38 39 40]
+  /Pg 1429 0 R
+>>
+endobj
+
+567 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 566 0 R
+  /K [32]
+  /Pg 1429 0 R
+>>
+endobj
+
+568 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 547 0 R
+  /K [573 0 R 571 0 R 570 0 R 569 0 R]
+>>
+endobj
+
+569 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 568 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [27 28 29 30 31]
+  /Pg 1429 0 R
+>>
+endobj
+
+570 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 568 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [25 26]
+  /Pg 1429 0 R
+>>
+endobj
+
+571 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 568 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [22 572 0 R]
+  /Pg 1429 0 R
+>>
+endobj
+
+572 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 571 0 R
+  /K [23 24]
+  /Pg 1429 0 R
+>>
+endobj
+
+573 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 568 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [574 0 R 18 19 20 21]
+  /Pg 1429 0 R
+>>
+endobj
+
+574 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 573 0 R
+  /K [17]
+  /Pg 1429 0 R
+>>
+endobj
+
+575 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 547 0 R
+  /K [580 0 R 578 0 R 577 0 R 576 0 R]
+>>
+endobj
+
+576 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 575 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [14 15 16]
+  /Pg 1429 0 R
+>>
+endobj
+
+577 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 575 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [11 12 13]
+  /Pg 1429 0 R
+>>
+endobj
+
+578 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 575 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [6 7 8 579 0 R]
+  /Pg 1429 0 R
+>>
+endobj
+
+579 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 578 0 R
+  /K [9 10]
+  /Pg 1429 0 R
+>>
+endobj
+
+580 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 575 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [581 0 R 1 2 3 4 5]
+  /Pg 1429 0 R
+>>
+endobj
+
+581 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 580 0 R
+  /K [0]
+  /Pg 1429 0 R
+>>
+endobj
+
+582 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 547 0 R
+  /K [587 0 R 585 0 R 584 0 R 583 0 R]
+>>
+endobj
+
+583 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 582 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [101 102 103 104]
+  /Pg 1427 0 R
+>>
+endobj
+
+584 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 582 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [99 100]
+  /Pg 1427 0 R
+>>
+endobj
+
+585 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 582 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [95 586 0 R 98]
+  /Pg 1427 0 R
+>>
+endobj
+
+586 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 585 0 R
+  /K [96 97]
+  /Pg 1427 0 R
+>>
+endobj
+
+587 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 582 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U3x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [588 0 R 88 89 90 91 92 93 94]
+  /Pg 1427 0 R
+>>
+endobj
+
+588 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 587 0 R
+  /K [87]
+  /Pg 1427 0 R
+>>
+endobj
+
+589 0 obj
+<<
+  /Type /StructElem
+  /S /THead
+  /P 546 0 R
+  /K [590 0 R]
+>>
+endobj
+
+590 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 589 0 R
+  /K [594 0 R 593 0 R 592 0 R 591 0 R]
+>>
+endobj
+
+591 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 590 0 R
+  /ID (U3x3y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [84 85 86]
+  /Pg 1427 0 R
+>>
+endobj
+
+592 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 590 0 R
+  /ID (U3x2y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [82 83]
+  /Pg 1427 0 R
+>>
+endobj
+
+593 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 590 0 R
+  /ID (U3x1y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [81]
+  /Pg 1427 0 R
+>>
+endobj
+
+594 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 590 0 R
+  /ID (U3x0y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [80]
+  /Pg 1427 0 R
+>>
+endobj
+
+595 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T (Correspondencias con el modelo operacional)
+  /K [78 79]
+  /Pg 1427 0 R
+>>
+endobj
+
+596 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [75 76 77]
+  /Pg 1427 0 R
+>>
+endobj
+
+597 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [67 68 69 70 599 0 R 72 598 0 R 74]
+  /Pg 1427 0 R
+>>
+endobj
+
+598 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 597 0 R
+  /K [73]
+  /Pg 1427 0 R
+>>
+endobj
+
+599 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 597 0 R
+  /K [71]
+  /Pg 1427 0 R
+>>
+endobj
+
+600 0 obj
+<<
+  /Type /StructElem
+  /S /L
+  /P 92 0 R
+  /A [<<
+    /O /List
+    /ListNumbering /Circle
+  >>]
+  /K [613 0 R 609 0 R 605 0 R 601 0 R]
+>>
+endobj
+
+601 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 600 0 R
+  /K [604 0 R 602 0 R]
+>>
+endobj
+
+602 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 601 0 R
+  /K [63 603 0 R 65 66]
+  /Pg 1427 0 R
+>>
+endobj
+
+603 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 602 0 R
+  /K [64]
+  /Pg 1427 0 R
+>>
+endobj
+
+604 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 601 0 R
+  /K [62]
+  /Pg 1427 0 R
+>>
+endobj
+
+605 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 600 0 R
+  /K [608 0 R 606 0 R]
+>>
+endobj
+
+606 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 605 0 R
+  /K [59 607 0 R 61]
+  /Pg 1427 0 R
+>>
+endobj
+
+607 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 606 0 R
+  /K [60]
+  /Pg 1427 0 R
+>>
+endobj
+
+608 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 605 0 R
+  /K [58]
+  /Pg 1427 0 R
+>>
+endobj
+
+609 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 600 0 R
+  /K [612 0 R 610 0 R]
+>>
+endobj
+
+610 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 609 0 R
+  /K [52 611 0 R 54 55 56 57]
+  /Pg 1427 0 R
+>>
+endobj
+
+611 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 610 0 R
+  /K [53]
+  /Pg 1427 0 R
+>>
+endobj
+
+612 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 609 0 R
+  /K [51]
+  /Pg 1427 0 R
+>>
+endobj
+
+613 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 600 0 R
+  /K [616 0 R 614 0 R]
+>>
+endobj
+
+614 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 613 0 R
+  /K [47 615 0 R 49 50]
+  /Pg 1427 0 R
+>>
+endobj
+
+615 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 614 0 R
+  /K [48]
+  /Pg 1427 0 R
+>>
+endobj
+
+616 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 613 0 R
+  /K [46]
+  /Pg 1427 0 R
+>>
+endobj
+
+617 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [40 619 0 R 42 618 0 R 45]
+  /Pg 1427 0 R
+>>
+endobj
+
+618 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 617 0 R
+  /K [43 44]
+  /Pg 1427 0 R
+>>
+endobj
+
+619 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 617 0 R
+  /K [41]
+  /Pg 1427 0 R
+>>
+endobj
+
+620 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF00590069006E002D00590061006E00670020201400200070006F006C00610072006900640061006400200063006F006E007300740069007400750074006900760061002C0020006E006F0020006F0070006F007300690063006900F3006E>
+  /K [38 39]
+  /Pg 1427 0 R
+>>
+endobj
+
+621 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [30 31 32 33 34 622 0 R 36 37]
+  /Pg 1427 0 R
+>>
+endobj
+
+622 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 621 0 R
+  /K [35]
+  /Pg 1427 0 R
+>>
+endobj
+
+623 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [27 28 29]
+  /Pg 1427 0 R
+>>
+endobj
+
+624 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [20 21 22 23 24 625 0 R 26]
+  /Pg 1427 0 R
+>>
+endobj
+
+625 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 624 0 R
+  /K [25]
+  /Pg 1427 0 R
+>>
+endobj
+
+626 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [4 632 0 R 6 631 0 R 8 630 0 R 10 11 629 0 R 13 628 0 R 15 627 0 R 17 18 19]
+  /Pg 1427 0 R
+>>
+endobj
+
+627 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 626 0 R
+  /K [16]
+  /Pg 1427 0 R
+>>
+endobj
+
+628 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 626 0 R
+  /K [14]
+  /Pg 1427 0 R
+>>
+endobj
+
+629 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 626 0 R
+  /K [12]
+  /Pg 1427 0 R
+>>
+endobj
+
+630 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 626 0 R
+  /K [9]
+  /Pg 1427 0 R
+>>
+endobj
+
+631 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 626 0 R
+  /K [7]
+  /Pg 1427 0 R
+>>
+endobj
+
+632 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 626 0 R
+  /K [5]
+  /Pg 1427 0 R
+>>
+endobj
+
+633 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004C006F00200071007500650020006C00610020007400720061006400690063006900F3006E0020006F00630063006900640065006E00740061006C0020006E006F002000760069006F>
+  /K [2 3]
+  /Pg 1427 0 R
+>>
+endobj
+
+634 0 obj
+<<
+  /Type /StructElem
+  /S /H1
+  /P 92 0 R
+  /T <FEFF00540061006F00ED0073006D006F002C002000590069006E002D00590061006E0067002000790020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C>
+  /K [0 1]
+  /Pg 1427 0 R
+>>
+endobj
+
+635 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [61 62]
+  /Pg 1425 0 R
+>>
+endobj
+
+636 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [55 56 57 58 59 60]
+  /Pg 1425 0 R
+>>
+endobj
+
+637 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF0043006F006E0063006C00750073006900F3006E>
+  /K [53 54]
+  /Pg 1425 0 R
+>>
+endobj
+
+638 0 obj
+<<
+  /Type /StructElem
+  /S /L
+  /P 92 0 R
+  /A [<<
+    /O /List
+    /ListNumbering /Circle
+  >>]
+  /K [651 0 R 647 0 R 643 0 R 639 0 R]
+>>
+endobj
+
+639 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 638 0 R
+  /K [642 0 R 640 0 R]
+>>
+endobj
+
+640 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 639 0 R
+  /K [641 0 R 51 52]
+  /Pg 1425 0 R
+>>
+endobj
+
+641 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 640 0 R
+  /K [50]
+  /Pg 1425 0 R
+>>
+endobj
+
+642 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 639 0 R
+  /K [49]
+  /Pg 1425 0 R
+>>
+endobj
+
+643 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 638 0 R
+  /K [646 0 R 644 0 R]
+>>
+endobj
+
+644 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 643 0 R
+  /K [645 0 R 48]
+  /Pg 1425 0 R
+>>
+endobj
+
+645 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 644 0 R
+  /K [47]
+  /Pg 1425 0 R
+>>
+endobj
+
+646 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 643 0 R
+  /K [46]
+  /Pg 1425 0 R
+>>
+endobj
+
+647 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 638 0 R
+  /K [650 0 R 648 0 R]
+>>
+endobj
+
+648 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 647 0 R
+  /K [649 0 R 45]
+  /Pg 1425 0 R
+>>
+endobj
+
+649 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 648 0 R
+  /K [44]
+  /Pg 1425 0 R
+>>
+endobj
+
+650 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 647 0 R
+  /K [43]
+  /Pg 1425 0 R
+>>
+endobj
+
+651 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 638 0 R
+  /K [654 0 R 652 0 R]
+>>
+endobj
+
+652 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 651 0 R
+  /K [653 0 R 42]
+  /Pg 1425 0 R
+>>
+endobj
+
+653 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 652 0 R
+  /K [41]
+  /Pg 1425 0 R
+>>
+endobj
+
+654 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 651 0 R
+  /K [40]
+  /Pg 1425 0 R
+>>
+endobj
+
+655 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [39]
+  /Pg 1425 0 R
+>>
+endobj
+
+656 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [29 659 0 R 31 32 33 658 0 R 35 36 657 0 R 38]
+  /Pg 1425 0 R
+>>
+endobj
+
+657 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 656 0 R
+  /K [37]
+  /Pg 1425 0 R
+>>
+endobj
+
+658 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 656 0 R
+  /K [34]
+  /Pg 1425 0 R
+>>
+endobj
+
+659 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 656 0 R
+  /K [30]
+  /Pg 1425 0 R
+>>
+endobj
+
+660 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [26 27 28]
+  /Pg 1425 0 R
+>>
+endobj
+
+661 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [21 662 0 R 23 24 25]
+  /Pg 1425 0 R
+>>
+endobj
+
+662 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 661 0 R
+  /K [22]
+  /Pg 1425 0 R
+>>
+endobj
+
+663 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T <FEFF0053006F006600740077006100720065002000640065002000700072006F0079006500630074006F002020140020006C0061002000660072006F006E0074006500720061002000610063007400750061006C>
+  /K [19 20]
+  /Pg 1425 0 R
+>>
+endobj
+
+664 0 obj
+<<
+  /Type /StructElem
+  /S /L
+  /P 92 0 R
+  /A [<<
+    /O /List
+    /ListNumbering /Circle
+  >>]
+  /K [673 0 R 669 0 R 665 0 R]
+>>
+endobj
+
+665 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 664 0 R
+  /K [668 0 R 666 0 R]
+>>
+endobj
+
+666 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 665 0 R
+  /K [667 0 R 16 17 18]
+  /Pg 1425 0 R
+>>
+endobj
+
+667 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 666 0 R
+  /K [15]
+  /Pg 1425 0 R
+>>
+endobj
+
+668 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 665 0 R
+  /K [14]
+  /Pg 1425 0 R
+>>
+endobj
+
+669 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 664 0 R
+  /K [672 0 R 670 0 R]
+>>
+endobj
+
+670 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 669 0 R
+  /K [671 0 R 12 13]
+  /Pg 1425 0 R
+>>
+endobj
+
+671 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 670 0 R
+  /K [11]
+  /Pg 1425 0 R
+>>
+endobj
+
+672 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 669 0 R
+  /K [10]
+  /Pg 1425 0 R
+>>
+endobj
+
+673 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 664 0 R
+  /K [676 0 R 674 0 R]
+>>
+endobj
+
+674 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 673 0 R
+  /K [675 0 R 8 9]
+  /Pg 1425 0 R
+>>
+endobj
+
+675 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 674 0 R
+  /K [7]
+  /Pg 1425 0 R
+>>
+endobj
+
+676 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 673 0 R
+  /K [6]
+  /Pg 1425 0 R
+>>
+endobj
+
+677 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T (Inteligencia artificial y agentes)
+  /K [4 5]
+  /Pg 1425 0 R
+>>
+endobj
+
+678 0 obj
+<<
+  /Type /StructElem
+  /S /L
+  /P 92 0 R
+  /A [<<
+    /O /List
+    /ListNumbering /Circle
+  >>]
+  /K [687 0 R 683 0 R 679 0 R]
+>>
+endobj
+
+679 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 678 0 R
+  /K [682 0 R 680 0 R]
+>>
+endobj
+
+680 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 679 0 R
+  /K [681 0 R 2 3]
+  /Pg 1425 0 R
+>>
+endobj
+
+681 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 680 0 R
+  /K [1]
+  /Pg 1425 0 R
+>>
+endobj
+
+682 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 679 0 R
+  /K [0]
+  /Pg 1425 0 R
+>>
+endobj
+
+683 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 678 0 R
+  /K [686 0 R 684 0 R]
+>>
+endobj
+
+684 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 683 0 R
+  /K [685 0 R 65 66]
+  /Pg 1423 0 R
+>>
+endobj
+
+685 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 684 0 R
+  /K [64]
+  /Pg 1423 0 R
+>>
+endobj
+
+686 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 683 0 R
+  /K [63]
+  /Pg 1423 0 R
+>>
+endobj
+
+687 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 678 0 R
+  /K [690 0 R 688 0 R]
+>>
+endobj
+
+688 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 687 0 R
+  /K [689 0 R 61 62]
+  /Pg 1423 0 R
+>>
+endobj
+
+689 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 688 0 R
+  /K [60]
+  /Pg 1423 0 R
+>>
+endobj
+
+690 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 687 0 R
+  /K [59]
+  /Pg 1423 0 R
+>>
+endobj
+
+691 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T (Industria y manufactura)
+  /K [57 58]
+  /Pg 1423 0 R
+>>
+endobj
+
+692 0 obj
+<<
+  /Type /StructElem
+  /S /L
+  /P 92 0 R
+  /A [<<
+    /O /List
+    /ListNumbering /Circle
+  >>]
+  /K [701 0 R 697 0 R 693 0 R]
+>>
+endobj
+
+693 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 692 0 R
+  /K [696 0 R 694 0 R]
+>>
+endobj
+
+694 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 693 0 R
+  /K [695 0 R 54 55 56]
+  /Pg 1423 0 R
+>>
+endobj
+
+695 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 694 0 R
+  /K [53]
+  /Pg 1423 0 R
+>>
+endobj
+
+696 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 693 0 R
+  /K [52]
+  /Pg 1423 0 R
+>>
+endobj
+
+697 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 692 0 R
+  /K [700 0 R 698 0 R]
+>>
+endobj
+
+698 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 697 0 R
+  /K [699 0 R 48 49 50 51]
+  /Pg 1423 0 R
+>>
+endobj
+
+699 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 698 0 R
+  /K [47]
+  /Pg 1423 0 R
+>>
+endobj
+
+700 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 697 0 R
+  /K [46]
+  /Pg 1423 0 R
+>>
+endobj
+
+701 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 692 0 R
+  /K [704 0 R 702 0 R]
+>>
+endobj
+
+702 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 701 0 R
+  /K [703 0 R 44 45]
+  /Pg 1423 0 R
+>>
+endobj
+
+703 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 702 0 R
+  /K [43]
+  /Pg 1423 0 R
+>>
+endobj
+
+704 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 701 0 R
+  /K [42]
+  /Pg 1423 0 R
+>>
+endobj
+
+705 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T (Grafos de conocimiento empresariales)
+  /K [40 41]
+  /Pg 1423 0 R
+>>
+endobj
+
+706 0 obj
+<<
+  /Type /StructElem
+  /S /L
+  /P 92 0 R
+  /A [<<
+    /O /List
+    /ListNumbering /Circle
+  >>]
+  /K [719 0 R 715 0 R 711 0 R 707 0 R]
+>>
+endobj
+
+707 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 706 0 R
+  /K [710 0 R 708 0 R]
+>>
+endobj
+
+708 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 707 0 R
+  /K [709 0 R 38 39]
+  /Pg 1423 0 R
+>>
+endobj
+
+709 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 708 0 R
+  /K [37]
+  /Pg 1423 0 R
+>>
+endobj
+
+710 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 707 0 R
+  /K [36]
+  /Pg 1423 0 R
+>>
+endobj
+
+711 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 706 0 R
+  /K [714 0 R 712 0 R]
+>>
+endobj
+
+712 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 711 0 R
+  /K [713 0 R 34 35]
+  /Pg 1423 0 R
+>>
+endobj
+
+713 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 712 0 R
+  /K [33]
+  /Pg 1423 0 R
+>>
+endobj
+
+714 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 711 0 R
+  /K [32]
+  /Pg 1423 0 R
+>>
+endobj
+
+715 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 706 0 R
+  /K [718 0 R 716 0 R]
+>>
+endobj
+
+716 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 715 0 R
+  /K [717 0 R 30 31]
+  /Pg 1423 0 R
+>>
+endobj
+
+717 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 716 0 R
+  /K [29]
+  /Pg 1423 0 R
+>>
+endobj
+
+718 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 715 0 R
+  /K [28]
+  /Pg 1423 0 R
+>>
+endobj
+
+719 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 706 0 R
+  /K [722 0 R 720 0 R]
+>>
+endobj
+
+720 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 719 0 R
+  /K [721 0 R 25 26 27]
+  /Pg 1423 0 R
+>>
+endobj
+
+721 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 720 0 R
+  /K [24]
+  /Pg 1423 0 R
+>>
+endobj
+
+722 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 719 0 R
+  /K [23]
+  /Pg 1423 0 R
+>>
+endobj
+
+723 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [22]
+  /Pg 1423 0 R
+>>
+endobj
+
+724 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T <FEFF00420069006F006D00650064006900630069006E00610020201400200065006C00200064006F006D0069006E0069006F0020006D00E100730020006D0061006400750072006F>
+  /K [20 21]
+  /Pg 1423 0 R
+>>
+endobj
+
+725 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T (Uso actual)
+  /K [18 19]
+  /Pg 1423 0 R
+>>
+endobj
+
+726 0 obj
+<<
+  /Type /StructElem
+  /S /Table
+  /P 92 0 R
+  /A [<<
+    /O /Layout
+    /BorderColor [0 0 0]
+    /BorderThickness 1
+  >>]
+  /K [764 0 R 727 0 R]
+>>
+endobj
+
+727 0 obj
+<<
+  /Type /StructElem
+  /S /TBody
+  /P 726 0 R
+  /K [761 0 R 758 0 R 755 0 R 752 0 R 749 0 R 746 0 R 743 0 R 740 0 R 737 0 R 734 0 R 731 0 R 728 0 R]
+>>
+endobj
+
+728 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 727 0 R
+  /K [730 0 R 729 0 R]
+>>
+endobj
+
+729 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 728 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [16 17]
+  /Pg 1423 0 R
+>>
+endobj
+
+730 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 728 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [15]
+  /Pg 1423 0 R
+>>
+endobj
+
+731 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 727 0 R
+  /K [733 0 R 732 0 R]
+>>
+endobj
+
+732 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 731 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [13 14]
+  /Pg 1423 0 R
+>>
+endobj
+
+733 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 731 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [12]
+  /Pg 1423 0 R
+>>
+endobj
+
+734 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 727 0 R
+  /K [736 0 R 735 0 R]
+>>
+endobj
+
+735 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 734 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [11]
+  /Pg 1423 0 R
+>>
+endobj
+
+736 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 734 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [10]
+  /Pg 1423 0 R
+>>
+endobj
+
+737 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 727 0 R
+  /K [739 0 R 738 0 R]
+>>
+endobj
+
+738 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 737 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [9]
+  /Pg 1423 0 R
+>>
+endobj
+
+739 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 737 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [8]
+  /Pg 1423 0 R
+>>
+endobj
+
+740 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 727 0 R
+  /K [742 0 R 741 0 R]
+>>
+endobj
+
+741 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 740 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [7]
+  /Pg 1423 0 R
+>>
+endobj
+
+742 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 740 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [6]
+  /Pg 1423 0 R
+>>
+endobj
+
+743 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 727 0 R
+  /K [745 0 R 744 0 R]
+>>
+endobj
+
+744 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 743 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [5]
+  /Pg 1423 0 R
+>>
+endobj
+
+745 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 743 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [4]
+  /Pg 1423 0 R
+>>
+endobj
+
+746 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 727 0 R
+  /K [748 0 R 747 0 R]
+>>
+endobj
+
+747 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 746 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [1 2 3]
+  /Pg 1423 0 R
+>>
+endobj
+
+748 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 746 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [0]
+  /Pg 1423 0 R
+>>
+endobj
+
+749 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 727 0 R
+  /K [751 0 R 750 0 R]
+>>
+endobj
+
+750 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 749 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [59 60]
+  /Pg 1421 0 R
+>>
+endobj
+
+751 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 749 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [58]
+  /Pg 1421 0 R
+>>
+endobj
+
+752 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 727 0 R
+  /K [754 0 R 753 0 R]
+>>
+endobj
+
+753 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 752 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [57]
+  /Pg 1421 0 R
+>>
+endobj
+
+754 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 752 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [56]
+  /Pg 1421 0 R
+>>
+endobj
+
+755 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 727 0 R
+  /K [757 0 R 756 0 R]
+>>
+endobj
+
+756 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 755 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [54 55]
+  /Pg 1421 0 R
+>>
+endobj
+
+757 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 755 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [53]
+  /Pg 1421 0 R
+>>
+endobj
+
+758 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 727 0 R
+  /K [760 0 R 759 0 R]
+>>
+endobj
+
+759 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 758 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [52]
+  /Pg 1421 0 R
+>>
+endobj
+
+760 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 758 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [51]
+  /Pg 1421 0 R
+>>
+endobj
+
+761 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 727 0 R
+  /K [763 0 R 762 0 R]
+>>
+endobj
+
+762 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 761 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [50]
+  /Pg 1421 0 R
+>>
+endobj
+
+763 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 761 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U2x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [49]
+  /Pg 1421 0 R
+>>
+endobj
+
+764 0 obj
+<<
+  /Type /StructElem
+  /S /THead
+  /P 726 0 R
+  /K [765 0 R]
+>>
+endobj
+
+765 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 764 0 R
+  /K [767 0 R 766 0 R]
+>>
+endobj
+
+766 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 765 0 R
+  /ID (U2x1y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [48]
+  /Pg 1421 0 R
+>>
+endobj
+
+767 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 765 0 R
+  /ID (U2x0y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [47]
+  /Pg 1421 0 R
+>>
+endobj
+
+768 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004C00ED006E00650061002000740065006D0070006F00720061006C00200063006F006E00640065006E0073006100640061>
+  /K [45 46]
+  /Pg 1421 0 R
+>>
+endobj
+
+769 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [41 42 43 44]
+  /Pg 1421 0 R
+>>
+endobj
+
+770 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [36 771 0 R 38 39 40]
+  /Pg 1421 0 R
+>>
+endobj
+
+771 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 770 0 R
+  /K [37]
+  /Pg 1421 0 R
+>>
+endobj
+
+772 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T <FEFF00540069006D0020004200650072006E006500720073002D004C00650065002C0020004A0061006D00650073002000480065006E0064006C00650072002000790020004F007200610020004C0061007300730069006C0061002020140020006C00610020005700650062002000530065006D00E1006E00740069006300610020002800320030003000310029>
+  /K [33 34 35]
+  /Pg 1421 0 R
+>>
+endobj
+
+773 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [30 31 32]
+  /Pg 1421 0 R
+>>
+endobj
+
+774 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [24 775 0 R 26 27 28 29]
+  /Pg 1421 0 R
+>>
+endobj
+
+775 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 774 0 R
+  /K [25]
+  /Pg 1421 0 R
+>>
+endobj
+
+776 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T <FEFF0044006F007500670020004C0065006E00610074002020140020004300790063003A00200065006C002000700072006F0079006500630074006F0020006D00E1007300200061006D0062006900630069006F0073006F002000280031003900380034201300320030003100370029>
+  /K [22 23]
+  /Pg 1421 0 R
+>>
+endobj
+
+777 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [11 780 0 R 13 14 15 779 0 R 17 18 778 0 R 20 21]
+  /Pg 1421 0 R
+>>
+endobj
+
+778 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 777 0 R
+  /K [19]
+  /Pg 1421 0 R
+>>
+endobj
+
+779 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 777 0 R
+  /K [16]
+  /Pg 1421 0 R
+>>
+endobj
+
+780 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 777 0 R
+  /K [12]
+  /Pg 1421 0 R
+>>
+endobj
+
+781 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [4 5 783 0 R 7 8 782 0 R 10]
+  /Pg 1421 0 R
+>>
+endobj
+
+782 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 781 0 R
+  /K [9]
+  /Pg 1421 0 R
+>>
+endobj
+
+783 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 781 0 R
+  /K [6]
+  /Pg 1421 0 R
+>>
+endobj
+
+784 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T <FEFF004E00690063006F006C0061002000470075006100720069006E006F002020140020006F006E0074006F006C006F006700ED006100200066006F0072006D0061006C00200065006E002000730069007300740065006D0061007300200064006500200069006E0066006F0072006D00610063006900F3006E00200028003100390039003000730029>
+  /K [2 3]
+  /Pg 1421 0 R
+>>
+endobj
+
+785 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [0 1]
+  /Pg 1421 0 R
+>>
+endobj
+
+786 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [52 53 787 0 R 55 56]
+  /Pg 1419 0 R
+>>
+endobj
+
+787 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 786 0 R
+  /K [54]
+  /Pg 1419 0 R
+>>
+endobj
+
+788 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [45 46 790 0 R 48 789 0 R 51]
+  /Pg 1419 0 R
+>>
+endobj
+
+789 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 788 0 R
+  /K [49 50]
+  /Pg 1419 0 R
+>>
+endobj
+
+790 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 788 0 R
+  /K [47]
+  /Pg 1419 0 R
+>>
+endobj
+
+791 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T <FEFF0054006F006D0020004700720075006200650072002020140020006C006100200064006500660069006E00690063006900F3006E002000630061006E00F3006E0069006300610020002800310039003900330029>
+  /K [43 44]
+  /Pg 1419 0 R
+>>
+endobj
+
+792 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [41 42]
+  /Pg 1419 0 R
+>>
+endobj
+
+793 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [32 795 0 R 34 35 794 0 R 37 38 39 40]
+  /Pg 1419 0 R
+>>
+endobj
+
+794 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 793 0 R
+  /K [36]
+  /Pg 1419 0 R
+>>
+endobj
+
+795 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 793 0 R
+  /K [33]
+  /Pg 1419 0 R
+>>
+endobj
+
+796 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T <FEFF004D0061007200760069006E0020004D0069006E0073006B0079002020140020006600720061006D006500730020007900200065007300740072007500630074007500720061007300200064006500200063006F006E006F00630069006D00690065006E0074006F0020002800310039003700340029>
+  /K [30 31]
+  /Pg 1419 0 R
+>>
+endobj
+
+797 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [27 28 29]
+  /Pg 1419 0 R
+>>
+endobj
+
+798 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [18 19 800 0 R 22 799 0 R 24 25 26]
+  /Pg 1419 0 R
+>>
+endobj
+
+799 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 798 0 R
+  /K [23]
+  /Pg 1419 0 R
+>>
+endobj
+
+800 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 798 0 R
+  /K [20 21]
+  /Pg 1419 0 R
+>>
+endobj
+
+801 0 obj
+<<
+  /Type /StructElem
+  /S /H3
+  /P 92 0 R
+  /T <FEFF004A006F0068006E0020004D00630043006100720074006800790020201400200065006C00200070007200650063007500720073006F00720020002800310039003500382013003100390038003000730029>
+  /K [16 17]
+  /Pg 1419 0 R
+>>
+endobj
+
+802 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T (Los fundadores)
+  /K [14 15]
+  /Pg 1419 0 R
+>>
+endobj
+
+803 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [12 13]
+  /Pg 1419 0 R
+>>
+endobj
+
+804 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [5 6 805 0 R 8 9 10 11]
+  /Pg 1419 0 R
+>>
+endobj
+
+805 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 804 0 R
+  /K [7]
+  /Pg 1419 0 R
+>>
+endobj
+
+806 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF0045006C002000700072006F0062006C0065006D00610020007100750065002000680061006200ED006100200071007500650020007200650073006F006C007600650072>
+  /K [3 4]
+  /Pg 1419 0 R
+>>
+endobj
+
+807 0 obj
+<<
+  /Type /StructElem
+  /S /H1
+  /P 92 0 R
+  /T <FEFF004C00610020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C00200065006E00200073006F006600740077006100720065003A0020006F0072006900670065006E002C0020006100750074006F00720065007300200079002000750073006F002000610063007400750061006C>
+  /K [0 1 2]
+  /Pg 1419 0 R
+>>
+endobj
+
+808 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [810 0 R 93 809 0 R 95 96 97 98 99 100]
+  /Pg 1417 0 R
+>>
+endobj
+
+809 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 808 0 R
+  /K [94]
+  /Pg 1417 0 R
+>>
+endobj
+
+810 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 808 0 R
+  /K [92]
+  /Pg 1417 0 R
+>>
+endobj
+
+811 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [812 0 R 87 88 89 90 91]
+  /Pg 1417 0 R
+>>
+endobj
+
+812 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 811 0 R
+  /K [86]
+  /Pg 1417 0 R
+>>
+endobj
+
+813 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [816 0 R 77 78 79 80 815 0 R 82 814 0 R 84 85]
+  /Pg 1417 0 R
+>>
+endobj
+
+814 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 813 0 R
+  /K [83]
+  /Pg 1417 0 R
+>>
+endobj
+
+815 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 813 0 R
+  /K [81]
+  /Pg 1417 0 R
+>>
+endobj
+
+816 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 813 0 R
+  /K [76]
+  /Pg 1417 0 R
+>>
+endobj
+
+817 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [820 0 R 68 69 819 0 R 71 818 0 R 73 74 75]
+  /Pg 1417 0 R
+>>
+endobj
+
+818 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 817 0 R
+  /Lang (en)
+  /K [72]
+  /Pg 1417 0 R
+>>
+endobj
+
+819 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 817 0 R
+  /Lang (en)
+  /K [70]
+  /Pg 1417 0 R
+>>
+endobj
+
+820 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 817 0 R
+  /K [67]
+  /Pg 1417 0 R
+>>
+endobj
+
+821 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T (Matices que la tabla no captura)
+  /K [65 66]
+  /Pg 1417 0 R
+>>
+endobj
+
+822 0 obj
+<<
+  /Type /StructElem
+  /S /Table
+  /P 92 0 R
+  /A [<<
+    /O /Layout
+    /BorderColor [0 0 0]
+    /BorderThickness 1
+  >>]
+  /K [973 0 R 823 0 R]
+>>
+endobj
+
+823 0 obj
+<<
+  /Type /StructElem
+  /S /TBody
+  /P 822 0 R
+  /K [963 0 R 953 0 R 943 0 R 931 0 R 924 0 R 914 0 R 905 0 R 899 0 R 891 0 R 884 0 R 877 0 R 870 0 R 862 0 R 855 0 R 846 0 R 839 0 R 832 0 R 824 0 R]
+>>
+endobj
+
+824 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [830 0 R 828 0 R 827 0 R 825 0 R]
+>>
+endobj
+
+825 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 824 0 R
+  /Lang (en)
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [826 0 R]
+>>
+endobj
+
+826 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 825 0 R
+  /K [63 64]
+  /Pg 1417 0 R
+>>
+endobj
+
+827 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 824 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [62]
+  /Pg 1417 0 R
+>>
+endobj
+
+828 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 824 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [60 829 0 R]
+  /Pg 1417 0 R
+>>
+endobj
+
+829 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 828 0 R
+  /K [61]
+  /Pg 1417 0 R
+>>
+endobj
+
+830 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 824 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [831 0 R]
+>>
+endobj
+
+831 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 830 0 R
+  /K [59]
+  /Pg 1417 0 R
+>>
+endobj
+
+832 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [837 0 R 835 0 R 834 0 R 833 0 R]
+>>
+endobj
+
+833 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 832 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [54 55 56 57 58]
+  /Pg 1417 0 R
+>>
+endobj
+
+834 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 832 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [52 53]
+  /Pg 1417 0 R
+>>
+endobj
+
+835 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 832 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [48 49 836 0 R]
+  /Pg 1417 0 R
+>>
+endobj
+
+836 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 835 0 R
+  /K [50 51]
+  /Pg 1417 0 R
+>>
+endobj
+
+837 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 832 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [838 0 R]
+>>
+endobj
+
+838 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 837 0 R
+  /K [47]
+  /Pg 1417 0 R
+>>
+endobj
+
+839 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [844 0 R 842 0 R 841 0 R 840 0 R]
+>>
+endobj
+
+840 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 839 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [44 45 46]
+  /Pg 1417 0 R
+>>
+endobj
+
+841 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 839 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [41 42 43]
+  /Pg 1417 0 R
+>>
+endobj
+
+842 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 839 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [39 843 0 R]
+  /Pg 1417 0 R
+>>
+endobj
+
+843 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 842 0 R
+  /K [40]
+  /Pg 1417 0 R
+>>
+endobj
+
+844 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 839 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [845 0 R]
+>>
+endobj
+
+845 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 844 0 R
+  /K [38]
+  /Pg 1417 0 R
+>>
+endobj
+
+846 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [853 0 R 851 0 R 850 0 R 847 0 R]
+>>
+endobj
+
+847 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 846 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [849 0 R 33 34 848 0 R 36 37]
+  /Pg 1417 0 R
+>>
+endobj
+
+848 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 847 0 R
+  /Lang (en)
+  /K [35]
+  /Pg 1417 0 R
+>>
+endobj
+
+849 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 847 0 R
+  /Lang (en)
+  /K [32]
+  /Pg 1417 0 R
+>>
+endobj
+
+850 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 846 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [29 30 31]
+  /Pg 1417 0 R
+>>
+endobj
+
+851 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 846 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [27 852 0 R]
+  /Pg 1417 0 R
+>>
+endobj
+
+852 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 851 0 R
+  /K [28]
+  /Pg 1417 0 R
+>>
+endobj
+
+853 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 846 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [854 0 R]
+>>
+endobj
+
+854 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 853 0 R
+  /K [25 26]
+  /Pg 1417 0 R
+>>
+endobj
+
+855 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [860 0 R 858 0 R 857 0 R 856 0 R]
+>>
+endobj
+
+856 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 855 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [21 22 23 24]
+  /Pg 1417 0 R
+>>
+endobj
+
+857 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 855 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [19 20]
+  /Pg 1417 0 R
+>>
+endobj
+
+858 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 855 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [16 859 0 R]
+  /Pg 1417 0 R
+>>
+endobj
+
+859 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 858 0 R
+  /K [17 18]
+  /Pg 1417 0 R
+>>
+endobj
+
+860 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 855 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [861 0 R]
+>>
+endobj
+
+861 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 860 0 R
+  /K [15]
+  /Pg 1417 0 R
+>>
+endobj
+
+862 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [868 0 R 866 0 R 865 0 R 863 0 R]
+>>
+endobj
+
+863 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 862 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [864 0 R 10 11 12 13 14]
+  /Pg 1417 0 R
+>>
+endobj
+
+864 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 863 0 R
+  /Lang (en)
+  /K [9]
+  /Pg 1417 0 R
+>>
+endobj
+
+865 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 862 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [7 8]
+  /Pg 1417 0 R
+>>
+endobj
+
+866 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 862 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [3 867 0 R]
+  /Pg 1417 0 R
+>>
+endobj
+
+867 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 866 0 R
+  /K [4 5 6]
+  /Pg 1417 0 R
+>>
+endobj
+
+868 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 862 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [869 0 R]
+>>
+endobj
+
+869 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 868 0 R
+  /K [0 1 2]
+  /Pg 1417 0 R
+>>
+endobj
+
+870 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [875 0 R 873 0 R 872 0 R 871 0 R]
+>>
+endobj
+
+871 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 870 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [155 156 157 158]
+  /Pg 1415 0 R
+>>
+endobj
+
+872 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 870 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [154]
+  /Pg 1415 0 R
+>>
+endobj
+
+873 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 870 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [150 874 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+874 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 873 0 R
+  /K [151 152 153]
+  /Pg 1415 0 R
+>>
+endobj
+
+875 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 870 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [876 0 R]
+>>
+endobj
+
+876 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 875 0 R
+  /K [149]
+  /Pg 1415 0 R
+>>
+endobj
+
+877 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [882 0 R 880 0 R 879 0 R 878 0 R]
+>>
+endobj
+
+878 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 877 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [145 146 147 148]
+  /Pg 1415 0 R
+>>
+endobj
+
+879 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 877 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [143 144]
+  /Pg 1415 0 R
+>>
+endobj
+
+880 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 877 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [140 881 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+881 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 880 0 R
+  /K [141 142]
+  /Pg 1415 0 R
+>>
+endobj
+
+882 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 877 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [883 0 R]
+>>
+endobj
+
+883 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 882 0 R
+  /K [138 139]
+  /Pg 1415 0 R
+>>
+endobj
+
+884 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [889 0 R 887 0 R 886 0 R 885 0 R]
+>>
+endobj
+
+885 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 884 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [134 135 136 137]
+  /Pg 1415 0 R
+>>
+endobj
+
+886 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 884 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [131 132 133]
+  /Pg 1415 0 R
+>>
+endobj
+
+887 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 884 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [127 888 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+888 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 887 0 R
+  /K [128 129 130]
+  /Pg 1415 0 R
+>>
+endobj
+
+889 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 884 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [890 0 R]
+>>
+endobj
+
+890 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 889 0 R
+  /K [125 126]
+  /Pg 1415 0 R
+>>
+endobj
+
+891 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [897 0 R 894 0 R 893 0 R 892 0 R]
+>>
+endobj
+
+892 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 891 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [121 122 123 124]
+  /Pg 1415 0 R
+>>
+endobj
+
+893 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 891 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [118 119 120]
+  /Pg 1415 0 R
+>>
+endobj
+
+894 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 891 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [110 896 0 R 114 895 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+895 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 894 0 R
+  /K [115 116 117]
+  /Pg 1415 0 R
+>>
+endobj
+
+896 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 894 0 R
+  /K [111 112 113]
+  /Pg 1415 0 R
+>>
+endobj
+
+897 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 891 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [898 0 R]
+>>
+endobj
+
+898 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 897 0 R
+  /K [109]
+  /Pg 1415 0 R
+>>
+endobj
+
+899 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [903 0 R 902 0 R 901 0 R 900 0 R]
+>>
+endobj
+
+900 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 899 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [105 106 107 108]
+  /Pg 1415 0 R
+>>
+endobj
+
+901 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 899 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [102 103 104]
+  /Pg 1415 0 R
+>>
+endobj
+
+902 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 899 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [100 101]
+  /Pg 1415 0 R
+>>
+endobj
+
+903 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 899 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [904 0 R]
+>>
+endobj
+
+904 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 903 0 R
+  /K [98 99]
+  /Pg 1415 0 R
+>>
+endobj
+
+905 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [912 0 R 910 0 R 909 0 R 906 0 R]
+>>
+endobj
+
+906 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 905 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [908 0 R 93 907 0 R 95 96 97]
+  /Pg 1415 0 R
+>>
+endobj
+
+907 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 906 0 R
+  /Lang (en)
+  /K [94]
+  /Pg 1415 0 R
+>>
+endobj
+
+908 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 906 0 R
+  /Lang (en)
+  /K [92]
+  /Pg 1415 0 R
+>>
+endobj
+
+909 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 905 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [91]
+  /Pg 1415 0 R
+>>
+endobj
+
+910 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 905 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [88 911 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+911 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 910 0 R
+  /K [89 90]
+  /Pg 1415 0 R
+>>
+endobj
+
+912 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 905 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [913 0 R]
+>>
+endobj
+
+913 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 912 0 R
+  /K [87]
+  /Pg 1415 0 R
+>>
+endobj
+
+914 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [922 0 R 919 0 R 918 0 R 915 0 R]
+>>
+endobj
+
+915 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 914 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [917 0 R 80 916 0 R 82 83 84 85 86]
+  /Pg 1415 0 R
+>>
+endobj
+
+916 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 915 0 R
+  /Lang (en)
+  /K [81]
+  /Pg 1415 0 R
+>>
+endobj
+
+917 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 915 0 R
+  /Lang (en)
+  /K [79]
+  /Pg 1415 0 R
+>>
+endobj
+
+918 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 914 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [76 77 78]
+  /Pg 1415 0 R
+>>
+endobj
+
+919 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 914 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [72 921 0 R 74 920 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+920 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 919 0 R
+  /K [75]
+  /Pg 1415 0 R
+>>
+endobj
+
+921 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 919 0 R
+  /K [73]
+  /Pg 1415 0 R
+>>
+endobj
+
+922 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 914 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [923 0 R]
+>>
+endobj
+
+923 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 922 0 R
+  /K [70 71]
+  /Pg 1415 0 R
+>>
+endobj
+
+924 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [929 0 R 927 0 R 926 0 R 925 0 R]
+>>
+endobj
+
+925 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 924 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [68 69]
+  /Pg 1415 0 R
+>>
+endobj
+
+926 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 924 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [67]
+  /Pg 1415 0 R
+>>
+endobj
+
+927 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 924 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [65 928 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+928 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 927 0 R
+  /K [66]
+  /Pg 1415 0 R
+>>
+endobj
+
+929 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 924 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [930 0 R]
+>>
+endobj
+
+930 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 929 0 R
+  /K [63 64]
+  /Pg 1415 0 R
+>>
+endobj
+
+931 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [941 0 R 939 0 R 938 0 R 932 0 R]
+>>
+endobj
+
+932 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 931 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [937 0 R 55 936 0 R 57 935 0 R 59 934 0 R 61 933 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+933 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 932 0 R
+  /Lang (en)
+  /K [62]
+  /Pg 1415 0 R
+>>
+endobj
+
+934 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 932 0 R
+  /Lang (en)
+  /K [60]
+  /Pg 1415 0 R
+>>
+endobj
+
+935 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 932 0 R
+  /Lang (en)
+  /K [58]
+  /Pg 1415 0 R
+>>
+endobj
+
+936 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 932 0 R
+  /Lang (en)
+  /K [56]
+  /Pg 1415 0 R
+>>
+endobj
+
+937 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 932 0 R
+  /Lang (en)
+  /K [54]
+  /Pg 1415 0 R
+>>
+endobj
+
+938 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 931 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [53]
+  /Pg 1415 0 R
+>>
+endobj
+
+939 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 931 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [51 940 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+940 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 939 0 R
+  /K [52]
+  /Pg 1415 0 R
+>>
+endobj
+
+941 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 931 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [942 0 R 48 49 50]
+  /Pg 1415 0 R
+>>
+endobj
+
+942 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 941 0 R
+  /K [47]
+  /Pg 1415 0 R
+>>
+endobj
+
+943 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [951 0 R 949 0 R 948 0 R 944 0 R]
+>>
+endobj
+
+944 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 943 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [947 0 R 43 946 0 R 45 945 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+945 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 944 0 R
+  /Lang (en)
+  /K [46]
+  /Pg 1415 0 R
+>>
+endobj
+
+946 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 944 0 R
+  /Lang (en)
+  /K [44]
+  /Pg 1415 0 R
+>>
+endobj
+
+947 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 944 0 R
+  /Lang (en)
+  /K [42]
+  /Pg 1415 0 R
+>>
+endobj
+
+948 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 943 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [40 41]
+  /Pg 1415 0 R
+>>
+endobj
+
+949 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 943 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [37 950 0 R 39]
+  /Pg 1415 0 R
+>>
+endobj
+
+950 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 949 0 R
+  /K [38]
+  /Pg 1415 0 R
+>>
+endobj
+
+951 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 943 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [952 0 R]
+>>
+endobj
+
+952 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 951 0 R
+  /K [36]
+  /Pg 1415 0 R
+>>
+endobj
+
+953 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [961 0 R 959 0 R 958 0 R 954 0 R]
+>>
+endobj
+
+954 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 953 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [957 0 R 31 956 0 R 33 34 955 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+955 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 954 0 R
+  /Lang (en)
+  /K [35]
+  /Pg 1415 0 R
+>>
+endobj
+
+956 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 954 0 R
+  /Lang (en)
+  /K [32]
+  /Pg 1415 0 R
+>>
+endobj
+
+957 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 954 0 R
+  /Lang (en)
+  /K [30]
+  /Pg 1415 0 R
+>>
+endobj
+
+958 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 953 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [27 28 29]
+  /Pg 1415 0 R
+>>
+endobj
+
+959 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 953 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [24 25 960 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+960 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 959 0 R
+  /K [26]
+  /Pg 1415 0 R
+>>
+endobj
+
+961 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 953 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [1 1 1]
+    /BorderStyle /Solid
+  >>]
+  /K [962 0 R]
+>>
+endobj
+
+962 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 961 0 R
+  /K [23]
+  /Pg 1415 0 R
+>>
+endobj
+
+963 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 823 0 R
+  /K [971 0 R 968 0 R 967 0 R 964 0 R]
+>>
+endobj
+
+964 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 963 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x3y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [18 966 0 R 20 965 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+965 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 964 0 R
+  /Lang (en)
+  /K [21 22]
+  /Pg 1415 0 R
+>>
+endobj
+
+966 0 obj
+<<
+  /Type /StructElem
+  /S /Code
+  /P 964 0 R
+  /Lang (en)
+  /K [19]
+  /Pg 1415 0 R
+>>
+endobj
+
+967 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 963 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x2y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [17]
+  /Pg 1415 0 R
+>>
+endobj
+
+968 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 963 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x1y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [11 970 0 R 13 969 0 R]
+  /Pg 1415 0 R
+>>
+endobj
+
+969 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 968 0 R
+  /K [14 15 16]
+  /Pg 1415 0 R
+>>
+endobj
+
+970 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 968 0 R
+  /K [12]
+  /Pg 1415 0 R
+>>
+endobj
+
+971 0 obj
+<<
+  /Type /StructElem
+  /S /TD
+  /P 963 0 R
+  /A [<<
+    /O /Table
+    /Headers [(U1x0y0)]
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.99215686 0.9647059 0.9411765]
+    /BorderStyle /Solid
+  >>]
+  /K [972 0 R]
+>>
+endobj
+
+972 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 971 0 R
+  /K [10]
+  /Pg 1415 0 R
+>>
+endobj
+
+973 0 obj
+<<
+  /Type /StructElem
+  /S /THead
+  /P 822 0 R
+  /K [974 0 R]
+>>
+endobj
+
+974 0 obj
+<<
+  /Type /StructElem
+  /S /TR
+  /P 973 0 R
+  /K [978 0 R 977 0 R 976 0 R 975 0 R]
+>>
+endobj
+
+975 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 974 0 R
+  /ID (U1x3y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [9]
+  /Pg 1415 0 R
+>>
+endobj
+
+976 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 974 0 R
+  /ID (U1x2y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [6 7 8]
+  /Pg 1415 0 R
+>>
+endobj
+
+977 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 974 0 R
+  /ID (U1x1y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [5]
+  /Pg 1415 0 R
+>>
+endobj
+
+978 0 obj
+<<
+  /Type /StructElem
+  /S /TH
+  /P 974 0 R
+  /ID (U1x0y0)
+  /A [<<
+    /O /Table
+    /Scope /Column
+    /Headers []
+  >> <<
+    /O /Layout
+    /BackgroundColor [0.7529412 0.2509804 0]
+    /BorderStyle /Solid
+  >>]
+  /K [4]
+  /Pg 1415 0 R
+>>
+endobj
+
+979 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [2 3]
+  /Pg 1415 0 R
+>>
+endobj
+
+980 0 obj
+<<
+  /Type /StructElem
+  /S /H1
+  /P 92 0 R
+  /T <FEFF00470065006E00650061006C006F006700ED0061002000660069006C006F007300F30066006900630061002000640065006C0020006D006F00640065006C006F0020006F007000650072006100630069006F006E0061006C>
+  /K [0 1]
+  /Pg 1415 0 R
+>>
+endobj
+
+981 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [14 15]
+  /Pg 1413 0 R
+>>
+endobj
+
+982 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [2 3 4 5 984 0 R 7 983 0 R 10 11 12 13]
+  /Pg 1413 0 R
+>>
+endobj
+
+983 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 982 0 R
+  /K [8 9]
+  /Pg 1413 0 R
+>>
+endobj
+
+984 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 982 0 R
+  /K [6]
+  /Pg 1413 0 R
+>>
+endobj
+
+985 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF0045006C002000650073007400610074007500730020006F006E0074006F006C00F3006700690063006F003A00200068006500630068006F007300200069006E0073007400690074007500630069006F006E0061006C00650073>
+  /K [0 1]
+  /Pg 1413 0 R
+>>
+endobj
+
+986 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [67 987 0 R 69 70 71 72]
+  /Pg 1411 0 R
+>>
+endobj
+
+987 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 986 0 R
+  /K [68]
+  /Pg 1411 0 R
+>>
+endobj
+
+988 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004400F3006E006400650020007300ED00200068006100790020007200650073006F006E0061006E00630069006100200063006F006E0020006C006F007300200075006E006900760065007200730061006C00650073>
+  /K [65 66]
+  /Pg 1411 0 R
+>>
+endobj
+
+989 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [50 992 0 R 52 53 991 0 R 55 56 57 58 59 60 61 990 0 R 63 64]
+  /Pg 1411 0 R
+>>
+endobj
+
+990 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 989 0 R
+  /K [62]
+  /Pg 1411 0 R
+>>
+endobj
+
+991 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 989 0 R
+  /K [54]
+  /Pg 1411 0 R
+>>
+endobj
+
+992 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 989 0 R
+  /K [51]
+  /Pg 1411 0 R
+>>
+endobj
+
+993 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [44 994 0 R 47 48 49]
+  /Pg 1411 0 R
+>>
+endobj
+
+994 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 993 0 R
+  /K [45 46]
+  /Pg 1411 0 R
+>>
+endobj
+
+995 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [39 40 996 0 R 43]
+  /Pg 1411 0 R
+>>
+endobj
+
+996 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 995 0 R
+  /K [41 42]
+  /Pg 1411 0 R
+>>
+endobj
+
+997 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T (Los invariantes como propiedades esenciales de un singular)
+  /K [37 38]
+  /Pg 1411 0 R
+>>
+endobj
+
+998 0 obj
+<<
+  /Type /StructElem
+  /S /L
+  /P 92 0 R
+  /A [<<
+    /O /List
+    /ListNumbering /Circle
+  >>]
+  /K [1013 0 R 1008 0 R 1003 0 R 999 0 R]
+>>
+endobj
+
+999 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 998 0 R
+  /K [1002 0 R 1000 0 R]
+>>
+endobj
+
+1000 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 999 0 R
+  /K [1001 0 R 35 36]
+  /Pg 1411 0 R
+>>
+endobj
+
+1001 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 1000 0 R
+  /K [34]
+  /Pg 1411 0 R
+>>
+endobj
+
+1002 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 999 0 R
+  /K [33]
+  /Pg 1411 0 R
+>>
+endobj
+
+1003 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 998 0 R
+  /K [1007 0 R 1004 0 R]
+>>
+endobj
+
+1004 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 1003 0 R
+  /K [1006 0 R 29 30 1005 0 R 32]
+  /Pg 1411 0 R
+>>
+endobj
+
+1005 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 1004 0 R
+  /K [31]
+  /Pg 1411 0 R
+>>
+endobj
+
+1006 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 1004 0 R
+  /K [28]
+  /Pg 1411 0 R
+>>
+endobj
+
+1007 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1003 0 R
+  /K [27]
+  /Pg 1411 0 R
+>>
+endobj
+
+1008 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 998 0 R
+  /K [1012 0 R 1009 0 R]
+>>
+endobj
+
+1009 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 1008 0 R
+  /K [1011 0 R 23 1010 0 R 25 26]
+  /Pg 1411 0 R
+>>
+endobj
+
+1010 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 1009 0 R
+  /K [24]
+  /Pg 1411 0 R
+>>
+endobj
+
+1011 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 1009 0 R
+  /K [22]
+  /Pg 1411 0 R
+>>
+endobj
+
+1012 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1008 0 R
+  /K [21]
+  /Pg 1411 0 R
+>>
+endobj
+
+1013 0 obj
+<<
+  /Type /StructElem
+  /S /LI
+  /P 998 0 R
+  /K [1016 0 R 1014 0 R]
+>>
+endobj
+
+1014 0 obj
+<<
+  /Type /StructElem
+  /S /LBody
+  /P 1013 0 R
+  /K [1015 0 R 19 20]
+  /Pg 1411 0 R
+>>
+endobj
+
+1015 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 1014 0 R
+  /K [18]
+  /Pg 1411 0 R
+>>
+endobj
+
+1016 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1013 0 R
+  /K [17]
+  /Pg 1411 0 R
+>>
+endobj
+
+1017 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [8 1019 0 R 10 1018 0 R 13 14 15 16]
+  /Pg 1411 0 R
+>>
+endobj
+
+1018 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 1017 0 R
+  /K [11 12]
+  /Pg 1411 0 R
+>>
+endobj
+
+1019 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 1017 0 R
+  /K [9]
+  /Pg 1411 0 R
+>>
+endobj
+
+1020 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 92 0 R
+  /T <FEFF004C006F007300200075006E006900760065007200730061006C0065007300200065006E0020006F006E0074006F006C006F006700ED0061002000660069006C006F007300F30066006900630061>
+  /K [6 7]
+  /Pg 1411 0 R
+>>
+endobj
+
+1021 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [4 5]
+  /Pg 1411 0 R
+>>
+endobj
+
+1022 0 obj
+<<
+  /Type /StructElem
+  /S /H1
+  /P 92 0 R
+  /T <FEFF00BF0053006F006E0020006C006F007300200069006E00760061007200690061006E00740065007300200065006C0020006500710075006900760061006C0065006E00740065002000610020006C006F007300200075006E006900760065007200730061006C00650073002000660069006C006F007300F3006600690063006F0073003F>
+  /K [0 1 2 3]
+  /Pg 1411 0 R
+>>
+endobj
+
+1023 0 obj
+<<
+  /Type /StructElem
+  /S /TOC
+  /P 92 0 R
+  /K [1208 0 R 1191 0 R 1187 0 R 1182 0 R 1178 0 R 1157 0 R 1153 0 R 1132 0 R 1128 0 R 1107 0 R 1103 0 R 1082 0 R 1078 0 R 1053 0 R 1049 0 R 1024 0 R]
+>>
+endobj
+
+1024 0 obj
+<<
+  /Type /StructElem
+  /S /TOC
+  /P 1023 0 R
+  /K [1045 0 R 1041 0 R 1037 0 R 1033 0 R 1029 0 R 1025 0 R]
+>>
+endobj
+
+1025 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1024 0 R
+  /K [1026 0 R]
+>>
+endobj
+
+1026 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1025 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1027 0 R]
+>>
+endobj
+
+1027 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1026 0 R
+  /K [1028 0 R 65 66 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1408 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1028 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1027 0 R
+  /K [64]
+  /Pg 1409 0 R
+>>
+endobj
+
+1029 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1024 0 R
+  /K [1030 0 R]
+>>
+endobj
+
+1030 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1029 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1031 0 R]
+>>
+endobj
+
+1031 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1030 0 R
+  /K [1032 0 R 62 63 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1407 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1032 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1031 0 R
+  /K [61]
+  /Pg 1409 0 R
+>>
+endobj
+
+1033 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1024 0 R
+  /K [1034 0 R]
+>>
+endobj
+
+1034 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1033 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1035 0 R]
+>>
+endobj
+
+1035 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1034 0 R
+  /K [1036 0 R 59 60 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1406 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1036 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1035 0 R
+  /K [58]
+  /Pg 1409 0 R
+>>
+endobj
+
+1037 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1024 0 R
+  /K [1038 0 R]
+>>
+endobj
+
+1038 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1037 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1039 0 R]
+>>
+endobj
+
+1039 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1038 0 R
+  /K [1040 0 R 56 57 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1405 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1040 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1039 0 R
+  /K [55]
+  /Pg 1409 0 R
+>>
+endobj
+
+1041 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1024 0 R
+  /K [1042 0 R]
+>>
+endobj
+
+1042 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1041 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1043 0 R]
+>>
+endobj
+
+1043 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1042 0 R
+  /K [1044 0 R 53 54 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1404 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1044 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1043 0 R
+  /K [52]
+  /Pg 1409 0 R
+>>
+endobj
+
+1045 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1024 0 R
+  /K [1046 0 R]
+>>
+endobj
+
+1046 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1045 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1047 0 R]
+>>
+endobj
+
+1047 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1046 0 R
+  /K [1048 0 R 50 51 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1403 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1048 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1047 0 R
+  /K [49]
+  /Pg 1409 0 R
+>>
+endobj
+
+1049 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1023 0 R
+  /K [1050 0 R]
+>>
+endobj
+
+1050 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1049 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1051 0 R]
+>>
+endobj
+
+1051 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1050 0 R
+  /K [1052 0 R 47 48 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1402 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1052 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1051 0 R
+  /K [46]
+  /Pg 1409 0 R
+>>
+endobj
+
+1053 0 obj
+<<
+  /Type /StructElem
+  /S /TOC
+  /P 1023 0 R
+  /K [1074 0 R 1070 0 R 1066 0 R 1062 0 R 1058 0 R 1054 0 R]
+>>
+endobj
+
+1054 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1053 0 R
+  /K [1055 0 R]
+>>
+endobj
+
+1055 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1054 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1056 0 R]
+>>
+endobj
+
+1056 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1055 0 R
+  /K [1057 0 R 44 45 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1401 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1057 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1056 0 R
+  /K [43]
+  /Pg 1409 0 R
+>>
+endobj
+
+1058 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1053 0 R
+  /K [1059 0 R]
+>>
+endobj
+
+1059 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1058 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1060 0 R]
+>>
+endobj
+
+1060 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1059 0 R
+  /K [1061 0 R 41 42 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1400 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1061 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1060 0 R
+  /K [40]
+  /Pg 1409 0 R
+>>
+endobj
+
+1062 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1053 0 R
+  /K [1063 0 R]
+>>
+endobj
+
+1063 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1062 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1064 0 R]
+>>
+endobj
+
+1064 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1063 0 R
+  /K [1065 0 R 38 39 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1399 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1065 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1064 0 R
+  /K [37]
+  /Pg 1409 0 R
+>>
+endobj
+
+1066 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1053 0 R
+  /K [1067 0 R]
+>>
+endobj
+
+1067 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1066 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1068 0 R]
+>>
+endobj
+
+1068 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1067 0 R
+  /K [1069 0 R 35 36 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1398 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1069 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1068 0 R
+  /K [34]
+  /Pg 1409 0 R
+>>
+endobj
+
+1070 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1053 0 R
+  /K [1071 0 R]
+>>
+endobj
+
+1071 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1070 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1072 0 R]
+>>
+endobj
+
+1072 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1071 0 R
+  /K [1073 0 R 32 33 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1397 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1073 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1072 0 R
+  /K [31]
+  /Pg 1409 0 R
+>>
+endobj
+
+1074 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1053 0 R
+  /K [1075 0 R]
+>>
+endobj
+
+1075 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1074 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1076 0 R]
+>>
+endobj
+
+1076 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1075 0 R
+  /K [1077 0 R 29 30 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1396 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1077 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1076 0 R
+  /K [28]
+  /Pg 1409 0 R
+>>
+endobj
+
+1078 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1023 0 R
+  /K [1079 0 R]
+>>
+endobj
+
+1079 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1078 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1080 0 R]
+>>
+endobj
+
+1080 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1079 0 R
+  /K [1081 0 R 26 27 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1395 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1081 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1080 0 R
+  /K [25]
+  /Pg 1409 0 R
+>>
+endobj
+
+1082 0 obj
+<<
+  /Type /StructElem
+  /S /TOC
+  /P 1023 0 R
+  /K [1099 0 R 1095 0 R 1091 0 R 1087 0 R 1083 0 R]
+>>
+endobj
+
+1083 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1082 0 R
+  /K [1084 0 R]
+>>
+endobj
+
+1084 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1083 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1085 0 R]
+>>
+endobj
+
+1085 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1084 0 R
+  /K [1086 0 R 23 24 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1394 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1086 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1085 0 R
+  /K [22]
+  /Pg 1409 0 R
+>>
+endobj
+
+1087 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1082 0 R
+  /K [1088 0 R]
+>>
+endobj
+
+1088 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1087 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1089 0 R]
+>>
+endobj
+
+1089 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1088 0 R
+  /K [1090 0 R 20 21 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1393 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1090 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1089 0 R
+  /K [19]
+  /Pg 1409 0 R
+>>
+endobj
+
+1091 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1082 0 R
+  /K [1092 0 R]
+>>
+endobj
+
+1092 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1091 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1093 0 R]
+>>
+endobj
+
+1093 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1092 0 R
+  /K [1094 0 R 17 18 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1392 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1094 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1093 0 R
+  /K [16]
+  /Pg 1409 0 R
+>>
+endobj
+
+1095 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1082 0 R
+  /K [1096 0 R]
+>>
+endobj
+
+1096 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1095 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1097 0 R]
+>>
+endobj
+
+1097 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1096 0 R
+  /K [1098 0 R 14 15 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1391 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1098 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1097 0 R
+  /K [13]
+  /Pg 1409 0 R
+>>
+endobj
+
+1099 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1082 0 R
+  /K [1100 0 R]
+>>
+endobj
+
+1100 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1099 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1101 0 R]
+>>
+endobj
+
+1101 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1100 0 R
+  /K [1102 0 R 10 11 12 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1390 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1102 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1101 0 R
+  /K [9]
+  /Pg 1409 0 R
+>>
+endobj
+
+1103 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1023 0 R
+  /K [1104 0 R]
+>>
+endobj
+
+1104 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1103 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1105 0 R]
+>>
+endobj
+
+1105 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1104 0 R
+  /K [1106 0 R 7 8 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1389 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1106 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1105 0 R
+  /K [6]
+  /Pg 1409 0 R
+>>
+endobj
+
+1107 0 obj
+<<
+  /Type /StructElem
+  /S /TOC
+  /P 1023 0 R
+  /K [1124 0 R 1120 0 R 1116 0 R 1112 0 R 1108 0 R]
+>>
+endobj
+
+1108 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1107 0 R
+  /K [1109 0 R]
+>>
+endobj
+
+1109 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1108 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1110 0 R]
+>>
+endobj
+
+1110 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1109 0 R
+  /K [1111 0 R 4 5 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1388 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1111 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1110 0 R
+  /K [3]
+  /Pg 1409 0 R
+>>
+endobj
+
+1112 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1107 0 R
+  /K [1113 0 R]
+>>
+endobj
+
+1113 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1112 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1114 0 R]
+>>
+endobj
+
+1114 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1113 0 R
+  /K [1115 0 R 1 2 <<
+    /Type /OBJR
+    /Pg 1409 0 R
+    /Obj 1387 0 R
+  >>]
+  /Pg 1409 0 R
+>>
+endobj
+
+1115 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1114 0 R
+  /K [0]
+  /Pg 1409 0 R
+>>
+endobj
+
+1116 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1107 0 R
+  /K [1117 0 R]
+>>
+endobj
+
+1117 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1116 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1118 0 R]
+>>
+endobj
+
+1118 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1117 0 R
+  /K [1119 0 R 73 74 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1384 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1119 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1118 0 R
+  /K [72]
+  /Pg 1385 0 R
+>>
+endobj
+
+1120 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1107 0 R
+  /K [1121 0 R]
+>>
+endobj
+
+1121 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1120 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1122 0 R]
+>>
+endobj
+
+1122 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1121 0 R
+  /K [1123 0 R 70 71 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1383 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1123 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1122 0 R
+  /K [69]
+  /Pg 1385 0 R
+>>
+endobj
+
+1124 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1107 0 R
+  /K [1125 0 R]
+>>
+endobj
+
+1125 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1124 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1126 0 R]
+>>
+endobj
+
+1126 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1125 0 R
+  /K [1127 0 R 67 68 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1382 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1127 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1126 0 R
+  /K [66]
+  /Pg 1385 0 R
+>>
+endobj
+
+1128 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1023 0 R
+  /K [1129 0 R]
+>>
+endobj
+
+1129 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1128 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1130 0 R]
+>>
+endobj
+
+1130 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1129 0 R
+  /K [1131 0 R 64 65 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1381 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1131 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1130 0 R
+  /K [63]
+  /Pg 1385 0 R
+>>
+endobj
+
+1132 0 obj
+<<
+  /Type /StructElem
+  /S /TOC
+  /P 1023 0 R
+  /K [1149 0 R 1145 0 R 1141 0 R 1137 0 R 1133 0 R]
+>>
+endobj
+
+1133 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1132 0 R
+  /K [1134 0 R]
+>>
+endobj
+
+1134 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1133 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1135 0 R]
+>>
+endobj
+
+1135 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1134 0 R
+  /K [1136 0 R 61 62 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1380 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1136 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1135 0 R
+  /K [60]
+  /Pg 1385 0 R
+>>
+endobj
+
+1137 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1132 0 R
+  /K [1138 0 R]
+>>
+endobj
+
+1138 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1137 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1139 0 R]
+>>
+endobj
+
+1139 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1138 0 R
+  /K [1140 0 R 58 59 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1379 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1140 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1139 0 R
+  /K [57]
+  /Pg 1385 0 R
+>>
+endobj
+
+1141 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1132 0 R
+  /K [1142 0 R]
+>>
+endobj
+
+1142 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1141 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1143 0 R]
+>>
+endobj
+
+1143 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1142 0 R
+  /K [1144 0 R 55 56 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1378 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1144 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1143 0 R
+  /K [54]
+  /Pg 1385 0 R
+>>
+endobj
+
+1145 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1132 0 R
+  /K [1146 0 R]
+>>
+endobj
+
+1146 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1145 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1147 0 R]
+>>
+endobj
+
+1147 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1146 0 R
+  /K [1148 0 R 52 53 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1377 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1148 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1147 0 R
+  /K [51]
+  /Pg 1385 0 R
+>>
+endobj
+
+1149 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1132 0 R
+  /K [1150 0 R]
+>>
+endobj
+
+1150 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1149 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1151 0 R]
+>>
+endobj
+
+1151 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1150 0 R
+  /K [1152 0 R 49 50 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1376 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1152 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1151 0 R
+  /K [48]
+  /Pg 1385 0 R
+>>
+endobj
+
+1153 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1023 0 R
+  /K [1154 0 R]
+>>
+endobj
+
+1154 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1153 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1155 0 R]
+>>
+endobj
+
+1155 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1154 0 R
+  /K [1156 0 R 46 47 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1375 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1156 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1155 0 R
+  /K [45]
+  /Pg 1385 0 R
+>>
+endobj
+
+1157 0 obj
+<<
+  /Type /StructElem
+  /S /TOC
+  /P 1023 0 R
+  /K [1174 0 R 1170 0 R 1166 0 R 1162 0 R 1158 0 R]
+>>
+endobj
+
+1158 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1157 0 R
+  /K [1159 0 R]
+>>
+endobj
+
+1159 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1158 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1160 0 R]
+>>
+endobj
+
+1160 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1159 0 R
+  /K [1161 0 R 43 44 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1374 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1161 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1160 0 R
+  /K [42]
+  /Pg 1385 0 R
+>>
+endobj
+
+1162 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1157 0 R
+  /K [1163 0 R]
+>>
+endobj
+
+1163 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1162 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1164 0 R]
+>>
+endobj
+
+1164 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1163 0 R
+  /K [1165 0 R 40 41 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1373 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1165 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1164 0 R
+  /K [39]
+  /Pg 1385 0 R
+>>
+endobj
+
+1166 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1157 0 R
+  /K [1167 0 R]
+>>
+endobj
+
+1167 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1166 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1168 0 R]
+>>
+endobj
+
+1168 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1167 0 R
+  /K [1169 0 R 37 38 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1372 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1169 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1168 0 R
+  /K [36]
+  /Pg 1385 0 R
+>>
+endobj
+
+1170 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1157 0 R
+  /K [1171 0 R]
+>>
+endobj
+
+1171 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1170 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1172 0 R]
+>>
+endobj
+
+1172 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1171 0 R
+  /K [1173 0 R 34 35 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1371 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1173 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1172 0 R
+  /K [33]
+  /Pg 1385 0 R
+>>
+endobj
+
+1174 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1157 0 R
+  /K [1175 0 R]
+>>
+endobj
+
+1175 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1174 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1176 0 R]
+>>
+endobj
+
+1176 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1175 0 R
+  /K [1177 0 R 31 32 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1370 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1177 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1176 0 R
+  /K [30]
+  /Pg 1385 0 R
+>>
+endobj
+
+1178 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1023 0 R
+  /K [1179 0 R]
+>>
+endobj
+
+1179 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1178 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1180 0 R]
+>>
+endobj
+
+1180 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1179 0 R
+  /K [1181 0 R 28 29 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1369 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1181 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1180 0 R
+  /K [27]
+  /Pg 1385 0 R
+>>
+endobj
+
+1182 0 obj
+<<
+  /Type /StructElem
+  /S /TOC
+  /P 1023 0 R
+  /K [1183 0 R]
+>>
+endobj
+
+1183 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1182 0 R
+  /K [1184 0 R]
+>>
+endobj
+
+1184 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1183 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1185 0 R]
+>>
+endobj
+
+1185 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1184 0 R
+  /K [1186 0 R 25 26 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1368 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1186 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1185 0 R
+  /K [24]
+  /Pg 1385 0 R
+>>
+endobj
+
+1187 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1023 0 R
+  /K [1188 0 R]
+>>
+endobj
+
+1188 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1187 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1189 0 R]
+>>
+endobj
+
+1189 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1188 0 R
+  /K [1190 0 R 22 23 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1367 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1190 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1189 0 R
+  /K [21]
+  /Pg 1385 0 R
+>>
+endobj
+
+1191 0 obj
+<<
+  /Type /StructElem
+  /S /TOC
+  /P 1023 0 R
+  /K [1204 0 R 1200 0 R 1196 0 R 1192 0 R]
+>>
+endobj
+
+1192 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1191 0 R
+  /K [1193 0 R]
+>>
+endobj
+
+1193 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1192 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1194 0 R]
+>>
+endobj
+
+1194 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1193 0 R
+  /K [1195 0 R 19 20 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1366 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1195 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1194 0 R
+  /K [18]
+  /Pg 1385 0 R
+>>
+endobj
+
+1196 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1191 0 R
+  /K [1197 0 R]
+>>
+endobj
+
+1197 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1196 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1198 0 R]
+>>
+endobj
+
+1198 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1197 0 R
+  /K [1199 0 R 16 17 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1365 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1199 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1198 0 R
+  /K [15]
+  /Pg 1385 0 R
+>>
+endobj
+
+1200 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1191 0 R
+  /K [1201 0 R]
+>>
+endobj
+
+1201 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1200 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1202 0 R]
+>>
+endobj
+
+1202 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1201 0 R
+  /K [1203 0 R 13 14 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1364 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1203 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1202 0 R
+  /K [12]
+  /Pg 1385 0 R
+>>
+endobj
+
+1204 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1191 0 R
+  /K [1205 0 R]
+>>
+endobj
+
+1205 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1204 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1206 0 R]
+>>
+endobj
+
+1206 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1205 0 R
+  /K [1207 0 R 10 11 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1363 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1207 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1206 0 R
+  /K [9]
+  /Pg 1385 0 R
+>>
+endobj
+
+1208 0 obj
+<<
+  /Type /StructElem
+  /S /TOCI
+  /P 1023 0 R
+  /K [1209 0 R]
+>>
+endobj
+
+1209 0 obj
+<<
+  /Type /StructElem
+  /S /Reference
+  /P 1208 0 R
+  /A [<<
+    /O /Layout
+    /Placement /Block
+  >>]
+  /K [1210 0 R]
+>>
+endobj
+
+1210 0 obj
+<<
+  /Type /StructElem
+  /S /Link
+  /P 1209 0 R
+  /K [1211 0 R 7 8 <<
+    /Type /OBJR
+    /Pg 1385 0 R
+    /Obj 1362 0 R
+  >>]
+  /Pg 1385 0 R
+>>
+endobj
+
+1211 0 obj
+<<
+  /Type /StructElem
+  /S /Lbl
+  /P 1210 0 R
+  /K [6]
+  /Pg 1385 0 R
+>>
+endobj
+
+1212 0 obj
+<<
+  /Type /StructElem
+  /S /H1
+  /P 92 0 R
+  /T (Contenidos)
+  /K [5]
+  /Pg 1385 0 R
+>>
+endobj
+
+1213 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [4]
+  /Pg 1385 0 R
+>>
+endobj
+
+1214 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [3]
+  /Pg 1385 0 R
+>>
+endobj
+
+1215 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [2]
+  /Pg 1385 0 R
+>>
+endobj
+
+1216 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 92 0 R
+  /K [0 1]
+  /Pg 1385 0 R
+>>
+endobj
+
+1217 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 1
+>>
+endobj
+
+1218 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 2
+>>
+endobj
+
+1219 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 3
+>>
+endobj
+
+1220 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 4
+>>
+endobj
+
+1221 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 5
+>>
+endobj
+
+1222 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 6
+>>
+endobj
+
+1223 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 7
+>>
+endobj
+
+1224 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 8
+>>
+endobj
+
+1225 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 9
+>>
+endobj
+
+1226 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 10
+>>
+endobj
+
+1227 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 11
+>>
+endobj
+
+1228 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 12
+>>
+endobj
+
+1229 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 13
+>>
+endobj
+
+1230 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 14
+>>
+endobj
+
+1231 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 15
+>>
+endobj
+
+1232 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 16
+>>
+endobj
+
+1233 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 17
+>>
+endobj
+
+1234 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 18
+>>
+endobj
+
+1235 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 19
+>>
+endobj
+
+1236 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 20
+>>
+endobj
+
+1237 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 21
+>>
+endobj
+
+1238 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 22
+>>
+endobj
+
+1239 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 23
+>>
+endobj
+
+1240 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 24
+>>
+endobj
+
+1241 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 25
+>>
+endobj
+
+1242 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 26
+>>
+endobj
+
+1243 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 27
+>>
+endobj
+
+1244 0 obj
+<<
+  /Type /PageLabel
+  /S /D
+  /St 28
+>>
+endobj
+
+1245 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /EWLGLO+NewCM10-Bold-Identity-H
+  /Encoding /Identity-H
+  /DescendantFonts [1246 0 R]
+  /ToUnicode 1249 0 R
+>>
+endobj
+
+1246 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType0
+  /BaseFont /EWLGLO+NewCM10-Bold
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 1248 0 R
+  /DW 0
+  /W [0 0 280 1 1 864 2 2 639 3 3 447 4 4 575 5 5 319 6 6 575 7 7 319 8 8 559 9 9 383 10 10 639 11 11 454 12 12 575 13 13 510.99997 14 14 607 15 15 958 16 16 639 17 17 527 18 18 639 19 19 474 20 20 319 21 21 797 22 23 831 24 24 575 25 25 319 26 26 543 27 27 639 28 29 607 30 30 639 31 31 383 32 32 543 33 33 692 34 34 863 35 35 900 36 37 575 38 38 882 39 39 639 40 40 575 41 41 756 42 42 319 43 43 904 44 44 869 45 45 575 46 46 436 47 47 885 48 48 786 49 49 527 50 50 1092 51 51 639 52 52 351 53 53 894 54 54 594 55 55 869 56 56 900 57 57 351 58 58 724 59 59 1189 60 60 901 61 61 607 62 62 319 63 63 1150 64 64 447 65 69 575 70 70 447 71 71 575 72 72 800 73 73 559 74 74 575 75 75 818 76 76 639 77 77 607 78 78 869 79 79 756 80 80 1150 81 81 639 82 82 510.99997 83 84 639 85 85 703 86 86 894 87 87 864 88 88 575]
+>>
+endobj
+
+1247 0 obj
+<<
+  /Length 12
+  /Filter /FlateDecode
+>>
+stream
+x���M?v
+endstream
+endobj
+
+1248 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /EWLGLO+NewCM10-Bold
+  /Flags 131076
+  /FontBBox [-56 -250 1164 909]
+  /ItalicAngle 0
+  /Ascent 806
+  /Descent -194
+  /CapHeight 686
+  /StemV 168.6
+  /CIDSet 1247 0 R
+  /FontFile3 1250 0 R
+>>
+endobj
+
+1249 0 obj
+<<
+  /Length 1850
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+88 beginbfchar
+<0001> <004F>
+<0002> <006E>
+<0003> <0074>
+<0004> <006F>
+<0005> <006C>
+<0006> <0067>
+<0007> <00ED>
+<0008> <0061>
+<0009> <0020>
+<000A> <00660069>
+<000B> <0073>
+<000C> <00F3>
+<000D> <0063>
+<000E> <0079>
+<000F> <006D>
+<0010> <0064>
+<0011> <0065>
+<0012> <0070>
+<0013> <0072>
+<0014> <0069>
+<0015> <00660074>
+<0016> <0077>
+<0017> <0043>
+<0018> <0031>
+<0019> <002E>
+<001A> <00BF>
+<001B> <0053>
+<001C> <0076>
+<001D> <0071>
+<001E> <0075>
+<001F> <00AD>
+<0020> <003F>
+<0021> <004C>
+<0022> <0052>
+<0023> <004E>
+<0024> <0032>
+<0025> <0033>
+<0026> <0044>
+<0027> <0068>
+<0028> <0034>
+<0029> <0045>
+<002A> <003A>
+<002B> <0047>
+<002C> <0041>
+<002D> <002F>
+<002E> <0049>
+<002F> <0055>
+<0030> <0050>
+<0031> <00E9>
+<0032> <004D>
+<0033> <0062>
+<0034> <006A>
+<0035> <003D>
+<0036> <004A>
+<0037> <0056>
+<0038> <0048>
+<0039> <0066>
+<003A> <0046>
+<003B> <0057>
+<003C> <004B>
+<003D> <006B>
+<003E> <002C>
+<003F> <2014>
+<0040> <0028>
+<0041> <0039>
+<0042> <0035>
+<0043> <0038>
+<0044> <2013>
+<0045> <0030>
+<0046> <0029>
+<0047> <0037>
+<0048> <0054>
+<0049> <00E1>
+<004A> <0036>
+<004B> <0042>
+<004C> <00F1>
+<004D> <0078>
+<004E> <0059>
+<004F> <00C9>
+<0050> <2192>
+<0051> <0066006C>
+<0052> <007A>
+<0053> <00AB>
+<0054> <00BB>
+<0055> <005A>
+<0056> <0026>
+<0057> <0051>
+<0058> <00F6>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+1250 0 obj
+<<
+  /Length 7988
+  /Filter /FlateDecode
+  /Subtype /CIDFontType0C
+>>
+stream
+x��zyT׶w�mu��B٘8T�QqBI�hg#�Cpj���{7�<	2�
+
+�R�c4F��7Fc��M�W���N�
+ԛ��{���[���,�����g�߰ϑ�:w�����=�f,x��q����t�Y�+��n��_�w�i:>���~>e�O��_�L�Ln�C���/�����J�h�
��,�6X��˕��i�~<�{�{G�1��M�7�v|��M'�i��~�^A�˂�|��z{��_��d��m�d������n���h��a?k��e��~����6z�y�;:��yx�{��3ztp��Q~��Go��
���P�h�{�΋.s�?g�̅Kg�
+���h�����4J&�$��V��%���刋�p]Iv����y�_�.iݺB�W���-�9%՜�ѭ*_����X2SS�u7wS)dr�L&[%����,���L&�,E��V�J��β�e�@y���Sa�G�8E|gY�:fc`��Lg	O�?��eU�G]c���z��n���u��ݹ���_��������EϨ�Wm޶ɵm߰��[ș��z9��TOUa���\o��n�?|uҫ����ZA��}.�}�/�7���_Y��������|��&���>k��
��x��a���g�3�
19�tx@�Q�I��M�(?���4tT�����03�M7�u�g�,��0}熫[�$\��p���O�N��8�,|��8g߼��0��c�}�`Hi�Сt݈��u�H�q0����Wo�JTR��F%�[�v�3�O�78�#�eI�J(c�O���2g2�]�5�3�Zm-hk�=��n�xOI���u�dTt���.h�m��R+`W��$�bO�1�Jv�w���p���j"��'t.�*�&���C
d�����*�"���K��Q����3��Bl@_5��������hϯ�]���c���^9�@e����+‘�=E{����~�U�Y��`
Qѣ �w��0[$(��b����֗�>�3_��zX�RiJk�q���8��z���P1��Q�r{�m�I�xB�.��[�G�ְޓpt�n%Ӏ�e�`&l(�PL�u~��-J��� :uu�w�az(�'+OLE-μ~��>��g��E0`-$�	�c���>�
٩��3[��+���p
�����
+JRr�Ęd�=	,���+��5�{�z��stc����p��a�:
w�2�����˩����壏!��.�d`��=D��7!j��R��K!:%��#%T��q��o>�b�b	�L_ݵ	G�`
+�Js�����
+��G��Va��h�cQɅ����e��/4C��f��	����Ǭ����-��î�BC���Zs.�+unS4^,'�g��@���|BS�~…��ڿ�ŹP��k�Usˆѱ�+7��
e�f�aQѸ��x"���3��<��:ʸ��#;�Y�����d������o�j(I*�%������Ľ:�~ouE���gi'ړr�]x���+�IG�`���g�=-OO
��e^M�׋}km�]|��n�;��WCsX��N�j��Iy�2��g�A���盶�mN*
Rie���"Ȋ�����
y��a+�Q�7l,�/�l�ܧ�C�K��+�����
�⚓� H�5TΆC�>R�d;�,dZ��B1�8����7�M������}����B�.����G�ɘ0M�J T���
��5uPh�3DDC!��ȡ۔��OgM�,%���#q��I��Qm��� /EV!^�6ȸ���Z��$�Q��(�e�
��@�О\d��	Y�=�=�
�9���Z�udwP���o��Ja�G�5��L�I���Z&#ߒq�
+���}P�n���~~����l�;��U
k�x�XX��{~<��-<7�)<�zfOi����_�dT%R���z|_*|)�i��|�z`$B�ħ����+�U
+ҷ�C���ߎġ��p�����fQ�96bò)@U�.�!8^�P�_Ȯ����]�KTN�F��ߩ���+���V��]�;�!��c�>*t��@��vD�'���P����U�7ջ�pBǎDmh���8
����PÉ#���&(S��KvW66�V�.�B{�=����8�^�ލ��P�i؋�!>� 5��t�⧂�Op�V�{�r���w
]�4{����x5�ˎ��b.ZXw\�f("��B�ٜ�O��MG�K���`!�ը� ˁ�dx�2*{�]P�����ǖW��g��U���3/��'��샳Po �X��^�@�黡6����-i�
��N�_�U��kz	��ع�����f�:WWOOW�:�f����Y��Y�F�������(xA��:����t�c%v��U��Ad�u	�N\��+wՖ4��|p�0*s�?�%h��<�`8�b�#:IH���3��P&�i�;T-U�Wa�8X����s�q�)����ZM:
z���YT��?�
+Zi`��-%�L��ؒU`����&�e1���*������+����* 5;��<��g4���� |Ƥ�f�j�Nl%��p\+p��20L]�h��c!L�����pL���+NkO�]����e����W9G�p�Q��R�n6�
�4�Tj�{��A�`4&$�9���g���PG:��Q7t��0E�����jT �z�l����o'P{�[
+�
���8y� �-V�	�"bͧ-{��Z���
C��B'�
tR����(R��T[����.�d{{E���2�p��Ο$�&�m*��	��æ��B�j�w!
��,�<��c9�%��θ
+��nH
+��>��
0�v��s���������G��� ��l{��pA���V�x׫
��p*"tp
�1k�8
�;�����9<�*��R����hq^G��d�m#��3���Zt$�
a��%!���$�<���BCi(�m���Dܚ��a
+Б�	8̴�B^����͂�����œ[}O/���g�y�Q�:2�9-}_�RQ�!AƤm�j�V���fX�1��y���m�.!��Lz~j;�螃_�rMC�B|D]Ԃ�x�2��l]�k���ޤ:��X�hZ�`~K#�+��J���0D�n\�R�ȂE�k8����CWi:8幒T�]Q�i%��Qd�����,e�^�2}���8������1������J��\'�X����y��/�t�C��pV5�IL��=�6qR�]�^�C1P�T�j���f��yk�FU�֔6� ��͕EM�-�-��A�������s�Ie�^�W�����}��8�P|L�Kt�u۟
�e�b��ďE[o�Rk�	Α�M:� ����
T�a=
Ѣ�z�ʴ�V�F]}�
������A��w����RI���i�)6I��A�Ť�Ɯ��Li1�ո �u�8G�N��b��S�2%��~���H�a�h�k���ƫ��a:́��w�@ș�& ?��-�PA�^
+��<%X�]�8.e�B�
+��������m��8��z��;��
+�`!�
+v�v�
f���=]��
+�FHs�H	!&��!�`? �Z�>�Ih��p�Dd# ���k1����X,&Yo2�qfS
�*�*ӈWX(5�wZ�-)��
+�FlO� tF^�]1�Z�!,~^L�N?��P��p���n	c���
+�� ��e:�vw\���C�嵍��y8�V�Z]�/�nC�,�
+��f�G�Z���Ɓ�H�8�U��H�6�BG�5�c����2��������=����I�M�M9C�W&�{׵����ڽ<|�<'�\��@f
��<��&��u��єd�f2n=�Ȇ�
%��0l���	��kv�Ϊ��&rлf�����7.B5��.���F��/V����A��!�U�?�j���N��3��
+�cܢ�f5Py����5��D-��A��C���fx�̺���`�L�K
�J��&Tr7~�eX�~�1�=�od���o{ �\W���%��JTr��;a8�
+,wo����ʘ�'U�).���s@�p����ig)���f ���w�n��j
+��3*�A�:
���3�	�����By�7� t$֝P~��:M#��rZ��p~f�H%�����;����
�(�ޤ2m�-0�����M����6u�#�b�N��A��!$����ǹֶ��XO�Xp��a,��d����=8 �[�
�`\��3�'b/�7���?���x����p�OT��?���5��ԁn�kp0�9���$���E���
$�k�V\˪0"��>���4��at�q	~���A���=��6$L
�Ub��^Xx�4����;��Gt�H-��������^�D5
Tb
�:{��Q�m��Q91|($��c��ǯ���8H��]>�@Z^�v"��y��DE������@��T
+��0XkaX;���z�*�l���@<�PI��:A��#YD�1E����,J嫩��0b�xq,ʤ�Wp�a/�`$K_C��F\�%v��h�;�É��ل�Ռ��.����NC�����q/jY���������ę,j���h���{0���ilžJSfh���Y����tD�� 6����(~���L��2��!�] ���J�(V2�
+����L#
+f+�����ܤ(����YkEj�b+�f��4AܗHU�=b��((5�r~�,����tX�F[~k-�5և[PT���(3�u� 8��sG	ZX��a��f���m�	��Uu��O�hw�IT�X�g|���-E���De(�,vS��c������B)l7�&C,��vU�E|$�D�姕�vRP��#�󒏼��o?8�L�����g��e�q�iy��U]���/��>೵r5mS��հKC�:��8��)�8_��S�5m��k�F�
+�Ջ?�2t�>�O!�����-pk��`�e�a�P>�X|�	�;�Z�e��������结ӈ1�䅶��m7�85�10a�[����+�i ِd2�κ�z�A�
��0\�ƃ��h�e��9����
+ r�	�����ԬKR�Ձ�
�>g��8P���e"���i.���Z����5ڀ�i��h����?��
+q�ئ�O\5+����š�g�X&`��Y����h/�S�����?����@+�\2��e���^�c��p~���D�e���.D%~ڱ8�:�ۗ���i\�6K�S
��ZG�T�R:S\�]��8�/������6
����y�3�Q/yz
��N����j�x�I/4��X�SJ�@Z����h��힝�˺�!5���2�	Q˔����
+ʍ�%�0�c�yPI�q�,�*��KR��
�G+����&n��L��}DBx�h�.�F�L"H���w��9�p��[vB$�
+���B'�f�{���de�"���@BY�;��Ő�@��t�k��4X7v��8��,3�8�<��LV\AH|��ZvW�W�4 ��Ǵ���f�$Zjԟz�|�����*�#|-Z,	T>�%4B�d0��l�ٖ���m��*�uW��:ǠN}�M�ª��=T��4N�fH޷j�i7��׸o�k�UG�
섊���fH��m���M��:���t>j��v&���7���df[r4�2_Q��$e��mp1�㞊e���<`.�F}�����S�_Bƾ$���񋣽"�]�9ClN�ZL������p��	����1��T�h�o��,
���z~���<�t$\��|������n����Fr��.��
�=�}&�-8��I�ǵjd'ޣCy_p1h�	���L�1�
�C�E�"�1�	t�KB\��0\���-&�
H�4� 
��|��fd6��nrƼ;h7M�Os;I���S�ˡ�s��"~�[��ž<�}tUR��
��n/���W���N�e�H
(/��I�k�k���G��TI53'��a��������M��4B��l��;���;i�k�xI
����:���W{ȅ�����fi�;%�W����[r�:׏5��?�9�7Tڽ�2E�(%���V������^.�KT�_�M�@�To��Æ�i�11#�i�\5,��1���l�Ͱ�4�R�Q�����'�Gs�\�.��‹�n�
�6�RQ�P�ߠE
h�;io�A��|8$�6� 	��QhK{��3��
t�J��D$:�d�$�e2����Q�N8���3���Ц6���'h�=O������`�i1����,c6�J�G�4C&�j^^jJN�
(���!�!#h��/Сd��^ ���[X8d9�!d�/9U�Dp4�f�s �
+�Z���~
q &�6��E�/� ��`��ɾ���\�^��Z�q*��T�q	���WY|��ͦ����{�ő�F>��I=��p�Džhc���h?��·�Rp�B�X@eIs��r�x�i�
{e��7'E��V�U���P^�����=���01��&ga�ՠ@P���G�=��R�C`�~�7$D��%�A��p�S4*Ԃ�뱰�����@ʢ��M���ݸ���,‰(+��#�F��9�[�_�uw%�s��G:o�'�Y0��NX�s��wL\��V���q@{�A<���EY�q%��`)���{�6ē�/��1��yL����CU�g�Jۺ�9c@åo���A��5�pƵ��Sb ����u:��:��+θkOHڻ��C����"c>쀪Ԫ����]99/8Q.~a'��҉�������q�
�$�ˇ\cnb����d��0d��iw�d
�E@':���Ld�fX��Z�@`"��ҥ�kJF&y�r�D��%%��¨0�r@Rb'�h��!8����&I�ATn3&K���Et']����d��pX�v����h�W!���u �0�����	��”�BI���,蠶��̽~u	q�q�>A��o
"�n��qfGzYF�dȪKB��������Т��K�̚м�Aw����8 +�Ec眛C~�LX�T�i;��?�mn�iVB����C�δ��|s��Ҏ�l���/�_�,����z�n}[_6޴,9������
+�p Ԏ��@YͿؖ�@�8K��aa������ow|�sJ�����n4��8}DA\jhA,��I�QDe�b�z�!
/	
+q�Sk��C��W��܅�0�9))�EJb ��* ���IRS�ȗ@k�2&�}˅��A/�T�JJB�XW�c�x�P}5��gu�U^�~�d�����F��	+��^l0�!�S�-��N�=!%)-9H�23==G�.v��Wt�W���t��S.��oӓ,z 	��~�f���iuΞ�
Ӣ�]�>��T�CZjJZ��cWΟ����B*�S
B]����P}	G�1��"�?DoA�����TCU�
��u�&��f�!:�9�Nr�����h��:�z�I��H@6���N�
+�u�"`H��!��7�H��}�ֱ�m�h��7��-[
+UP�m�E�Θn����s��YS���	�1��&_@Y����L��gB#�|�Sh�6��Q��b'���N7��\����-��V�<uLb8ha���Z��{ZΓ/�&�x7b}g�a���ً�}\u�)8Cn-�h͛����w��~$J�h�Q^�
!�hSuݜSc��"*4Qt��x�^45ڞE�Q�u谵����%A�	;"�h0ƚ�L�[���~�	g�q(������+��C0�/����̝�)�xȏu��S
u�ǒ��F�	ٽ��lX	k�`
�ͺ�$��
G�uh�{/�����~����Co�m�	��!u�nk�?�e���U�E�5;a��x>�8�r����{h�]�~����3+|ǡ�~h��G�*u�O������G�w�PS�O���l�?^V�?wC
+endstream
+endobj
+
+1251 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /SAXHEJ+NewCM10-Italic-Identity-H
+  /Encoding /Identity-H
+  /DescendantFonts [1252 0 R]
+  /ToUnicode 1255 0 R
+>>
+endobj
+
+1252 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType0
+  /BaseFont /SAXHEJ+NewCM10-Italic
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 1254 0 R
+  /DW 0
+  /W [0 0 280 1 1 774 2 2 460 3 3 562 4 4 510.99997 5 5 256 6 6 510.99997 7 7 460 8 8 307 9 9 358 10 10 460 11 11 510.99997 12 12 332 13 13 537 14 14 307 15 15 422 16 16 409 17 17 486 18 18 818 19 20 510.99997 21 22 460 23 23 537 24 24 307 25 27 510.99997 28 28 460 29 29 307 30 30 743 31 31 460 32 32 627 33 33 510.99997 34 34 729 35 35 678 36 36 358 37 37 897 38 38 307 39 39 716 40 40 678 41 41 653 42 42 716 43 43 386 44 44 562 45 45 704 46 46 613 47 47 464 48 48 743 49 49 664 50 50 460 51 51 769 52 52 562 53 53 409 54 54 562 55 55 767 56 56 562 57 57 307 58 58 613 59 59 583 60 60 767 61 61 307 62 62 999 63 63 525 64 64 743 65 65 755 66 66 743 67 67 510.99997 68 68 743]
+>>
+endobj
+
+1253 0 obj
+<<
+  /Length 13
+  /Filter /FlateDecode
+>>
+stream
+x���~,��
+endstream
+endobj
+
+1254 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /SAXHEJ+NewCM10-Italic
+  /Flags 131140
+  /FontBBox [-35 -216 1124 716]
+  /ItalicAngle -14.036209
+  /Ascent 806
+  /Descent -194
+  /CapHeight 683
+  /StemV 95.4
+  /CIDSet 1253 0 R
+  /FontFile3 1256 0 R
+>>
+endobj
+
+1255 0 obj
+<<
+  /Length 1570
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+68 beginbfchar
+<0001> <0047>
+<0002> <0065>
+<0003> <006E>
+<0004> <0061>
+<0005> <006C>
+<0006> <006F>
+<0007> <0067>
+<0008> <00ED>
+<0009> <0020>
+<000A> <0063>
+<000B> <0070>
+<000C> <0074>
+<000D> <0075>
+<000E> <002C>
+<000F> <0072>
+<0010> <0073>
+<0011> <0079>
+<0012> <006D>
+<0013> <00E1>
+<0014> <00BF>
+<0015> <0071>
+<0016> <00E9>
+<0017> <00FA>
+<0018> <0069>
+<0019> <0068>
+<001A> <0064>
+<001B> <003F>
+<001C> <0062>
+<001D> <006A>
+<001E> <004E>
+<001F> <0076>
+<0020> <004C>
+<0021> <00F3>
+<0022> <0052>
+<0023> <0050>
+<0024> <00AD>
+<0025> <004D>
+<0026> <0066>
+<0027> <0043>
+<0028> <0045>
+<0029> <0046>
+<002A> <0054>
+<002B> <0049>
+<002C> <00660069>
+<002D> <0042>
+<002E> <00660066>
+<002F> <0078>
+<0030> <0041>
+<0031> <0077>
+<0032> <006B>
+<0033> <004B>
+<0034> <00AB>
+<0035> <007A>
+<0036> <00BB>
+<0037> <004F>
+<0038> <0053>
+<0039> <002E>
+<003A> <005A>
+<003B> <0069006A>
+<003C> <0051>
+<003D> <003A>
+<003E> <0057>
+<003F> <004A>
+<0040> <0056>
+<0041> <0044>
+<0042> <0055>
+<0043> <00E4>
+<0044> <0048>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+1256 0 obj
+<<
+  /Length 7699
+  /Filter /FlateDecode
+  /Subtype /CIDFontType0C
+>>
+stream
+x��z	t�U�~Cy�E�jcp�� " 
+�#� ( Ⱦ��}K�5m�&m�<I�����M�Җ��]A@@QT@����,"޷�V���+�
��|�zNz����������<Ͻ���Ã���brg���C��ϊHJ���j.ww'F��]�� 10�9j������	�s���]AA��7�?�n�+(��5��ÂQ���f��7(X 	����|tLJVBVެ��Ԕ��1�3S�R��2��Y��9&75C2&!sLFLRLDfL���蘌1Y�1c�X�|̳�)Yc&DŤdƌ�<y̘̘�1�YYi�|0+;nJjF܃��)Y�&
�Q���&?�x�����=wѲ�S��YcbS3�D�dE$$eN	
+$��9n�$��p��4�\�	a��<u�Hۨ�0�z�u�����S5jԾ뻫L%&���duCp� (((h��[L�ɏ�]AAA������ÃN
+�
5�<9�x�}�O?���)�6b���ԍ�Nr��ןe�aj��C7�oZz�77��Ƈ}$� ��Q�hf��[�[w�n�m��~�����;��y�ž'�����>6��{J�=<�zoڽ��'<3q����{��C�VN��~<>y�`X(���r/w�W@'}<�e"��Z�&!A/A8�[�AT��U�Q!v����Xf��;��C�*5��295<z�"��7Ao#��@=C��}�y�������1�k%У�+ܰ�J��A�cL%&J����J�p~|I�>���ټ������X��x���˙�s����&;�2�Vovm)s�ZʯWzh��>�А�T���_YUn��Ѩ*�5�]>�C�6z���fo&=!m8j�Qsʇ#���؞c�{��s�����-g�Z��S��n�	$�9![贯-.�M�i	��	J7�#8�Ak�U�����{D.�l!`�����[�+P�FW���vo/���p�g-�xiq�T�3�-q�͒%K�
+
CP��d�zH�[5;�hR��L/���L~/c���C�(�E)�����>�X=�:��l7LU��
�N���0�G^��_/�*L�#�lh��q���geC���cT��#��\~�FM�|�taҁf�ÎJ{9F��r�aE]����Z�I������?�7��x#��m��^ەӉ.�5Vx��K��'�$/L_�uP��6��ґ՛����V4�B`T4$tp��3��\��l
��g��z-Qg0�s��?ΰ��@�=�1�d1�85����z#��bF�R�t�|�Y]�����7��N�����ݨm.k��,?��-}+��Q��O/����c����p��DEt,����!<{�dq�����"�^���&�{������D��HB��|��.�pc=a{�ӎQݛ�/������ϔT���!x�R�R�i-� �p~?������N�wcKoi)ə�(�hrR��E��HB����u�֡��פ�%ɣ�����o�ޛ5�]%RI�X"�r��=�;��Ⱥ퍯l��r�9U%�˙P�^*���}�_-�s�O��e.+�JU\L���X���r�a�N���72N�ۉ*��:��m��v�� 7�Me�e]�0��j5;PC^�i����ߧ�o�&sI�����~���"����$O����ctJ���=_��I%l(��t��2�h�L+��
����������7�*"�
�
"s��<�;û\^쉯?��!R�Wz������������]�s1��q)	���K3�rSk S�>dp=�JR飍�X
�ڪ����ji�z/��5�E����%���TRI6��=�1r���Dz����N����KPtXL��l�KG��^�6��B��4���$�?��p�y��tlg0�~$
Uyy/E�\�4���H�]���*7jJ�!?�.:�g����]
��}L�R���Ә�ob�:�r��=j��vH��Y���:=ag�{V���Yn�����1�LVX�VU�"���<�oc�y0JSS�7`=����?m<樳���w�J�O��'j~Rgo�u�G:2�����s�ga۟�
+�^t�5�U�آ��Dh��������4k��]�$�ʈO��ձ��t[K
+�^�`]�Ba4BG�����Ho�׳��ڕC?=](���`N��(�3ߝ�>��5�l�����Uh���� Ghc�l���`Ej��J�ZR�6U{�{�Pަ��;C���	{�z
{m�;�o����"|<�:ҙj�s`ꎝF3����DxQ�S���(m�
UM�[�"{��GF޿���G�-"��֢mEX
Iړ㐍%V���VlHMaYNVj��E���f�n�'������}?�D~��`�>�E%��^��2&��׋��Q�D-~�����7�w�
ee1��6�
^P�$ﶊ
��n�ZK�QI�d�<E�f٪ބW����ƿ�P��ee�]RW���Ԯ��5�7����Y��X����U��"&��7���0�ֵ�ڲ�8���(@�:�PL��e���Uʋ���}�]���H>�C矡�3��u���<יtV8��u��rl��YY�rG^7a�O�eD�a?�����@�
b��5�m��!Ҟ��܆�Y��,��dY��Ζ6TÞ���֐�Hɳ�q?��}�;,��Oc�����r�D��ꌄ$IRnEV�mo��+��
:���g�.wv��d#�.n�*Ie�C��xY��5��_�֋�����&�>�7�I��b��.���Ҏvl���:�x�%R%-z�P:Po�Nu�\)�荺��V�~��|翩��ԥj���L�����lDx�Os��yo���6�ŋaC20�
+^�	���������c�N,�G3/�+�$K$I��ʜV�����C�B��z��]�{��I����Z��������tԠeU7�Bg]bH�F���W��r譫2`�W:F?���\�8~��Y�2޴戫&��r�q���8C�{��ʳ/P���o|_u
+�������>�O��dž��iʚ���U���ƿ����CO2�`��ȔL���#����ِ��$�:v��mo7�����a�ʣ�S��u5*�S]~x����;�*�L��|�`����>9)1%!�&������C��e�G�Fӎs.K�$�$H�[��P�8(f���HĐ���C�ӑt���D}g �'����JC�
0WԶn��辤�ɏ���)���e���wT0��;��\t���1YJ*QN�'oY6q%?A}i�d�l<��=h�9y�Z���"ƨI�oxZ�#���6m^07%?~�n�F�,���9I���;�\io��w@��x�r2���@�K��R����18r��b�ڠ"����b���TM9 ��!2��w���
+Tbz}�7��ȗ��l *�1/�I��a�‡΀�)Eu�ڤ/5�P
l%����R*�����p��d�!tw��<���j"�RS��-8۾�`�>���`��%�TxS�2��X��6uC�E����/�G�P�jg���q�	k��5��H4����XG:��:}��~�Ć�
�}�<0�E�
+��իWK%���dzH,v1l�����2�N���+$`��Љ�c��eD"��)��ی|1�0�䄿cp�JШ��.�,nf�&����G%�/���n5ep�"�8�큗���=�0.�O���4�'���\d7�
+�B�zE��dI��o�M�K�X�[Q�}�WWӱ�}�wȩ?�of��
+��
���,��A���^��25r�����rr�V� 0���f�!�˗a�r�N)�uPr��o��y��A���bErʚ�y �q�V2�6��n��}�
9�&�C+��M��W00���)�O�?�0/`���2�^K6��y�&� F��;)~�Y�܂�b6�G��.��p�63΍v��|�����P�5�x��#j���%��f���<˯
+Ʌ
�B�F+�D�Y�<�VXJI(U\���բ�sa
8J*���f>��H��e*��[B��n"����X��g2A�ۮ�-���|Y�b�5��m\�	�f��PW�V��M�����
+�.�X�э^}���!2�>�I�t�j� ����̭��D�G��!��N 9����Ȕ�U�P!~��d/EjrT	���P�EY�f-o@-��d����t��Kc��[��L��L3࣏3�.�ލp��P&K��3�K���*k;��W����=7�<�{`���r�iߎ�,]E4q����5PC�Z���o^0
S��>бx��>|��i|O���
+����{���)Dn?�)a4g-������ |��x�8m��1�,M��@�B`P�.��nX�%��U�v��*�C!���H�H���L�Ajd�vS;�9�5&�#�P��4j���C�F	�a*1W�z:��4i�2�_��B;j�CU��C*]
�ol�5f���d�W�Oh*c2[\~\/��}^��T_��*.]���_����%�	�-�Ї��m ��KZQ�3p�
��
+r�Zς9�,H#_'��j�������Z3U���_�(�Q��˛#˘2�i�P��֓���z�>�e��L�����^
���u�	��Emtc*��2#�Br��&����c�����o),(A	,��dr5m�w��QHGH=tj1Ԇb�ƏBS_\�(�,$��*ލ���؃W��8Ժ�P .�!ɬG�E�_HK����\g�^&��U��z9^�Ң
��l
+��I�`�X�I(�:DQ��Q�
+�/���S�}������<R�
+d�:q"���D�%D�Zˎ�'
+�z��9�˨�Uρ�#Oi7�kً^l�o���o�{��T�=gZ<a����=�6t�9�����ޣ���-;@�j䳼
+���?���h�v}c�S��Q@ ��Tw��C�̊N��&y���$v}����F�^�2@:
+��K�J;�O���o����(]���v�`���fx���
��}��G�����"�}�o������w?���ċ��w�����lKG��I�c��O�ˇ�����\�b���~׼$��oK�wE������7��lޢt&��ߧ�C�����陸�����
���7ԇ��H���D��2�ф����9�JG���׶h��(�)H
+0rs�Ƿ}U�2�����ϐ�qσ�/�
�g�c
+�+\��k2����ϡ���F$�	*�<�kF>+�2S[�h5r�[RP^o��Ac����l����n�8����m�Q�׵���\J� ?RկUV����wo�M����J��L���f#ʯUz�O�������������¿s?-T�,��k�#���R�A�{���tл��=ͯn��2 '�o��s�(����%�)�r$�-�fX������5v����Z��ў�MhEM���X[K�bd�s�\=2�_N�mg#*�?Y���I�ޜ���@{wU[�ҝ�&��w��k����
+���� cWA�04�����1Zd�A��(	n�C&��x��h��{S��h�4���-�������ʤ1ř��Z�"'h���Ogr3�JE��,lW@[vb��cH���9T+�c@��L�����`پs�d�p���ʕ J26?2{�$�e�����45�����)��qmB	�����*���2�Tc�D�4m�\�[�╦w��k�~?#�su�f
'ґ�GBS&�n�W�UWY���Mi��6���Mڎ��)�{)�;�a��[��M�W~�ܪ�'�9����ۀ4�Q�z{��J��Z�`B��m^�/��k��y��oj�Pjm<T?�\S^�zv%�#��>/��瓿s�or�6�[�N��n`���s�j��r"��C�"�ޞх��k������@��K�wM����l�IaWx	+�+]H'3�5�^{��0�L��]gU#�,�#ƒ0��N�U\n,�B���Y!���N�?U�n�Y-�A�>Q��D_��7����Y�I'G�t����
+��Hb�����ok�F�X������(��N��p�p	"�	f��MW�a
!z�F�KS̘���`��H��`2l��0�6�hKtV�b+)��-��>�>���t�&t����@�3��n-���{}��\�]�ܺ¼5/%�#2{�3�LQ�-T�펺ƖB��0�����p��A�X�K�2��%mRT��K@f�O�-�bNo-uxZ�P�j������
+�+e2�O�2˖�4E(r�����뚷u���:�2�B���:�aZj��[=���jE0MFsě��k��u-t:��0���g���k���n4�є�l�9Ă�ɲ���N�i6�-�~����H��d7
LkЉ�r1�Z�Etڽ_�a|����_��w�u4�fS
��f��o[��|>��g�!,��z���1o�н_|����'����N��<�"x�T�1��X0�I�EhU���2�
+�-Rk�SkN�����|��'�h�7�E�Q��&F9��K�d����"�ߐ(��eL��9�\�7��?F`�g��4�ޣ�Б�rfO�C_�疻�G�1ᑙ37^`Qom����Ƹ��a0td��F�_[u6�a2;\��&���k��=�a���vа��4��#�K�$&'�W�7ux���е���
+��S���FpQ��X~D��~�R����䀕85�go
��W�����g���"/���Z��٦s�:���3Z�F
)������"ت�1�����_�=�Gx����$��Bc_]ߟ�k�wS������~��+��t�wSI�����V>0x3S�2z�ȃ�a0u�->X��ە��}��ԝ#��������S��G�\���~�4�=\�[���C��;��Ev�Uê`��Ysפ�"
����'�
t1�<r���v��QgW�2�@N�������Exto����yw6�#b)^�̢6k�g}�8�H�?IG���Ͻ�.>����_Ox��ZQ�B���(��i�o��W�r�*(���¤�����J�s��O�_�b��3�Q��6�ИU��r)T��s��Q����Sթ�_�7�p+�,��3!A�E�C۳���$A�-�K1�`l6�d���h)��Vs�D��u��q,�����@�1��
�x3J4��β�o�a�^>8����䌱��x�DʇBT��p�\�
P�2����S�YW����
+B[8x��I
+Mp�UM���L�
�;J��Tℋ��we{�|�ɽ�/�6��/A�D����,��0�!Up����ߟ,�ߟ��n2�Hg.����oT���f�A�=�M�e9����W�����'=�:�Kb�-v�6�
+��l�	�/`np+�L��3��DKJ��A
+�c��A�L*suj�o�]��1m{M��8����� 
�ġt?��.������ c����(���&����̦���.����k���t�sg��|�g��y���w �.��h�O���~���.��$ȸ�G�,��qv�/CΜZ��*=����+��O�xl��D}$H�����#�Ew?4�����c�Y=���kWd��fTd����y_�
+��`t3F>����"���s�W�����W���w9z�{]TSX������[.mf[Z[Z��A:w�~�WN�/^� �~C���
/��
+������P]��f����D
+Տ�Fڌ�����8��
�t�x
+endstream
+endobj
+
+1257 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /JMTFQO+NewCM10-Regular-Identity-H
+  /Encoding /Identity-H
+  /DescendantFonts [1258 0 R]
+  /ToUnicode 1261 0 R
+>>
+endobj
+
+1258 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType0
+  /BaseFont /JMTFQO+NewCM10-Regular
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 1260 0 R
+  /DW 0
+  /W [0 0 500 1 1 514 2 2 444 3 3 394 4 4 556 5 5 333 6 6 681 7 7 444 8 8 392 9 9 444 10 10 625 11 11 500 12 12 556 13 13 778 14 14 389 15 15 736 16 16 306 17 17 778 18 21 500 22 22 278 23 23 472 24 24 556 25 26 278 27 27 528 28 28 500 29 29 528 30 31 556 32 32 500 33 33 444 34 34 472 35 36 500 37 37 278 38 38 833 39 40 556 41 41 764 42 42 556 43 43 528 44 44 500 45 45 681 46 46 278 47 47 785 48 48 500 49 49 917 50 50 556 51 51 686 52 52 722 53 53 278 54 55 500 56 56 750 57 57 500 58 59 722 60 60 750 61 61 333 62 62 1000 63 63 750 64 64 306 65 65 653 66 66 500 67 67 528 68 68 556 69 69 361 70 71 556 72 72 778 73 74 389 75 75 750 76 76 528 77 78 750 79 79 556 80 80 278 81 81 778 82 82 750 83 84 500 85 85 1000 86 86 778 87 87 556 88 88 1028 89 91 778 92 92 708 93 93 500 94 94 611 95 95 750 96 96 472 97 97 583 98 100 278 101 101 750 102 102 500]
+>>
+endobj
+
+1259 0 obj
+<<
+  /Length 13
+  /Filter /FlateDecode
+>>
+stream
+x���
�
Z��
+endstream
+endobj
+
+1260 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /JMTFQO+NewCM10-Regular
+  /Flags 131076
+  /FontBBox [-40 -250 1009 936]
+  /ItalicAngle 0
+  /Ascent 806
+  /Descent -194
+  /CapHeight 683
+  /StemV 95.4
+  /CIDSet 1259 0 R
+  /FontFile3 1262 0 R
+>>
+endobj
+
+1261 0 obj
+<<
+  /Length 2075
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+100 beginbfchar
+<0001> <004A>
+<0002> <0065>
+<0003> <0073>
+<0004> <00FA>
+<0005> <0020>
+<0006> <0050>
+<0007> <00E9>
+<0008> <0072>
+<0009> <007A>
+<000A> <004C>
+<000B> <006F>
+<000C> <006E>
+<000D> <004F>
+<000E> <0074>
+<000F> <0052>
+<0010> <0066>
+<0011> <00B7>
+<0012> <0032>
+<0013> <0030>
+<0014> <0036>
+<0015> <0031>
+<0016> <002E>
+<0017> <00BF>
+<0018> <0053>
+<0019> <006C>
+<001A> <0069>
+<001B> <0076>
+<001C> <0061>
+<001D> <0071>
+<001E> <0075>
+<001F> <00660069>
+<0020> <00F3>
+<0021> <0063>
+<0022> <003F>
+<0023> <0033>
+<0024> <0067>
+<0025> <00ED>
+<0026> <006D>
+<0027> <0070>
+<0028> <0064>
+<0029> <0044>
+<002A> <0068>
+<002B> <0079>
+<002C> <0034>
+<002D> <0045>
+<002E> <003A>
+<002F> <0047>
+<0030> <0035>
+<0031> <004D>
+<0032> <0062>
+<0033> <00660074>
+<0034> <0077>
+<0035> <002C>
+<0036> <0037>
+<0037> <0038>
+<0038> <0055>
+<0039> <0039>
+<003A> <0043>
+<003B> <0054>
+<003C> <0059>
+<003D> <002D>
+<003E> <2014>
+<003F> <0041>
+<0040> <006A>
+<0041> <0046>
+<0042> <00E1>
+<0043> <0078>
+<0044> <0066006C>
+<0045> <0049>
+<0046> <00AB>
+<0047> <00BB>
+<0048> <2022>
+<0049> <0028>
+<004A> <0029>
+<004B> <004E>
+<004C> <006B>
+<004D> <0058>
+<004E> <0056>
+<004F> <00FC>
+<0050> <003B>
+<0051> <004B>
+<0052> <0048>
+<0053> <2013>
+<0054> <002F>
+<0055> <2192>
+<0056> <002B>
+<0057> <00F1>
+<0058> <0057>
+<0059> <0026>
+<005A> <0051>
+<005B> <003D>
+<005C> <0042>
+<005D> <00A0>
+<005E> <005A>
+<005F> <005F>
+<0060> <221A>
+<0061> <00660066>
+<0062> <005B>
+<0063> <005D>
+<0064> <2019>
+endbfchar
+2 beginbfchar
+<0065> <00DA>
+<0066> <00E4>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+1262 0 obj
+<<
+  /Length 9172
+  /Filter /FlateDecode
+  /Subtype /CIDFontType0C
+>>
+stream
+x��zw\��>�2;ǶQ�11�Y4�{��Dc�5��ADPzY�,e���.�ׅEp@@
{�5*��Ē`�EM1jL�;x���gAM����>���~�Y>̜�9��}��yΑ98::�d�.��C��>l�"w��k���t�5�'t�^2���
+���ۓ&�ͣ'�EO�r������K�ߠ���<������N}�w�� ��Jq�F�
�6��o�?�-�Æ�<b؈��'���nr^����9xs��%����C���79�{���otV�ntt�tw��t���~���lvs�
rw<��9����38���j�!~�C7��
�nk4�������-������[����`�M~��݃]7{
qph� sp��ݚ{�j�45}3��
+����N�;u�N�V��d6�R��N�.v��Y����%�Sg��������>��L&���rppg��N��q�px"��n\�	�v���w:*�O2�3sCQ��-���$��A{e��~�;:v��R��N_w�W�J��/%v1w��nw�t��m�6�[6ߕ����^��̗?y%�G��=ξ��Ն�<_��si�ozb/�Wt�ۂ���ꖳ����{�ԇ�������S��~���yC���7W���Ͳ����~��/�>���i��T0�jp��_
�CzY7�a�I8,�:,;|���.��v��Un��Yx�e�������K�Qm�BD�:t���nr+%�fgD��W	��]��Q���5�Е��	\-��S��{4sGi�_�������^���K󛄟����R�;����ˆs���jl�S�S���F�m�K#�#�gNb��5�y�1��v����&X��Z���һ�+%C�Ү��_ ��}(/P��<�Q$�[7l;�<L��mT)i��j.�z8U�
�rc���p�gGt斊�����3ff�)�*��	�����M�@� �@��̓������bbuz��
�
��g�b���ۮ�(��%�p����"`*��t����d���,cb�(�W�Wح�!$��TP`6�9�`��>cd���r�ƪ��G��w�v��"���h����@��i	�D�� T2i)��I����ox�}?.�
+n}S�̧��Tt:���SF��N�_3\]�������	0	'`oY�Gd8��� �R<':U=������!�k������W�K���KW�UK	�aм'y+�,��.��C���A�{��j��;�×�U}��
�@G
+�*`��Ҝr@hR�H�J�ǥ���t�[j��,�t%F�X�5<��9KRW�»�E�3#]`;��HU��r�N3�F����P�'b;�u��\�{�M_)`�Qo���W��*��2+�c���*+��˥~8�T��!�)
B;�R'&-̐H�� \��ACGـI����
AЊ��L��@"S�@�F�2D
�F�Nسɭ	�m��f��
+M�HڕN2V�
[cH�TI;؝���P)*<��`l�	�i!�%,w���-�+��4�5X;�n�ۛ�hg�u�Ѯ��y ���~DN�ʃ,U�W�[�\`}�����p�a7/ٻ}���Z {a��ԥ��;Qy���P.Ut����pD��ȼ-�a
+�q�K�{`z�(0�� 235�O��\�ϗN�?��t�@#׊Q"�.mc��x(�(ov�;<�Ё��Mg���m����:��l�#q��7��������l��7>i�_�F�%�����P��P��/��U���.���%�d�M3�v��"�.����{;Pa������YN���+�K�FW��.�W)�Q[A�cV<ږ�𰛒+h��+�Q� ����� �H�$�:~~U咂�%W��U�Uܠo
+��ߥ2��,��-�`7�KP��dNF�ͅz�ڼ��a&��:w����P��ռ��TT-�����*�LC�Ւn.�B ��QG��������7R�2tl����_�k6U.{��e�ʎ��F�����"m������{�@N|�.<
+Ipn����A�<�?�a�#�4��˃J|}��|}K���KJ�%\�{���@t:�ˏ��m�������1���f��	�28�3��>A���TI��CG,>���̺��RO!�7_8^s
+nÞq0�p�ݕ��%t	��h�	D$V�T[�_��v7bj�\r��<\O��������|��a���	�.LJ^7qϸˡ���/��~_q1�*\#c2y�34s�
Z�6�[��3pN�yhH=^�?O�
�������a"|@��2�Oq��Z�S�-�|���"�E�?P�06�]MJXL�.�G���Z7�mn�λd�5wG��3�����k��
�ɴoU(��;|E5���jX���g�ڄ`�|�hs��uȇ��oA.�q.��+`Q�(Ո�Ce���\�]�b��3jT�)�����9#��7[��h)i�a=���/{��j����e�d7�qɱ��O]������G�#��#��и���;���4r*�38;X�L�
PqETX��b��U@�����!����):הr� %�13�@�A��`aP��#	�����N;iQ�|Ux�ͱ�*�,�vL�4�Ω8��Q����ų"�+��8��y��\��<���jt!��Dfn���p
+>=~�*�`�c�י���x�Ypmi��1>�x.b|�7ԍ*��#�7�96���H��[���/�T����i+��|abҲC������q���q�5m�%L�Kb�1q�y�&��p�G�v��/6�Z��������"�
+� �ݎ�}�#rK��[��%F��G�#��������,]�nHr�;I��_HӮRQ���Y�*�2�4�%�~{p���P�@�����2�JI;��HCEY�)������|��@޶�7Sc��0�5�tҏ�RG�`�Y�@JwC�"��p3���7{�Y��Rd�#�	Y��i��_ƫLZqޞQj�[E̽u��Tq�.^�qQRh"p1�0P�٢ۗ$��a�G���+sW;}�,�­�e���*�L�\k�^u(�ȡ�G�6�+���?�9p9�'�f���sB�-8@�)�n�]):�����{��L��n^�jC�a
>�oı�DI;�e"�Jl����{Q.��e�� I'լ�0&�lùw�P�g�N��}��S��c-'t��w�0�x��p?uPT��� S�5!�?m��z���^Lfg�V��Km��5U�Y�k��k��<퀌���ڏ�PS�������4���8w��%���Ў+�L���1r�cWp�=ޟ�AE�_�F��O���tx�;{�䣙O���2�ɓ�:9�G���O��ݖ%�u
+�ʪ+�ԴF}���z+�vʬ3h}�̎�d��=8
+GUL��U�Z�"
��e��y;>���=�u(u�]h�{����)����<p{��O�N'�=Z�h��Fn��4R��?�v��E�Θ-p='m.�,���}�)��N���f�F¹h`ADxx����0��c�A��i�B�t�RU)�^XT
��:�d���e��`�b���#��w�N7�k���D�����E_m?�)\#��F������X���Bk��5�M5dU��V��z�
+�
���ެ4���g�7BѤ@��W�U͌�h���U.0)��tU�h�]�ڿ�7�+r��og�L@���?��tC1.:��8��8��	G~�������H�R>/�b�
+뵮G5���5����=Xou��Dm�����.�����nd�%��?�*@U�e���<�7y%��P
�P�W�ۯN_�ۋ���&�&��zguR>�e
@
+X�]���֬�\!� �XE�Q��^�㢃���m��P�M96פf�٫/�	�۵t�N[��ӏ`����rZqa��7�o
+���%���sik�hU���@���h@���zr����K�0�������u�W_��*�ЋU�?��WJc�ST�!Fڜ]��F�����<r�QGa��Uh`��m5�R�v����wh���/
��4&)�~�@���?"�ӝv~�d�
���"����X�[��C�G�\>;�2O!���Zq��w6bA�.�<m�g���]%�ެo�K�R�F���)�cT��þ��
+2��&!��`�l%�n����.�<J�v��OBԘϔ�+wU�*��&�#[��D�mtB���C�1꾼y96񸈾���D������ǣ��q����@�a�=V��/Z��7mw��9�W|�G��Y�6
+&���O���X]�7RӾ`"<�(�E埙�ڙ�Z�o�UZ/:��������ܭ1FCB4y����.ݐ���
+������$8�E��Q/�>w%
�!,_LL�B.)
�
UG����8,����%'	N�&2�玘r�N��UQR����%��ˋoJ��BH��7.��3-���f<��B����r��A�oB�v��S=�^jϴzKgě"�eس	W�p��b�7;���tH�:	��4,�����zGe��@�(��a8Ʒ�G�+p0l_�KU�~�n9�]L��(z�6>S�
+�{�$�����ǚ�=�"F�R?����$o��7��3n�xX�w{ˢ٤��RG���ͦ���ŕ�g�_���B��9��@r�R�˜6���}WNf�*���Σ��ԕ���t$.R]��;Og�k�2��Aj�
+����.�k��q`#���ˍ���Wb��:����=䛸n]����ǎc�Sa5�Dxx\�r��P[_
_��vV�>)^ŵw�C�>4vU�L \;7X���(΢���ܮ�1����p�˶�X���ܚ�_�Q��M
��d�+���m���Xtl�i�M�d1Z�@*�l%%[m["=�ת��WF]����~@N�Ģ�	�\�9e�����3�A*a��W��NP��	�k9C}��L�U�@�琠g�aM_{���`{{���UR��E���1IQI��
�������&��K�3���I���1�����i+n�wz�0�͍܁���=�&�˚�p����[[��\��ߐ���&�|�u!��[��6������"�@>�
,�)#w��SK`T���n�,�
DY��E��$�S�]����tX�줷ſ�A�[��"5!�g�J-g[
'��$<3���nT��U��qlP?j����C$��#��'˛����kTb�j�5|A��(2����
+F�i4Cپ
+_�3>��c5�kH�J�f����Q4a2	�����T�o}�a��gzt��2I��8�.3''e�X��a5m�,�Z�=~����Lv�J'���G2�e��z>���3�?  ϿZȁ��t�x�y��h�F#�H|�.)��5d�1D�y�"T���ȥ��Wx�SP%�t$��
6`,�i�2S��Bt�HG0��k[�
+��X�?�K���������7�~D�5\ĤهZd�j�_���F�a����P���^V�>��/5��B�3��7�M"@�y]����@�օ��C����� �}��4���EW��$��uf��N�#g/�_�|~�T����'OZ>v�ԏ8Uq��_������yD�@-J��`{l�CQ��ҶQ�AO;�#�����}����5��
	R�gɫ��Owp�IO]�t����H�{
+�­o�Pa)�4�%x���}�hW}�F�9���]v۪�v���?&�ůx�R���X�nú`W���_�BG8u�ŧ�K�^_�P�
+K��;����<1(G��q��
_��G%�D/��;���F
+�˚X���k�0��p��f��]YU�*7���J���e��<��8��7�H>RZ
X�1&���<�ʀ̣+v�ք�;ϪPM�
+���(l�^qx :���#�u�w�p���D��2�Ҽ\&ʪm�ڄ�6y��͓�|�[¨�L6�ʩā���n(ҁ�IK	�"�n������`��} {�o���3�ۡp �lɦ����S��qS���̋����������内�@���������ƒL|:�3�Ȗ�OY���v��w>0�@�&7��&�=��̋��`�=}�n�R\�35gO'VQҋ8��,��X���S4!�ğ1�X�)��#�����W���PY�kw1���)�JB��&!ː
��&D���=��m�R#
�,����@[�
+,�t���Ǘ�̘tf-��b���j˧��i<U:�UFI���ܿJA���۝#��j��M+y�4t��P\I�
+gv]�9���d�$f�vR^�gw"���;)|����*�[s�8\� �
�?�߹�h��m\��������/�C�*�U�
up�u���ܢA�U�B�M���g\kt̖��(LVpo���N
+�ŹTx����?�8�(��|�o����BAtʔL�^��.�܂�n.'�U���ص�PI�X]�s�DD6��cXf�2o�s.2�FAD�D%h��	0����99i�&�e�H�L�1���٧`2�O�Q�Y3k�^��bw�K͢������MR[� ��g�B�x�I�dE�bc��	�ot�0�a�q����Du�	����m�kϜW-��Ig��� }�l��?��<���E���-�(�.�~��3�ۘۡ,6V
��-OK���)6�l����b;^���Y2�	�v��Ʒ�){��T��5�����`7\k���H��2ܑΛٰ/���\�s��J����&���A������zœ9�Qo�{1����N��I��ObV$'&&�������<떴Ю���<���>7��:��A���*�?�⽽��O�>�Xm�.Y0+!�bJ������/�LfX}���=J��:�N.�������Ϟ=~���׮]�p;I�Co��;%�4����Ѯ��
NمS�%�egH�Y��g,�YA�E��`B��!e�Зn�N�dH�@
+���qJ��h��;/Af>�4�BhA�-��P�c��t$Pw�7�i�u"Q����!n}r�#�.�p��ӳL[��v�����i�L���\&6�h
+B��.W�@�1�X�7!���]��&%���Z"
j�ˤE��)@�!�@%b�Á�B+b�3-�n<j��aw�(�796�܏���hNH$^:��z�H��Ȏ��(��k��#G�1]QOho�)��^l#�,��7s����+��C8	۪�7p�0�Pn,7<����E��-��0v9�$���z�_��e�;+�LD*�k���f!�l�Ѫ�`�1�5�*���R�`IH�Dȴ�����S?ܟV�`>���?�P���	Ld���7���E
|�ZTص���!�,�)��&A��΢=D����7[e�l/��Ɛ�wDYA�jy��sf�*�>j,
+�`��k
ZB約�Œ��f��Ք�[+����I����	�Np[�c&%,I�y`NJ�&$}�d�l/��D^�M�A�g|IH��wx�:�(x� �ۊU8���/گ	�l|����ٛg������|�᳥τJ��B�)Ü�~0a����m�V����H�࿡sspw�]
+!�>{	��oq&NfRS�������<`ln[��11���@�^b�=���A�S�`;�E���ش(�&>!T���V"���>qg��8�$ovy����(z��Dq�ٙ�<Js^�F��^J���B��NH/˥�;|Fl�F�]�>����>�I��Ф�f@6$�[�	~�x
+��ID�?[C0�έ�C��a��b"��I�Iz|����M����wN�J�&%��7�o:#�����s�w�Ygd�Ƅ�D�����3��*
+��~�8IO�HLO�׆�!�P�
�2����h���a)�
+�TdCAh(�E
+��V��aZ�*����H+,��l
+���.�B���8Hs�_���W��!0gk�nx�Vp�0���$w�����-/f{D~|�]P`=r�p�?����o��^2�H?��b���8�[�X�
+4>o��6���p��Vg�
-��RcL4$C�q
+J[�fF+�(�pcNBdz��T�2���4��bL
ˋ��27\���Xb-1Y�
��,�c�$ג��W�1�X���ܓ�۾^x����%Qm�\�S}�sI�_lx���;��%��A�g40��v� &!ި_2`��eQ9T���#{��'_@��8m��%�pM�Z5���[�K���>�_��=5�矎8_����$P�?g�:X
+ˏx���������&ii�x�f�{;(������FmE��
BIK&?g��x?�w�8|����Q��=s����l��.{xѶ���U*�9�������~���h��b�t���`�3�h������t�p
�����
p ��i�u8���قU��i9���~��qD��n�L���{OP��:�V��l��0�ai!u���`:y�.�ESڗ6�ütj +���U�V�k�B��G-���hlϴ�m����if娤�9�L���`���b۳��~P�e�i����E'T5⩻S�s��U8��o�~N�w.����7C���_�!�S�=n5}��]�?g�h*h?4>�λ(�V��LUB�i���hק~{).`����R*�a��wQvB2ɥ�X�*�*l0���
��1��
&H!�ZZ����uHAfǗ8��-��׀���$($W��
JeͱT�K��*��cBйVzR+�
+���y�ha�� (ޗБ-}TB�!�,�
+�@�M���v�;����O*��E�I�T#ISX����Ov�,52���tS�Q�nJM�t��ϭTK4@C���/ͤ�%��s�ba
�wO���,N�2�����O5Ӧ?;�Ɗ�����}�h�G����|�>�N������(��hBY�����W����`�	��ǃjeءV�����x<�n-���m���v���Hj�o�����}�7���١�����T�}B��eg�c��w�w�p���q��������A�����9�z7/u����
+���C���ǯ�������aj��O8�����E��-���}��
N���
+endstream
+endobj
+
+1263 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /HVAWYA+LibertinusSerif-Regular-Identity-H
+  /Encoding /Identity-H
+  /DescendantFonts [1264 0 R]
+  /ToUnicode 1267 0 R
+>>
+endobj
+
+1264 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType0
+  /BaseFont /HVAWYA+LibertinusSerif-Regular
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 1266 0 R
+  /DW 0
+  /W [0 0 500 1 1 307]
+>>
+endobj
+
+1265 0 obj
+<<
+  /Length 9
+  /Filter /FlateDecode
+>>
+stream
+x�;��
+endstream
+endobj
+
+1266 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /HVAWYA+LibertinusSerif-Regular
+  /Flags 131078
+  /FontBBox [0 -112 500 698]
+  /ItalicAngle 0
+  /Ascent 894
+  /Descent -246
+  /CapHeight 658
+  /StemV 95.4
+  /CIDSet 1265 0 R
+  /FontFile3 1268 0 R
+>>
+endobj
+
+1267 0 obj
+<<
+  /Length 619
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+1 beginbfchar
+<0001> <2082>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+1268 0 obj
+<<
+  /Length 434
+  /Filter /FlateDecode
+  /Subtype /CIDFontType0C
+>>
+stream
+x�cd`aa`dd���LJ-*��+-N-�L�
JM/�I,ɹ���!��#�C��A��A��B��<��=EFB��OVY�����\�����(��>񨀨W<�̌�l|ގ)�I��)�y%�%����E��%
+����������BHF��Y
+
E�Y��%
+��%�E�zL���8Θ;����u3VNe��4�]��y��n��\��oM����O�����Q�����
��4@�!!&FF�]��{���ؽ���^�n�?��t���+#������v�'��a��W��
�}��>	,��g����E�}��������,���Zɸ�����W�3�N��7�O�����_jb���	r�'L��Jz�ꖘ�Ⓒt����h��IlS�M\տ������A:��6%GڂmʚY˶��8�y͚�R��E��G��l�.�HL��
+endstream
+endobj
+
+1269 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /GLERDG+DejaVuSansMono
+  /Encoding /Identity-H
+  /DescendantFonts [1270 0 R]
+  /ToUnicode 1273 0 R
+>>
+endobj
+
+1270 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType2
+  /BaseFont /GLERDG+DejaVuSansMono
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 1272 0 R
+  /DW 0
+  /CIDToGIDMap /Identity
+  /W [0 38 602.0508]
+>>
+endobj
+
+1271 0 obj
+<<
+  /Length 13
+  /Filter /FlateDecode
+>>
+stream
+x��������
+endstream
+endobj
+
+1272 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /GLERDG+DejaVuSansMono
+  /Flags 131077
+  /FontBBox [0 -235.83984 602.0508 765.1367]
+  /ItalicAngle 0
+  /Ascent 759.7656
+  /Descent -240.23438
+  /CapHeight 759.7656
+  /StemV 95.4
+  /CIDSet 1271 0 R
+  /FontFile2 1274 0 R
+>>
+endobj
+
+1273 0 obj
+<<
+  /Length 1138
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+38 beginbfchar
+<0001> <0053>
+<0002> <0074>
+<0003> <0072>
+<0004> <0069>
+<0005> <006E>
+<0006> <0067>
+<0007> <005B>
+<0008> <007C>
+<0009> <0020>
+<000A> <0027>
+<000B> <0041>
+<000C> <0063>
+<000D> <0065>
+<000E> <0070>
+<000F> <0064>
+<0010> <002C>
+<0011> <0050>
+<0012> <006F>
+<0013> <0073>
+<0014> <005D>
+<0015> <0076>
+<0016> <0061>
+<0017> <003D>
+<0018> <0075>
+<0019> <002E>
+<001A> <006C>
+<001B> <005F>
+<001C> <006D>
+<001D> <0062>
+<001E> <0079>
+<001F> <006B>
+<0020> <0078>
+<0021> <002F>
+<0022> <0068>
+<0023> <0077>
+<0024> <0022>
+<0025> <0048>
+<0026> <0066>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+1274 0 obj
+<<
+  /Length 7988
+  /Filter /FlateDecode
+>>
+stream
+x��z	xT������=w�L�N&;	YL!�
�Q�@��$��B�,$@H$,bJ��,�Ѫ(H۔RD��K]� 6"K[E?�m��Ri-6R_�br��s�dCj�|���}�o&3��{��/��z��@�.[�2�£67��-��nq����
��S����|��{��1�gU��r��DwQUU���t f��jV�m��> f:���j��Q6	�i��5��o胀�O$-_XS��`
+b�@�w�j�W�%�Ȕ���VTԍ����l���`"�JAܯ�����NBS��$��^�I�C'Nb#>eN���h0[O�I4a�Z�Db����*��*���Ӌ8��B2{Al��B�؂cԊj��l ��^�4�Y-�k
j3Є7������a/j�l#{��s؆<�]�\<�qBO������p	���Z%������F��^sB�oӆb/��n������$����O�@�4��T���m{�
�J�s�
3�N���e��T}
]@�Z�%8��?�l�h2*�Z}�z7H��5�v�m}�-'�[i<���%�D?�J��n��d}^�A��j�M���h4�A��ǰ� ���<,A-�v�8���vh
,W���J�
M�%��Z���D��M�9��֡��
~�`3�Nð$�A-���`���I����1��$�-� �[��re�l=��9��4�A=-��?�y.c��鳓�,�h\�lтq�&��>��ɳqs��3��S�$׃<� O+^p0��*i�wkꨭފQ�P%v�U|/6|	:�0�����FMG��3���=�~�=;�L6Ӓ��*����G�������`�v�}������ܚǕ�80���lδ���ā:�|��Ḍ&{O菦�0wNt��ې����d�v��l?o������{�����{��E3:?{r�l����:�IyƮ16/���l�46�Xd&KM1"�QY"���)��YnΈ�ҳX&˻�ܜ(�����}5�,y��������7�vX��_���ޟ����qS���ʪRc���|��32�)���nuRlVk�/�{6����p{8��dG�-N�ˡ�0�>��{��xg���SV{G���3??��ɑɦ?�&����g&�%�Z%�#��_u�ܩ��
��G������}'�M�0��7�
01-a�\���:�ڜ�
��{�x{�pd��h?�o�gҼ�2���$�df#��z��l�7`N�t6�;���crf��y�͉��7�tO�S-�e��q���6��_�������5�S�������3�p ��X{�ǝ�㨡��4��>��6��L���΂v	B���v��l�,1�K6�Ynd*{��t.g5��:����P{A��1@���a���@�h4[�G�L30)�a�����9��E���v3Z��jސ�_/=�В����_��'���IZˣ�y�']���/(;��ju���`b`P�;1~��|�"���Q���x]kb���G��=����'^��8��ٞ�#���q�ދ޶�ҧ��-Q�Q�9#"�Z�oެyXj�fz}��\�{����i����?���tQ��G6{���G_�#+��z�S�~sV��St�o���}��;�(fcv6j��U|�S�SZc�aCb \o1^�Z�3�v���R�����l�)Jf���O>�
��;�$
�'�V#
����5ݫRD��
+a��)#��%K0�K�V���������O�.YT�i��o�'�A��u-��4�i��4r����Ҝ�D�"g!��(;H��a�*
;x���}�=ZF[���v_(w�?�y��6�����gl$3I��hBtb.M'������.3%�\�Jɔ�U��⧟�WO/��l�i�ڥ�ӗ��M�A���4�
��
+Ĺ���b[쑨F�6:Zig\��
c|���x��|G�Wt\�~v1;��R�`z��ݓ�rsF���&}o���4�ߺ����y����<6&�Pv#�3?���a����_�:w�-ʇ�|����
(
+�#��s�c��Œ�0�JlcDk���xQ������ǥ�Z��W>�q�'s	�L��-in�Ϧ��;?x���Gj���!�gY�'o�ܼ�)�6���/o�a�P��"����N�y��Z�,}��/ 70�����7�[=;�1:�`�\�u�x)]�
+q�����2��zb�gzeT��3��С̝�'?��k�;�'�o�o���6ݿ��U��Tz����H�U�z;����ﶝy��;��W2e�zd��ie���f��s�����#3s�Լ>!͔(yJ,�J�.|�Z�.�yӦm�6m�L���o�^z;��E2���]��>�ޙ�w��8~H��>eY��	��G�c��`�*����l�I��R�]e,딸���t9
�x���@{ԋG�}^�]��s�n4����{;e�4U�t���̐�i�9z�_��T��Q��voS�s��&�|��k<���Û6=��w�����-W>�'����@���C��շ3��G��[bp!���ͺ�wC���%xZU�4n���>��хK����o���n��gw�7��9��`��g��k��n�
�U���G�C�Ht���„�ՑDb;���E�i4�ƴVs���%�ox�m�?�h�����$^�T_t���ז�����o����c��?ޘ�\����&m���fi˒�/ͼ���E?���������̼�t�TO��u��������0\�|h��+Y����SҺs;��������t����\��n&�I�$M�������]�v�z��*C(Z��h�qy ̀��
�P#�'�q�E6o��a��0�Riw����� �a^�#�3;�u�ki�6�͋�u�4Z�2+b�_+>�tΤ��l��MW>�SC5wy�z�
qI�pZbl���mI/'4^'}(���������\�퐭c�l��#����N����ǧH�������Ҍf֬5;�]{����c�c��<��dڼ���Mϻa�M,�JnyVK�ȴ֛��{�=�4;t�g7���/��9���G�<<{ۉ�ӓ�ܻ�*��zqȔ�{���=��&ܷv�
�A�f��yȊ��l�����<>�֛
�v�tY�TY��o�&&G&�֓�|_~4�SZ�
tיx}�U=�N';���$̀�:��0'�E���j=i*�Xpiݷ�:4�ɻ����˦H�D��/��_X.["0v
��Ȇ��m��%�*�Y2����Bg'.N)ٚ
�iD����}�Z8:���{���u�'�lW9�_/���c�;~���
{����v�9J	�iQB��2Z�L2[E&�駿�5��C
0�z�ހ�(0(���mH�7l���$j�?���O5J̩�%b���z;&tϜ|Мy��C��r��9��RF�j)
+�w���ϗ����l=�gٶl{�#ۙ�5:zt����q��G�0:qm�7�
�����WSTStSLSlS\S|Ӏ����T6�)�e}֫;WmoJ���k�獟~���I��O<�\6zJ]�XtZ�ߵF�����?v�����N�
+F������?\V=�q;�1���3�5��X�|b܆ק���g}�i�:;H���c�US7w}0�����|�d���`P�
+T���ū��/������֙�J�������@QX�'L#����h"͌����BLo�',<��	+�F���h�3=G#Lox�����bOQ�L�xy���������Gm#U�w�Ou~�w�W��v_�;6,ݓ~�Yl��9e�g�4��
ey2=0��hn愲�̭w>\s�Ĵcl
+�p�U�zY|1��d���zI�ô���v@Fa"0���s�ۢ
�;�I�G���h�k��R�B;׈��q�a�:f�P��v���X幠�|t�����~J1�M*6����(-������x�R[���n������X#�cL��`�Mt�>���F�'Q�m�1Ƕ���j�ڶ���k���
�0/"ה(�2�����>�����9�hK���l`��ܮ�Z�b4��&}
\x20�G6C�&�&\g�hMc�h08�N&.��nsD���B�Mg�G�i�vC"q�ӲuV����h�j��)kR���JVm[�����g�s�~�s�>�i���9�TKuH�����DPjcAQ�:�����d����D-F,e�Xv��R��T����K�f�%BK$�${v?r�����˼��5��Z���}����׶_s_k��}���ګ6��,�����(rɠs8I��S纃s݈��m<Z�v�M�n�Y4܅N���u��.��n=�rِ��e��ie��X�U����W�<�,��r讁��8���
�xg�k�;�l��·Cl��a�|m�7�m��w���1�5�=�>�q�k��R��J��/qT�Vk����t�pe�;c��4P����C����^����\�U�*g�������{�7��1��<37��?��b�-ǚ���X��|��g�:�h�]����f��k���l����6$<�mѱ�]3tـ����^О��%ߴ��ӡ����2l
�
.���G�����4l�ȇ;X"��VW.x������\3Y�m3�L�rRr�d�;'y[�n���P�]��/J����@�2���`�x�cuj68����tJ��^��I��LM1�l>Re=�]�>�t�Tv�X���|���uqYSg�n*	���ʿ������+�^�Z��::���[�o�K1���u�� }�˭W�����>/��F�a#[�gp
��4/��Z���	lA��rl�lcd���*�c
+*%.l�|� {�;����KZ���ek��ZM���Y�B�ї�������1��1�+�˶B[�>�����v������Sg��x���P,�

^��ۙ	]���;���5fHBkh���߅����g��?��@��v����؅Q� 4�Hg�Cc���1���Z�@5�
++���(�$!��F.��k��BTc%�+P����0$��Q�L$a�a�0��V�:�@=*��Q�rd‰q��,�,�B�P��X��jf*�I��r$����P�2$����Bu�j:����0�X�Z��X�Y�=��u�Y&r1������k�ӮTW-�V����DF!Y(�_�U�D=j�
++P�
+�v��.�Q��߇Z��ݨeyO"(g-B��k0Da�Ii�59[�-DR?���5Nd�o��������!���ic��
+,EjQ�Oe��MW�j�^�hW�{!�+.e)��S��V�p�,ly����J��j}]��-�X��!���.e!��i�TR���(S�jP��MAζd�<�BE���)}�$EYN�-WG�{-�P��!�,,�*�(*R֕=�T��B~<�G�^2ޥ�+�&��c/&�JV��X��앦\i -V�EX�����s��%eX�U�����U*�W�����Q7�^��v��pX�ȱĥ�ֽ�[�*u�Zz��3K�$Eي�vu����z��������Ұ���j�F�Q�/q莆J�3��4�|Ģ&=EjR���WH,A�=kN_?^ʒ���K
�{�Q�Q*:g�V-�"Ԫ��k������fY/$]�e���v�J�5s@�uR+ۖ�d���kV&_�5����|!m_�����_��J�U�V�, ig�C���JL���/�K�e,wg4)���=W,I%��ZJ{�����%�Xx��2u֫�\+�]�BK��X���
+][�'�J�,I�{x\�O�?թo�+��aR�kK�����w5.ג�Bq��$5��Yݲ��������}���3����*R�wR�^Nk�V�j}�5�bJ��W���nJo�bg�UuF�bզnY�7,�c�ըE�5���
+g�{�u�*&�_V��}�o���#�(��c}HF�������Z9\�]�f�G8�k���~6�?�Y�c����ި�(�AX�'�U_��(הa)�9��U�W}�������VV�e�(�beRQ��LC	f*>�03q;�`��W�R$a:f`f��P�q�.c�y?EE��(V��6Eˢ1c�9HR�e�Z��&�%��a��Q�REuf(�S1SP�xZ=f	�b
+n�85���Q�_	����?U�bI:��p�/��\�#�Ta�bb���Xѓ�K��ո�G��!I�(�$eIs,n�u&�ކ��i(UxJ��BҖ(�cFH�"%�e	K����阣fL�D�TRHN3C3��L����T�u��jI6-de9�"�
��%��V�R��LQ�J}K�"����ݾ3AQ�r[hܦ���+
%
yO�(��3sF�H��Qv����R�H�55���:��n����Dj��]��(�XEɺb�c��ul[�����OXs-��}J�eo�m�����"D�߫�e�1��}0뵾�(e�Gz�����"�OJ"gI/�g�(H�V�$�����m��B��y%W��G�����a����߂O��%��IJ�%�V�*��*�Յ�Z�W*w�+���k�vV����\5�������ջ����]�ru]��vwV��F}��rէ[�`}OWbՏڞ�Dv�V�нk��AY_����_K3��WiY���gq���_W���!JY$F��5�ϒ��a��#����\�+�t��mЭ�?�_^�WUT���
+a�Of���ٟ�b"��n�\e�^��F}��,�#���Y=��+$O��|-K���R2ʵw#Su�uȺ������וoan��u�� ���]U�[1	�� 8q�c"J1
����BsTE�-�+�*�މ�
&c.8������E>��<>f��������"Է>3�ث�ñ�<j��_&�1i,�@̍4s!Ĝ��CͲ�b656��ƺ�Nꊦ���ABP�:�����9�����:��R#�\��G�K��������t���,�w���>���L�f��D�_s袟>�A�������\��2�>�P�?�A��ς>�S<�PП�鏂�/���0��},?w�~K���	����ލ��	z7�����:��FЯ����#�u��(j���m�茠��t���ӂ~%��V��@�+����:�-�����W��/�B����Ə:&訠W���G����Zv��
+��K�����6�/N�/ͣ���4zQСl�_��4��\��w���=[Nϔ�O=t�G�%�i�������耠'�{��9��CO�3��i�I?ޛ����fЏ��CA�?������{�b��^���vz|W\Ю0��X&�)�Lz�it�x�0�!葇��G�#�H�ϣ��C��+�������4jږƛ��w���w���E���xc9m�j�mi�դ-�6�_�}�L~��M&������m�o�ҽ�6�M뿵����u�.��)��C�Z#h��U+�|U8�
+2��+ݴ�^����B�]���./�;hy�`���j�2AKsh�����L�S��
+A��%�2A���i����!h�\��y�t�/鎹.~���h���~�]�,A�����rh��RA3ݺ����ASY�*h�a�<�&��I#�x����Ģ>QЄ�>>������(�ƍ��q#il����hlPz�p^hRaPC���	x��pdG=0��
+ن�C�v��nd�@�~���Y��2�i0��o��r�qx�q2�4"��GʛL7d��&Snv��c�A�3�|xe�QV��g�P�#�g��a<�OAM��5��&�ݡ�>�t�#�_�FC�Q|����	J���B�VDׅS����p�"(9)�'���8��B�((AЀ�X>@P<�<>���
+�Uȣ�STd�*�H��Gf��K��?��B�d�n�7<�{M�Z؅{�<<��-�<aN�q���.���aN
+��s��An�[#u� �#�;9���%� �epC�i�8]&�epm1x9� x�Y���l��?/�O�o��
�q77
+endstream
+endobj
+
+1275 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /OQQYZV+STSong
+  /Encoding /Identity-H
+  /DescendantFonts [1276 0 R]
+  /ToUnicode 1279 0 R
+>>
+endobj
+
+1276 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType2
+  /BaseFont /OQQYZV+STSong
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 1278 0 R
+  /DW 0
+  /CIDToGIDMap /Identity
+  /W [0 13 1000]
+>>
+endobj
+
+1277 0 obj
+<<
+  /Length 10
+  /Filter /FlateDecode
+>>
+stream
+x����
�
+endstream
+endobj
+
+1278 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /OQQYZV+STSong
+  /Flags 131076
+  /FontBBox [25 -180 979 787]
+  /ItalicAngle 0
+  /Ascent 860
+  /Descent -140
+  /CapHeight 695
+  /StemV 71
+  /CIDSet 1277 0 R
+  /FontFile2 1280 0 R
+>>
+endobj
+
+1279 0 obj
+<<
+  /Length 788
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+13 beginbfchar
+<0001> <9053>
+<0002> <53EF>
+<0003> <FF0C>
+<0004> <975E>
+<0005> <5E38>
+<0006> <3002>
+<0007> <540D>
+<0008> <6975>
+<0009> <5247>
+<000A> <53CD>
+<000B> <7121>
+<000C> <70BA>
+<000D> <6A5F>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+1280 0 obj
+<<
+  /Length 3475
+  /Filter /FlateDecode
+>>
+stream
+x�eX{��y�zz�{vfv�5;s����ٝ�{��{�w���$N��$$��Љ�@���-
+"HD�(Ab��%�	v�#�Mbl��e�*���e�#�(��R3{w�U]��{�ׯ{
�
������uÃ9x �w�~�����Pw;�U� �ٽ�8�B���
�u���S�/*������a�!����>8u`��ei@�)t��CG��CPB����1|;�`�����!����zfv���\x��
+��Y|���`F`��V8
�T��	�ؚ�PHiC
+�"=��Lͮ�BUK��m#C-���]1J�;
��(�P����[�P��L#C(�� ;D	��a��T+[)��":��՞m�I�a;��� lo�p��#�C��9D}�H��녛�����+��i�yၖ�O.d��@�������M�'E���4T������!·������,3Fh����6��v<^���?���ۈ��X���l����U���K33'<}��W�%%vV۪U���O�搘��@��ȶ��%���,�f�*q��8W�������;�Ϙ�!
6x��D!EF��,[-�k!��h�su]4��j�ZD�5l2B�y^z.4�g��&fzed�)+�,Ƙ���ӄf΋���1���T�7��)?���34�
+-:sṞ�9�g�G�����psHN�{�^��qC��6 xb�j�� 
+�
�Z��Ԫ�JY�*2C�H�4�I�#�n�,%����~5��d[J��;���<lȹ�������������5-�!�U����l�R���祠�cw
+��}�[�Rй�5 ��w�w&b�f>����ڋ?��W�	qZ��^/GC�����G���F���hH������Šl�Ṉ�,��b�]>���p��_��	�`@7�p��HU���e�222f�j!��F��(�ۖ����6׊\[��h�Z��L��˵�՚a�t����.<���*^�ž:-7</���|��W���r�Ad�V�U���;�@��(�ͽ��ï�
̱T�[bHG?3Q��C��� ���o���X\C3�Gd6+~�j� �U�>���XE��F]��/1�Hn_[O��븅%�q�.T�i��ᕵ�P�6�q^KȒ$��}`M�
+*�'N=�گ�y��;��ҌO�ɒ�}����ta��mZ�dI��;Y���'�/�C���˪B2f����U3C�����.+����J�������+.�W�+�e��3�������9��\X#��FtB�zDS@�R6���4�[�f�w��%8�9?��,A�_qN��~��_2B�����T��t$����l`Qm4��>�l>�jQ�F1�L_�ʯ��_m�L꓂��M}��~x�T"��|��c�ċ���T�0��w�0E���/���/
"�Kt���.CU��e"���ӑ�B/YU��Ҝ����*ʹL�{�i�j��)�t��鶑��8ha����*��#�/�1���꿱�g%�z~+�F���<�g���n�/�|�ч�
�F���	���+�d2ݾ1�G�K�=�h�r�-��lfL�s�_���H+��vg��û��Ͻ�-���fB�*�kf�Ф�u���Ej٣h�[għ���-SE���7��,��8㐳�'�"m)1 �At����b�^�`�"�8thc馎h4�޲��{}�P9iv�O�b�@�)��bj�C���ܜ�>*f�^１�߀��yT(C�a6�����,YES�S��L��"k�՝��\d{UA�nm���T@�R��/3D�5�ݽ*�^������4bL�y���I
����oVFL2Y�<�w�}��Lѓ�]s_(��oa��3jF�oi�Ba��ydt]7�ю*pagS*�'����(I��J��g�>s�܅'�D?	8/\��	�b��ׅ��vꛛϱ�!�ۖ��ߜ���k�������s���Ɉ#cZ����in�����Ϻz˘F��J3��{����uU���v�㤋�G�_bD�)�ض׿���Bx)u�ġW��TJ��7�O��s�{��Ml8B��fGw{��7�
+�P,���J�V�Om�.s>��N��ug]}���"��o�/b�Ԗ�o�	v�^����4U�Qs˝���6��c�]rX��5�6]�2����V���z	�jZ�%�z�����蠐�,��R_�Y���m)V�G�y���]V2�������q��cG�:?ZD�=��J?���dZ
+g�#wʹ�q
=�kC�Mҷ1ۦo^�S�P.k۶����+�.4$��c/�9�
���ݲj�S��+|o	���.|}IK�c�m|�F#����tcH�ߊ/1�0˼�6��Cp'jF����R�F͌�
�Msl��N8['L�^r5����N5�b�V��ꩶ:�j̯[��dR�� G�
+B–�e�h!��#��ӗ��DF�3(�H�S^ϘHȔ91�܊&L�,ڟv���VlQ��]L]�R�Rw���j�Ƶ�yS�I�~�TԤ@w�kt˟��q˯��:�J��H������:�{�L�8urU��Q���ʩ�q�fK=�t[�s�����ӣiq��Jc)�Zّ��ž�X��nIN�ݳQr���@����?�믟��%_u�ex�_LJK̄O���`7X��"��k&���JU��W5�dF���TB�掺=���ד!���6�E��w�A��xB�k�e�.�ň�e��	���d/lܴ^\�K��x
�Sݢ���A�í*6$�a�n����#����j�h��^4-ã�j��0���Vz&7���oí��'$Ժfl�qa>X<]>����?�||8�6V��*�b���<�M�l���~U�*�5v����ZD���X�jq�?T�S>6oݙ�_�o�*���V��!lJMHtYP�C�{r���.�xg����W�ьRR�6,�a�j�F���.>�k�bWW)�&ʹuw�HCg��9�>�����P�TkN���o*}<��Z�/�D)�{��-e�'{]\�������7?fӜ�o�����[LB(�i�k��#!UJK�_�,_h�F�����}=9���]x���>��?�-�²�������XlPْIIъ{�r���s���������ʳ-_q�g~>9n�_�
�'�����
�?�?7�
+��4�X~�A��	0 '�b�j�CғQ���֣��cX��_d�=B���{��3P\\�S%?�u�c�����x�(�y�i��@
+���	z�C
��p�{���x2;=�\3W�Ǽ�z8�q?����.��f�m�����eX�0
�� �	���,�)�a��g`?L}b�۷0zhatO�j��ŷ�����\<��}�f�V��N���Av����{�'�6��{�2�������
	7�
+endstream
+endobj
+
+1281 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /ZPZBJK+LibertinusSerif-Bold-Identity-H
+  /Encoding /Identity-H
+  /DescendantFonts [1282 0 R]
+  /ToUnicode 1285 0 R
+>>
+endobj
+
+1282 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType0
+  /BaseFont /ZPZBJK+LibertinusSerif-Bold
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 1284 0 R
+  /DW 0
+  /W [0 0 500 1 1 1123 2 2 877 3 5 871]
+>>
+endobj
+
+1283 0 obj
+<<
+  /Length 9
+  /Filter /FlateDecode
+>>
+stream
+x����
+endstream
+endobj
+
+1284 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /ZPZBJK+LibertinusSerif-Bold
+  /Flags 131078
+  /FontBBox [0 -82 1072 712]
+  /ItalicAngle 0
+  /Ascent 894
+  /Descent -246
+  /CapHeight 645
+  /StemV 168.6
+  /CIDSet 1283 0 R
+  /FontFile3 1286 0 R
+>>
+endobj
+
+1285 0 obj
+<<
+  /Length 675
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+5 beginbfchar
+<0001> <2194>
+<0002> <2460>
+<0003> <2461>
+<0004> <2462>
+<0005> <2463>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+1286 0 obj
+<<
+  /Length 1079
+  /Filter /FlateDecode
+  /Subtype /CIDFontType0C
+>>
+stream
+x��R�O[u>���#T�t��8eS?0���,#Nh�
(û8h�-�B)�rk�=�^(���\ʥ\6�]$8J1!��1�>�eK���&�����~L�����}��}�7ϓ��1��q��@]��Ե�
r�^]y��V]��4�̉>E�1q���U�(���(�8�t 6ð�WR0�5��'���iQ�K|
+�8��Q�+W��P��㻺:�^]�2�>�2�d����:�Z�R�>הZ��i��ԋ��Nߐ�a1�	M^����DOhl�M�����R�8q��q�I���%�p�0"�꘍�lQQ}�Ѳ��ѓ�c|!�S��-�6��7pn�O�pB���p�J<�g(�7d#w��oR�E���p�7$B�A�BQ	�O��i@s��e�"-�$���$i`�;��Xp�;��O�����=�ڧ�JU��Vݬ5��m㭦�:Rb��p���D�LH�j~�W��Pe6��%����j��_��ɕ���b04u{��Ռ�&��D�*o]�C��4�~�&�r��M�i�fU�	p�\NM��;#�F_���o���(�A.�t(r��P:D�Ay��=iB1d�$Grte�"$G�
r(�t��l}OI�I\�w���b�Z���6��4��{��Ќm�v;��2��誖�b��~���Or��`paK|���|�N�ʮ�oI�ר�0ܛ�7��j�w�B}X���l��T��efb�p�N�Ű����ttz����~��
�2�N�aJJ�g��l�k��Xp�Y��s�t����鲎c
�a�&VB���ޕ�w�Ǚ��WAg�TȞ�z}�������t�p>K��̖]��m5�(*�t�X��J{m=�>F��>��C/}�u������N�H��c1��fS��s��g��f��I�mVo�&�}���u�7%NF�?2�!���'g�m�f��-�LGg��G��h,�GNfuh�>E����
>�0�a�ӃĽt*|E��Wr���<ɲ�~Vju�9,��ku8X?
�z���w۱��/��޾%
~/�x�.��"CiPL�b$-mm�Z'�*��]
+endstream
+endobj
+
+1287 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /ORLEIQ+NewCMMath-Book-Identity-H
+  /Encoding /Identity-H
+  /DescendantFonts [1288 0 R]
+  /ToUnicode 1291 0 R
+>>
+endobj
+
+1288 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType0
+  /BaseFont /ORLEIQ+NewCMMath-Book
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 1290 0 R
+  /DW 0
+  /W [0 0 500 1 1 361 2 2 569]
+>>
+endobj
+
+1289 0 obj
+<<
+  /Length 9
+  /Filter /FlateDecode
+>>
+stream
+x�{��
+endstream
+endobj
+
+1290 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /ORLEIQ+NewCMMath-Book
+  /Flags 131076
+  /FontBBox [23 -20 516 664]
+  /ItalicAngle 0
+  /Ascent 806
+  /Descent -194
+  /CapHeight 683
+  /StemV 95.4
+  /CIDSet 1289 0 R
+  /FontFile3 1292 0 R
+>>
+endobj
+
+1291 0 obj
+<<
+  /Length 637
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+2 beginbfchar
+<0001> <D835DC61>
+<0002> <0030>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+1292 0 obj
+<<
+  /Length 672
+  /Filter /FlateDecode
+  /Subtype /CIDFontType0C
+>>
+stream
+x�%�]HSa����\�8�\m�R�b���l8�.�׎�uN���Y�M�|��J���ԙՒy�2�>!�.� �o���8�q������?<�Pa!�(��������H��8���-��@W��€*
�H�˘���R�\�SkA��Q���Y�ܤ���9ڪ/�i(Jg|P��}���p�%W�������Y���u�9V�^��YŠvl�v�"��B�"��&2~����I�N�x���nd/0��`�
��;�T�p�p��
G;�������>��5z\��-�v�Ep;/`?C�ؐhG�Q��{�\�e
+�Ȯ��AjA��F�*�^�ޖ�
@�3c�x�v"Iӹ�\r2�H���$]�ABiT��##�4B��d�v�Q_�C������=y*��V��N�*?ڰ���L�GZ�����L/S��4�7��i�'�YQ���=��0�׻�/������pgt|`��U9#Z�Z���`�3��e���x*��QH����a���B�$�Ý�T|J�Iz��s7c3��@$+�Y�ź4���'��&���������'� ܀U
+��rP�Ƃ����뭢��@�����T����@���$�R�I7X�j���q���
n�th�k�omhv�9|��z��x6�!����3��Y��?�b�
+endstream
+endobj
+
+1293 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /FQRALA+CourierNewPSMT
+  /Encoding /Identity-H
+  /DescendantFonts [1294 0 R]
+  /ToUnicode 1297 0 R
+>>
+endobj
+
+1294 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType2
+  /BaseFont /FQRALA+CourierNewPSMT
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 1296 0 R
+  /DW 0
+  /CIDToGIDMap /Identity
+  /W [0 27 600.09766]
+>>
+endobj
+
+1295 0 obj
+<<
+  /Length 12
+  /Filter /FlateDecode
+>>
+stream
+x�����	��
+endstream
+endobj
+
+1296 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /FQRALA+CourierNewPSMT
+  /Flags 131077
+  /FontBBox [8.300781 -188.47656 593.2617 648.9258]
+  /ItalicAngle 0
+  /Ascent 612.79297
+  /Descent -188.47656
+  /CapHeight 571.28906
+  /StemV 95.4
+  /CIDSet 1295 0 R
+  /FontFile2 1298 0 R
+>>
+endobj
+
+1297 0 obj
+<<
+  /Length 984
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+27 beginbfchar
+<0001> <004C>
+<0002> <006F>
+<0003> <0020>
+<0004> <0071>
+<0005> <0075>
+<0006> <0065>
+<0007> <006C>
+<0008> <0073>
+<0009> <0069>
+<000A> <0074>
+<000B> <006D>
+<000C> <0061>
+<000D> <0068>
+<000E> <0063>
+<000F> <2260>
+<0010> <0070>
+<0011> <0072>
+<0012> <006E>
+<0013> <0066>
+<0014> <0064>
+<0015> <0067>
+<0016> <0054>
+<0017> <0041>
+<0018> <0044>
+<0019> <0052>
+<001A> <00F3>
+<001B> <0078>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+1298 0 obj
+<<
+  /Length 13078
+  /Filter /FlateDecode
+>>
+stream
+x��||SU�����l��I�W���6}$}����
Ry��H�<���P
+:��:GQ::*8rF�5�΀��:8#>�9��9jGg.:ރM�o��Т3w���ַ���z}��^k� �����׬���x�l�r�U+�������V\�b��{�>k�����=��9ky��Y��g��I:���_�r������n��/������]��r�Wk5�@������i,�;	��МY��o--^
�>0��������&�
�}�'�p��'�E��S��GɛޞJ��;fg���l��=J:�LI�X��X��	f�j�[<�L���`�܍k����ِ�0�@�X�JŒ
H��0��0�o��i����@PB*�rJ�l܋\� (I��~�G#���W�m8U��9&��Z���}[x�a��$�ڜڙڅl�eyC�IU�Vb6��p6�?�뤓��GSw���`���$$@�31���p��	���	!&RD�ɛ�
+CǓ�S���V�S1��D�H
i���<�{g�%O��0c6��q=�a;����?F�t6�Þ��1��w�x��ђ2���md/�F`C�� ��zLRvv�a<��po�wxE�0�$!2��'7�[ɝ���Q��<I>�*�.c�f�E���)}��cp�
$#�:\��x��"RB�$J~OC4̈`J&�SSR/��A���x��S0��?�����:~�?��F��B�Dd&�E֑��	��v�G��
+:@O�{]�+<9t iK$�H�R{R��oR�)��:4c��}
+��^���"�����I���!�������j�Mt/M��l;{Ep
+�%�&W&�K�Oդ���A'jP�q�`:ч[��q��~����� ^RA.!���M��Ud5�%דȭ�1r�!o�?��P�����h�.��л�
z��M�0�Y��������{��Y��P!L����u*��ڮ}���oW-z`�7ɲdK�����o'?Je���>���W�7bnÝx��q���	>�_�7|E���B�fr	�B.#sI'��\I���d�'{�y�#ϑWȫ����%�F�hm���z%��>@��8}��G����Y�U�j�ȺY/��ngw�����lB�0K� ��b�%�{U;U�U/�>U���32�
�^��	�lvc:e�S�{!7�s�4�<G��yl:�N�i(9�IX	�f�ZV��
+Q�e"������̀���y�6ڍ��gq�N"3�5�u��.`;���F�6��P#�Mh"��1�&z�۴�=%����ҲoU+�1�Q�DE��
a<��?�<2H�S;y�4�; �H ���a
+���N�-t2��w���k�V�#�g��$k�t��U�&ҋ٨�r�c��j��
���Mn!6��sdɧWB`F�'i'"x�Xh�	5X��d�d��ktƐ��o�CE�|;H��I�G�	��P
��s4�T-��C<����!� CT4�.t�R��W����}�ȣ�	����h�7���Īф��I���Z�"�<�F�	��W�*@�T8�����Mv6ՙ��T��p��$��l�~L���N� 3�mR�˰�>%|��%"�w�
+rI� ����DzSYdn��dž�6�
+���9\��p��x���B��R��|,�2��U�E1
-��b:.�<tc	������ �bbh�\�E��ы>\��qn�Fl�܋��w�q�����z
]���>{���e8)�!l�,�c��2�E>l��ԛ��b����`���g��S�:����w�'�3u3�0�|-��*�4;miW_7���������4*).*������y�.�#�n��X̢)�h����J`� �h����L�T�ˁ�R<�pDEw\Z(��F��%��¸43�P�_yf,�;�ID)�HiXj
H��[R�̛���[[�R|P��(�v6�:e�4,I���-R�tK��k�nj�n)
�}Y��@s��4�}���@sVi����}$��(�m��Bk,
��]��ָ3��gg��ħ��hmq�rgi8N���7�4+����q�2������}�c��$D,���,��g;��P<7�Ͻ��B�4�4wl��f�Z�$^ܴi��=�cd����NGi�4�mݛ�ⱅ[&���gI�;���Ύ8���4,��U���h�5�˥�.0!�t���Rܵ)����
�+v(u�Vi�쎀���[<���4s�~gLr�n)
���ݗm��H��|�)�j�y~g	�Q��x�;.-����ӂ:���a��:��?��4�_2��uY\�ܽI����qU��6}�8�~>�fa�F] ~r>9�jq�p��B��"�渚��Q)ז��I�e�բ���Z�8�w����q�Ұ,soNİ�4,��gt���+u�i7o96�b��[��[�?��K���[\<��$�sZ����?i�I���
+�Ϙ�!�n���m��Q�t{����i�`n����)�qK���ȼ�a�q�@�0���F;�#]C����=)�v�e�_|(���?�d�L3>.4��0�<jz�M�}v\����6mҏjk�uo����6uoZ�H�/
+Hb`�!d�M�[��)�H�쎷m錋�Kɸ��>�	����b��Y�:��t��
Jhs����RP����T�m�Ԛ1�*�z�*AO3F]:
�{����v�#4U<�2�*~�"E�E�����,�d�|��o%v�ۘ
+� 	�@��]��"Vg�ʮ�X�d���s3����鬮�ڜ�<
f�]4Kk��+�ͱ8���1����F���� �:�Δ2ye1��1�Uv�Y�kk�TW�i@��۬�y��9r,�o��.���yJ>�����L�����KN.�47''�A�#uJ��~d!�L��Y�ݚkg��W�ޢP�Q�V���ef�C{�e�e���果+s�6��d��i2�c�MΨ�g�*y�h�����@7tc�a��1�3��e�FW���O�O��T��	��~G��	���"���PWo��F��TV�=�5�=�?c^��SgaM�= Z����S!'��~����|:�Ez���<�#�e��'�D��X��Ͳ��Q���jxbN�>��Y��k�%���՜�h�I��d��c�5��Ĩ�'��|Bh���X��U��~��j��#D���O��X�?Or�|�{���K�:�����#�����?O=���)P�'og���\z�����b��:S������6\Kp�k�܇\ה�[]�����I�c��ៗ�	���{�=Ōj��X6TDU�S��:?�8xU��e���p�-�,^��l�]�������p�:,Y
sv����
T](�n�&�8uN�_OL���/�L�,��}�}�}q�p�w������%�i�N�v�N�*~ͳ�� zf0:�F#���R_/��ԗ;��U���PU�:{\aC�!\VN}�����p"��@�5�H�q_q�<G(��y���dU�p�P���WW匠������^���j~�^��ѳ�!�~}z����;ɿ����
<�Bn������g9�t*پ�ĉ;�x��N	�l
+�-f����N\Z�N!��\k0�~�7M��\�ۑ��ʼ\�|��&��˒���MVIn@�>����zMZ]�hR[e�%I@�=A��E��=�!��l�ą�Xd("��p���~("FČ����p��>5m��ӓ�;��HfW
B��ʊ��1ђ#�U9������i��f��s2D�l��aO}K����k�
a�vX�U�9�^(f�n�����Y��/,������b"O���ܱ�z91���\�l��o�.��*uJ(g�pB��ez��6ZE'Йt1}������-�;?p�/�Ǿ����O
��Nv_���[�^�ɽ�}��~�3*�:�a�qv����Z���%I Ĝ'�jٜe��j�
�	�q,�/5���V��z�z�z�*X�r��Q�c���3�0���rN����n���/��>/M�>x�����vβceaը�$Q�s�F(�������M9٢C����w�������\�[w�}�E|���j��l���+�u�߼�ɽ[
�#�`>�]��E�>ۏl̤3t��_G��B�Z��v���w@�ՃHfQ�&��t�ؠh�.���DG+�
+����QgS�
��L�qȡ���]>�s+9}��9��\O�N�$�G�k�vk��D-���Z�J{��{�k9�L�Ϝ��o7�a�g9gj�<�ۜ�5�L�d���F�_��nktO�쿧�)knμ�eu�����Ꮊ;�X�w6���Pq X�q9L���5U
�l�)���`fg���Yn�'H�
&����7�	��e4h���Ӽ�����j��h(��1�E�&g4ֹ��;[�5L��YWd�)��h���r��š3\�NW�]�8d�/�bՒ[o��֓a���R��Ӻ�,�� � ���C���G�J�>b�`��u���h?./�#nWC��7�k��G
+q��NHZ�)�ۜ��
�����'�M}�����ֱ�������gX(�u��^�v5��8����1��m���:���-�`m���[�1�b�2E=��SAz�V��P��"i{�s�?X,��͚k����Z���r�⼖��ښ�`~Pa�*v�Ò�p�s��n�ƭS�*n{�e�߾������ř��#�g�L�t��'�~���\.�v����Շj#ES�#pä���o�vy�𺜶2_E������R��Ȋ
\2�
+�*3<�+�;��1�L�b�Z�959F��j��]�J�7�=�{l6���@��O{�F��YJNDzA%٦Y�ՙ7�܉C����z�����g����3���
�O�W���ｗ�����$�C|��I�+uJh�B���Ϲ���r�G�����ԟѫ�ɾ-���g���v�:WK4��Ț�M+���V��l���i(N��cfoC~����
�3�z�� �Ŭ�V'��=�g��G�IЏ���<����Ld(�*�5"FġA~2,���ݔ��r볲\:�nCZ��|37#z����(:j����2�E�&O֭���X�Qt��������v�}b����KUKF��d��lZG�y�7��
��~�ђX�4�T�]og��ƛb�fS��4�1p��B2��9�E�E�����M��U�ꙣQUhǎW,�S�ȸ��q��7���L��+Y�t��Ӗ/-,�%fa��l�Ŕ���8����6������T�+��jZ�V^]]Q���=��Is[��9�������Ң<��hJ��Ц.��K��f�16�1[ʵ�|��~;���J�A^�i��F�K�Ƙ�E������9���5	)�}�+9�!�X5��"�/��z��柖�z/nM�?��N�oϯ!���P"uz me�!�����UVA�P�#*�S��#E�q\�0%!1�%��7���qwF�6�O�B�/Фއ:�>H�͌��%��9
���Q�g`w5��]��HW���`�1�U�v۰	0+&�ftY3��"O��?XѴH��w������sjuٕMm
+8��4<�Y���'i�hN_k[[kå�r�@���3���h��W�$]Ȉ��)��N	��~aY�����e,�)�5ks�u]�����z�k�f��ד�5�f�p�}�Ҧ'NRB�����e��v�=����ʷ*?���9s�
+����/������1���j�.<&AN��ۋ���Y�NFX��pq��Pd�%�y�`�$ɜM2)���4y�����|T>%keW]�
+Y��W�RU�Rj�ؒ#l=�2�'�J��b�q�o��k�����͹��Ro�X>�=�̰�QhRߠ&�%jS_™:�ߢ-Ӟ��{3޳U񞽩/��:�pP�R��³�֦�@��j��`F�*�u<��O?�θۧ��/�'��\}��]�P���g��z횕G���U�=moo��^ﰘ�&WɃ��Nrʓ���i
3/]:w
�psa��	�ȚX���s��,Q�So'?���#4N�S���k�j^Լ�9�Ҹ��\E��>+��wX�����\���+���
+��O{�Fb��3�:��憬��n����@myemmU���H�Y(.*2��:Q��I�S��V�8�R�ъ���X��X<�v�DbD4�t�����O������8_.�J�.p��>��x�|v�,~�R��A��z�;��	���w�/�
G����w_���W��N_�N�+~�n~S�����O(���й�[�M�����.��T�v��Lx�=�SB;�G	�<(�LQ���Lїl�Wx�wZ����BM����"M)�R8G�*�W�ܴܹ�����H�-֗c�̹����+�v��.�hs�Ŗ�&q�x��ף�G�{ h1��V7
�f;=�&#�9���bM�~A��Y���n�v�����c>�s��r0f�Ew�)�n��3t��.	
�<�rw���3\����f^��@oW]�D)2�bP*ޛm��1�۱��@r;��y����:�&�s{~���9���s���T��ϕ����O>9i�5����^z����I%�e�����6R�Fr���|UD�ׇn��l��6�f}>���
!�>ߜo5��ACi8��vW��=	�M,�I	q���9Z��bɩP�������!w�����)�*ߘ]ʾQ<�7��D"%�� �D3lw����6O�9�bT���ˆ���S0�+�n]�$Y�g_0�-�A�̋���x'9��!f��&^x��$�S�4eٓ�3��4y�0�����4�ԛ������d�2.!=�G��Zr}ɚ��ߨ���Ӽ�{���?�?�k�,̮�la���Lm�(��Y��tz����E�eyy����/�X^�X\nj�y�v[v���/��݂��uP6i��UF��5�M�[��*O�sV]1�1�����)�-#<�����Yht`��Pq8u���B�Dxd�2��}�E�8{�2��K��\�#%�{������Я>��B�U<�U�
+{����?y����d��|��k&S�$���8���,�llj7��x��-�0�]��Ŭ7d��n�~�e��˽�s4��b����
��m�>���b:��a��ݡ�1�hj~QQA�?��mU�qU#Ɂ5[��4 ��GEY�����ǭ7i��P��V)���l�|P��;/�i�������'Mڒ�m���t�@]����X��ϰe��а�6����<��~�3G���q��y��U�Õ���՝<e�N�xN�+ǩ�	�:�$k�X̹�ٌ�_
(���l��}K���W�҇��
-���k�b��7��~TM���'(����(�@�W��g�&��Fmm����V@:ȝ̓�m� w�
+Q ЋL
cm[��j�S���4$�8��'��"�Ŭ��%�)'qr�@T���+h��[��-�vJ��F����g	X��V�Uk�ڣr��[���Jk`���YI"�����4�{�����Օ�
0��W0�{�����qgg����?�C�m���u]�H{ �#l�l)�n�d��@��ͷ��"���3+|��+����U?2�\��0`8z>t2��՚t��y4����2A��
ժ$s��d~�]�r`��e��N��'Ȗ�~�Dr��� �Č嶘m��
�`s֮;D^C��SZF�?�0J$�#�Cg+}t(�kdL�GJBnO�'�C�]�&��77�Y=:zp>^��B۰N+'��R��O���w^����js~���k���&���
�wٵv�%��hݐ�[��R��=�
R��ç/)YZB��Q�(�k�W��<s5����t]�Y��H6�dqL��V/YV
���I���JB\��_��-�C×B�׽��谱<xF<s���!�
+�=n�p��l�E�y��6:�������p�S��h�n�3������[=��#��vh��p��֯�]��*����s���{ϻ���_�����?<u���}�����|���?�7&onb�(D5��*[�����͂��3�����,]\���:�
Ž�[���,�Q��xw�aov��n�ȋ;xC�I̕�f1�{w�,�C2�j��Q��<�Hz��ۭ��I7M�@���N�sՖ�i�G'�_����d�wo����t�A�<8�}��ŷqG�N��+uv�D[�H}3��"�:;�V��P�+K��ԧ��i��E�F��Y5�tX��r���ӊ��X.������w��?E-�|m�}�[O�;��[_ݺ�?�c��W��(t94{B���ܜ�����or� A��ǯ�~��_�!"܅�t�ZQ���=�1�@��jU���is��1K�3`�]T��:d�Ӣh桺Y�DO���s�#�<Q�4��j�6�S�S��
+�_�${�*v�E�E��wF���\٬(�����"�������ɞP���o�â�}��#gcrm�%�����](Ay:c��Ի�cK'Z.q]n+�n�n_�Z�^�M�)���p��T/&�#1�q��!#=e$�b��(���fK��7e��%�`qI^�$�cJ�Z]M�jF�t�ԙ�T��Y��K��b�{x�$|�J���#�b���q���.W����vY݊O⥥VJK�
�N�
��|e��L�,
]9A�����Ic�Zts
fbr�ܧ�_�w�����Asi�r�4œ:�߬�w��b�I5�	���_�Sf�� ���+���ܚ=;(*�v$4�}�Ƞ�k�\MD�Xۨ*�(�X�m������]��g�������wĿJ�^f�~U�\��X������wE)���&ҧD��/�MJ�K�;v��g߇dc��������e��.z�둟�5�]�C�b��;�n�ܑs�u�m�o��I�Z��x[�!����J"u:����L���y��Y�b�\y���
+�X��X�>yZs�����ڐW��^-��L�d�L��*{�D�4P8K��z��c���y�{;ܦ�����w�j��-�d4LY&A,(,(*(.�9����}~�l�2R`
+���W%�
+��pC��Nq�F���o����5�1cj����e��XG���ʦ筹u�-Ɉj%��u9�'�o��<1;M���˧,[{���M������M�9=|	(:���](�X�� ��ǦJrcȚ�h�����u�LW9�r���r��6��ڭ���<^y"���M�T��/�fSPW��k��շ)���3ߞ���+�BF��7���)g��I�ӥ)����P�a��[�l�Y�y��O*	�ji8�+Ø��i	:h)��e�,K�,�e���iŁ`I�N�����PJwnj�X��<�`�t���桞)��'����o����͋��1���_�Tt�M^1B�E�u���:�ܓFx"m�*'!~����rg�j�E����ԍ��
�w�e�-�����j���i�X	��#�:m�i���a�H��՜ij8��٩�����KL�}�[w����ջ��J~����g<�>�*�''�>�/�8���]����u���+�fߗJ��SU� (@���_(=C�ޔ�zu����B�x�-(&�[�Q�UˣI���-B3�+v���PK�y����[a>7YiY%�
+t7���F����k�ת��y�٤�������li���(�TW��* ZD�TY	��i�X,�\c���@L
b�>��� 7H
�ʆ����@CɄ��چ���Xss��.��F;U5	Rv@j~ *򈿛�A���
+vb��L�U*�r�V%����
+-
+��@a�)�<TQ�9[�z��Dݠ��a�I�W�g�g�N�C���S�8�^0ח;����x�נC<�+yE&w�!^��6���[K�?�;���D����R���o-����-
���Ƴ$R�9��4f��7ިh!�b\@㬊���Ҩ�EE�%�ٹQ�k�GE�����L&�)�p4;����D��FS��T8}U��(��dE��,K�:��(69�5����f������D	O��z�Q“�n1Jx�lu���'�>Oj�1�I�UV��Q�oT���W��͉������y�Ekc "Ze�d�J�I�#���ĞF(���qL5!i��z���a=��P�nrs�jr���Gos�P��Ee~�u�r������9&��c_B�$o	�&������4vcF{�'_JG�]Cz�$���#V_3�9��aC̬��r�&�n���3{��
��ߡ.D�`We����	��7�v��ԉ���L��kV�E/����=l@�e�YL���:�����f,��Y��iJ��ۅs����?����S���s�[�?DxW�;��r��N�W�6l��AJ�+�n�3~��@9`,"+�J:�9YB�V]�zS}��S�u_��5��d�{���YbH���ˎ�}9��~h��2�<��%�T�Q�o��C���LC9��&�2���4)k�+%Sd�w�a������sV�C�X��?kQA�qt�PG6�h
��@a~���l��X������k�rX�V`]f�i8=�4��N�?
�矆��O��������pz�i8=�4|a����R�fW`È�l��
+,򹩪8UL��#�mJ?3�>�ީ<�H��
+�j���
�+��}(Q��>��ȳU�8�1툱#�
�ky�P�
+�A�l,E$L�*\�UX��X��4c�`��.�Z,S0� �	+�fb��R�E�R�Az�נKP���k�L��0=�V��
+���;��O�Sz]���<$ԢLY��"L�2,��BV�J�E1.��e��/%=�t��5���@�2e�!a��'KЃ��*~I����W���<7Wc�*;(a�*�=�����+=����px�2㫕y�Y���<�IX�k����у�X���J�������N�tz��]�R�c�2K^�#��Je-ë�̽��V���h�<�fD�j��K����;t�2�b,���w1V`�B>�Z���%J�j��֟���X�2=�]�P��rN���r��B���ŐXt~�������_ߥ��\�� a�B�4�������w��WÈ=�+I���a�y����׺D�
��U
+��J�;�pԮ����wu-V+)��(��9��\���MU����s[��Ĉ,���/����'��W�p���~Ŏ�_�C�0�;����|�C�5FϨ�ҧed��X/x�J�]�(�څ�Q=]�L�B\��b�T.%q��
+m��k2'y���
$��/�M�<n�k����$L�e�s1-hG�a�`.�DL�X��X�˱�0f���cs�c	vl`����ί��y�Q�t�Ѧrv��1<Ŏ�;
�1l��0��1D�1�v�Ҿ�A��1vo�#J�a��ac��;
Q�aϰ��}�;�ߙ_�E���G��e;�f��+2��L��mF	|l{&��64�LM:�_��H�͠�cl���iU�`lD
v���?6R�kr�]��]8�v�
��`�@�c;g;q���l�R��_V�]rq���k�d�Ԙ�UMz��.C|�#��e�
T��6u�9 xJIw��<Z���t��nPZ7(�*^��Q�f`���H}Jj�)��f�>6�MV���il��Oe�(�6Q�/e�ဏ��VX�c�Y�R����>6��)�D�6��hZ�ڰ����x}�Y���R���`7k�)�f
k�ւ�L�$����f�̚XL,�����X��X�EY#k���a��x�XDYc�,�(�`�`
����
��EX�E�E��b:�E7��
+&F9c��0+E>|L�[`��I��G7����^_�IG`:=�nz��
��*���J7���͘F7c
݌
t3����-��X��(�F�1!�`��#�*%���=y���25�a�����+����#z��cŠx��(+�	V�S�P��BLc�X�
+���!V�`��/X!R��X!6\�÷�>V�����"��",`E����+�
+�X��b�J�tV�m�G3m~��ye~Lc~P�3�R�⇏��Δ���8S�X����H���nE�n�B�!.�[�`l�[��;��bV�
+Y!+bE���Lf2���|�Ǽ�K���t�F�wҭt+�B���ts��T�h�.�]U��v[�C�O�����B�M�cz��,f��I���H�[I�P�5JS�ܘk���|�����7�3��1�8u��m��|�1A�rC�?���C��B�1!cm�X2��Mf�I�ˆ_+�%�RR��摹F�%�C��H�
�f��rB ��	-�ݒ.]��x�Ӿ
+�*_8]Lg��搽АP,�yE�@��k�4��"M�&��i�Z�V�fk
Z�V�Uk-�Bk���6V��3��SA�EKS�
(�RLF<����YH{��b�/��_�
+$�~Ƽ�*0��-�h�=�jOhR3�u���n�����3^����;$ūnu�!�[��3yg'�c�@�n����#ji4׷�|OҝIG�O#�)|&y�{�gu���Wq ��ٞ��!ZGǴ��cy��qH�O�Zg�z}K�<H����d�)x�8���t,�+�Yϫ�yG��/����a��
+���8W�ƹJ��*���8��i�
+��9�￀S�8#v�gB�|�!L&o�k���	��@kOw��;�����x�"I:�f�v��-^��=	�v��%�h��M�����x��@�>\�:�c�u����ɱɭ��-��'.,yb�pw��d��t��wV�ǚ���4?��'��c=�ǚ����p��}ZL�l�����,�����-wN���nn�7� �DV�3nL���Ҧ�&�$@i��)"�丩Av&��4��	qs`��ZZ����e��__��+�����o�����J|�}k�<��ɠ�7ݪ���0�-��f}}�k�aȾu཭�Ʌ��C������Q�K>����#D�֡oQb��[�a�>"�>�>�L/��8�
+endstream
+endobj
+
+1299 0 obj
+[/ICCBased 1301 0 R]
+endobj
+
+1300 0 obj
+[/ICCBased 1302 0 R]
+endobj
+
+1301 0 obj
+<<
+  /Length 314
+  /N 3
+  /Range [0 1 0 1 0 1]
+  /Filter /FlateDecode
+>>
+stream
+x�}��Kq�?�U�X
94%MQKS��N�#Ԧ��Q�v��B�ˡ�h�Fk	������!
+��ՠ����ZP��~xx^ޗ��Q�����x"��^P��� c�!��H0
+ ��0l+��ߣ�y7���;�׫���;Q�?V.�_�tF��3L�E�K�)y	��z�80e�IP���k��T�/%[�p���:8������+%��d����"��G���	`d_�{ٹ�֖gz��m\��8r��S�i�������f�����1\���C��U`��S����]�
���@�o���
��_�
+endstream
+endobj
+
+1302 0 obj
+<<
+  /Length 258
+  /N 1
+  /Range [0 1]
+  /Filter /FlateDecode
+>>
+stream
+x�u��J�PFO��U�v�D�D@�`]\�
+F��Mk�I���R�B|��&إ�n�ऋ��(�H�r�Tų��9�`��ըa�&�ʺ���:�l
+�3Ŭ�*ުn��h)�&
C|>b��������纝�黓��A��v��Cƫ+	�y')�
+̵�8+�/<����a�Rkp��ش�A�mW
�Q�Ih��
�%��l��E5VW~��d���X��)�ϴ�:׺
��b���k#�&�_¤3�0��
�I�
+endstream
+endobj
+
+1303 0 obj
+[1411 0 R /XYZ 79.37008 679.7396 0]
+endobj
+
+1304 0 obj
+[1411 0 R /XYZ 79.37008 449.1866 0]
+endobj
+
+1305 0 obj
+[1411 0 R /XYZ 79.37008 197.98663 0]
+endobj
+
+1306 0 obj
+[1413 0 R /XYZ 79.37008 781.0236 0]
+endobj
+
+1307 0 obj
+[1411 0 R /XYZ 79.37008 769.8236 0]
+endobj
+
+1308 0 obj
+[1417 0 R /XYZ 79.37008 439.4796 0]
+endobj
+
+1309 0 obj
+[1415 0 R /XYZ 79.37008 769.8236 0]
+endobj
+
+1310 0 obj
+[1419 0 R /XYZ 79.37008 710.5156 0]
+endobj
+
+1311 0 obj
+[1419 0 R /XYZ 79.37008 521.2377 0]
+endobj
+
+1312 0 obj
+[1419 0 R /XYZ 79.37008 353.56464 0]
+endobj
+
+1313 0 obj
+[1419 0 R /XYZ 79.37008 217.4176 0]
+endobj
+
+1314 0 obj
+[1421 0 R /XYZ 79.37008 733.1076 0]
+endobj
+
+1315 0 obj
+[1421 0 R /XYZ 79.37008 565.33563 0]
+endobj
+
+1316 0 obj
+[1421 0 R /XYZ 79.37008 413.42563 0]
+endobj
+
+1317 0 obj
+[1419 0 R /XYZ 79.37008 558.90967 0]
+endobj
+
+1318 0 obj
+[1421 0 R /XYZ 79.37008 262.15564 0]
+endobj
+
+1319 0 obj
+[1423 0 R /XYZ 79.37008 519.45764 0]
+endobj
+
+1320 0 obj
+[1423 0 R /XYZ 79.37008 335.92264 0]
+endobj
+
+1321 0 obj
+[1423 0 R /XYZ 79.37008 173.1336 0]
+endobj
+
+1322 0 obj
+[1425 0 R /XYZ 79.37008 733.07465 0]
+endobj
+
+1323 0 obj
+[1425 0 R /XYZ 79.37008 586.04865 0]
+endobj
+
+1324 0 obj
+[1423 0 R /XYZ 79.37008 557.12964 0]
+endobj
+
+1325 0 obj
+[1425 0 R /XYZ 79.37008 277.92963 0]
+endobj
+
+1326 0 obj
+[1419 0 R /XYZ 79.37008 769.8236 0]
+endobj
+
+1327 0 obj
+[1427 0 R /XYZ 79.37008 730.6196 0]
+endobj
+
+1328 0 obj
+[1427 0 R /XYZ 79.37008 474.20563 0]
+endobj
+
+1329 0 obj
+[1427 0 R /XYZ 79.37008 186.26562 0]
+endobj
+
+1330 0 obj
+[1429 0 R /XYZ 79.37008 481.18863 0]
+endobj
+
+1331 0 obj
+[1431 0 R /XYZ 79.37008 781.0236 0]
+endobj
+
+1332 0 obj
+[1427 0 R /XYZ 79.37008 769.8236 0]
+endobj
+
+1333 0 obj
+[1433 0 R /XYZ 79.37008 730.6196 0]
+endobj
+
+1334 0 obj
+[1433 0 R /XYZ 79.37008 594.80963 0]
+endobj
+
+1335 0 obj
+[1433 0 R /XYZ 79.37008 338.59363 0]
+endobj
+
+1336 0 obj
+[1435 0 R /XYZ 79.37008 408.01263 0]
+endobj
+
+1337 0 obj
+[1437 0 R /XYZ 79.37008 781.0236 0]
+endobj
+
+1338 0 obj
+[1433 0 R /XYZ 79.37008 769.8236 0]
+endobj
+
+1339 0 obj
+[1439 0 R /XYZ 79.37008 710.5156 0]
+endobj
+
+1340 0 obj
+[1439 0 R /XYZ 79.37008 488.14362 0]
+endobj
+
+1341 0 obj
+[1439 0 R /XYZ 79.37008 252.43561 0]
+endobj
+
+1342 0 obj
+[1441 0 R /XYZ 79.37008 665.07263 0]
+endobj
+
+1343 0 obj
+[1439 0 R /XYZ 79.37008 525.8157 0]
+endobj
+
+1344 0 obj
+[1441 0 R /XYZ 79.37008 477.32663 0]
+endobj
+
+1345 0 obj
+[1441 0 R /XYZ 79.37008 241.9226 0]
+endobj
+
+1346 0 obj
+[1443 0 R /XYZ 79.37008 598.99963 0]
+endobj
+
+1347 0 obj
+[1439 0 R /XYZ 79.37008 769.8236 0]
+endobj
+
+1348 0 obj
+[1445 0 R /XYZ 79.37008 730.6196 0]
+endobj
+
+1349 0 obj
+[1445 0 R /XYZ 79.37008 548.3192 0]
+endobj
+
+1350 0 obj
+[1445 0 R /XYZ 79.37008 164.35822 0]
+endobj
+
+1351 0 obj
+[1447 0 R /XYZ 79.37008 561.16864 0]
+endobj
+
+1352 0 obj
+[1447 0 R /XYZ 79.37008 289.22266 0]
+endobj
+
+1353 0 obj
+[1449 0 R /XYZ 79.37008 618.42365 0]
+endobj
+
+1354 0 obj
+[1445 0 R /XYZ 79.37008 769.8236 0]
+endobj
+
+1355 0 obj
+[1451 0 R /XYZ 79.37008 730.6196 0]
+endobj
+
+1356 0 obj
+[1453 0 R /XYZ 79.37008 437.35962 0]
+endobj
+
+1357 0 obj
+[1455 0 R /XYZ 79.37008 553.6016 0]
+endobj
+
+1358 0 obj
+[1455 0 R /XYZ 79.37008 355.83762 0]
+endobj
+
+1359 0 obj
+[1457 0 R /XYZ 79.37008 735.79663 0]
+endobj
+
+1360 0 obj
+[1459 0 R /XYZ 79.37008 553.5686 0]
+endobj
+
+1361 0 obj
+[1451 0 R /XYZ 79.37008 769.8236 0]
+endobj
+
+1362 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 422.33746 515.9055 429.85046]
+  /Border [0 0 0]
+  /Dest 1307 0 R
+  /F 4
+  /StructParent 0
+  /Contents <FEFF0031002E002000AB00BF0053006F006E0020006C006F007300200069006E00760061007200690061006E00740065007300200065006C0020006500710075006900760061006C0065006E00740065002000610020006C006F007300200075006E006900760065007200730061006C00650073002000660069006C006F007300F3006600690063006F0073003F00BB0020007000E100670069006E006100200033>
+>>
+endobj
+
+1363 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 406.57446 515.9055 414.08746]
+  /Border [0 0 0]
+  /Dest 1303 0 R
+  /F 4
+  /StructParent 1
+  /Contents <FEFF0031002E0031002E002000AB004C006F007300200075006E006900760065007200730061006C0065007300200065006E0020006F006E0074006F006C006F006700ED0061002000660069006C006F007300F3006600690063006100BB0020007000E100670069006E006100200033>
+>>
+endobj
+
+1364 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 390.81146 515.9055 398.32446]
+  /Border [0 0 0]
+  /Dest 1304 0 R
+  /F 4
+  /StructParent 2
+  /Contents <FEFF0031002E0032002E002000AB004C006F007300200069006E00760061007200690061006E00740065007300200063006F006D006F002000700072006F007000690065006400610064006500730020006500730065006E006300690061006C0065007300200064006500200075006E002000730069006E00670075006C0061007200BB0020007000E100670069006E006100200033>
+>>
+endobj
+
+1365 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 375.04846 515.9055 382.56146]
+  /Border [0 0 0]
+  /Dest 1305 0 R
+  /F 4
+  /StructParent 3
+  /Contents <FEFF0031002E0033002E002000AB004400F3006E006400650020007300ED00200068006100790020007200650073006F006E0061006E00630069006100200063006F006E0020006C006F007300200075006E006900760065007200730061006C0065007300BB0020007000E100670069006E006100200033>
+>>
+endobj
+
+1366 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 359.28546 515.9055 366.79846]
+  /Border [0 0 0]
+  /Dest 1306 0 R
+  /F 4
+  /StructParent 4
+  /Contents <FEFF0031002E0034002E002000AB0045006C002000650073007400610074007500730020006F006E0074006F006C00F3006700690063006F003A00200068006500630068006F007300200069006E0073007400690074007500630069006F006E0061006C0065007300BB0020007000E100670069006E006100200034>
+>>
+endobj
+
+1367 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 343.52246 515.9055 351.03546]
+  /Border [0 0 0]
+  /Dest 1309 0 R
+  /F 4
+  /StructParent 5
+  /Contents <FEFF0032002E002000AB00470065006E00650061006C006F006700ED0061002000660069006C006F007300F30066006900630061002000640065006C0020006D006F00640065006C006F0020006F007000650072006100630069006F006E0061006C00BB0020007000E100670069006E006100200035>
+>>
+endobj
+
+1368 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 327.75946 515.9055 335.27246]
+  /Border [0 0 0]
+  /Dest 1308 0 R
+  /F 4
+  /StructParent 6
+  /Contents <FEFF0032002E0031002E002000AB004D00610074006900630065007300200071007500650020006C00610020007400610062006C00610020006E006F0020006300610070007400750072006100BB0020007000E100670069006E006100200036>
+>>
+endobj
+
+1369 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 311.99646 515.9055 319.50946]
+  /Border [0 0 0]
+  /Dest 1326 0 R
+  /F 4
+  /StructParent 7
+  /Contents <FEFF0033002E002000AB004C00610020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C00200065006E00200073006F006600740077006100720065003A0020006F0072006900670065006E002C0020006100750074006F00720065007300200079002000750073006F002000610063007400750061006C00BB0020007000E100670069006E006100200037>
+>>
+endobj
+
+1370 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 296.23346 515.9055 303.74646]
+  /Border [0 0 0]
+  /Dest 1310 0 R
+  /F 4
+  /StructParent 8
+  /Contents <FEFF0033002E0031002E002000AB0045006C002000700072006F0062006C0065006D00610020007100750065002000680061006200ED006100200071007500650020007200650073006F006C00760065007200BB0020007000E100670069006E006100200037>
+>>
+endobj
+
+1371 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 280.47046 515.9055 287.98346]
+  /Border [0 0 0]
+  /Dest 1317 0 R
+  /F 4
+  /StructParent 9
+  /Contents <FEFF0033002E0032002E002000AB004C006F0073002000660075006E006400610064006F00720065007300BB0020007000E100670069006E006100200037>
+>>
+endobj
+
+1372 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 264.70746 515.9055 272.22046]
+  /Border [0 0 0]
+  /Dest 1318 0 R
+  /F 4
+  /StructParent 10
+  /Contents <FEFF0033002E0033002E002000AB004C00ED006E00650061002000740065006D0070006F00720061006C00200063006F006E00640065006E007300610064006100BB0020007000E100670069006E006100200038>
+>>
+endobj
+
+1373 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 248.94446 515.9055 256.45746]
+  /Border [0 0 0]
+  /Dest 1324 0 R
+  /F 4
+  /StructParent 11
+  /Contents <FEFF0033002E0034002E002000AB00550073006F002000610063007400750061006C00BB0020007000E100670069006E006100200039>
+>>
+endobj
+
+1374 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 233.18146 515.9055 240.69446]
+  /Border [0 0 0]
+  /Dest 1325 0 R
+  /F 4
+  /StructParent 12
+  /Contents <FEFF0033002E0035002E002000AB0043006F006E0063006C00750073006900F3006E00BB0020007000E100670069006E0061002000310030>
+>>
+endobj
+
+1375 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 217.41846 515.9055 224.93146]
+  /Border [0 0 0]
+  /Dest 1332 0 R
+  /F 4
+  /StructParent 13
+  /Contents <FEFF0034002E002000AB00540061006F00ED0073006D006F002C002000590069006E002D00590061006E0067002000790020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C00BB0020007000E100670069006E0061002000310031>
+>>
+endobj
+
+1376 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 201.65546 515.9055 209.16846]
+  /Border [0 0 0]
+  /Dest 1327 0 R
+  /F 4
+  /StructParent 14
+  /Contents <FEFF0034002E0031002E002000AB004C006F00200071007500650020006C00610020007400720061006400690063006900F3006E0020006F00630063006900640065006E00740061006C0020006E006F002000760069006F00BB0020007000E100670069006E0061002000310031>
+>>
+endobj
+
+1377 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 185.89246 515.9055 193.40546]
+  /Border [0 0 0]
+  /Dest 1328 0 R
+  /F 4
+  /StructParent 15
+  /Contents <FEFF0034002E0032002E002000AB00590069006E002D00590061006E00670020201400200070006F006C00610072006900640061006400200063006F006E007300740069007400750074006900760061002C0020006E006F0020006F0070006F007300690063006900F3006E00BB0020007000E100670069006E0061002000310031>
+>>
+endobj
+
+1378 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 170.12946 515.9055 177.64246]
+  /Border [0 0 0]
+  /Dest 1329 0 R
+  /F 4
+  /StructParent 16
+  /Contents <FEFF0034002E0033002E002000AB0043006F00720072006500730070006F006E00640065006E006300690061007300200063006F006E00200065006C0020006D006F00640065006C006F0020006F007000650072006100630069006F006E0061006C00BB0020007000E100670069006E0061002000310031>
+>>
+endobj
+
+1379 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 154.36646 515.9055 161.87946]
+  /Border [0 0 0]
+  /Dest 1330 0 R
+  /F 4
+  /StructParent 17
+  /Contents <FEFF0034002E0034002E002000AB004C00610073002000630075006100740072006F00200063006F00720072006500730070006F006E00640065006E0063006900610073002000700072006F00660075006E00640061007300BB0020007000E100670069006E0061002000310032>
+>>
+endobj
+
+1380 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 138.60345 515.9055 146.11646]
+  /Border [0 0 0]
+  /Dest 1331 0 R
+  /F 4
+  /StructParent 18
+  /Contents <FEFF0034002E0035002E002000AB004C006100200063007200ED0074006900630061002000740061006F00ED007300740061002000610020006C00610020006F006E0074006F006C006F006700ED006100200063006C00610073006900660069006300610074006F00720069006100BB0020007000E100670069006E0061002000310033>
+>>
+endobj
+
+1381 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 122.840454 515.9055 130.35345]
+  /Border [0 0 0]
+  /Dest 1338 0 R
+  /F 4
+  /StructParent 19
+  /Contents <FEFF0035002E002000AB004400410047007300202014002000670065006E00650061006C006F006700ED006100200079002000660075006E0063006900F3006E00200065006E0020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C00BB0020007000E100670069006E0061002000310034>
+>>
+endobj
+
+1382 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 107.07745 515.9055 114.590454]
+  /Border [0 0 0]
+  /Dest 1333 0 R
+  /F 4
+  /StructParent 20
+  /Contents <FEFF0035002E0031002E002000AB0045006C002000700072006F0062006C0065006D0061002000710075006500200066007500650072007A006100200065006C00200067007200610066006F00BB0020007000E100670069006E0061002000310034>
+>>
+endobj
+
+1383 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 91.31445 515.9055 98.82745]
+  /Border [0 0 0]
+  /Dest 1334 0 R
+  /F 4
+  /StructParent 21
+  /Contents <FEFF0035002E0032002E002000AB0050006F007200200071007500E90020006C006100200061006300690063006C0069006300690064006100640020201400200065006C002000700072006F0062006C0065006D00610020006400650020006C00610020006A007500730074006900660069006300610063006900F3006E002000630069007200630075006C0061007200BB0020007000E100670069006E0061002000310034>
+>>
+endobj
+
+1384 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 75.55145 515.9055 83.06445]
+  /Border [0 0 0]
+  /Dest 1335 0 R
+  /F 4
+  /StructParent 22
+  /Contents <FEFF0035002E0033002E002000AB004C00ED006E00650061002000670065006E00650061006C00F3006700690063006100BB0020007000E100670069006E0061002000310034>
+>>
+endobj
+
+1385 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1251 0 R
+      /f2 1257 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 23
+  /Tabs /S
+  /Parent 1 0 R
+  /Contents 1386 0 R
+  /Annots [1362 0 R 1363 0 R 1364 0 R 1365 0 R 1366 0 R 1367 0 R 1368 0 R 1369 0 R 1370 0 R 1371 0 R 1372 0 R 1373 0 R 1374 0 R 1375 0 R 1376 0 R 1377 0 R 1378 0 R 1379 0 R 1380 0 R 1381 0 R 1382 0 R 1383 0 R 1384 0 R]
+>>
+endobj
+
+1386 0 obj
+<<
+  /Length 8867
+  /Filter /FlateDecode
+>>
+stream
+x�ݝ[�d�u���W$͡���=�ۦi���
�7Q�1ǖ!e����X�n�Y�'����`�TwU~q;q"־}������������ů~���0��������/���w�0���yС�]�PcH�Q������~7�Bۧ�h�^����_���_��/���o�A���o^>_�������ߌ1~3�|���LOk�_��k�}��_k��������o/��|��o�~������������w�������8�yaT���uȚ�zT�yXfui̇n��~�2J:�΋}~:�����KjGC�0$C��������\OZ|6���4ͣi��C絩�KS�C֘ՇT�\�N��O�?������ي��1�������1�����x���ͷ'��'��������p�t���G��L�u��N�����'�s�/ߜ�{sa�.��i�|`�|�������~����_�-����u�Ni_�O��������_��wC�����W��A��9�uj�����C��.m���ë�.�3�Թ-سe��9��e�g�-P]X�'O��o|{��N��'����F#U��uk�]~���m�����|5T{[��p���'�y�z�fO����;�g�������_��0�b5�:�������{kE��"{�ܳ���]�<�/�k���N��Yyް��^*W�R��&��^My[6��c���x�t��^��M2������������~�3<�O�;?��Om�3֥��|���%O��9:y<���9��NO*���ׄ�k�����w�'���ǽ�}d�b��5����×��
z��1E#��H��͉ȘSƜiTF�Z=�c�C*�;����!����a!֭T��ʢN���p-&#����p���و�Z���t�t��F���Ee���v*�E;��X.�ܩ�X-#�C�&����%k0	1Lry�q�L�.��c��k\yu�Ͼ~�/���/���~�˿��/�С#�O�>�m����_}�L2�藾��}����?#%��t����м$�ι�Wާq�����'ŋ8f�e����g����|�}���|ESl֍�����v�&c鰤"B���������<�
Q͢"B�{S��ZV�R��]rM����o�P���M�"2Z�%w#I�O�t�*m&�L�'�"c����c,բn$k�p�@��JW.�$��>T����5D��̡�2��:T3�#���)U��E\S�����M�9�:��
+�����]�����
ɒLj!���X����q=���DFȴeNe��f^�!ʥʹG�DۤՂb9��hQ�>���
.c��T�P��}�E\��ڷׄ��7uO��漯ܧo��lW�C��TN��q֪�'~xz���R?��e�6w����{;uM�<^TN��O����3�yi��.D����˵C�m��2b��U&C�L5."d؊�2p���N���l�X�ʞԱr���p*�%�J�d�K�)�����H�XQ
瑦NG�L��0-s��X.�ۋ��mQ�MVX%����'s��O�wF\>{����	ߑ���V�uǓ���|���c��-������ޡ��x������������O~�4����{�Y�tA�9f_�|�_6�[��~6żn��k��͑W������"Bj�0&#�x��MD��R��x�2�g&�eE͠2.���åM2�E�J&'�E]UkI�.�j�XcPWU/�,�v���n�U�m;���Ŷt%sU�,��E�y���Ad�����)�J��,��N�p���ቌ��|�{L���"d���#�cR�T��*j7jl
�	/�N�HUL��5���-"���u��e�A��]�\U/��}����Wm���Qm֮j-�⎪͏ެ���jss�~|&I|���dg}�0��?���'�V����J��2�C�����6������t�]��"�,�7N�5!�$/R�Ul�S��%#�����F4��S��������p��2jH��8Ad��eS����;�2�-�����
��h�R	P��-����1|��	�<
��0���$F2g��
��ByŨ��r��yS�&/)+�ͤ�C��^F}9A�)ۼox����b 2���x�ҹ��x�Ԅ�:�QC<'�"cs��U�y��\�;��X�ꢞ�7%�uR��r�2u�9{M�4��]����3hn��	9ov�W�K4x{[�8�Ҫ�K����}N%�g���?~��>��.��~�ǿy�a�В��l>"�ƒ}�]��LO�	��Du��c�]�E��m�����J�t��:������v|/x��c��Od�h:w>�+�;8mWp磐yy�
���-��XC�O�|,���n���*s�]l �4��2҂ɀ���rV0!����6g���K��$_CU��aTFIt�a*ʝ+D�p��'�hQ�s(t97�l#�X9\=����Tj����G"I���G
�N�s^Ȯ��lLFKSO��@�J�H���ŝte�ɀB�M}�_$��".�F�/��5�Ӑ�Ƶ���n��w��>6�ƾ��͍;M�p���+)Ίp�Jq��g5��^w��K
+܉��Ë`':�LjkZ!iz���?����BB|&�\���I�!�]�6W~"���L�
G���2J��1
'3沂.l=��P�-X�H(Y퓉0��@�
+"�pHV*�U�o��DF���TFLɇ�["#��H��͉KKq!=tQ5$Ú�@6�T�2#�[�B�Ē-�	o�2JS-3���Ѓ{n6"c�F�TL�s��#2���Ee��<a�DFIfA6�1`C�kr%3[�ؐu��R��s+�
�M��FD�x1O�æzr!YQT65�2��[	�`m�+
+����hd4Q��|S�|���l�/z�}���׃���i��[	�+��]��	'��Au}s�4]�.�77�)5�+�����H��'�����q���U�a}�{^U��Q��f���O��٨} o��o���{�|�z2h�
��{)��c���-��ڹ�dE��c�J�C�%2 �Ad�!Le��O$$NH��c���B@6��⛏,SԶ��DF�
+Y���!U�"�̛Ńǀ�Ȧ"BF�
���;��#��ܧ�r��������7�O"cIOh�4�˙���˧��?��5��8�_롍=m�Q�{�?ց����-��/�B�� (=
+�zy������c?�����B��!d/����:����!��FP2#&�z���H��TĜ�S��Dv�iL�N�5�S!]�eN�9����L�ï
+'g&#d��̍��{-���>���}NE��J�r-:P��(��62c���Gd,��ԍd-�J�H�Dg(u#�%c.en$���t��@��T�'2��ks�B�T�\��$��1�.�./K.æ��1����j��çl
+1�Ą���!+�'Dd��p����G9u#���Ա*T-1�n$����l�����2Pu�ԩ	�.W8u#A��Z�)G�E��=��wF\V�tW��m�wD��]Qr
�w%ߜ�|_���e�*.�P���2��o3��U���xzF�Ի��gvK�l"B�C�(�2Z�!��1M�r ��%�h_"BMl��2�L�eL���ʠ2J��Cd ��j�Py��
��@��M��HI�Ee�wL,*#�Gݬj �����H�Ed �cQ�@���9�(g���u�b��bݰV�[�1�9�1`��1r3��8b�}lQA�aP������&2w�<�"��Gs��Ž#�:������c �8�O�u܃:[�qCC!2Bz��N��7[:0����c �8��DOkꦋ����C �xFR�c�;���K�����s��bN�C�1u�naNJxLF!��BaǪp�!2
+q�L�t��L�����<o�"Q�I�-踨*�C�1�u��-�xu#ق�����hi�h%"��c�o!����C��n�1��	1�q�^��½&|GL��k*L�u/Sᛣ����77��N���z1�I#�N���u/F�T�=5;���%c䟐M2\F��뚒�N� 2�$��Bd�����(�67����y��b�IE������S*�BD�xn84�r&*2!��t�c�!93DF�M�N�j�c��2��@d�d�¯��0؇Ȉ%+����@�l�[�|�Vs�*q?\D!w�J��0K��dv��,�����y�'w�Ae2Ub�3ް�@�&"B�&u6|kPg��5(����cL�3t����U�j���&��l�1�dxp�a�簫���lJ�|��D|<�������S��Gp �#��2&|�x�R�a�"2B\��\���w2��r���>�E}�����������1^d��".��k_��k§!�׵����K��^2���*��ؗ�on���䩜e�����Ǭ½%��f�*y*�x�V\�u�zn�ۃ��1%�*��-d�&2��+
�۱���!�a�DF �*�(���@�\r(�V"��tKa�c��pgn$��B��&�FS"c�n�:�pD��e�hZ���1J���DD�X���ה5���.�҃��)���d wޖ�ư1$[����7��<����Q������h���<��h��OD��YH��c�I��3K<`&�!�6�V�P9������r�3l�注�x� � u�2���<DM�a�)��N�5��gDDHfN*���5��z��
͘�hY��Ic����]�_M��x�i�=�+�ߚ������ԭ
+�_�*����\Z��ڲS���0s3/%èw~�~+�zV��׫�W'�~gSCWOԠ!2�����1����XއD��FR�
+�_�J0��׷p7"cI��%�b��/�r�[���iX�Zu��k&s�{YǛ���o���Y������q!�y���C-������ץj'oh�5�/���R�c,]
���2�)��h��_��$� ��:Ll�1��c�Ȁ��c�-�?�3%Q͍�P3j?t���|�cؔ��~XJB/f2|�E�Nd�̹e��1�-?{Q��t�0
+��;�ޠ0�H��Է��K��3��'�t,���͆�@<��k���$"c�������ON��FI;�y������(�@4�*:7?"#��¦�*��({0��(�c�s�00��C�	������I�!Mf���ʧDDKդ��Q�/��*"�Eז��X.c��%"���:m[n"b�m!'<�j���J��ԗ8J�E!Y2Qb��_������#"RPڄ:6%x""�b��1|ʜT	�=d�őLj��TB��A�&#��d��=]V+�%��(݀��<F��px���L=��B�g�i-�R���mȢ
+�C"c��l��E\��]*�y\6[�^����ԎC����U����b�S�s�`�|{���4���KC���eO	+�"_�?86����<��ן[P���v������gǿ���߼�l??���� �'����5�U�N{���1��IR�'�^H�yC80J�d����i��ǡb�+<�t��p%2�������
["�6E��0�T��2Jj޸�\�zxR%3��<r���I��@Λ��-�MC�%2R�r.r�x#P����C ��jd�'2\z,ē-����.��DF�H�����j�xND,���y�i�Zܡ��An	CyUI�@�2Jl��<�)B��GP�rK�L�C���AQ�F3��cĐ^�LDFH�Dq##Q#��8�8��-+z��m���QC�Z�Dƒr���c@9I�"�T��".�C����5��S�^	+���v����=T���]W�z?���=��������I-�����kNK�<iu�J;�xjˑ����:ک�vQһ\�	���'������7a
+(r7��u/��J��%lHY<Ƅ��0R%�e,�!p��
U"��!���s|�\D*CO�<�OIw��)��_y��2DO$���1rH��I��@��V��Ǩ!s�J)�LQ�����-���X.f��!~k����&KeӉ�%�p�!��ٜ�KCd,6x��R���\�
����� 2��c���p�/��_<���1ട邉��16���ؼ����DF���B!���
+��f�� �=���p���>�pܧ��j5�P�
6��V�x�(6�������b�~s���Q�ӷ�~�����-�ͳ���\,L}�~uI�;�?\����F��tg�5�����C�iw�-r�%�矱�j�ov�CZs�7N쵻�ɈEe��eØ�R��TB	b����I%ⵛڋVS۩T@Y�\Q��Zz+���=�	�9�f��2R��b�+�!�ܡҐ�+a���n�n!�YT��B�U&#d�*��-�D�DD�TP_����I}��^��G<[j�Q.W&b�Z#��2�\uY�%aԱj�ԇ��|g��+��^�w��]�*��UV2�{]��s^mU�ߓon�i�ϯ����0�+~s�Y�խ�ȅƮ_���zɽ�9��5����O'��\J}�4i׼Z�P��pA��h�^#���>����tJ�4���Xb��FD���3%�{6��T��LDI�!2Z�'u*�DՒ��2l2��Dzf&<��"#��Oo;L\�#`��M�pM�EؐUE]S��O}�B��V�ꈌC���\+�.=qo$"`Q7�1
�Gئ�-f�;V�2=�����Ad�1 y��C�Ȟ�)��Q� Od<I�E\�m��ׄ��Q��G�X�$�7g퉵/y�ܬ�x�]Q�������$�g���o5�^j���N�^��S������`��H7����5�lDFI/3&�'�V�p�Uff���|���SP1�����
�.���5$̒�p+pov�2Zz�R��\jdJ"�����c��,M�a�c�w��@�L�bN9"g,��CL��~MDrqoCU��(�,(�<�M�I�%� ��>Ń��B��,ܜy$�Ў���4W���X��%2��u��2g"����9�Y5��-S�S��9c”Fd,)]�DyD�����qY��}}l�	��>�ה����-�a����\
8�>vs�^�cG"�WG�O�Q���,g/=��zNZ�o��s����Ŏ��g5�.t��yt�y��PJnɄo��k����\N;sQ8�jQ���Iel�
Q���e�&2����x����Id��nFG�@kS݈�%��9�[-P��3�q��ND y�	GR�,�DD�+n�<�*n���FP��R�En*DB�k�0�c��W5w�Q�Ƥ2P�k
ě�
Me��ך�e���=�ђK'���"��e�����Y!���qKrYQ�E�Dē.r_��+���"{M��t��_�l���T�����d���B����CK~������^��SE�BƐ�^;��'��	gU�.��\�7|֔�/+��J���f�ڥ��r2!�bjP��������2F ��	A}^�����+��H��%u��
+�>LF�8tsʈ�_8� �pF*"C��M꣎d Q
+����D��Y�
�2ڹ��V��d�T!��.�����ŝ�t�^�9ϖiP��2�5�s�� ���L����{��d�r'
2�ݥ������/]���m�'v����f�:��Q�sw�U��Uo��־��-���-l�����m�I:�(I�+���z;g7�ӻ��˟y9K�g
+�<��P��L2&�v�)ւa�����X&:K"cKAbF��pY'2�$7%†�M�K����v�p-�j+��b"#%����c(aW"wc�+�6�>���
+H<�#�w=""P@4#���p#2�8�3��hI7��EbF����h�]0b���=��$2�Tl��<�2	m8�Klli�y�F��SO�h��/4���@D�߭��I��zZ@J��-���)�s?�
��6��vu'2RF5uY���<.�9)��0
}R�pN�k��'2Z:�z\�j[N�0C�Ž��rk��+���#��H�+P���X��ԃ��4�zZxQl,�}�v�	�9�����F���\��jG5v�ڛ��1�ć�$���k�-����y���JA{%�~�
�6�����J��ٸ|�Ӆ��k�W�C�ե�R�H�]�F`������M�j.e"JְH"�U���[ JP���z1g;
+>��DD�t�2{����#��2ȢJ$��ps�.�#�bND,�:�m2��nD������Dž�{d�k�w��V�wb���y�ۛ�n���nn�C�����?ڏ�p�����{$�Y�n�����c�b"2�2[!7d�!"2�3�����m���^<�6W*��٨��0����Š"Ztx2	WGW*�[��C�I��1�$��X&shR(���D�J�\TD�;�y��)R���(cwAx��$�D4�lo"��bvBS�l�8��I��"�Sg1	�wŜlQ+�C?�I<������A$l�]��.Q�a��8`J�!�?4��9Ւl&b�x��"�h���������DB[�CѨ~KD�X�d"�8���D�#����Df""�>aS�Ƥ"R�7��m!DD�^c31��
+*"d��W��0�UL����x����F6
w��AE����D�[ȣ�8Q�is�f�N�%\6z���i�����gO���&��ۯ���������޿�������������������w�_}�}�������߾��?|{i�����H+S�a�;������c�
+endstream
+endobj
+
+1387 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 763.5106 515.9055 771.0236]
+  /Border [0 0 0]
+  /Dest 1336 0 R
+  /F 4
+  /StructParent 24
+  /Contents <FEFF0035002E0034002E002000AB004C00610073002000630075006100740072006F002000660075006E00630069006F006E00650073002000640065006C002000440041004700200065006E0020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C00BB0020007000E100670069006E0061002000310035>
+>>
+endobj
+
+1388 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 747.7476 515.9055 755.2606]
+  /Border [0 0 0]
+  /Dest 1337 0 R
+  /F 4
+  /StructParent 25
+  /Contents <FEFF0035002E0035002E002000AB004C0061002000740065006E0073006900F3006E00200063006F006E00200065006C002000540061006F00ED0073006D006F002020140020004400410047002000760073002E0020006300690063006C006F00BB0020007000E100670069006E0061002000310036>
+>>
+endobj
+
+1389 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 731.9846 515.9055 739.4976]
+  /Border [0 0 0]
+  /Dest 1347 0 R
+  /F 4
+  /StructParent 26
+  /Contents <FEFF0036002E002000AB004600ED007300690063006100200063007500E1006E0074006900630061002C00200063006F006D00700075007400610063006900F3006E00200063007500E1006E0074006900630061002000790020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C00BB0020007000E100670069006E0061002000310037>
+>>
+endobj
+
+1390 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 700.4586 515.9055 723.7346]
+  /Border [0 0 0]
+  /Dest 1339 0 R
+  /F 4
+  /StructParent 27
+  /Contents <FEFF0036002E0031002E002000AB0045006C00200067006F006C007000650020006D00E10073002000700072006F00660075006E0064006F0020201400200041007200690073007400F300740065006C006500730020006500730074006100620061002000650071007500690076006F006300610064006F00200065006E00200065006C0020006E006900760065006C002000660075006E00640061006D0065006E00740061006C00BB0020007000E100670069006E0061002000310037>
+>>
+endobj
+
+1391 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 684.6956 515.9055 692.2086]
+  /Border [0 0 0]
+  /Dest 1343 0 R
+  /F 4
+  /StructParent 28
+  /Contents <FEFF0036002E0032002E002000AB004C006100730020007400720065007300200063006F006E006500780069006F006E00650073002000670065006E00750069006E0061007300BB0020007000E100670069006E0061002000310037>
+>>
+endobj
+
+1392 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 668.9326 515.9055 676.4456]
+  /Border [0 0 0]
+  /Dest 1344 0 R
+  /F 4
+  /StructParent 29
+  /Contents <FEFF0036002E0033002E002000AB004C006100200069006E007400650072007000720065007400610063006900F3006E0020006D00E10073002000720065006C006500760061006E00740065002020140020004D0065006300E1006E00690063006100200043007500E1006E0074006900630061002000520065006C006100630069006F006E0061006C00BB0020007000E100670069006E0061002000310038>
+>>
+endobj
+
+1393 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 653.1696 515.9055 660.6826]
+  /Border [0 0 0]
+  /Dest 1345 0 R
+  /F 4
+  /StructParent 30
+  /Contents <FEFF0036002E0034002E002000AB004C006100200063006F006D00700075007400610063006900F3006E00200063007500E1006E00740069006300610020201400200068006F006E006500730074006900640061006400200073006F00620072006500200073007500730020006C00ED006D006900740065007300BB0020007000E100670069006E0061002000310038>
+>>
+endobj
+
+1394 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 637.4066 515.9055 644.9196]
+  /Border [0 0 0]
+  /Dest 1346 0 R
+  /F 4
+  /StructParent 31
+  /Contents <FEFF0036002E0035002E002000AB004C006100200063006F006E00650078006900F3006E0020006E006F0020006F0062007600690061002020140020004C004C004D007300200063006F006D006F002000670065006F006D00650074007200ED006100200063007500E1006E0074006900630061002000610070006C0069006300610064006100BB0020007000E100670069006E0061002000310039>
+>>
+endobj
+
+1395 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 621.6436 515.9055 629.1566]
+  /Border [0 0 0]
+  /Dest 1354 0 R
+  /F 4
+  /StructParent 32
+  /Contents <FEFF0037002E002000AB0050006F007200200071007500E9002000650073002000720065006C006500760061006E0074006500200065006E00200073006F0066007400770061007200650020006500200069006E006600720061006500730074007200750063007400750072006100730020006400690067006900740061006C0065007300BB0020007000E100670069006E0061002000320030>
+>>
+endobj
+
+1396 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 605.8806 515.9055 613.3936]
+  /Border [0 0 0]
+  /Dest 1348 0 R
+  /F 4
+  /StructParent 33
+  /Contents <FEFF0037002E0031002E002000AB0045006C002000700072006F0062006C0065006D00610020007200650061006C0020006E006F0020006500730020007400E90063006E00690063006F0020201400200065007300200065007000690073007400E9006D00690063006F00BB0020007000E100670069006E0061002000320030>
+>>
+endobj
+
+1397 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 590.1176 515.9055 597.6306]
+  /Border [0 0 0]
+  /Dest 1349 0 R
+  /F 4
+  /StructParent 34
+  /Contents <FEFF0037002E0032002E002000AB004C006F00200071007500650020006C00610020006F006E0074006F006C006F006700ED00610020006F007000650072006100630069006F006E0061006C0020006800610063006500200071007500650020006E0061006400610020006D00E100730020006800610063006500BB0020007000E100670069006E0061002000320030>
+>>
+endobj
+
+1398 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 574.3546 515.9055 581.8676]
+  /Border [0 0 0]
+  /Dest 1350 0 R
+  /F 4
+  /StructParent 35
+  /Contents <FEFF0037002E0033002E002000AB0050006F007200200071007500E90020006C006F00730020004400410047007300202014002000650073007000650063006900660069006300690064006100640020007400E90063006E00690063006100BB0020007000E100670069006E0061002000320030>
+>>
+endobj
+
+1399 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 558.5916 515.9055 566.1046]
+  /Border [0 0 0]
+  /Dest 1351 0 R
+  /F 4
+  /StructParent 36
+  /Contents <FEFF0037002E0034002E002000AB0050006F007200200071007500E90020006C00610020007200650066006C0065006300740069006F006E002020140020006C00610020006F006E0074006F006C006F006700ED0061002000730069006E0020006F00700065007200610063006900F3006E00200065007300200075006E0020006D007500730065006F00BB0020007000E100670069006E0061002000320031>
+>>
+endobj
+
+1400 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 542.8286 515.9055 550.3416]
+  /Border [0 0 0]
+  /Dest 1352 0 R
+  /F 4
+  /StructParent 37
+  /Contents <FEFF0037002E0035002E002000AB004C0061007300200074007200650073002000700072006500730069006F006E00650073002000710075006500200068006100630065006E0020006500730074006F00200075007200670065006E00740065002000610068006F0072006100BB0020007000E100670069006E0061002000320031>
+>>
+endobj
+
+1401 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 527.0657 515.9055 534.5786]
+  /Border [0 0 0]
+  /Dest 1353 0 R
+  /F 4
+  /StructParent 38
+  /Contents <FEFF0037002E0036002E002000AB004C00610020007300ED006E0074006500730069007300BB0020007000E100670069006E0061002000320032>
+>>
+endobj
+
+1402 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 511.30264 515.9055 518.8157]
+  /Border [0 0 0]
+  /Dest 1361 0 R
+  /F 4
+  /StructParent 39
+  /Contents <FEFF0038002E002000AB005200650066006500720065006E0063006900610073002000790020006C00650063007400750072006100730020007200650063006F006D0065006E0064006100640061007300BB0020007000E100670069006E0061002000320033>
+>>
+endobj
+
+1403 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 495.53964 515.9055 503.05264]
+  /Border [0 0 0]
+  /Dest 1355 0 R
+  /F 4
+  /StructParent 40
+  /Contents <FEFF0038002E0031002E002000AB00460069006C006F0073006F006600ED006100200063006C00E10073006900630061002000790020006F006E0074006F006C006F006700ED006100BB0020007000E100670069006E0061002000320033>
+>>
+endobj
+
+1404 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 479.77664 515.9055 487.28964]
+  /Border [0 0 0]
+  /Dest 1356 0 R
+  /F 4
+  /StructParent 41
+  /Contents <FEFF0038002E0032002E002000AB0049006E00740065006C006900670065006E0063006900610020006100720074006900660069006300690061006C0020007900200072006500700072006500730065006E007400610063006900F3006E002000640065006C00200063006F006E006F00630069006D00690065006E0074006F00BB0020007000E100670069006E0061002000320034>
+>>
+endobj
+
+1405 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 464.01364 515.9055 471.52664]
+  /Border [0 0 0]
+  /Dest 1357 0 R
+  /F 4
+  /StructParent 42
+  /Contents <FEFF0038002E0033002E002000AB0049006E00670065006E00690065007200ED006100200064006500200073006F00660074007700610072006500200079002000700072006100780069007300BB0020007000E100670069006E0061002000320035>
+>>
+endobj
+
+1406 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 448.25064 515.9055 455.76364]
+  /Border [0 0 0]
+  /Dest 1358 0 R
+  /F 4
+  /StructParent 43
+  /Contents <FEFF0038002E0034002E002000AB005000650072007300700065006300740069007600610073002000610064006900630069006F006E0061006C0065007300BB0020007000E100670069006E0061002000320035>
+>>
+endobj
+
+1407 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 432.48764 515.9055 440.00064]
+  /Border [0 0 0]
+  /Dest 1359 0 R
+  /F 4
+  /StructParent 44
+  /Contents <FEFF0038002E0035002E002000AB004600ED007300690063006100200063007500E1006E007400690063006100200079002000660075006E00640061006D0065006E0074006F007300BB0020007000E100670069006E0061002000320036>
+>>
+endobj
+
+1408 0 obj
+<<
+  /Type /Annot
+  /Subtype /Link
+  /Rect [79.37008 416.72464 515.9055 424.23764]
+  /Border [0 0 0]
+  /Dest 1360 0 R
+  /F 4
+  /StructParent 45
+  /Contents <FEFF0038002E0036002E002000AB0049006E00670065006E00690065007200ED006100200064006500200073006F0066007400770061007200650020006500200069006E006600720061006500730074007200750063007400750072006100BB0020007000E100670069006E0061002000320037>
+>>
+endobj
+
+1409 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1300 0 R
+    >>
+    /Font <<
+      /f0 1257 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 46
+  /Tabs /S
+  /Parent 1 0 R
+  /Contents 1410 0 R
+  /Annots [1387 0 R 1388 0 R 1389 0 R 1390 0 R 1391 0 R 1392 0 R 1393 0 R 1394 0 R 1395 0 R 1396 0 R 1397 0 R 1398 0 R 1399 0 R 1400 0 R 1401 0 R 1402 0 R 1403 0 R 1404 0 R 1405 0 R 1406 0 R 1407 0 R 1408 0 R]
+>>
+endobj
+
+1410 0 obj
+<<
+  /Length 7528
+  /Filter /FlateDecode
+>>
+stream
+x�ݝۏ\ו���W��.{i��މaC�m`H0z΃�X��X�x8@����]dW��T�U_@�TS]����sY�o}�����Ow��͋��/���?��n���o���7/���w���z�Rf�1�\RG����_�0�~����_����߾w���ŗ?�;ջo|�a���__���/��V��W����o��?|�⿾�ß�y����̚T�Ĭ[,��O���{��߉�?y��u�Kz���?}��q���}d�_/ǰs�Ϗ�������w6�_�ݛ���x!���_�7{}@��������/O��ѷ}��ݯ�|���ǟ�=
�/}s��W�W{��������:�|��Ǩk���e[}q�����>F,� Z,Z��&#����YÜ�P����D�D�3��T��d���YA$���lQ)�V	1$�7vD��%�CT�2	!˲���ݔyKTH�l�]K<ҘG�!����SVNg��ӥ+�0%[�y�Ż�XStƭ�g�l;ce�P����<���~x�o���o��^�����ǰa��o���Վ�_ퟞ��K����6�t\�'|�lj��������Kk�}��E������'|��_��v!>9��䞻򟾸�~p^{Q�ȯ��~zΣ���Y�����ǩWg�8�x��q����k6ꂃ�-e#���D���$R���P�!
+�B�Zh�R1��$��LgnE���LDJ��$�!j��[{��Hcn�\��ͼ�W�-��c-j�BX�.��L”�Ϋ��.:� ^O�-kXo�0�b�a-�;;�d�0x��^��.B%��Pb��8i��r1o�L�i�DԐ���R�1\bɌż�:$۔�Xb+��3dh3_FsJ�Bp��X.Q6���� �M	g�����,����bu�`{���I\e͸"$1,��s��y��|�V�42Of��~:�z�+�� �����7?���ӧO�d�gY��������+%���_.�L�!�
+'��Hq���ibF��j9�d��$XȜ��{mK��2�C*'� |IFP�"B•y�b��d>:2�Ft0K�R&�\VOD�x�)����ҩ�c�)卺	b��E1	SB՘���b�Q$�C<��Κ�s�s�Y��9\4��nzUK��6nÉCs��t�gY�]ۯO���6��|�~���o��z���՜��5}���,��,���{+��l���᝭Y}�ߏ�;����C���[A:���G��ؕ9{�OžOU�v6β������B
��7̫n��k���|z|�:���[�Dii^w,��j�%ݎ87
�Tb)jKy��
󂅰1d��Q�H�(�
4��JDpy��19,b�
+������暇Xb>���N�S���L"\|9"<�D��"M�g�E�d�b�E�X5�Q�d�t�Q�JG\��(	3%��2@<B���ED�!٭̻b��JD7XC�:���#��<Ē��|s��hM��uʜ���R#���!���"�e���Q̭���|o��VW�o��H���|o��ݜ̣(Os�mQ%�QN#����k<DJ�f�ܨ�5KT��!+���Fum�1]nT��ȣ�H-��SQ��#LI�7��bɼ+����zFiT׮�t�Q_c1_ܨ�Uk���3��Q���[{����bE"U� z�(�m]̷�V^�̗�V]��4�����	�k��!Pr��@�V^k�_b������k�!<�!I~S��Lم&�|lYr��B�R(̾a�ܞ�&�oF�~m��K��J����;��_��݅������I"�����<7Wj5dd�5��R\:d�,g"��T#:Ќܓ�X��<���2k.�VL��)� �s�b41a��a�tQ	-5���P���b��`*�Vom�Bb�;wT�s��IJ�x�y�]ͼ�"��9��ˁ"s�%�)>��۰�[
��-�w����w��,�n�
����U0m2wVAd0���u��%��ě���]�Ͽ�U����
� ��i�1�[
��sDBIG.�V�J��̣�U���������`�1�'9's+`�:�ƃU0}��ŜQI|�!&]ccI{k1�Y���
+��DĔJ�Qx��3_��;ĤoK8
��|�[�G��ݘtMYn��I���
��H_��'�1}�[�t�Q_�!��n�
O7}�~�i����N�5��|���;?y� qQ3�?^/;�rh�?F�"D@Rs�^s�\2sL*���N$��!:���^x��e2��Ih�+���!�d�/&�e�-�ACL��en��2�d�rQ/X�<��Ĺ)���u�9}o���M����J�u;��3�l�5����5q><	�#��?�{�?�^~���v��K��3��G�S���#!�:_�poo��Qo�k����zw=�U2�_��’��DB�GH��Z7B7�i�4x��t���Uρ1�b��'"B�N=
+4\r�ڤS����s
������,E��`:
Q�Ø��'«4D�L�6<tb'�&z���fU)Z�R�Ի�_pS�y�T����|l~�ٮ_��e���Ϟd��\���~�Nu���өӃ�}TZ{����	<,s�S/���!t\�{��'1�����s����P�O�2���"���>.�6�L&��g"�CC@����LD��J��4D�����-kl�C �9!�DDL�0�h�71ِH��c�e�o�3榄����7cvV���1뢹�{��e*���B=�l�=�������#=ȣP^�ג�/T�ͷ>�Ay����#+)vÔ'?�)=�B��?A=�V=�����o�:,��[�Aѷҵ�$�.>���/WюY0rh�e2{�s�!Z�F�u�����|k}�!�d�*�(�d��hS��yV�S�6BC�B��}b�̣�!�LBJt�y�����):��<Ē	�-�B��jbI�	�!����%�QuE#��֜��ުh�izl���(h�!.�M	����w�vV��`
�����`2�s��{�>�+[���kvv����g=�O�KO����=�;O��?%0��~�d�������'�U�ɟ�2���Cw�5x��Hh��G�[{j���<M5%�_�!��0O�H�2��<Ĵ�f"No
�$j�b�Y8�a��M�#�ء
+1C�]�=DyS���;a%:b44�Y��D@�yg�iO##�9ZP�!r��)������C,�O�!*$��s�C,�t�V4��&��x�)���4�t�B��G���Y<���-qp�oJ8�2��w�wV�:�{�{�!
+}�[9��Vdt�uܯ^ء
+�ԁ~+�z�m<����&�?�09��[�~�/��o���?�_�E����J�������T�2	-ᣜ�P��D`�q��t
cJ�{�����l/s5s�cȬ2�
+Ls���"�]����	dJE�������
+阓�ِ�CK
�%���IX2zs��BL���M�7�����G<8"�%�7�}��[���Į#��`ƍ�g��x�{!׮�6U��~�[��/�~��?�<���	�s�ȉ�ҋ�jOn^��BܠRe��������!��/b��j�
+��5�kbJ�i��h�D�'1%�ډ�B:mLDĔQxE�mH�� �!6��b��i�f�Ih�0e��e���x���x�
+=�I�ld�K� �'d�����@?�u�H��B���H��
>�5�/#$���
I�fR���.�0�Z���NLR�A��Q�ͮ��5�穖X-%:6<�yBRxt2���VD3sK
+GQ�EH
+W�<���ܔp>P����|���ލ��\z�X̳�}��b�]�mc1{Eݿ?A9��;"9����~�
+x������4T�,.8����������|
3Yfz͵t��lY�����ah���-���ԣ�dy���2$�x�t4�`"���#4DA�èG��Q�t"�]�j�<<�ՠń�!��CLă��	�;�ԃxmޔp�0Y����
+><������ص
�R�F��5�B�m�kWu���gGf�XW��|�>o����uj%��`�_��
+��J�����^�̂����z)	S��P�FCZ_̨��a(�z�z&!%ʢ��9D!��D��aeLĒ2����ɭ��;s1�yQ��CLɁ�XB]̜�����W�:/��3�y�������m7����m/�{_�~����y��-�NC�@�8j'x�DM�<Ր0_LB��kKf�猇萜h��������	1�#���֖��!��M	�=I�w�wV�1:���[���o���Z��bߟ�vaO���w�����ow1Om�+�f���.Yk�mBCL���$��ʋv"��#���B�R�v�)�aB�@��<Dl*�ċ�tI��� �B̩[aj��x登T/5&bJ��""�E-�y�-3Lj"M�͙w^��J�W&C�"�U�
+aOc"Ze�F$����ӄ0���!R�1h��X���(��!R�}���X2G�#$m����Xb�ut�겶�vaJ-��ᆡ�hMD�D ��FC�ɬ��g���_LD��&�J�h>�/U�\��z�fILXh�R�1�;Q)�u2��T&�^F��{���ȹm*"0�i/#�1�CC�k-*b�b>����)�G��"�NC(*�bQ-󾁙�@�U�i�l)W��D��ȩě)מ��9ׁ^r�Č��2����
+2�Ր��j�/�'""R�ԩLqf|yK�V�o����s5&IW���d��[Εi�I�ޔp>ٗ�)ם|x)W��ڬ�y��[ah�{�6��9��Xy�B^=��N�8� ��4�TN�l����l�g���ǣ���}T=�����;�m�2�+O��hoKu���1�6��Ѩ"S&b��,Ⱦ%3�kh�/�)"�$�Q�c��9���'�Lِ����d�������S��5]܋ʡ	�A��EM=U�$�@ʅ��ﭺ��X|
+s��((С�����	"�]2�����N�ںXn=��V{���ˤ
��DDKjS/*�л*����n�.BU�ԋ
+�g��j�@-0��(�^��p�9����V<2n<F̒�F�D4�x�b���)#�;��֞���km�b"c�H���dk-�Ua���W��ZX�H44����[dƈ�C������ù��[�����X��듉-�u����ηα�(s�ʾ:��>!5��>�{VJ���p2�	��3��[��=;<�!a1��sy�,z�
�"��bZ"1�b�D���b��1�͔�p���uw��`���T�B�R	-��&fÒ��Z�&��*�yuLFI���r)���>���c0�N7&Ç�Ջ���N*#�D�R�<R̋��!C+�[�!�0F��X��E��1ě��%�1��vY�S����+��ho�U5�XB*��X.�mQwC[��<�幌��;����������*�[uuǭ$ ��d�b:¢"J�]�Lo��|n�[�l�E�C|Ng�����Ŵs���Ie`��Q�N��A5�h����8��U��@�<��;�A��Fel�\��8<D�o�8���h��>�hy�n���t�
����h��G˯^��N��_O��3j�{5�Gbʇ��ӱ!�����ו�g�G5лR�{�������Sr�'8V�V4b���?=��YݻO�Y�Ƽ�ڸ䲨t�QP���OC`�b�

ёL��Ac"�%��3aCsj����x1(6NE0��H�	�)"#���Q�Y�N�p"2�+7C*4��%�����vsK4sR�"�K$L�n�DL��q�#( �z��~
�L��!�-�-���>0��W���u�
+-�~+�����j��zY��~�[6��OO!<qNg���q4��W:����s���ꛓ5�{c�֒_���L^�U	��_��%^�0<�Qm�T��)�Q+c��+�
+����t>�nx��8T��Tb�#����a&ڎ�?��L�Y�R��z�:�����
9����r.&"����d$
+��b>�Q!2B:y���1J=�������IH��~OR/*��!��c,��j���y�h��/�D��1P��c�If3_��s��w�L���#�Me�J�GJa��N��@2
=DD��BI���+B�n���"���̷8�y�T�a,���ɨ��P�Fd v]̷8��K)"c"�L�76E6@0���m磎���[��#�s7���������qߵ��zY/Ǩ�a���-��٣z�����!��]�IL����z|G
�g���U���^os"zrŹ�����M*c|��;#U�P�JD��.�QʜC��H��b�<Fq�p�����[���>`���C�����F+8��B0�˘2�Wh�(20�ɘb�y<�n�
�E��=3IG��b3�L�c`��ch�Q[�6�9�$E�'���9B(<F��"DFJפZV�S�&j爌+�"z@�iY!Mщl�n"TC�3�rKS��3�x�c,��-KHdL��T[y
+��`����ڀ�P��@���h��T[�
+�żt��X]���1�ϹW*�ʰ��<��D�,��(P�"���3P�\(��1P�0
+RADF��MՌ�X��|�����*�*V�]��@�"�MCD@�d0��-U�M
X�IU�q6P^c?U����-UQ���p��}�T�xn��l?Uq���zTs�1��?�
��}��ֳ<�DzTtf�Ч����E�S�����Rp���_\F�mDU&�1
���hq(��D1���(Y��d`�Qk����(ɱ��
�F�H�Q2�-*C�&���f5z�ytƩ3�ԍ�2b�ȭ���5.cI9$fy�M⹑I"2���Iel�3��d �����y^	��@U�
���>s1�Q��C`�Q9R0DFKf,&c�r�m"�E7�bB�Z�3�%mF}=A��\F������)Gk*���^0#��"Odl�`���6������
�e*}�ql#�B��X�h 
�F�ne"dé���	G���@�wj��~�F�0��m�Ï��[�~9,.F[c7��q�0��0p�^��T���~�������r��X���\����$D������Bk|���1���[51��(_�[���!�?Bd����%"�,�5x��v�@K�6ef�M��X�Z(s�10�Q
"aJ���~�c
d���o:�<�tk��&20�l���c`V�C3(�(e�1DzN��DFCH�I���*����t�w�6�;��"�Ķ�?a��ȥ)s���d���%2R¹n�V5"2R�%ҵDƒiqE�,Eȁ�X��d �5�{�oú�I-<�t�5�w�6�ۡ�c �5ꢈ�C(불�!��e�-�eՓ�%�z����������W�~��?�_��߾����w_���_��o��o����w��/?}��/?�t��[.�`���dG���&v��
+endstream
+endobj
+
+1411 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1251 0 R
+      /f3 1263 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 47
+  /Parent 1 0 R
+  /Contents 1412 0 R
+>>
+endobj
+
+1412 0 obj
+<<
+  /Length 4221
+  /Filter /FlateDecode
+>>
+stream
+x��][�\�~�_q�fM�VwU����'�
x�:B*�����:3�n�Z�93�3��V���˧����_^�u�~z�u�/������uO>z���NvW�Q� e�'�V�w���~-�׿^H��V�N
+�d��d��뻋'7��ywq������f!8���|����K)��F������ˋg7�\<{���s����g�N��f�.����|{�]i/�Q�����t��2�z�w�=}j�e������Ҁ'n����F5���l3��x�5��4��>�!���i$;�:��y�g�r��H�0��;��MM�5��NJ!�ӧ��I@᲻�Óaכ�B)w�����;J�]M��5^2
+'�Ȯڱ�XS�RU֤���쮔T����^	��9��qs��YA��S\�wZ8t+��N?B�쾹�0�A�{��-��
+�{�2��F
+٘�����I���z�Hw�z�$�������g�X�5����5 �{M��<�1�9	��2v�2�ш`�e<{�IW����Q�����+�d��Q�ݿ�����J�8�k��q%��ڿ���>��^~:Ep�:Z��E�]N܆4ri�2ow[����6���j�{���b}*|��_������ͯ�T�Ћ^l���zϢ�̃J���B9��	��������yP���4a`��/�X��t}���͍5�X���^JW���R5�5&��+۴�te~'=�.�p�^&��.*a��q�&.;��OY�(e.;'E�1.1�d_@4�"�����`ml�KYb��d�n;�c0��V3y��kS�kS��Q�D�U3
f8�	\/B�Ѹ,��X�y�9���~0}�Q���N�0�������'�l�/��O�nm��Sk)L6��Q���rV��,F��bS`����֖҆4q��2�0�B	�Jp�pU�ˉ8��3ͣ+I�K^(Y�EO�w�б~�Q����eY�Li=~���տo��w�}�����R>���D�����|�Gl^>�x�(%�2���Cgڷh��	5JXi��}1mY3X��4��U)JioTi�?�>��X�t9j�����sh��X�O�/�_\v���*$���e
a�M��Y�3��|:]���ѠJ�%ag�H������{O�M��&�_3,,��&���Uh�~� 7���z��}���QdFq�l�"$�R3�==���1xf
�E�e�����t#��>1�;��\
�U_�ō���x���T��rRdf
ю)⟢�!�T2/U\&�P�GA��H�'��K����s*�Ȃ_fK��]E#��>�\�5����Rͅ��fP�F�a��}���7��o��M���c�f銧�E*	\�iHsI��@�b'�)��v'���_�7��*wA��x�/ǻ:��>N}kj��sO=L��(�#���B����
+^���Li
�GK��u;Au��ʴ	�I�nPhź#���ҟ��*���B���y��=Dy�&�M�p��%�/��o��H��4"Dct�8���	�gG�����G����h�J��/�D�V<`���.0��8X�_[�� �5�i�k��2�*c8`b�w�l��9��5NI����-3�$G�h��h�É��=k
�z"iWI��/����/=$Ro��j�����|�B��}W�!��.��P�%����١��Y�����
�����=�!0�B�O�r��uU�d���4��ɰ�a���S��OC�-��XB��2�
+>dJ�У/��g��_Iw\�	"|'ǧ�!6@�Y��y�ڼ�� �\A���Jm�æ�6Z�[�<�iM������|u�^Y%l�R
���>m.UAUΩ�^���4r�
�H����D��B;����D��
%9e6$@�4']��U@�������j���b����(j�L5�Y�����^��P`��e�H!G#|�`�PI��0V��4Ж`���� � ���-�����*M����BD�w�y�zĬ�#��	U��J�
+Q���1��Ge\s`�
�I�+K0Z[l�j��_�uRDi�l�$�+�M�of��'�� �8,j�li����0���|NS�|�l���gK���<�`��'��f3���CI��|���(�����at��
+��R7�g�W�+��B)��V���ݕ�(=.��B�Z9�\��y
�g��*������/�E
+#
���ؗ9P�P̘�TW��n.�<�^��P�(����!�rr �U��;�,+���a�ڙO�ĺ�N�ЯY��-y�
Dפ�PZ��)3y���N�;�`Z���ϛ�:˱a�k�B�u��l�0݃�pr�����{а�8�
���S'-��OD��٧������퍈��m�Z;���r8�J��T��
��G@L��+�j ���g@����iK��@
�l�%>nI�E��aچ��)���0{���H`�s���5�y�r'mH)�V�m�z�w����hU�)\�%B�24)1�Q�ԡ���r8=l\/zOv����k���/���H�C��. 銧Ir��F�nV�@�D9��q���!�yb? �Ks�%�p�.W6˃��Br�R����V�&��B�g��!y��ͤ�V턩~U��ݽZ#d0+J��D�B�����m288�Lv$ns�qK^���!J�4��"s��s!��x���Hy�� 2;�?���ID�_a+�,�2�ʱF�h�,<I�X���޴�=M	/v� j'0�)��ސ��@����C�:{Y��AuT�s^q�U
+�ol�(QSA4�f7o9K?	\k�)Q:�I�̬Kb|�ě�[��\�
{@hC�_�tԷ��p0���p׬/
+㠋���9��@h)@m�Bʲ�
F)��"�������5�厴��;��fk�Pۛ%~�a��k���E��1�&�:J��b���|}�F�����
+..Ѽ��B�X�$e�i��
��nɢk�֗���ì�B���;�u{�)�p�����lQ��8�<�J$������$sFw��M�#E��{x0Z7�QW�U%i��zgN���	���qx���
YI��X,�>�yG˱:8!͚B�S^^3�%)�1!�-Rv�bC50��c� �P\/>
����'��^��sG��3��17��H���%Y�ůdQa6�ܸ��L��
+�-#O~�cki�p=o�N�Qw�$�Pm5�W�h��mP|W���
+���V06�Q@����V �)�%��z�
���U�Sh�VA��m��᷇��l��p�[�����5|W���㐖2
P���!'D
��V��*��g/����Uݗ��W%�.s�#ι����ܣ�r"{������7ϖ?ik�Sz�; �Z�F(~�h��:P�?�o�\A�jN�+��C�(��its
;��sv��
F��Tl�C%�Ǭ�{���}� �Q���AN�B�j@���\�"�U��?\O�
K}<�I�֤Q��1#K�e@T�Ҍ�N�T�H��lʋ�(G����D��(��_:s9>0���t�En��V�A���nl^��BD��u�tӒ�b�S��Dy���膧�)���ҹU�'�6،³�����`f��G(P�t5B����B����n�B

l)����BU��K�U�"8;!k�?�6UaT�m�كogg�>y{&ydk�֤�ut!i��T{5�@�e3���r�y>�ȗ~����a��~!$Z���E��19��i�
�Z�:!T$�p������x��y���z6Q(�ncKJ�w<|��#$V㛔�z�	�ʼn�p�\F�Z6�?_�5̎`�M��X�V[��ceA}��\8���)������	� ���U��_c�$e�P�_w�l�f�(���c���t�kk��
ӣ�O��
+�V0��щrq_I��{qg����E�o;FF���	�?4Y��HS5X���ò�M��'��<�mׄ�R;[E�<�^�,�t/�rk�����oL͗2d����/W�#D�2G;�JÍ��l&�,Oa�i�G����������� uj=ђ����
+�E��g�5����o�ͮ~��wzU�@����BQ�~ȸ����E�N��Ӄ%6�����+#�-�2������ͫ��iwr�}���o��}w���v����O���v��_�t��߾ݾy7|t����W?�t�j���;��J�t7�=\`~4�
@��_�
+endstream
+endobj
+
+1413 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 48
+  /Parent 1 0 R
+  /Contents 1414 0 R
+>>
+endobj
+
+1414 0 obj
+<<
+  /Length 1474
+  /Filter /FlateDecode
+>>
+stream
+x��Zmo�6��_�ib�3�7���+�b��om?8^�eX��Q����D�G�(���C[���sw����}^���iB����+’�3rqu�|I8a��)'�Riˉт�5Y>$�%#˧�Q�˟��>-���<ad�If+F� �U�{{�?$�'c�d���W�-O?����z��I�_]&��l��͙�J�awZ�ɲ�ZT�D�㲾[]o�'&5���e����HA&��n�ު�Tk�ݍ/{�*?�mR"9��Z�W��5�]�n��[{z�	�B-������h�*A�ɵ���˨a������}P)58ϝ�)�*I��MpjW�`H�+��kC������j�L�\���IM̽�kd
����n��)���k���e4���^��l��Ff�fH���E�_�sp���p��[h�н�����.����p�N%R�WTd��x���Η���_�rzs}s���ՙS=Pѥſ�d�gib��*�)�P��=8ţF�M�gT0��r��~����H��΀Kp�H�ԫ�"��a%�d�I ׬+�
+�h�`���t]��^W�r^dz�"�u��
�!
����8�7�>�[��t�I��#x�_y�_�h�rN�z����@�po�Q�-v\�kM��
;>�j�������=�Xs�s�4z>,�ώ
eJ�FS)��M����K��&�J+��y����="m*�R'F�u�b�_�c�Ί��N�p���Ha�tU��`>TJZ�r�Řp�h8IJ2�@��$�Tf�G�@L��|и��n��wTC��!H��6�:ӻ#���Ȝ-�F��P���@`n�4��������ӗNI�J�j"di{ZF`Cn5\�1��y�!��Dj�ڭV�wl�4����{�w	���g��x'��w�3P��
+�g��~��B�p7�f��`�r�!�c�1�?��7���~�u�ac��[�o�|�?�J��@@���J�Q�1��Vb�`ۯ��jf*��\Z���i�dئ��]ƍQl������P���;(�5]ӔWῂ's�1'��,�L85��֓ll��e�@ڗ
���ElT:�?��:���Of<:��ZQ.�^-4���=�.�{�`B��"�p�[UH|��&��G���'�]��*�
��.��g~t���z6�
+4W��AG��E��G�".��g��
+�9��|��(��Jm�a��1��ڍn�khN7\u�Š�ٖ� ���
+�ƌ&Z�3��
d�S-H����qMV }�l|p�
+���m��v]c$��;>E��� �m�_vy�� ���ʨ`{����!�_)6E�g4��a��4�
�n�r��{������ZsN���Cs��_nC�i�A������'\�㷯%C<�7��j�,�뽢X,������]<���������|�#�����n��j^~~��t�^��k��"��e� JPf�rK�^�V�'�O
+endstream
+endobj
+
+1415 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1251 0 R
+      /f3 1269 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 49
+  /Parent 1 0 R
+  /Contents 1416 0 R
+>>
+endobj
+
+1416 0 obj
+<<
+  /Length 5252
+  /Filter /FlateDecode
+>>
+stream
+x��]ےܶ}����X�DY,�ۑ#�vU�(G�K�������W�zR��}
+@�
���e/3$t7�����W?���>�h�u�/_���>��b���p���)���XwJ`D��ݛv�op��FJP�	�0��y������_�pwu��������ؠ�u�����_a�i|���]}���j���ӗ/v�i�H�kF"��}u�ɱk�c�=|*��|�-��*�������AX��&��nr/�8��%ǧa�����c�&��C�J㛇�(zs��n�<���II��3j��rP&镙(�t��+�I��މ}wA1�冇��Œ�f(���n���F�4L���E�.W2R�����G�k�\�
+4*�����(�"H�u��ay���w?ω�A}�~�#���4�Xt�A�_#qI�/zP�\���%�y�fl�	�>7���������NҜ5�w��p�옰p�X���ϳ��:���,�,�s�M61���c��*eh�:���*��-Q�7���P���d��.�9����p*��D�T�V�������7���˼�I#�V�	��v����6�}�c�}�����n�")�\H��J�Բ��H��0�/N����S\؋�N�?u�HD0g"9�c�ˆH&����Z��+��sQ��B�C�%���3œ8�K�
AZF��H��r���u�Z$B#͎��.y.3b�9�&2J���5�Lm*�V�Pn1��˂b���geD(���)���0D��f�g3ic�+����DГǖ�S�t`�vOOmrk��Y�Ԃ>�^mV�}���&�vO�Z�i�Ǯ\2�Ƕ��{�ւG���68�{�Mɟݓ$Z��=u���6},��V����ؕ3ڤ��:�&�?��6ic�~�M��9m��N�v��{F���O'ݤ��=�M��Ӊ7i�s�g���t�M�a*�8��+g���N�Ա�Ǟ�c;�386��srlc�N���s�ȱ�}:�c{>����Ďm|��86ߧ;�(�&��W;ܽz���^����HI��b�j���mG�m��\��M"��B�C�J#�	�f��L!�g�	-h:۶���
�N���m'.~æ�m۩ް���n�t�2�l:r��6=�n[�y�M��8�-�&1%׈$e%�+IK��4AFfxH5)�zRfĂqJGcI�5��2j�A*���b��m���F鄦戦l�E0Z�c��H3�)�4J�e�ެ�oU�>Ϭs�e|�I>H!�	C$�$4|{�w\D�pʹ�tK�Rq1�WB�yy�侻�'�|
�K�;,��K��&�����
�X%@b!�{�br����$�?�������۟��j���KF �!i�������T�5攅�
+�H����h�x
+ە̓����.f=�;��$X�����~���W�+��=�����=����0hIہd�D�hʥ���h~��I�Ȏe���u�?�+�F��.�#S��F몡�Z�d�bGxQ�\ ���@��%� YLh�z.�^{��F�����SpD.0T��3�Eb��z`Z���U��aح,:�a_�&у�"��p|Q~�m�8X
����~PL�H��`
+!�5;���P��@�>&�@�s=�4E�HQ#�)DA�2�-�,�X��rY}kı �Y���#%�)1����A�r�EYw�$(�3�bn3��C%9.w��K-{��|��}n����i�P!EH9�B7p�m��t)����"��|��햇�̈��uvknK�r[�I�����(�++�@
N���E1YB�9I֚�>VN��0"@�=�P���4��13�$��A���
P|	��P*>^
+1���@��_�&�S���ݧS��L,s��"��X� �q���W9��OJ�R���� ��SϜ�He��d�ƞ�9����Wo�s�>����>���'O]Ӱ
+���戚_���j #h�c`Vs���a�{U�h$舺�sG��s��)��K����^��k�t]��]Q(YTd(���>s�I��D�i<��$'��OzT�VIw7�����D�Y[6����)�5�`J��U)����7����mh�{�XT6�xE265Y��J2YjZ9l�)A����w��w�Y����d5B"��j0�#\��)zm@��Y����P�B!����P�H�yJ�;CDȲ`�cDM� �K�G�/K��D&���
+�5�x�Eu|qC��0m�/@�3!��2�gI�j���j�2W?�8�P	��2�8�	�������)SZ�f�vcfy���9�ߗͤ�]��+�oғ��Q�v���7X���!�c�K>GC��#J�D��)���P���ܖ�z��,w$>���6�0n�.�Q3C)&���S�`�f�|��N7�tTq�9�O4Gc��q���B`аA��K,.cV<50�����pW, ��PBҵx�Q4�GzJg���&^�)}q\�D�*Z����u�&����u]X�0��W��*�v��9L˃�BAS���l�O���)��P#�E��2���E����	E�j��2zMp0��|-\�<�-v?���0n����6r�ٳw)�ˍ�������)	l��*���F�Y��
+Y���Hj��Cy����$c�x]>���'#F��c�0&|�>��F�iCÙgQ^�^^RL2R
+lBiH�FR�ql���Qv�ae�Q�U.��[MhZ�T	��[4��s���
0L��9F!y���w�h�6�S���LI~��ntl4A¯���r�p+���%�iF_�[B�3�Z�P�M��R�\ ��j���b�X����$����y=�����â,� E".W�mj�E
E1�z13���4����Ӝ�8ɩ"�����9
�jx�x�h��8�v9������G����A���ڥAa)���s�S��K��2Y��HkW��-�>��Fw��=6��5cZY������z�S)�1�4�l�D�8D��+�����(�'�����OR�@����MXb�?T�0��0��:����X�_�v�X<tgO��T ��䩏����<G,e
*Ua+;��]�rϘ�)#L�8���m�ΰ%|�lJ���$&
gy4O��C
�y��O��Z�9���T���(o�‡�ː=%f����t�+�2%�\���չ��
+	!���w�J�A�
+�
++"�r� H͇l�i%���>�zd�D:J�Ctl���2��8G岨����
y�\��<�Q�����UO.�2]O�`Z"��������갌��r�֗�E���｢�U��P��[�VK���������kS֙�%��j���~�9�G����}d~v��3>	1�ŸZ2�T��>�kL�I�RyN`�yI��)\��LW�ć�02W�?H�l��3[��4~��
+�­|�n��U�~�����9Z���!�B\
+�z���E&�"jÁ�^�eL�7a��3~O��m�w�ϯX��N���O�UHG��K߯Cb+�M'�n��P�d��[^�A�ο<1�Y�:Hu ft=�i�ׇy̼!8.�v�wS��i��[��F���0�4ƙs��5�Dʨ��˓*�5n�$���_�{���t�"�
��F�6��8
�gN�`T�X(������>U�?�����UM0Lq��3���ŠU��X�U�`mR��0���o�_�:�x'�JP7}���:�M�����hP�����Ah��dz�-���	�'m�W_wr&-->̜[Q��J۾;n�
+x�K���ԙY*��Z-��p����/y�k�2����ʿ"{��	��] �f����A�#"*��z�gs2�V���H��p"
+�'�r�J�C�l}�M0�c6/����9�³pJ�ffp���ÊёDuh��"M�J��}CD���ӯ_Ň�Qh��"�>�z�
���H�d��`lq]����Y�	A�,E��Bmib�>��E��%H'��Kf,{���0�2eEjd_TU1䒂���-Ymī�Ui���M`BN�I�PF~k���8�խu�YO��~"0�d:�$)8qS[��d��z������
+k0ڇ<�4H3SN4�e��k�z^����WY�ܔ�2<9�͖�]^ë��9���)am��<�,4��3^Z	���S�yr�9�k�(>0ZԪ����tYx��,��$.����Sm�W�&̷��{�Y�mZ�P�nO�0E�f�1�
�
+�4Sf{6K =��Ƀ�D����Ÿ�ɦ�6T��FʥQG���IK�[zb_Q�l�&5_u�k�=��
��!-�W�+J��^����݃s��ߩ̃�̥K�kYR�z�^Nu���;C�}%pwf=�ϖ$�]�l��R�!�b�v��+�<�K���,J
+�z@�'w��Z���+&�b>�gB�2��$��<�y�=fP�s]�Z��N�Б��
F�,���w����'f�д_c9Jٙ�EY�1�v��r�,U�Vn��;� (�à&�0Y��Y����o�BT�D�16P�_��[���e�XA�L%*��v��
+M�Uj=:'����ԴҖ\֤�35�CGJs-�)��K��I~Ǥ�� �o���V�	��䑾$5^�H��KԵ�}�$\X~�ܳ&FH�����A�x��z����E>��
+��n����QJ!J8�$�gu8I彝�N
+R�q�/9�(s��]>��ȇ͆p��r���ff�� �
�|�[��{��N!�<~���w��SC��[$h�(�$+�lj �n�kK�w$���+�P��3�S:M^��mu9�$[
ymyޖ$�b��d�	.�l$׽O��k;4�n2���zmP������SӋo*ڋ7u�`Zd��x?�d�{
�LPh�������0X
����*r
����T�J{%|S���v+wM\Oz�$y���e �|+��H=�K�3��$�{Y���-a�#k�p…���
�_O�Z�u(�����7[g����;e,�``FUl={��}�ݼ�/{�3cn�o��q[}��#VƯ��a�T��b
��1�d;B4bJ6���r[�wUY�0Fk�`d�
_��|[���hX^
���X{�*��9t���e�=�*R>�mr>����̍�j��8�8�ҿ,�_�5�*f�"�K�|�]���oc��0F�}���
>����5�
:�E\g��O����5g%:#�����<�;|w������px��۷�込|��px�����W��>��Ƿ��g����ُ��������n_�{w�S%��ؔ���
+!2BfU�#/�?��
+endstream
+endobj
+
+1417 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1251 0 R
+      /f3 1269 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 50
+  /Parent 1 0 R
+  /Contents 1418 0 R
+>>
+endobj
+
+1418 0 obj
+<<
+  /Length 5399
+  /Filter /FlateDecode
+>>
+stream
+x��]YoG�~�(�I�����[�Dۀk�Y�c<�֜�–f�`��/��*+#2�&!=PlvVqe��Y�^��~�}�~7�|su�?gb�.��S�s?8'���;{���?����g�������g�ÿ�	��wޏ���|R����_������n��t�ݛ��g��(ǂ4�%)��v�,X��	�oZgMǤ��']3)^Os��C�-�ZY�Iтs&��Mc4V�T.�g^�5��fN:�ddP7-F��:*��׎qEPLj��1
+4�!��T�#�5��}��t��,sj��}��E[�kN
e��a��*=�#5u�RKU��q��J��.|�2�Ohc��z؍u��X�I=�����6�8��X5��K�>
}]��	�k�����2�'��ˤV_㸟���9=�����D}���3>����3=����⁹}�o�yا����خSvl�a�s�g��Ȏ�r�(��^pVt
��I��F�]CCwھ��:a��&���̴��o+�t������l��c^�|wxy}Ƈ뻳g�|��������og~���'.����������[1��+�
+���~��f`O���6i3����_���g�]7
糩ܸ����E+K.�K�,�2Y!�3��L-��e�ʂ�L�V^Lk2pm�����V~Y�L;��L�����p��`2����`�b�\��⢒���R�'QA1팱(2��]������\kh�׫ͤ&���l�<L�F��I�RP���}���nK����>_}u6��\��v��o?������Ͽ_�=>���`��f�k�5�5kvC�Y㒧�>�e׸��Q���MZ��8"�.T�V�E+�5��9��X�������y9�H�e�m73�9%d�E_E,"�BF�:o+��E��b�*n!ӔRE�a�
+��i	�a��y�� �bQ���G^�V'�,z
+,Ȩ�:&�����b0"2n��;Ix�D��d��C}��L�v.0�������f���6x�|���B�y�o�oG�(��UF��<9;�a�������'��f4�.0/�;|s]��'���`,RX����<�j�kZt�U:�E�J��@N:��;���Zs����k58&�n�
n���
Jn�+eҍ�@w.l!�hJ���2�a�:�7��$mc�f����p���1��� �X�����B�4�1c�7�E�}�x��ǻ_��������}�=�/�}>�\��~�_�ߛ��TeU�6��A�λy��3��ࡍ�l8<1j����B#6A8�m�ɪ��
+��ɺ�䎍��d�����
dL�3d�l���Um��!��D���C��rw1(�B�Nxw����AZ��6�剡8�Ӗ�_��������vPd�>�H65f�7��ݣd4��������c~���
��A��p�k�����-����>��<]#^}���&a�
+š1��E+�i�`XC�:.a��?,
P�s{1�1�������$J������e�hI��X!���w���ឈ}�@tD�]�&�$�����uKC.E�F��	�&�i,��n�[y�۾H�p�}�3�t�n!&�.u��m���N>5�L�+Z$⣋�֢]���|�뻡�����!��1R�]�j+���N�l
+=!	�[�E�I��#[��C;����vl*uq�ĸ���ˢ3�x��n��7���Ɍ>���[��X���j�	o���(.Y��k��{G���i��i��"�~L�tC����,h���8�s�
?��Ww�F��#[%���L�1�!����5�k�đ?!�ί����.�$���������|��{+�]6��^�7]`d�y�N[i<��߀�#�T�8枭�.k�}�H��tnU�?����o�X;�~� �9�y���dO�Gk�~��e
G23{l�ޯ��׳B[V��o�i��^�@�GMҖ�B8���J�}ݻ���ôUL�֢55CMMq�]p�knMy�%W`����D��6��T|P��n��Q���~_w)�0(��jT���p��8���W9�� 9o��w;{�[�?��埑kG��M��#C-�>��C��=K���T����$T�zZ{��]G���W%2����Kŵ��C� �Y�,x4禚h�z��� $
R�Q�3� ������X�l��yg]�d�
+�$��b����fNio�*�W��u�w�5�nȾ�o��M�ʪ/[��C��ZU%��)

B t2���RY���&o2��m��&��%��v�,�1�.�����-�i�w�[D4U"n�
+�z��eF�9��
�1�z^�4t���_�2�'8=�MG[��� ő�x2�XR��,P\���_�,=6)���8�Y����}(�G�.?��y2#������GZ|ݏтq�7d��Z`3x�:ci��a�&����B�L�����J��#��“��>�o������I�7����@6�x�p�G^|����
+��
��k��XA.��+�
b2�6'&���V�zFm��9���qǃT}�yԜ��p#Tw}S~�v���@S�87*uR�GlKUW+ns��=ʭ2tI8A"&�s��Ų�o�_����ܗ�sH��E
+H
�i���T��UJ>t�d��V�%+g~L8<��{S4��8*tZ|�����C�e�),H�}u����B��_���;+$B�IYah���۽�^2��sտX9ʆ�5����bK�^��[d����I����{�I��JM����ȧ]�ji������%�BΞ�KLF�s�!�G�Y	[��6��?����P����
++?
+�(�?zT� ��Q5�4���!��y�yɰ��8H��S�U�4y�#!�c�����p��� r�xAT��/��U?O��"�-)Jh
aN1=f}�1�4hG���+ZK�<���K)���Jp����~���������Y�;\3-�CL*:<�A��v��!�f���o	��*�`����5e�jҼ��-���\4ΩQ�&�-C�8Y�FwXx�K	Պ6�l�f���_�fJ[E�$�D��ա.$�h��.n�*�b��Q�\	n��I�<nv�Š�Y.�6gTiuE�1��y��ӡpw�H5rYf4�<�uSnBM˕7�*�MnJ���$��=9
,�����h.�{�sB����&t!?���%�'��Ha,4:�,�., +8�g����C
��6J�!��w��k����?�Z��Q��;&�
+z���Wwݨ�rtU�d�{�}�ä��s��NE_2t��Y�;�uɕ��3%��:�����>Z�t�����3H��3X͂P*@���X5�l��1H�T/���]*��x8,��+mIe��	7]��N%��6��r¢Ʒ��٤`
ߏUfƿ�*�ʜr��d@"�s2�����D�q��7�j����0�<�t[Ö�?	�i�o����I h�G�L�Z�4�z�v���e���l��@L�<e]u�u�-a��*q+�q��==���27J�=�V��ff�5��Ɍ���A)Ŕ���?&\���'T�Or������K�p�fw��H����BBT<<J-��&˓�S�	�<�/b,��R�/;�{XP�&<��s�0�=�]�ۅ�˄�RM��o�D0]��Rh�
+4J���g���}O8f��D�ž<0��v=e-崨��x�0-`��CI�	��`C��	��jԻ��<��#��]�͌��K����{�J%P��+k!�p*r�V�:8��*vL��tRVv��h0�˩�KÍ[��5DŽ@���u*���+3��v�ULZIܮ	c�t*� f�(��Hzd�@t5c>t,P�E0y��X��Q-9
2����E43Oq����kMyMۥ��;�m3&�%$n�ڎQ(y0}�3��L�W�+y\7Eڷx[ŜR�t�Gv�*$oΌBn�SL��ݟ�YEjP�M��#w�6
z�����k�^I���)��C����D�nl��9L8eש=�1D�ݟ�����hg�	:�Dw����4�!�/�rnľ�8���뵗�*L(gJ�
�JU�4okW�0(=YsY^'�8��R�0TJ�SO��L���@}4�V�b�v¨Gy�E0�h��8g}��+�=�ۿ��Xl�+F;�4�NQ��ȓo'��Ǣ��% �㒮���Q�&�����=
�91x�4�T�ֳ+�����+R� k�B֔o5��Grz� i�ЂKy�����l���������B7|ѳ���|�J1���{���ʶ�ǜ�r��z��;z��jp�5��Qi�B�������5ph���!��+�~c����Q!a����d��c��m�(5�Ful�G4����=	�]m]��c�)��x]N���!��wj��i�H"���p�z��
8<�ݘ*H���>
+Ym���~�ed�6�ڎ�h*H/0��̋U
+	I�d���X#
+�V~��u��^m�a_�����������(��-?Ou��0�k��3��50hvy�f���<Z@E�a��k0ej��gm+vuŲ
��\���˦G�`��]dAԴVxˌhܷؗ@/���@�x����l����@��j*
+H�Kā]G�׶��)	�P�0�x�`v;u�ݨ���<���a�����{��n��H�nY�����|��
<#���;��U�
|�$�3.��],�J�(�t�M—��њg�ͣ�3kl�[@}<Z�^�Y�n��e�Ri���"O��#�	J��NH�<!���X&Js2[��p�W�I&��\�W���H�w&*���eBx��M7�on����3^>��#c�fM�&W���Xb����l�96p�Z���Q'i᫆�'5���P�-����Y��"(�H�ue��<Q�;���l����z�貧��	�!��V>
���㧶ʺ��*֛ͤ�h-��i��$K�d�-�����K��F_�L����(H(kD���:��$
���a_S����
+`Λc<߶r�ϕC���y��}��C�F�u@��Lj<"�Z��oM�t_���/!<�^l��c�Mw)�eR8ѨI��
+eXi��>�2ΰ��yPGL��D��o	�oe�=��E���v�h�LW����V]�K��x��t���=�GEzEZB-�R�dkS���-v���ϛbW_V�?i�ʡ}N�r
�����uN��ߛo�G����$�c����	PS�s�?n�BUsQ�rV�Gw�f���5l�H�Q�P�����A��*n m
��Fua�
����#�xJ&�
+��1Aw�gr��^�OQ�l�q������݌��v��������������o���7�����ó�?~��|7��z���w��û�/?�%̜����d<(%����nkp����Lˤ
+endstream
+endobj
+
+1419 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 51
+  /Parent 1 0 R
+  /Contents 1420 0 R
+>>
+endobj
+
+1420 0 obj
+<<
+  /Length 3900
+  /Filter /FlateDecode
+>>
+stream
+x��]�s�
�~�:ue�UQ|s����:Mf�F3���Zi:��:�������>�nOn>�,A7���r}�|���iοy��U#WO�6/^�\�{����L5!
+�l���*���ϫ�y��J��t�J7Rh'c+m#�_�ޭ^\�ds�au~+e��ۉ����������������F�h.�Z��\}�z����9M�E/d��K��h7���NLi�O�n�n:�+)
�֡�}���^7gK��O�<�����$4��ȧ�������X�dZ7@���R����4y��m�`y������>�m��洩ٓf��H��;�=�U���?�^Mnù��Z��@����k����o�?�\﵄�JZa��b_�I����v�3��	�k4������D>pܓC�
�8cl���Q[ǚR����nQT�(H�Q���g�I7g^)��4;e��v�G�d1r�)9
+Vï�o{<��q�cX���K�1F }����	T�b����]7��L@��C����w�����?D�a�����
��^"ϕ��Yf�;�p�[N~_!Ib���H��BV��:�������R��)�n���UԘ�_P�i���;����2۰����w?6�����6��4�ϙ�d���.���nZ-�W*�����܉�X�??��1�V�I����0���Ѫ�Bɥ�6e���8�>%~t�\[�X\c[��v�a�����9#^�s��/���8N�
�I	�Z�
zm��AW�4���J�:|Eq��NKG�u�} �{\FJ���I�q�b�iyh���=�o��Y����cS����+pȌ1�#�����<~��./��s
+���D�q��Y�QN��-<������)��'n*&�pc�"D^;�l����UwSqz���ə$:��ɡ@����A�i��겔�R��������HwB�F���u
^��lK�f���, ��e�9)�{�D�[�T���=��yk�
e��*�Ye�u�;�RF)p�w�}
��1i�L�+�3glx3о�[?��)�o\�
+�Hx�ę�[|��H�+ez���<4��PAF
�M��<�}���ޙ�H``"�#��Pq�he�)W����m�e�0�a'*�;��s
0
+��8�W&�T�Ɗ���Ӳ�ZD}X���v�������j��1L�a8�'^y�٣��ӑ��p��}��";�;pv;4�F�O0����T�O��kY˵R�>N���슭c�>�-�Uؼ�i'����EɪX|�I#|T^�R�������eRZ���"�O��(�Gk)�U��47l���V���%o���Sa8p���/S?��x1bv���?�N�^T���ƶF8�z�,A�IV�Pډ�x�v�V�;}Z�O���tݴf�OW暥~��E�b776��#��Sbm�v
s���#��2���X�~Z�*�����׉��K�fd�fU�H���ܓ���"xdžK�=��$�e��L� ����LCT%���w҃�nj�+��Z\�,IK�4�q�
�S(��)���r�mI֪���v�-��V�ºɬj*a�ߎ�T�0G��ѽъA^���}Z�;�ސ�)��qF�שXZ[U[�)N8%��l{Mmm��7�
BW1,MaXJ)�c�)b5�p��amU�^�nx���
+f2�lF=j�ψǬ�!�B(��9䁫�Цf��uB����U�5O�0�e{�>����Y�1q�S �P>?�"#�TUu�6Q
+c���o���A-�;��Nٱe
��l��D�u��_> �Ϫ��"5�RٞDk���/8�<M�r'�3n4'��+#�p�m�w��K"��<gý���h��#�����><
����*��gO��qɱ���<p�t2
+9q6D�$Pؔ���0�|�g1�6���8l_sf��Z
+k� �D͓�6LJ(�#��!Y�N)�^�L]H\�좪��M�"����7ړUV�!ЁG)��48�:��Cr ���G��̨he��,�"B)Yc�%��c}Aܰ�=���\��=;k-`J�'�*1�;?ʱp�i��>,���G�3D41�"bT�c��ʼ!	j��n
+���%}�5&p$�9$hw&'�g{�rR�$�c�G4=�"���4
�rc�B��Ҽ(g�,���!��f^�[��=�;:%��]�`�d��W3�<�=�4�)��GG�Ĩk�x
+nŮ,�ףy����a{Y��B��^ʳ�K�F����2����
+m�X�����4;���ǫ#3��f� �0O��[<.���h7a���.��e�UCѬ-+'Z�;8d��ڜ�V�V����͈�=��<A`G�3lƞ8��>�
��T�oA�V�V5K�<�?U&�٣�~h
^��0�����"�TQ����<u��	EL�}r��X֣4�
�M*��n��wcp�ĸ��
��u��`��hC�����7��u�ϊ�ϒB9���X��9v|��#t;=U���)�tq
�v|xq���hET�S�����!D
+�ZS���_�.H�*�R����l�BSh�VVxm�֓�r�p�(�v~p����E�A)�aQ$Uz�yHU�{+�[�K�XYw�Ю-j#�o3eӁ����3��iX�E�V踤bt�+K;E� ��
��"!V��S��R�TjL�P�3�� o�E���=w������6�Y�Ǡ"�Fi	s�$vS��|
+���A�ΪҰ1��̮2�N����
�BJ�u)��)+���2{�v����.��g�|~���B�篞�+N���us��NF�
F,Z�}+�YT���vc��`I�KP��B��!�%=�MnYQ[_ɰQl©��|k��CP�k�LN
+7d���cW񲊭jD����Y0�Be��a���r�[��J�r����g#T{/���ie��.G��r}�9OT��|�ߑQ^H^
+]n��޲��+\z�d'I=��%8������>�[�F��(��̽td ��tT�C�;�*F���{�\5���Tk٪�r��n�`��zЮs�^�`�6�N�{\��Vpe����% $�/��>z�놃韏kSE�-���E��7�l�A>�:8��ZJ��E�s��\:tE6TM�Dv^D�-,Ah�}�_���ޭ�)�1p��c��̉��"��������In������'*�f�^�
'q�L��Z4�Fe��f�w!�<�maT~[�Xr���
��6�d=t�B��-���#An�x�"�5��_��z�Vf�&��Y'�$�C��uX2Jg�=��[ۖy'E�1N�0��)惟N���m��#���j��ԕj� .����99@x8>	~G����R���I�S8
+2��	�
+G�j���TŊ5�N�VI���/E���c�<h~�u�ߟs�ˍS赭P�^t���T¦��Øv��X2�V#x�2;���!^�Vt����:R˪%��%���������W�%��~>�}@�'�v7���ujI����RR?�:��eJa}����e�w�?�<3���ʨ�MR8e���0=����L����w��LΚ��
Ix�hV����������A���ԥ\�=:RJ闸�a*�^�ض��vig�y��[�I͛c���h�D�n�Vk�.��(��Í9�Lz�.<rC���+�_�w���Z��(�D����(U���_�H����S��L>"pd4dF�e�AC��BS?1�7
+�#oc$
�2}?5HI�e�z`���D�����K<ݳ.�&\ɪ�;��7T����{�#W�K^������	��2I�`�دi ���9J���[G4>��?l~��~�Z@6���x������7��?����w��l��˻�����ͻ�?]v����ǟ�7?��#;P���9�X-d4F������Kz��0��u
+endstream
+endobj
+
+1421 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1300 0 R
+      /c1 1299 0 R
+    >>
+    /Font <<
+      /f0 1257 0 R
+      /f1 1245 0 R
+      /f2 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 52
+  /Parent 1 0 R
+  /Contents 1422 0 R
+>>
+endobj
+
+1422 0 obj
+<<
+  /Length 4202
+  /Filter /FlateDecode
+>>
+stream
+x���r��}�bXѵ���@�%�t$[�$���"�NI�C����O�pt�q�Έ��.wh4ݍ��9~������/7Ms��ɋ�
�<x�<~�d�hxÛ���ǹ�����m�����o�����o��7�O6�9����F���l�ݽ�|��p�
���{ysڿre�w�m�ʶ�~ù[p�m�
+˙��뱍����Jw�7�����/W�;w��Y�sB�p�o"|o2�/��o�����7���駞"m��{1��O�V7
��7�I;d$9
TJ�����X�\�ˆp�bB@�����ӏ���ͳ���6�^=�cq��8ϔ�m�v�iw�������$h^��g���=0O_~�ֽ\GZN*�PO�j�%���m&Yz~�lG��|�Tc�q��]�?�,?Ì�dBd�z�\���
�Q���"PB2��*8�i��t��K�D�r/�F��h������o_n%�6C�ϟ-X�+�����n@A
+����8���yh���th��hS���y�����x���g!R�K��s"`�aʷ���oC%r�ؙ�]B�L(��&�?Fnûˠ�B�&	
++#�������������:���=8Ȩ��?�*7v;���m�*����o�I�%T�g	�sR�ƌĪ�E��bp���oC�z�dyFт�8s2X��"h�"�<�\��epU�Y*�]+̴H��6��J�x��
<�>FMi
l��wo�jn��mK,�-s��jE�{�එ��j/�1?nAqdu��)�#-ǍL��i;-R��U�3m��k�;�+/��'�˩�L(r��~K��N�%]d(|�6$K����l0nX��M��㝳�~mV[Y�Zn�^�|���B{ �|n>i�`�`�5��C_�>�e�iO���e�Ù����W�rh-�!6m�L�u�k��\J�Z�ϙ����D��(4�a �x�X]~{"n)�Y�&�
3Z���p��(4G�|n!��0'��6g~�6�r<
+us��]1ɒXo�1����<�n��s[Y$q�
��mc�В�ˎp�������1�y��<��V�;��h�`Ԕ�V��l���6VE��$/.-�Q���K
Ž�Fz���c��!�k�_f(�ibuwH�<���m��JHǔZ��x�T����`��M��u�g�
+a��.��V7ID��p�`u�џ�ĭ9���(�׫@B���u����Po�0�`��� "��8��ރ9�͍�zAf	G�r�i�4��;˒�v���K$�4�P�������90x�2�n��6��D"��r�{��S��Q��e�K�jX��
!e��gG����@�w\d�$��h��J�=�bB:e�`�xAgw�I�R�K�>����|�m)	lO�K�ï;jۨ`�l`�xÜj�����F��M�ߖ����#(O(
+8D�S����P�	���^���V�J*� �gB�u�7^���p�'�m�?��&4D�E�Oy��2��'�h
+�4/&�VJ�*��LX۴�2,
+����wf����I��o����� 
�&�c*�V�d���#(~�P���K����0�j(�d+RLW��z�
+R���R�8�и<���Y'�0�לH$�b�<�!�N�\�:��sa�$��Sz�i_@�#��s3H�����9�N���<����� ��MU��̆fL�6\��I]��5��z���s��Ic�Q�T|��EwNx�a�(5���qzm'��ZZSt���2��fm�f��"v�B�8ƅ3rn�.�Y�I��tg	81���K��%��d�T%�^���
��A�L������dQ+*|��K�ᲊD֪�C<�7��Y�,w
+���V'	�\Κʱa�7*#VG��d�/nF�&ܦ:�M�4��?L�y�4Z�E�5������T�?�z
�Ό0`Q��,
�Ӟ1���ho�n�L�G`�Y<W�vs�T��F�uT��-W
+�&!��}8 ������Y��b�Z� �{99��@���]�^)I��Jk˙�z���K�V_#�Ƴ��=m”Q��v�����*'��A�<*);�*�Oό�[��&��&��L	>��r8"�e9D���H�\D�L����E�����x+�|"�����<q��ՑE'סb��6��F��`8ʖ8e�o%�@QQ}�f��PI��!V�ke����t�Ô+7C'.5S��s��dp�|v/��:����^'/�cBO{>�
+�H�������@j���L�j�3���Bc�P
���D�*��Fʵl�f���Z�q������ 40ѡ
+8�/q�^uF�B0��9�`"�M�)EFP9FPް��������3�:�3Y��b���]p�m����@���F�������N��w�ܺ;��S���h>���q}ݶ9�wO�dG͔�ā�	����<7��� ��
+���O6I�1c���L1ܡMp �s�L�������o����`�pEE�_��z�t�x����&��-g�UV��_ä�L�R�fu���K�E.�܅�����fm���xHX�ZQ7A��צ�,�c»�L4E"@8�Q��+�]�L+�M���0�7��C�
�t�u�U�I�� 	|ۘ��$��&52�s��Rk�U�1o�֭���J�HT[��J3` (X75����vdB*�pO�2�O�BE����q6H�[�r-�U�Ԕ�L�l�
��C�T�H
ܕ�0ϕ�m�\�bi�7'h�rZ�g�������2��{���Ta@]f���B-+�a��BPK w��ҫ�,ͳ�%��=��x�9�@)���	5�4{*�[+e��ptu6��t��Q���F�ݮW։�Bk�&k��h�����v5���T��Jɻ1��3���O䲉�0��f��e^�n�h7����Uv�	W�Á&{j	�4����:\lM�#�a��>UPl栾�,<4��������&`�t°�2U���TL�ͤqB��'�������9��
�[��G��)X�A6w�^0/�$J�Hw%g�i9^�
��0)
+ck S�2	/3Rp�zf��s�����̅=�A���F�[�����l%(�6*xI��O�I�Y
S	-�"������b7�TI�LY:˄Y����:�h�#�`l_���� ӓ�!�n��C�8�KG+̸�!e��R�ކ虇I���
�o
+����Ci**�fZګ�z����5ĥ��A��4��
+R�x5,��������>��j���)�죋�/go��ǯ��R���
�tFz-dG?�}�'@:v�?nDw�P��Kp|>I΄�͇
���ysV��Lk���{)��c�jǍ���B8k*��nB��w��LqiE�_�U��D���!�R������)�2!(��b�.�<�)�Ci!ږ����CDGI�4�푙?(:
2Kd'��
�6cU��7��Ia��cJ���ȱ@~V�F��H&f��͓�޼~�n^
c�Ft�0�j>l^�#�+�6���>dײ�ڈI�Z9fTwC�6T�ŽD|M�P�ׅ�eqM�@�V��>�j IdeS��Z;c-�IW�V�O�w�^�Ř�M]/8lo��f��YF��$|mp������xW�c�gI���p�^��#H�#
TƝ*���J�D�����|0
f���.�▒9hP�,��u�VD�jy}�E_DJ��})V���ᇿ�����p�h�g�Sg�9}K�J��F��AO��!�FMTqT#$Խ����N⳪�p�&�"��d�����
+kF�ߝ"B:/1��h�����K'
�mk|�G?�q*���t�4��猡�)&Q^�h�By[�����Cn�!��ʦpBp�j�R�UTW���U���WS�gm"G{O�$�>�j�)#��Ej�8-�U3R�b����*�����y)F�W�T�r�w��(�PWv>$���*ݘ,������f�"/�Y�����t#C�t�&jY���,�
+
���@���L9��G����%Ơ�
"<�F�C�j�N3@d0[�P�{SA���̧P��h� ��*G�#��g|~O�S��DW?bV�
+�aO#��ɗL��V���w?��W����O�ݧ�?v߾��t��_�7��?}ڽ��:�?���O������霬��Lscd�e�(%�{���p�l���-��
+endstream
+endobj
+
+1423 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 53
+  /Parent 1 0 R
+  /Contents 1424 0 R
+>>
+endobj
+
+1424 0 obj
+<<
+  /Length 3972
+  /Filter /FlateDecode
+>>
+stream
+x��]s���4�e+� |�:N,�n�Ɲd��L�<H���۩�̴��s<�v��8�H��-���|����׻��us�����+���7g�q�)�y�8'��6�oW緼��cř3�k!Τ������V��oW�{&�jmÛ߂O�3�t�ۊO���[}�z���꼌���~���P�*Ÿ���5�p��R�{)���b�jǍ���B8k*I��H�͟sp��1kڹ�X<�s���E
+㘲���B9&7K�#-����	��-"�m����]����i.L��e����pɫ�Լ�i�@O䈱�NH���]�)�T^Un�%��`�ZR{�xf��y�zś�O���ͫ���9k}��<-�O�F�1{�Z��	X+nj2Zm�ah��	h�����j���J�/h�X�	����`���LHo��`�vm�/.W����:��o.�g��˷�\q�?�^��^�8ק?5�/W�.�zƛ��X6�O-��B�����G�VMs��鋯�z����:A��r�xgeZ"����{���'%��P@���oк�S�{��͙l�-�\�#�9mάb�	/�!���\A\�������ۣ�C�N�i\��Œ������5�N�i�‚߸�����@���c�.#Ez��O�e X�,�� ��$fziG�t�W=����[�#ā�J��_�o�c����?���v���o�o��n=��9�O�~<��K�f�_�6���`�FU��Tj`�g+�u����DO��n\�@f�rs%|L�h����
�N(�61{�����B���2`>	̇���vG��mը��4�	-G��=�ي�!�1 ?j�	SD�^��K���L<�^�G9ت͏/K�ds�m[
\��{��л)!�\��+�/�H^�`����;dF�?��ah֓���������&Y��4��vd�_�nd�{1P>&�?L40	�����	�?z��o�,i5Ε�%?+a����-��
o
θ��ܐԠ�$��PeD,���`����ָ�O��ϓ2�=�
��Ԥ��
HGZR��Jn1���@Vk-�Q�nP|����TE��!������ܺIs�HќN����ǭ%]	�R{�`��e���A�x�(��N�F�����QO�{!��-5��S!L[�tD>~X�Z�s���x�A�(��\:)�'��^k�Ҁ���Pa
+B�
��''�g<�M�+��(����;1�]ѐ�@���pT�����ɐ�I$�"�^f��(	���Q�e��$����X�E�
+�#�`Pw"c=0*/;��h�w��O��>W`h����wӫ�yI8m#�"ӏ;�Ki��ۼS	��cJ�P����J�1dr��/�Ld�7aHҤ^17�U�=����g�:��@Ǔ�%��'�kamg�Xd�)0=M�!��Pp1�(�{�Sak��ЍȬ�N!oRZ_7����*�Q-k��
�n��
�'.a	���{�KT�odx�5��~�GOsKn��o�<%��S��T��M�y!��o��M��;��IF�B2��^����<�A��!C�����6��d�ي��-OL	�b�����M�O+����r�<)s,�[��n����里VU��=n e��:�,c���z��%�Q�S�A�"S�}��}�45�`<P�f
+]��9���"��,�Z��*�in���M_�LϮ�?��~�s����*^1-�
.(��H�H�y�yY�9!��J���FE���F.>bgS��
+���x^�(*ؒ�3%�SU��+��[Ahf��v$,�����\�c��D]��$� �����1I!
�%���<;����2(�7��L�4�EQu4l��6�ȽT n�f�8���;��"��j�)��*"m�T~����˺$?�3����E �9����b��吋��Zd�g/+���$���iUc��"y��F�n%�j�9	��#H��6�A�}���A
Y�>rS��0όf5ְ����EՒ
���L;���ֹ0�ޅ�p�*���]���I�:��>�6�s8�������|S"��\�$k��j�^�u�{����4a@Ç����p̆@C�TVT���5�ې̓R�̺�5P�L*I-d�,��e0�#U�	�|�c�B��EVD�b��K�s��#*g��Um2��i.m�Ȟ�~�R��%mG萔������&(�Mg�K;��
_(�i�z���e��I!�����tU6J�j��Ҷ:���5�78'=e.����h��U>Xk׏>��ɧk�D�s�xq��R�R�+^I���0������^L)R��}�4~
+Hn�Wen?�e��3����c3��0�1�0�>i�\�F"�-OD��ѝV+1�}�*�$U�uC(�U"�dB�EB�3�j�k��Z�L��u��k�i$6[h�`%c��]L�s��z�<��8):�2U���Pv���u�AES�1�� 5.��J�**Z��0�r��6�>�Q�)��C����8喒�� 2�c�2��n�5�̓4�������g��pf�t:���&���~��
���a�K�J�̊[`=����Y~���D���N#J�`Y0�wh\J�$�vKg;�h�r6�e�s��؋[�1�^�ɡR8R��A�~�3�"���mjI����z�~�v``���J���)�I6”�eR��eAu}���"�l;yw���\D����*�?�(f�����L�Ŗ�L<0���nR�T�	/��!aMz�4�FK��4s���G1�d��}8�'�5�\
+��۫w�8E i�O�y�4났n��XO��i���}����o��yl�5Gp�" P������zt�6
+��Ԉ
+��*L%d�NW��7�D��]�~Q6���2k
+�F����p~�J!���o��DO'�'��LJ��n�ģ�^�5��V�<#���e���x�I{����t5

L[7�i0�ӹR���y�
�E��S�{�&4兕�z؟�`�::�B�-LY!��-L�guTwy܎�O�”�v��d����YZü��q�B]��>Y80̓���B�2x�@���wܤF�O�W$�r;I��q
љIfm��Y�����쫖�Ju�b�l��T�qo��Ï�XПS�u�L>
!<S�tW���H@Ϥ�7p�ڰ�R��Q�y�ݲ-���Z�hDon؅2uɒ���8�]Ւ�=8-X۝�%��TN��6
FݐĄ{��b�;_��A����'�pReqW;_[�
Ky��J/"bG��%����&Q�|�+f�n]��U�+�ޗ�P��u6��}'�I�Y����?��G��Y=	Ƅ]*!�<��8z$|�^�Tzmܨ�#���sL�
+x���GA�N<t�BI	l��ڜ���1����*y
+�jF�_�3.��c�i�a�p�OE�ln9��K �(�*U�Z�>�A%����}���XW)��g�R�ew��R˫��峗1Vm�pe�JSR2�Z.��}o��x[�q�-�d�ws}�T,��,%<k�Vm�I��%���^st1��0�<5\���^s�v�ٌ�)7m���vƓ���l���Pq�������T�Ȇ�r��Q��Vf�����½�5G��yV�G!�s���޾����I���3U�R�Uu���OA����e�W�U���`1�"u�f+{76{�_p�Z{�3������B֨�$~;��[�fp��w��*3�
+f�ࡕ{6��&�Q�x��,k�͎f�<� �h�6��^�']6���(6��&�N�M8�.I�z@�X�8as���z����ږq�L���b�Q���4��#�T�!n����T�{I��o���~�U�N��H�~F]����{l|'I���[�m�v������?��/ޯ�����}��ϛ�Ӝ?�~��C�����w�?���z���wd��3͍����{����<����P)"
+endstream
+endobj
+
+1425 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1300 0 R
+      /c1 1299 0 R
+    >>
+    /Font <<
+      /f0 1257 0 R
+      /f1 1245 0 R
+      /f2 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 54
+  /Parent 1 0 R
+  /Contents 1426 0 R
+>>
+endobj
+
+1426 0 obj
+<<
+  /Length 3668
+  /Filter /FlateDecode
+>>
+stream
+x��]o#��]�b�^�\M��6p��\Ѥ9�E�@��=�)z�ԧ��/V�]��CjW6z�%q���|ϐw��/7��矯�����o�n���y���տW���l�UZa���	���o�ޯ.�ds�q%��w��WW+�\=�.6�Q����ޮޯή��\7W�\��Z�e������\u}ws�Ss�����5���cbC�**��;\D�r-�޿��oy-�߿+��޿��]�߮�s�[�]�x��Fu�o�ä�������[�~u�s���?k5 ��������&����`�()�(����[l��B��E�2�9���†�����@�3pG���]7#�������݀O#��m�����{��pT1;n�"�g��;��%�v
+�E�UB�̈́
+�e���tM
ZP��+�6���|n;�jO����YL©d�dp�R�6����3l�g�FgP�{4p>`���=
^u/:�>장@~B���9�
+�4�K�`Y[��p:�W����џJ(:E��PA�G�i�A��{�5=�0j�����&j��y�}z�yM��­�l�<��u�x�h�E�v/v���읿�:`s6�'�b�D�"=#ſ��T�˨��m+����w�P;�T��s�2N��J�us>(d`���f��2�*
��7l8
+�|>ճ��8
+���]v�vC(C1�)�(�ئ���#pXUAPZV�C+�o�Ij���<�
d �[���'���1��鍚��G"��NK��R�"�A�;��K��p(6z�ވ(��]�_�����|��z�d�b'q���$�n׍�VH�:�\�1��e���.�
�6�`jHWm�Z)��.�=��J�w�:O�>��2�K�w��닩�
D:�C
+�!�HlA2�88MA��tn>@�TLA��9���ɭ3����А�"�D��M�b_o���rg��g����Ѹ�?��=�̓ �~����p�����-֜�C"��F�
s��B��x)SWos:T����1M>|�3#��	c�q
+9�0��͎��TQ��VQ�h��!_},���fW ���$i%S��<5�ώ��	��c�	��pv$I8��%�D/}Gbr�_�`9M�Ƴi6�0Ap���j�U����"�V�֩���[y@,��3�=�
+$�q�?3��� �fr��y'���+pa���yj������<n7�;Kz��'BA�g6���s%�
��zF���\���J���Ǹ�ڇ.Z�eF'�O�q*I���
�! ��'촑�Qgi�\�����p
[�t>
+'��T:^�:�d+�q��zO�!�0�5��-C��!o�};�F�S7OM�O>�A?n
WLj)-��Ƶ���o�d��2ъe+�1�|er�(��\i%�UN�*�3N���#�h�I8kD����)Wi���K
�vFD�F_���$��Dt(��us�dQ�i*���2����6�K�;wTbh./C�ؽ�)��쒛h,�I�zx�E�����2t����&G�����!�R�hT�
�
+ƀ��
����G���X�}9
+ �@a9J����,=<KlS��+��xv�4��n
�d���lva
4	2^�y�e�8;V��Q-����UVҎn�ɳ��e/_S����`R���%����=�[|N��2~���	��\?`:
WXJ�r4��i�r(D2�9<��
8��gH�>gR@�b֜xA}^C�q���nr���J7�Gb�xɈY��@���s�':p��F)�	$�O���R��)���-�6���
�����/��d���
+f	��MZ�����62�
+[*x]�9'`���-l�wŧN6��vV�J�pV3�gL�',��
�.?9�n�!u=|��nd��[�] �tn&W,ȗ�E�Wyu�kD�qI(���Fa]t:,5
P"�-���RFC-l�Z[E�Rb��L)��EP�5���2T�]���;�d���N5�O�����b�A@}��@��"��Bw v�
�
@g���"&;�1*cM
:Q��,ʚ��:
+��"!8�κ�A��K�D������,���Tb���Q&}W����p'I, b�R�n�JH4	I҉����?j��Q0��)MH�W���ؽ$��[���KN}в��jj'X��CP)��,ȹ#P�f�tf�Q��q8c�q��5s��������q�4;7��$�M�;����`I�l�ߵNݶ2T�^��7�w�Ęt��JD�?�"ऀۙ�3�'y{�д^8��(�B���'vъ�U�>�(۪ş���ÇA�1�d'��%�4+���;�m�J���S��0|��'��ޜ�d�J�
�Ȝ/�o�*5�v�ϨO
�N%ֹcmHC<~�o��ܩ=��	���N� �Q�y�=h�J�Y�ȪC�� `������'�fht��v�U����J(Et�r=��$RAj���@Ҟ@*}jf��'8��N%i��}�%�2����:+�K(���
+��,��=[/z�:���n�C:v�R�o̔8�����rGH\��k�㫜 ������1��vXiΘ^��K���
�2�)��y��&qK3(V��s�@G�z��������wO����d��L�C%�3�8J�����#��vz�nf��=�f����T�'c�ә���H
e���@ZSz�/�-Y8U~[�Z�X��A�f1m�p�^����^��NY�a�%ڶm}��/�
Y
+�������������t�EW�=������6��SBʀ�pP��Z�Q5j�*�d-��O$K��U5����v����4��vD����&�xE����m��nl*K
.k�����BY������洉n�hu�$-u���86�Ѯ���[��~�/6����ԕg�O:�xFw����)�.ݒ7�~�a�!���:پumZEPd��
Ɵ�f�<�g���R͙�]�Š��B�/��-�zp���#��M�8�	����Ҩ֣dP���(���9/m�J2φ���J�2�X�|�r���RG�QU
�X|����-\���J��E��o�يz _:�����L��$4'��ӷf!�����Nj�Йk3��^���5�����?q��c��h�q��v��cws���i/�E��`vA�5'�D��
8ɔIy����Fc��S���4�(qެv�2K!L$E��'���M(�œm_Ϙ2�����J�� Q]�B����� ���Z���[d�$�P��2o�
7���O�T�:��d6�Nk���e{J��w�(M@c��+���t�/�h4�y%�c�*x�&_��[��Beer_8̥��)��#��L�����$*�4�y4f^3
5ל�,���S]�Q��/���o�1[�?IDZ�,���h�
�_�L�������B)��\t��y�Z�!��
+�l��H�O�!c���D��D��Ɲ����
��vr��ɲ�4%,xvRV� R�&�P	�@X0pɌ��uv0��c�EU�A�!&b� 4$�9I0D�g���5���p!!<F����x�%9f;*m�k�1�ܮ�0�3��Rm��������v(oo�7w�x��懋W���������y�\��Ç�����?����������ڝ������j![c����x�*���h�
+endstream
+endobj
+
+1427 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1275 0 R
+      /f3 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 55
+  /Parent 1 0 R
+  /Contents 1428 0 R
+>>
+endobj
+
+1428 0 obj
+<<
+  /Length 4762
+  /Filter /FlateDecode
+>>
+stream
+x��=]�ܶ���+��-y�[h|笽�d)����E��A��&N�R"o����̐��8�nE��%�h4������W_����g/�����{�ͳ��v����.T��^��y+�躷������_wRx�(�+c���ݯo?�^�dw�i��Fv�tW7�������QJ�������������ힿz�{�qSn�	�]������;"x}DS�ӧ?~*}�����Ï������=83��oo����ջ�v_\�Q�0�Z�#������ˁjŷ��������>݄��Re�F���pz_J��Z��ed�� 컋	$�ӧN^���9"�
��X	X6W t�Ng���/t��[A˟>e���Ja���2���O��Z�lކɵĄkbf%�ln���yd��a>Q2U׎l�
r�3`m��j.��F���:L��)Ű҇/i�݅�Bk��0̓���wƿ�\~
�x��Y���M�����A�#�H���P�`�?`@����
��BP���S_�	�I��I��{x�r��m�5�=r�H��!G�
��8~����y��O��~�̧M��ޓ<��3��v�B�al�
+4���g�S����5�%EPVmA�I@-�(r�JmM]�@oya�_����1�>�6`	����N��l�ί3�;g��_�l�i�i|�|㻮*Ӕ�o�QZ����J����b7F�7v�,h\��%����BA�/����2�
+c{�۔H����nú}n@pS%x�B*�_�L�֣v[�%�
F
+g]Л�����L\m� x��f�2s"�}i�E�I�J	#��b��'�?�]Ф>Z�f����4���f�~�*�t���=)�#"�6�\-#�Y��O�꤅�� �RB�2�W�D���^<�Bʯ��D��k>u.��n�N���Lw$�Sznʴ?��
e�}^2y�(�-�����|��P$[�珲}�5��g`2��)�W,������z��K�]�J���3��"_r�jR�起�0?��1�f�F2d4�6t�5��
8���C�2�=7ϙy�&�x
��/�LAi��p�Ahb
+�(�����X������w#yjĠZ-����
��Cu�36�2�:�η�<?Jm~3�����a�W�[ǁ=)�^�����3�
����ެ`�l�{D�çj�X�RP�*8�8L���}���y��v� ��?�"#���@)�"H+��L�0/)��i��
do��ӾSny&
>��=|>���IDq����F�(����8ƎB��sgl#�r�j�]�ϑ�?��+����� ��F�������K`\�I,��,~�`S�3�r�b|�����̕#_�9xGv�)�
�!�=�,L$;��Cgˈ0�r c`=C V'�����¨�:�AH�7��Ŕ���)/!�qX`9<�	5y�|��f����*k��(�]�M5�����F3�'��g^8�:S��6�HpT�t��0�'�i�GmO\��ɯK��y#)d�<�n9���������*�.�����>d�x�}�@�Ўr�V�c-���`�Ru�V��#�����w�>J�a"x��
��f�t�$��(@9P�NH�B�"�5̑2~�b�YF��2���4����LjZ�.����7K��!�,�۫��A��̙9nϞ��g������u�sK��]�Z��FE�Y�L���m*��M��$~HXҒ��r7	��L�3��j���4�c�f���:
Be��8^f>�<�U��7�^>�lc��ygrb�,��DB�|��>���4뮳R���*_�8k�Ľ.�(��8o&S�Le�!
�#��Pr��(��HU��<[/9�7u�A���V�#*,w��zB�4BFM�S��#s��T�d��D�S�q8�"DG_�<lmΟ7"�~�.���Q����f�V���L��Ԛ9�)���JH��d��+�3���_��i�İ�.�:�{N*�H�A�����F�ms�������v��EYd�T�4e
�h��<[���վEJYN{D8��)�)%	q�{�mb̕��X��VD�﫨g�T��lIϝ��AyZ��АL��r���i���* CQ�!�\I�L�;U�?��#�����}��|^רL�TpT(���h�Z�Ė}��`�*%Pƒ
��R�	&h���C|Hd��[����;Ce���6�u7�f��z�&5��#"u<�<�7A��'��Kk��&���ҳ�Gf[�5�xI�ˤ:4�!�m�HhN��Y����UC[Wg$�}S2��~��
m՗�;��}F��,z8��S�)����ҵZa0�G�Bk��jD������O�9���w�i"F��!U��"
Q�&R�缓;GY�>9̲��|���F��󿣌dGX.�/qf�5fu��@�������q��Q8�p��|[I
5�yz8� ���a�y�;�J!�bJk�3Ɖ�Y����cg�PR5
s�]q�N��al�I�(�v��-q��(]�]lk�fU��v҄I�?S�J�)�KC�H0�$�x��G�J�����~�"�f��0�%�D'D��y�.��P���T
+Z��N3&�|͕Fa��q;H�8��W%�|�]�ƳHִ�X`g�y�;�]���O�����{9��n�٪1��Z�FBUcHc}5��P^���YzUꘂ�*�ڄs��*&������5aR��NI���y(���i#̴��0"k�q��*9��u�Et�����ɭz�
+�ݾ�ҏu�kB9s��DtdEt߽V��0NK3�D#�]�at,�D���,�`��V׃�u�����&�֏�!]��Q���hm%=-��M�B?ŞRl�N
+G�AF+���m�,�0-�J>����.
������w�l����	iHD�y������,���;�<ž���/Y�+	�4+!����3��oW ]�\��A�&$ۏ�dŅ
+��
fr����.��	�9"����2?Ĭͨ��zi���^���`և�z'���E~T�Nᆥ@	�!dE�'�p+����'I��PO˿���hXv�ᵓo��4�,)-))�8%��D�c%Ha��J��
�(U��T*�QW(��7G�6��%�V�.�ct���I�\�t��,j�)j��x�ʎt&�E��c��q��yc�*'�J������^��+����ڋ�R6|�Hb�b}�[e�\���?>�,�pƤ����(�dք���4e� 2���7�q^N��Z;����ʆGM
+<�W-��܀ɣ�1��r��x(_�,_W��q﩮��։�����<�F�i�R�g�&$�g�����-/�<G7אY��1(��
� ���b�֪����.�߼��v�t=�N�a��j<k8O3���~U��<:�g1[�,�8Nh�ם�zY��J
+6g�jPp�P��|``ivJn��KS�T�?L@;߈~��jpǝ��!�����۵y����}@������F�ա�0
������;���L����R+�DQ���WuTJ���g%��2}P���z M���d਄iIXP�>KE�`���y@�uo��k億uU�yfi���Ĕ�^�%���f��-V��/Se��u$�k(K)���o����aNfh�TAM�R/���Q���F�����Oޙ*��5����	��R1�U�x��)͑�=P�s�	�Ʉ��
+�P�EU��%�l��̩������|uC'��Y��ۜm�c՞�&g���d�|��-�0��z���_,"����2U3G<_�{�hKȮ�Lre
+93��Wlw��\d ���[:J~zD=�����yw@~����7oo���
+\i��Z�c����_v�)��a��~h-d������ywӎ	�pz�O��ʺ�:)�3^��3Jyg�('���Ư
()��>��#E�y���ӏ3�����<$Ng������(0ƒ1=��ɯ3�3ᲆ<%��$��J��Կ�&\���H'`"�ꞽ������w�{u�[�G�+����KV��p�,����"P{	�h/��F8g
�]7��8�bг���W�y�0¢� 5Ӎ���2�����\����
+wy��wWR��*�aݭsz�x�I��b����p~]r}��.^�$��i�J��g�$)^�b�#=z�-�5�����*15KL��:_�F�s�(�)�i.��[YJ�Izo������ł��\�8�R��Y���w�����Y*c��y[�gZf34��(O���B�w�n����*��5"DTЊ��@���N8mM��0�� <Ȟ�Z��ê�EY(X�K��J�i..�䮭��F�
�ԕ�sG,:��=��[�	=t�[�o󀢕›8�w��=��=Ϟ}�5��F��U��f�<�,��'����s�}�LS�LU�Y^㕱�eq狽%kU���7R8��o�\s�����ш`�K�țЪ���TD'lT,w��Y}#[,~Ak�:~/�תՎ)�%� �2X�xGVG_��)�V�Y�z;&t�1�d�A��e���}
&�������栁�kSP����W)|�C���ZVJq�o��d����EB��A��l?�Ȑ�OUc����J[�h��E��� OZ��N[!��<5���
�wg	�V'�.X�ֺ`Yj޷N�$�{X��lw}ʾ~<������?����~�����O�__�����}�=~�����O����x�?�����r����
�
!�ְ��������Y(
+endstream
+endobj
+
+1429 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1251 0 R
+      /f3 1275 0 R
+      /f4 1281 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 56
+  /Parent 1 0 R
+  /Contents 1430 0 R
+>>
+endobj
+
+1430 0 obj
+<<
+  /Length 5661
+  /Filter /FlateDecode
+>>
+stream
+x��=ks7r��W�["���a�[��H���TRuvR�"�9%K�W������h4��#}�\����y�����ׯ^�v��_��{':���Lt.0�8��s�q�l�����׼{���3gd�Bv�IÃ����������8w�u�{;P��໷;>������ݷ�/v�(ǂ4v�D�����3��N�.Yպ��w�$�A
+c��8V;nB���XI�~����E )ń^�m�����E�w�7�z,<�i�_��2X��/�eʹ`W	g�����V
+DIM"�eWJļ�6�@a�L��L૮��y�mBQ��
+��`�I�1	 �e�X�F(���I����F�V��JE*J뮔�i�MbQj+]*rQ\v�dL�n
++Dc�"G��
+�	/�G 3(w+Ӻ�l�@ݩlL�~L�1u���xf�
�]�����ŏ;�=��t�ye�����̴cΜuǜ8�ǜ��7��;�<��k�QF�=�,�j�#�
��q�J�S�qܹ��qn�[u��}/�{61�yr��rǻ˛��kޅ��z~��q���O'?s���~r=����7÷b�y�+�*}���
+�4�,T:��������˪.�j�ew�O�]��I������Ϳ�����L�q��7�v%���Vҝ�Qf�i`�+}�Y=9�Ƥd�����ooO�3��bNHa��ߤ+KްJ
+�m�>YZ�fF{v|���-��*�k"��鬂�������S����k���<�&0�1ee6X`/���L��jn�*w�]�W�Œ��,��qb`�/���]��W��{����������ݟ��7����ϻG�ڼ�+����4^�b9��,h�9>+}�c�*!���N�9�9s2ݴ;AK7=�Z0��0��_=8����{/5�%���ˠh�㩾y}��Wo/���m�Փo�<��������>�������/H�FB0o�]�g5���
WF8I�K(�
�B87��M+a������8��L���I�h�kz��6x�K�ץ��pBM��N�Y�ꈵbg{�%�W3藧�W)�Sn<p��C�؁k��L5�H����Yʃ:�J���Yɋ��[-��v�N� 
�	����; ���g�Mxr{6�[���e��	��O(:}9<t��[Gt�F0��MV	h_AC����t-(�J>ό_�\����M���0=�ρ�uUh13`��7��s#�Q,�"e��{�S�k�0� ү�	|f#f��_�����UDc���y%5�e����䉍�j��;e�b�<(�)@?�d���,f�d���u|&C��Q����jn�<E����=+>tY�ft�Ǚ���}��YYCb�/�h��ʔ0̺>ϕ��ս�J�T������F,��-�����5�P�Z�����:#.������S��������ʙ��I��8�9����L��G��K�ܦ���n�
�b�+o�6���+nbx�V�
+XJ2��UM`�JxP�G��,Dh���cpɂ�oG׳2�fR�Q�,��l	�I`)�d����!���Uu�����#J���}���$�<l�4�FU]
əՂ��I����.
+��ʬB��y=z���1�Xۏ^�ŸV�l|XGE�(��(D��O�3-��4��W��af�?�xD�
V_�"���T���\���yO4Xƹ.&�ʓ^:���0O���G�.|`�;U@:�4�j�|W^��U��~֌C����B����(!�VHro�ի��)Ӧ����ik��tSx=�q�Uՠ�h�$.7x�5�P��	
+�w�y$Shk�����v���
�A�b�!�y�i�Ԕ�22`İ��AU�A�lw�D`>��=�OM5U��V��x�F�W�93V8��"�毸`�p�tX�h8L�;�>f@	������4{?
+ME��牥9�@��6	��㸶
G%))��\5zRT�e!lIa���r誴��q��V�I���V�MV%r �L��9EdtZ����I�;
�)}A{�"��\���C�1Ɯ�Ɇ���Rj&m�/5����Ht�
L�t��U�M
���/�#f�g��&�ѭ��S�)�d%��$F���y�+,�}��0I�N+��'��
�����l�,��
�;���|le����#7Iև�E���F��%�\̂H��PW̲j6@�,:�kx�.�9����ǽ�X��rs���є�C+��!I�%�����=���h�|}�`d�4]�6���Z�3�$N�5�+�&�f�J?hұ*Z��(}w'\�f���a���)ǂ
´`��Wif�q�qU���I��Um�k�dՎ)�����4(���o�����Ѽ��uu3&%�I����m����
�%u�Ս�f䞁��h{�ܳom�**��m,<1>E�u�izg=������P�����؈�7�d5Ӟ��s`څ�5m��p��M�p��	ܚj��!�S[9�x�+�F����V��~5�G�
���2M�Z7���lG"{��6g�_9	w��?�j��U�DP�g�ʽ��S�
�İ�� J,}ż�|�5����"�HL�t��lpX*�{�o
�d�n+�K�Ui����-J�T���
+��\O�s��[
�Y:Ga��c����5�<
�8@\�����3��č'���&�!�4V]=�m{Ƭo�Ƭv��n`�d�P�����p�\%�z/�o�t::i��҆V>[���1/�� ��80��<�0��'4K��)k�"��&_�P/����A
z�,�c��y�̵�ƠY�v��ҍp{j{�>rXQ<ةr�,B�������n�x6�H
�5E��D0,=W��} �Iɲ
�����E��HO6�h���5���y��TgO�P�M4+���n%��1�����C��$�y̧�z�1�C�upn4ѐ��=1�B��	����D�I�K��D��z@��dI��L��H��'e�h�
�!� ��1Uy@�<@!��y�l�3\��=�/t�
�u�����S�+[IG��\����t��sb�*�Mo��ဪ��0�s�JO��;�}�WP��?��B�ZȺ�5�c�T�
�fZڻ�;�c���R6�\쪯�������:�G�RE	�����
}3��9ѷ�Q����9�^�s@Fu'��+�|Сi'��������j�VΤv����
+�/Ô��C^{6q�A{����&%���Z�m(u��1�T�}Ϟk�R_�[�ㆾ�ӄ�[-n��-o>*�ْ0�h���Q�����q���Wh$�?�k��ʏ�a�x��cH��,�=g:([�?A���n8�6��QV�)���3X�FL�u��M�l�
Fߢ�?ݐؘsz9�&
+m?̪vʄL%א���+�^�E��닅4'��)�9X�H����5�^�b,3^o�&B���BT�C��XHu[2zj%#�6_Ff\�	T���:�=�-���o���G�!���W��ШC���
+HW�ODȊ���)b
7J���}�,i��4L�$[�Z�j����
+�M��'��<O�ϔ��R��V�
+��c,��_T��JGc�_aDb�ec�dl�u�d���g���sC'�*��2Y�8<R��M���Nt�T}hA=\�6�&E�sƥ�Xld'������(�ءuS�茫���\�:����Єs:zCI@m��XCPɣ��:0�D�-�< z}��'���c��$ 0�W?/~sM���~�' �#'��⬱�#u��L���e\����Z�e=)�wLЕ���Q��Yϥ�cYR]��nݾQú�H�)���x�y!����Wm#�1.��.�T����o��&�}4xױ��x�H��D�O�pe\�x8��D4F#ʳ��BWU�v�2'���L�75o0�Z!�f���%��%1�.K��T�;���[~�l⒦l���G}�s����,STe�f�\7MD����`�8T�Φ �6�l��cN������	�gV��j�p}�0���j9�P�Y�Tۖ���B�W����Fxꉿ�/�֦9���<�B4�
^4�nԛ�.�W~
���X&���	=-օ	�Zm�
K\|H�q�s���Lb�;�64m$�
��G�˱��Ez+��ˑ���O*}le��$Vwc�f�8Qa��Y��i)�
+�48�hxUJ��A����FO
��~2t�U}飈�B0?Ao���pM]2����*�g�I��.:�x)ļ�Ǿ�����ƴN��j�%J22F�KKj���.�@F.d�L�IP�IR�r�ڑ/X�:��r}]�ri�H������w�佯��߷�}���:�y����s��]�������
�C!g�Z)r*(��H^��$"� ٞ�lϙVb�[:�F�S�nt���O�����^�KX9��1�V�82�P��������J��˦J^��S���&ŧ�X,���U�0@).E�5Y�,[ͬ�[d9K��,x�z
��&z�E�F���Y�
fkN�(\��6iV�D��R�d)쨞�p�j��
+=���o�	R��@!/�����&���V#��(����ek[�}��Ï�0t<�hߙ�9/���������	�҂}��p�
���������z�,��h��[�߈5'�D��!()8Vզ�dR��BĦ���~��}O�lP�����3�}>s^I��ImlU����w��,�*�uNR�׭lH{?�^��覩%��F��L��5���M�>�ഛ,r�-���9\+�p��DZ)N����к�P滬c@29RY=w>t���ѧ�D����F6�J!�
+����!Ee�	(	2s��66��z#0���w0Jba�4�G��i{�(�!Ӟo��������={Y	T@[�n��$-�dC>����5j	��ۜ��?�S�w2o%�>_�^�?O^�ƌ2�'��K�uQ�������(���E
�N�l6"U
�W�����Kb1�|��H �v�k`0$=P�&��ƻ-��I��Y���W�	N���R���T
���`"��
�Y!rJԎ���?��R�$w��l���E��Tl����9_��,	�{��%7�|{Nح=��ך,���/���q˭i�ysxMT#�B�oW���k����\jq�EЅ￮�T�#�V2g�oځ"QZ{���x�S����|�s���	�� i	�i���c���/G��<ug(+�&Bڊ"Ǭ�{�L��y�1)\PL��c��I5t-��x��I.H_���o�?j"c�]!�U����D,��ߢ&7I�2\pQ�`wZά�/�_����٘b��X�́��U��������Rk<`+�����4�ף�9N��)��
��U_�
+'J�&��B�A������-�F����Tr>����'�|K(��$�M(ek}���|W�8|

G�� �Wβ򓈮CJ����?�586�/����X?zydU��g�Wv�
M�ҰR\��������Τ
�����#~ݡ�����Uo?.���(��tO �����������۱�����뿼���O���}��/�__�������{�����77��.��_����^����z3�f�#;-J�-L:��\R��6u
+endstream
+endobj
+
+1431 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 57
+  /Parent 1 0 R
+  /Contents 1432 0 R
+>>
+endobj
+
+1432 0 obj
+<<
+  /Length 1840
+  /Filter /FlateDecode
+>>
+stream
+x��]o�6�ݿ�ES'n�ߔ��X��Æ�P?h; ��u�D{ؿ,���x$�8k�H:���Z��z�aϟ�[�=���/�٫���3��T2_s텨�w��Z:v�e����n&�r�_���6���L���l�L*�Z�poV_f�O>!>tW����,�,>��O�׫�o��o�gKL�$h��p��7 �QK�e{m��mG��ݯ��o��j�.@tτ0p��TC`�ij�K-�V�R4����

+�^�=����hGW\�T���i�}�L�j٪�R2)3�!E-�]0m���W�����2H�}>�L���sDA�킩����1�$�=���L(���{���{m�*����.s�,&Ծ��$��G��/�8�.����*e��aF�c�|���M�
+�8�U��ю��\�)�Yؒc��?_l>���E�K��Jڬ����>]0�냱�Y0�k�k�C@`E#	䬾ľ�R����rÅ�Ԫ�rkd-��^0X)G
+�f`�<�5��+��86!,�Q�i��&��+@�W8	�P��s��[�=�s:vK�yGxف������8�I�3�Y炙���lp��Y�C\
��8��Qhv��(��Z��8��{C�U��}"�y)�Q�|j�ރ��#�PEk�O(d�S��`T�"{�f�d�h�_&*H�J���Պ[/�w�Ղ{᜿�O�w�OWUe��Dh�Y$�Fr�`��t�&rST	�tR�t�6E���Q��=@q������3�_<�R����J7V�$�gR?��'
h��t�m%��K~o�sy�y�k��lAs�h��6ց�;*�A��p���D;.
+h8lTO�V;xB�RH�M��-��x���r� �͊--��L�(�%�'>�nmp[�3!@FU���"+0b�Y0m$�����!dvr�l�e%�
�%�JrQS����
�8'��Yl6�C���%��DTE
0�K��1PGܓ�1��Q\��\I���2�H1��"�5�8F&H���9�1.�k�Y�3�ˬ͍������Up(�~+Z�GA#��f�A!%D�@�F�**���T����Q��a3A*���m�h5}�^Q�x�s���Pd7���vQ�3�	h����h��NJ��'��^g��l/O���XU
���l��z���!]�M��"�����4^�;���k{��c0%��[�Jﺛ��e�=�i]."L���'�c�?U�! ��g�ģ}Bp����&N0sQ�X'.{�~Prx�M���ZT������<�$��*,8�-3Os�

�@����x=e
�G�ro
��\�Ҽ�ک)�.̅{3W!�����т�F�e{�6IW{���A7�~�$�}Hh��h��DQ�Թ�n��n�|O�P4)$p���2Y��
�N<�b�dT�i��g";	'a���,��╭%��������Χ C�I�#S�#@�贠�l���g�H����r*�U	݀c�YN��Ɣ�0Q�'%$\R&zݨ����M��Q��ҧ��=��
[.A4q��p�����{�u�*N�j�hW�v�N��[o��5*w�J�;P�Q곪xa|è�����8��t��))�?�EzV�>`�MI�ԛ�IF�!�ے��u�Ԉy��K[e����5
1\'��X-�yL3E�?��QT�H���T�
+�w�I�;����n��L�z"e�Qp�h wtd�Tm
����� o��m>�/���F_6���_���˳�����q���?�Ϳ_�����Ms}�}�j����ys�|��Vm�ᵰV1����V��D�E�ǐ��K{��
+endstream
+endobj
+
+1433 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 58
+  /Parent 1 0 R
+  /Contents 1434 0 R
+>>
+endobj
+
+1434 0 obj
+<<
+  /Length 4249
+  /Filter /FlateDecode
+>>
+stream
+x��]���}Ŝ��3>��5��0�M
+l\I�T%e���&���UɿO��Ύ��jiw�p�;{�j���RwKs����W�ݻ��;}���ם\ܿߝ}�h��d'�Յ(L��B��W���d��ׅ��h���N�^�Nv���Z��/dw�aqz);e���	��q�n���Rʳ�Ҩ���o�^|������Ң���E�����n����xgM�2��������~�����/6u�O[ɐ@m0*?}������Mkd��%I�u��
��Ҡ_w v����;�P��k�4�d'����q�f�Via��
`
�W���~̸D70��S�2�!�V�
+��Ǡ��1XܑMFn,` �����7r��b^��qmy֑������I�����ֻ���՜�IU�R�I]k���N�#�o�w���<�N\/z�B��vd*
+�F*$f�:	�����D���q�.�0�RR�]�
���L�\�z�YG�]Ғ�V_茤l�uy�63yQ��в�e����€A�z���QlZ��S���@����Nw���"Af�Ȗ�Ey;V ��R32WH`L���_���v6j�.��b!��
+�"�Y�jRa��xU!�Q��2��HiGC�ݶ��%ͤ2�(�
++A�1���UۙI}Uhоy9���:��h�q!�$Ѽ"՞�j������E���"��I���B��Ϧ���
�?cR�*`�q7ܪ#دHZ*��v.e
������'���j|�%�ڈ��luf�
+"oa�)qTr�Đq�0����I�H��P['t��[��Z�6:�H	��9�F�Ĩ1Xқ
}e:�L�|W��~s�d`�rϋ ���Y���'�T��xe�6�Fl�	EXZK�a�؍��C;ʷ)�`+�5�쩨ذ��2��k+\�ca/�Y���$E	gd0
+AjP��S����� �;V��t�|W�@�n�It��K)
+@��E4�y�[��� 4w�9d��j8n,��$����y������-��"ޑ5�����8!��`T8]"�pqln�����Yj�()>�W�*6��.�ѿ��8�欈b���N�y�Q�'��t$��m`7
�R��(�l����l\�oyz��)�{�FKW�7Z�����&�CqB����uk6R��6U�bs+�y��x����U8m2�	�M�	��PP��,0��F���U�.���+�1[��v�}I���G��XT���߽�������c�'u��z���dy�(�#�#B
B$��,�E�J)���z���p�6x�D)�z�L�T�r�)W^
+�z�[h�,��s���[ҵ:��<T�CJA
+�����C��ѽ�4���jT�:�T�R{'���
V��J~��70��`��*�Q��,N��n�;�y8�L�6 �噰�n�6F
+f����7��y�:€8MʨdG/u���H�SVh�e�*�����o�_�{��}����R>������}+��w��\���ڳ^�3n�v��'��dQ�a
�Z%'��-�<���<񂂠��<����n�!�PE�,�c�#'�M���x�MA5�'2����o!�c��y��W��;T��N���ظ���C`o�&u=�[�R��0*F;�y$��EP8djC�̶��U�v��z�8Snډ%a^���_��ё���18��L�]�
��4v$�g���I�U>�!+��"�8C�D�n
�5(�@p�i���i���*���Wg��F�zja&�.�6/�j�}NY]@վq�>��6���?I�u8�
+;�.��D�(�ذ��-�sCu�v��\��nc��{Ag�\{��.qJ��b�k<
(�=���PE��M�ʄ��L�4�
�3��R�y��R��A��!�K�T�%7��JmX�0�����q4�ۼ�-Q�9-G���$kT �&����$׾!$�c�kQ�n[��ZF!���i��Pܧ�8�
�%ф�P���a�^�`���׸���+j���U�5U�kTL�i��&�����cL�`'5f� ���:{X����ϑH,���D�a�0l���5&�+��m
+��C��<�Z��E��I�"D+���1�V_��oR���Vo}jn=����^��i��$������P�6�%�<����K��r(�&D���;�C#��>�B���D�J�ɲx���e��`�9��K���.�� ��`��|��0��Z
+��Y�>;���QG�"�����%�m����l��&v1����(@{BG;XBhc]
ڻVx=��ő����
+9�p>�O�3H��{[����>��ݮ2��U+��ڡ_>b'�轛#�;��ZL�pE�ش�\���'����5Y��6���H��M�‘R��>=�m.ڍVC޵�4�
L��
+������6�W���'�n4N�(vK�f2��RX� 8\�q�z'�����F�$���2(vQwY�٥¼9�a
O���U)1+��AQN\���I=[�i|/��sL�Pj�Q��V.��C�':tCjn�#J9Ix��mdk�P���*��M��]��5f��(��v=>�>�{*�
�*���	�e�z�O�k�-�$B���&j��|#?5��ȩql���=Hq
m��ta��0Q���.��*�8���e��^ag����i�9��ϙ��=�k��o��|���+�z��˵S=�^`����Z�K�&=�r�'���˗���ٳ���^H��E�r;����b��F��?���h)T��ۅ�^\6S�� ��R��r����� ]��R��*I�i�$mw"�Y�[(�ګu��^��;�*{�FK�N�
+N��e�?JGU����=)R��^k���(oE0Ҹ-ӷ=f"f�Ĕ�ړU�*BS�x��̒�k����\�t̑�edzv��-��JG�"f����L�̒�U{2LBUE~J�^�Y�sm�Q��k��9�s`����&�Qݣ��=���v�6}�Nk'dﺷ��Y����BP�N9�s��z�S~1�ٚ �qv������9,n�=�*���������$uoD~�n��ζ�T���� Z�2[-�nGc�`�w�7��¹捧5����No�8�{��ZUEe�����;��f�����p:<r_��
����?�
��y�,g+���A�X�y
[����n/��F����t^f��J�Ô|G'���
�ʰ���j���m���`�咀��b�xM2�3����&�����~��U�$PiuJ6�rà����Wv#d)��DILf0�3����
+�;�s�w,#�j��2
6�<�U��܍i��T�Ow9-?|�(��ڿ����z�@&D
�۳�*�����M9�jVW��E޴�%��qף���uF�^�f��d�^��V��1�&k,���Fʢt�`c��n�M�e6��W|�ټs�P��ժAD3i`*��`�7-�.�CV����^!��ߩ@�	�N��xC�{b�T�1�1�~�d"���Fx�@V���m�Ӷt'����Q�M����<�aJo�1��=�I>���,�Ы�_~ړ��~�J��;�Yz���/���{���*��@���J�rq�W���W�pY-���+,/�4�%���K�'��+��M��z)�9�[Q_��&N�u���iD�����ڙ��_&�pkA�"ݳ�\� ���jo;��1G}C���DRY}m�d-)�&ےz~��"�m�������<W^���WqVf��*E|�{6I@�!?�$	��4���1y��	y�Ӗe�/�"T���{��'�{´�]k@%�0������� *�v-��e� `�-4nN�B�7������?P6�^���sz}���1&�`�2�n��/~J_�(o��6�<#�U_9Trrn�(?�~��
G��l�gX6��	�~Pƃ�]9��k0�ll^)/B�3�"�Q^-o�R���g�ԣ=�*N��pޘށSN�k�
��7���6f]O�h�h��d�M��Q[���'�O(���Q�R%�胙�݁/٢}e���i���>a6���A�Ü��h���/'��Ѽ:BKQ��t�}�v�S���2��J�Ks���fs�Q��_ƴ�:x?^�S��#x�
2�*w�8B��Za�_�^�SM���+��(a����ub��T�o��ߖrn|���嫟_�������r��ݏ�_��v���/������/_�t�����?��z�|���t��YwV͐��q�x�0|
��/dϙF
+endstream
+endobj
+
+1435 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1251 0 R
+      /f3 1281 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 59
+  /Parent 1 0 R
+  /Contents 1436 0 R
+>>
+endobj
+
+1436 0 obj
+<<
+  /Length 4769
+  /Filter /FlateDecode
+>>
+stream
+x��=ks����W�|2%Z!��ñ�����O�J"V].f>���8eS9zSu���ٝ��x��Rq���;@�
�����w��o޾�v߾���Ft��ݙ�\`�q�;��R��ݏ��w�{�ӆ3gd�Bv�IÃ���O�n7}�7�0���a� 9Nw?l���/���o7_�bs^G�p�u="�:�6,h�tz�hP�Y��>)����,X��	�;-�����R���2Lo�`d�`�ꏻV����&H�]�z�B��敱�x�vo�P�=��G�mF���/\�
�VJy��V3��2;T�O
+����ʍ�~
V�R2͌TEf��b�[��5�u�^��I5�uo�aG�J�Џl��$7,���ӁB,¥,�r�_�B,�ԁ�3#Ub����ؽ�FM����k�ؽmRM������ؑ�>��d��^����͋�����ah�)�=�7�L�cB�r����0,pc��y�	Ϝ��cF������?&h ۏ
+K�c��B���QzL�PZ6�yG����qaKɸ��\o� ��-���
�.�6�7����ܷ�w����W��_�/���+�����o6__V壑�۹r��N�n���+; �F������=����ƽZ�����uӸ��A���g6�6�d�Y�i߁�܌`��^]ߙ���y�h)�D���
+��S�>�;���5^�7{{�}�Ŧ�ο}�꫎o�>�.��Юr�q	L:�V]TV}�_�W.�pNA�E��=2b�\��K���z����I������܃?@^����
+�S�i���a#]���Iqe|��ig����2� �@p��6S�7Sj湱+����o�	�q���~�IOO�A��P���п�-�"#�q�ٍ��l|4ɏ���3�D𣐹�R+{Ѹ�${���P�G�W1?A
+g��L�����
X�2���L	k��t^�g���el4)���Ɖan�y�&���g��;��I����˧u�#Y!�I4}H� �m���)r�lX(\O�
�Hy�u�1e�xS�I�M�c�Ѻ���2������-�*�֦lKV�
�9+�Q��ֿ���Jm-BC��l<�IX�LPʧ�z�3��??���`�m@h �#�:9a��R[h� ~�m��x�f�
+卉A=����.��������/_r����#���&��iw���yLC��z����Z@5���Ɯ�[Ȯs_l/F�a�U��(�ֲP1#�l㢗
���9j��	��
a�
�`�������#j/Xt�( {L�_j�dNj��,��)R#Q&���j:c���.Q���9CZ�������yG�AIx���i+�+t!��x�v�����>�PEJ�e�Z�b��m�=�%d1)��@ص{
���!E��X��W��G�Mɸ��AY󤁙��g�U���}j�[@' �RQ;L̔�.bBj��Z�&�6�K�B��;�]dE_�sb��4�s�Q�
�!\�q:]�yR�Фz��ȓ��jୣ�>�+��&N�l�������],ÐV�D��Z��7KL�
�֤�\Q	ɜ���.w�
�"��zh-FW)�<�)�&���C^�̽-e�<M��9�q�տ�U�-��w���"�U�����A�ԏ��	��0���k�|O��B�ؠ�o��,�K9�Ȣ��t�M+��Z����Cٷ4!$�J�(�Y^xs"��4s����

+�
-�J����<��cB��F)J�Ѹ����O6�*F�|�P�Ԁگ�����)���jSQ���oo��=~��)5/��Rڦy�Ҽ�c�wڙ�X]#���#c��9%�'�0&ZxDZ�>�Ȕ�V
�'`I�7ʚ)�=������]f��i�d�.���k�Ӈ��v���#g�paS%��
��R���������!"6뜛�DPT3v����q$�F�j
�9"~L�H�
2��㔩����~��a�>d�NȁɴAN&�܏+	>��$\u��CNZ$Kk[��[�@�ŭ��p��*%7j|�
mn�
�=��l&ӛ��+�8�m��f�����Yz�"���/�a֣�̚Xd?(�b?X�j�HQ+<�z,��t-�"^��Dʡ�՛p��/)~��<`�X�
+��D��i�G9�B4�8�1�
+%����Q�A��h��(A�)A)�šL?YY�s�nAt�X(�AD>[V�߬�P����=G��q�Ȓ�)���D�5tF����"i�˺�S���8gN;+�A���Д��j܆��N��n"�Y��d��+�.�>M�a���(h�.{�uYOC��N�=�_.
 P*�������N�
P3܆L���7!c���/@�F�4&�INݏ2�Em�)��pnj�ޔ��c�$)_�O�󆯠O���*0f�b�L'��Xӕ������q����'�Z�_Qb�&F{���BUηh�ၭC��h�b�����m�Jl������.]>��_���fJ�D�j�gd�S-�����;�:Pm	�<U���i
J
+�6��Al2d� H[a�C���5a��rV�
7�;�EA�u�-~������)3R�px�?	��
2��^��$�6A�}��tv"v��<�F�z�L6�B7-<�l�j�>�՚������y�.��v��;�->�+d~�/(���aEM�[�k���h�ҐaϷq�t0��f�@����dK�<rt`x��m�7:���U� ʾ~|9ec�]^��ͨы���g������CuBdݱ�R������s���:�0�S�Bz&�����.LF���6B[C���$�2�h	?#�{]������$f<8��2a��F��)�ck����f�s���&BeB�R!��^L�����d6�@�5R�تL=Uj���δ�+pS��q�Z��v֓Ő�L8�f��\\�D����km��\�̋���80ZU����dԤ��)�E&��m�&��EC��G2���(��^净�0�}����b����I�|�-E�-�g��URt!uŔ$���A1�J���2�
+W�ԭ�ÜC+D�p�m�$�UG�N2`��D�~^,�e���I�v'}�2gajJ�Wz�����+�t���������i�nY�n)��fwGg�/(�IYA
�d�;O��tID�]��:�*_ʒ�~�	�cP���O��_�8k��0��k�aul��+Aȳ�^��zl

+�ʒ�B��Eh���dh���U���e��U��!uQ>�S��y�Ns$��n�� ���@��f���J�|3��#��Z�$3�sg;�-�O�Ղ7��1��}.�0��&|�ћ��=I�(��Id�ٞ��CѼ��S� @�s��#0I�����nƆ���b��k��1�i�Z�"���e�PlÖ%3�������Uj�����iԲ'ɏ@���L�`�
+�s׿�(ɕ�t�)���ǹi�Lpƽ�<��m��|I����r��r
8�>I$h�f�ڔ0���
��
���K��
���DǓhA�4$ؙ�D��U}�j�닼kt_i�J	�n�0���j�!���c3
�QHR]	�8XKI@w�Ų��B��p�%@��TN@^�&k����3I��3
���Ѫt.��i�x;��C6�d��>`'�"'+ϼ�����-�5ށNQBUcL�6_���C�3�E�����?d��
��l��LJ-��	�0k{�\�T��=�S_���:��S
+�Y��4%�EWP���Fڄ.�܇™�j�����L��;���Y*��/��O�5ZS��	�}�%(S���Y�dR�q�J_��mm��a2�%Sʯ[)�
+����^@5��?��89���ܐ�iU@�%Ϥv�������6\�'B�U�;�m?�Q�d�ȸ5��L��5m��[�P�@g��X7��P��P�!
+gV�!$��ҹgxJ�L-��v@�˧ᣖ�6�G��]zy�
!�O���7��勳!��,��A"��^U~�
/z�f�I�=�D����/>�(-�q^��Y8μ��4�*�ck�6���h�.z�SfDO@I�|6(i����L�(E�[���ǎ��LH٤ѽZ�}�)�p��������d*j�y~���E�6�qV�����@b��/|HX2Փ��R5���$��WeIG�6A׀R�Q�g�2WZԃ�5
���̰J�{�C�`)lژ������o��~*0�����܍˻_��L�Lrf��!�vHo_c�Qƺâ�C�^e�$qV���� ������T�Z0��:��a�p��[L��ˋ�:!�������誕�>�-^���29����ڏ�,-L���Xq|�Of�� 
+=G��ꙔОI�sΕ<��I1�L��화|Z�
+Ϥ�nȫ�9չ����i�!;P�9��`,NU���G��mi_�]9�r��9�@���� ���X��k�Dy'}������w��'B_!іו%�d�vҨ��&���ǵP��Ek� R#��I����?�G��Ħ�b!8�F5��ݣ���&�E����{��%�/�W3�I�:)W7xfx�jL����vJ!jxR-Ck=����e.G���TTl��4��ws�N0i�*�A�U���v�b�a��7kU�D{~�������x5�v���_��������v���?�߾����{ߝ���a�����r��7o�������nɛ�4��i�xPJ�9��ut��?
zF_
+endstream
+endobj
+
+1437 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1287 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 60
+  /Parent 1 0 R
+  /Contents 1438 0 R
+>>
+endobj
+
+1438 0 obj
+<<
+  /Length 2306
+  /Filter /FlateDecode
+>>
+stream
+x���n7�}���cGo6G�%ۈ�5�E�-ʃF��^�R"O�ݿ�{��"���4G�źYUs��������!g�.߼$lv~N.^^���q�#��XG�e�#�j7����솑�O3F���:���ts?�X�Y<����,�&�����OO�c�WLr�a�3Y���Z��9{��rv���9ST	�h>��\��W�L��?�cbx��{տށ�����w
+y6��ӷ�ω3�I@Ĵ��m�����No��D�
�W��g!
RA��� ��=RS5)���)A��!�^���8�</�S@0czN���cLo�7��
￝ޞe�G�zi3];�\����������9���S��S-f�_1n&)���i65ZF��T[����(����5ϴ�����#�ȣ� �
+���}�HY{ �<��rmhPE���;p���=�2b��(�1�{"bdg��8��)�m�&R>)'��'�	�"?*���ړ7'���$&m(#U8O`*F��>LX�;��ͼ�6��izi���rȁ��)�dވ&�8	�:ƀ�y{��,�H>���V'�ِ�9	Ѫ~��Yz�eQ���>+�[��D�e!�����;�Ì�X5Ĭ���2f
��P%��MN�0��.L�<��i���s��/r��ܗP�1�h������N(���J�3P��Q�Z�N;�Z�(cVB������!�yY��H���Tb��T�����2T�-|c�Wx��`��X�C��\)1�Qc�*�%H!g@��+?F�m0��o�������c���(/�QR$)��x�o�0r/�d�c8+cu����b{@3Z���[� =Rɒ.P�x���j�sX������"�������4u�*b^4H�Pv��AO��s��wiwQ��NR%:O
+W��!o�g���.��e���P0��d��-�S.)c��M��(����Ө��+��^M��� �@U+Uy�aL��s�)&�Rl.uF�D�~�aEa,9�M��0��V�Uʚh�N`�FV&�;�O,Jt�+���!��j֥�BJ�47%�%gu���+���,���0jeѼ�Ԗ(5
+�{g�Jgd�.&g����B�Õ�)D�J��~Fp�-�5%+�ɸ�����HֵG�5)Z�s(;�iy�(�%�\*�YG��[˜0�1�y�Z��.��cZ�ZEM�`������r+!E"Hn�-�8�����b��
+z=.l�
+1�Pȋ�m�*��sݧ�5U�����6dE�g�vN�g?��ĩ�N�mE��pmJ�4�O��(�D�.Q�?j��Ê�	�/����,3��A�FS�a���]�:q
+�W��Q�:�/�>#�o��J�HnY�@���p#�q�]
 ����%�՝�\�}l7л�5�}ŖҞ&��L�6��N{Ⱦ�)��:a���V�<���3���=����Qc�k�]�D�i�XG���
�f�g�è�֍�)y<����:�
v/솣0������@��	0?2��� vu��Ҏ)e���]h^�Q�$2Js�@��;7��.+C
+F5�蚕MJ9�;�O�	-�􇜎j�oR1�7j=(�)��*$2�\���{�66�y�^�ﲙ�Ю��>A.4�>�)�M�yd�"D�]c�);ʅ�M���YK�3VN�8��q����P�D4���4;��=���-�Ka��q^U�a���rU�Z����������Y��_�~�؋��uKQ盡���DlW�FR��;)ڪˈu���TK�d2�x-�E���Q���n�*6ԭ���U�!�g�-���k7�n:�Y��y����� f���l��$	�)�"͑jn�Ւ�9���3-�C1qU�f�y_�ܽ���
*jj6LSK�7
+\P�亽�@aS�RQ.�O쮢'�Nck�6,��:E5fU0ȕ����ހ��!��h�1�GX�[
�B ���J��dU(�,��"R�<�q�Ml�ji��Z&\�vd�f��b
TȽ
+ɘU��!���ю`��j�}��*=��0yZZ8xƯ� O���֌��p�Zf�7��d��p,o�[C����}�yT����92�_�NY�N���>���A�t�w ��6���'V�J��=�D����hLJ�P)�R��4M7���C��t4Σ���UA�vh*{�%
��8;�sW��y[F-3���f�㟦m3��,���ú�H����������W\V����B~:�xX�>�����?����~K�^?<�n�_-����p���p�]Θ��(A��R��J1���/���S
+endstream
+endobj
+
+1439 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1281 0 R
+      /f3 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 61
+  /Parent 1 0 R
+  /Contents 1440 0 R
+>>
+endobj
+
+1440 0 obj
+<<
+  /Length 3858
+  /Filter /FlateDecode
+>>
+stream
+x��\[w�~ׯX��$&�;���6���N��4��C�I���4r�0������3@.���%�2�
�p�����ӧG]w���뗝<��螿|q��#��Nvg��0A��'�V��n�?:����GR�{�t'�v���v���������.���d�lwy�p�q���קWR�7�WҨ�7�団O/��<��틣sL�"H뽐�(�~�!M�ͧ2�Omן�7��Ewf�P2���{i����0�^tF
�,xQUn�H5Цu:�4�M�g��n$�PN(�/7.��a�֡�Dž���c�W�Nh5|�9%��	�&��)-��Z*�MXt��d�2��)�r�ՔI�6l���g9C�{-t4�F���bG�7#W7؜�S�� GP��V����Ѕ%��)��/S���%4%�b��Ⱦ��$�|=����
jԫђJ>Z���BQ$����*��2�d,�>6�]j��'4V�\�������E28!���C�$/7�K)��7�����6ѻ͂�z�hE�S�Ѳ�q[&�� ET�K?9�Ar����M9����W?�4=|9��h�x���1H�lںߣ���6[y���t�o����Ɛ�|�� w���O�k"4�+p��l���,����55!�h���ڗ���6��5k��S�bUQ�Ȫ��
n�ʎ���^J3�Ja��JO��8*�u(�7��j��@��LjF5��#�I�2�����r9�v����ĭ�����U���)� wa���$l7'���n�9��8�hu�j�6‡�쳏�.x�(�0�#�![���p/
+➚�T�;�w.
+5�B���������ߴ4B���q��qc���S��`��h#���@��f�ɢ;K,�命օ���;���M49�͛)�Ò�)9:�l(�2�i�R@�����N��.H|��1�o�m�R�z��赍=cu�6���[mx���=��Ff<hN�fJL0@OV����J�
��� 0�Y�n�]�Lw[q"s�����>L�1Y<_��,۶����5�Έ�7�L�-�m:���@��"z���F ��J[[���]B�����|ʠ/��(��}��=�Є�R�|r����_�����N�&�K�X%�'1ַL''\0p��᰷|%�(�G��֐8�"8׸u���μZ�~�p@�Kp�M��P�舡���R6�0Jwj���F�
e�`f#N��m��}�
iih���
�H�(���{uӾ��4��*�ٳ�)(a����`%�Jn5ƬQJۢ����JR�̵zҶ(�$�2q%�	R+��!)�&$�G����F~7m=�`��E���ۇZ�y`�%`B��x1�7U��0}���.(�+p��?}��WR~��y����yp��,���*.���� �
�9w@]tڊ���R�}�k1�2Qׯ6���jl�K�\)L�
��j��*lQ��*��o�+oq�-�>A�)�c�N�͑r*p�9~	m�΢���q��"%L#���X(���TN��"N)��X�~�
+W7;�7��G�Uj��Xkfz�s"��J���Q�Ԝ�H�=n�?t
+���i�2]���g�y�Z��
[j���~Dh�XR�t�,A�s[0|��x�B�lZ��5�l6��fk�l���s0Z
+�O�,$��PA��+�_��
+�Pn�K�&FE?�E碰�D]D�P��R�G�j�*k�S�VX�[��Z��WR�J)��s�A7�P�݌c_`����rU>XwE�t1JkCJ��n�q|�KR����{S����Lp��Fi��P���x�W��	�*ܱp䓲�v�d�.����Y�T;w�Llr��ǰ���X�����gd�f%1��N8����:Ï0Έ��8L?�B���)���
C���Q���8�ԶHU��V
+ich۰�D��(�@��Sk(z�E��VKe\ًem~y\�*���	������8�v����W'���!O2Ė		�.�&ոjXC���V�`*�k%��:2훦�\��gi�e���uuRҞLgI�DM�.Mw��4�B4�h6��"!
��4ֿׄs�����#zEt�4U*W�!�;����o=n`�tG �p;�1p����$2�F��Ϗ
@DK%�4�(��\z
�%�
7蠈[�c$#����hb.X:�R�/Hg2�����Q�*�f�T�d8���%U���ȡY��k&a�����hRv�/D�P��;�-Pe�g����KÞ�*��gZ'�N�ԨjԊF�snv�D�?>I�A�=���z�Fl
+`��-�hpU�f�'E�&��!�"_no2T��r��i��dB��'�`Aм=���5zVD���Z�j��^w3N�h���%(D��S���;W����O8�ǹ	�E��\=.t���Ps�l�����4��&���%�{Б��	�Ne�Ƨ�R�FP�v>��X��(�Z�k�PFjפ6�U�A�&+x�¢sj�o����- s��Z{���1x5k`�5[�a��!�`��=X��g����������KVsX%l��v=�p�$�*�!|�
wv�-�6��q
+}a�K�Wh�G�V�Ѣ��x;'^���z/�4=:C��D�z�iy���0\�����K&�/�U�Œ^6=fJX�Z�?8��W�k��0��Z.j�]IE2�x8��P�do(yA�a�Yv��b٢�Gᥚ�y�f�
+�i�ҪƠ&{`
+,vQ��pc�DBي#K�
��IO�Q��Z�+T��s��͜������U�u�M��E4����y��v����lϗc�Q]=��^��
+���c���&x��H�rj��і����)�i�|?��.�mM���	��Rh������"��[|�
+z˦f�B�9)��n����
+�_�;0mYd^+b��We���l.����0�Oc^Yx�1-rE(�N5e�x ��4���)��md�ܲGK�io�
�0��<tA���>��m����L����:?~ڟxd߇�ʊ��dΈ+-���q��+%[�X�{�8��FM��\R�~�Q~�}�b�d̺��y�
+A�n�3��s�c��#أ�.���"Ouĉ&'�	A���7Jv�P��G�,8 �>�.T�-��-w�0�����jŧ���7r��?ZVW�L�������N
+90�xԄ�R����lr�{���s��E�~��6���Zbkl҇�"��X
+��6z���yd������D��{bW�\t�mQF_F��ACS-�e�V7H7g�T4B;�����δ�7��yf
+P�
+wy����4G^�z�퐡i�5�	�0
+��W��<����O%f��]l쪏�(�]]�9�8=��4)fu��`ʢ3mKkz����e�Ckٶ��7P�񈺃���rq�nUX�-cߘY��dY��>�ڒ�ps��t�M+�wqƫ<-�	j%��
B�ǧ��tqBpl��4
��\�>:hP���pYw�f7�p}��<�o�w���
s����Lna\h�P$ܕ�M��>yP�I<QĚ�b%dM�هn��"�n֫Z�a��K�6�@�#n�5� �RR�����5=����0po��q�Ů^��7�����	0����AW��<U�n��rݍ�'�o�Rm3(�����2�.�P����G<���9��>�N��W�V�����D�צ�)�iV�)�������FIG{aJ�|��w:��k�K�ʫD�Һ�Q	F-���-��I��h�yR�v��!L�_!�������v���8$43�7�?yX~ww}��X.�o���������/���f�ׯ~�Y���w������V�\����o���^~���z7AZ�K�tg���1z�W��t@������d
+endstream
+endobj
+
+1441 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1300 0 R
+      /c1 1299 0 R
+    >>
+    /Font <<
+      /f0 1257 0 R
+      /f1 1245 0 R
+      /f2 1281 0 R
+      /f3 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 62
+  /Parent 1 0 R
+  /Contents 1442 0 R
+>>
+endobj
+
+1442 0 obj
+<<
+  /Length 4291
+  /Filter /FlateDecode
+>>
+stream
+x��i���������A����s�R�yU�`\	��l�lp�*���ݝC�j��;v���C�j���������O>�tݽ�����yw���oT';��U]��)�.x#���{����Kٽ�y#��_��ܿ������ޕ��.��w.~��z.���Yw��Ë��6�>��Ó�ݫ?�x�mw���g&���
��PQT���rg��s����欻kt��\j5�9���x� �	\��B�Ј�d�D�;���^8�R#�ԟ_
�,7V�{u��~�̃�Q����>l�-��׬�-�xs?~wm34!��f#�tnrN<��5\�dO>:�p�`<�Y1�o۲0�`~9���C�\fP*�dX��AeOW������)���[n�oo�VhY���Ryw����zh����(����Y�ZF�iDhؔ�� �O�Ip�M�U�Q�h�6I����?0s�&&�PBj�͔(��̸���I�$�$�52z��{#Ì]&�/$~���U�V�u�E�����#g�X�Z��6��p���<`�$(�֎b8������L��O�"�>S���:��Y��|�����w8s��Loc>�>��E��B
��
6
?i��r�c�a4�ژx�Ɇ�F�
�nd�+j�9���	�jj�L�����PF��d�Խ��|�6a�Ag�A�O�bD9�H<����Ŋ��Di�'3n�y�*���1-<0hh��U�>�ftT)ӕ
�$`d��Ԙ�l�X�ou_�T�V��n�(��K����
lG�f��񚤠���]�
��I�')ˮ�oF�\��T�Ci��^��g��CīG�8�c/B�7c�F�D�F���
��WS�g۳�[踎`����fӾ�6����K�=�R�2m�B�WB�V�����"�}�@��9m���[�$�"�f)AB�`��Ȍr�k���/�pdU�j��x�����6�E�bt�oJ�F�A��ܖ�CR'�j$]j!?���qD�\ ��̽N}�/@{��488c2h��
~�Y�'�9`�3^�;0�c�:�Q]�]���"C"���a�T)�\�|��)+�
�������?A���LqC�uj���UyT�7*�&Z�J6�Px.�L�1���Ҽнa��1Jg���]�v�K�%���b���˕��cXC�H�>N�ʦ�r�\j7m�!�#��%��n�Mv��9L����d����s-A��d��W��&S����-sщ^��`�yO���y_*�SJ
+���ɺ𥯼&�����Q���E-\�2��'dT~��m�Lk-b�Ѳ"]�I�>lH
��K6,)�1y4�Kd�`h�'�M�e��߻1O�t��
$�e<�M�o�Ћ��|Jq�����	b����� �b���)s����5�.
+��ӵ�Ū@
�T���{acPQ��4�mD��ږ.�Sc��$\�Sz
+��FΊ��L‡���UAACn���9W���3p̘zrh�O8&Q�|UN֕[�vW��'G"�Ϻ&TG)�3}င�(�>2#�Bu�T�OY'��mKkv�F�	њ�1J,����p�S�E�8B�r��1�%D�ɤW�)_%�X/X�P�^Rh�}O�y�Sk�W�	yM�\J�h�Po�Ƣ�U�JS��:4��+CE�����	�䃯A��Ȑ�K��=-�i���Xp��#�Ȓ7ɼB�^t��I�o-{�}t2�Go��H�ԫwBJ=�I�iJ<�k�"��R٥���H}I;x�4�h��1RT��w�by�O��J������
&%�gf<��I���]'�I

~{*�AF�V�j�w�F%��t?���\w4��iZ�A�d�4�?i]2ZoTQ{������~��8ucSEs���˔��j�>�N��X���g]�	�~ܓ1�V�;�[�|m�z�c`4�ШӃj�`���-���|�6Ad��&���J<F�^�
�<�r�%bޏkb�6JY���'[���e�d���l������B�Z���eȨ1~�h��M~D�\6&)rE�0b��*��Z>B� zg�i�����"��
7>�o�(8����[�LZ���ޢ���6��"κ5�;Ɠ��|���0J�*�,��RD��z��R�񀨾t8�!R�E)*pa����z
�أ�Va{����8iT�Q�ך�F�]���Ĥ��Q���,x��	����h�!?�T��x��v�	���2��`����7tl#���
+
+���u��z��j+z�n҆j!�v�PBsܾv�U���,��FG)����.��d�v�&Pk���z+��}��6�4��E��
��zX�P	���^7�^��@J�8!����%�q�� ��
��9g�2m{ReK��MN���X���#��oqs�����oBiI�b�=���LM�ؤ��JD헨{���z�z�A��c���2zx�p��@[��?H��h�!Ӟ՚12��w���Vf��ߒ8�5'ۨ�bTl%O�XG���> M����^C(ǴƠ6[ڝ�7�HS���d��kt�5�L��3%ȣ�G�-M� �h���#��9a�\�����ţxJ����Rg�Qu�	��OH�+hے��^���;�R-1~��
38�x��'�B&C*����f���:�f�����a����>�rI�P��$sz�EE�|��ۨ^İ�˽�^ ���1:�ďF�\�f{�w����w�9���ET�ј��B`�Q5�q��g�C{6%~0�n��x��z���t�m7��ݠc�("�-�p��LD6��QRv]�*z^^uO��Ԁ��% 8Y���ά�1��
�.�F�1�㽧Y�-[��}�.8rbe94��4�w�㘄�6S��� ��gr��
+{*"��p�Xh��Y�֡�����\��y5�S A��5s�M��TI"D!��z����2�ዯ$)��K��|m��㠄���E׸.jG���q8�r���p�ҋ���sk^�0b4�~|9
+B�
+;�m�Z8�Z�j��d�?}i�28��1�,�{L�=�WS_�v휈Z�����W�J%`�E���{e_��������Qz�^:j'�T[j�e[j�9�yK-j{���#[f�.n<MZ`o����u
�?dئ�=hcG/��@M�.�x��'s,4��D�N�������^��v��dY)a�Tc�"8��;Ap2��v[�{�ǍRQH�r��~H�#d�6r~������7:�Fl@&F���}'E�6Hw��R��f����ڿ<
++넶=8�����J-�x�ͯL{a�ӧ	��E�R�2ӻEf�e��08-�OU`�ռ��e]��w���eF�7$.	R�W^��C��~��=x��ݳ����0��1B��ͳ��֚����	�)'�tnu��=��	�g͞������kʚ��2�
+;D����ؙ�O�n�~�b/�򣓟��ԋ�e'�w�1՚�I�ѯ��jhB5�h�}k)�s��Ơ|�p����픚�>\�>l��D�:	���1^J�N9-Hfne�!��������>UTp��N�;aƨ�>��*#�V���
;�K1Jc��?m#/
+��x}�Y�,�����N([;�� M����_�����I)qk��i�kw�+�j-]0y)�.IcCt�o�IѼ�ƀ^pGu	U[<�
+��B)|I]x���"N��u:�ʏ�/n 7�-0Ul
o3�G���tچ8(h���
��1<f�F��_����U�su�rݘZ
+���8����ʜ�@�K_d�P�kM!�h�,�=�]�����f8��#&��
oz�ݦ��2������	��Ⴡ���WRۂ ]f��ފކ��}r/�?����������?\�Fd8�;�9����
+�RY�=d�b{�+x���>��������>y���#)?��|���q�����՞�$a@����K���5���Y��vٝM햧V,U-��Nڢ�BNOaK��  �
��.�eEx;=�69U{�sY på-�/yV� ��8W
��"G�J(���qW�����tF!W��LB��z6Y���ֺ���
+F쬭�����B�}�$'�Q��j5ۓ*��_r\{�ۥ��|�娏�rة�Q%�c��R���P2
+�b�}2�R-0:1����j��)̊��6���:,ݏG���ڿ��=��łY;����/?��!.�ol�3f+_l	��v���w���}}�����͏��>}�����?���=z�f�����������/�߿yM����l���G�h�^r��n9t����=��
+endstream
+endobj
+
+1443 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 63
+  /Parent 1 0 R
+  /Contents 1444 0 R
+>>
+endobj
+
+1444 0 obj
+<<
+  /Length 2982
+  /Filter /FlateDecode
+>>
+stream
+x���nܸ�}��.ǃ�4���M��@�u��~(���k�H�]gh��Ќ.<�!)�#~�G#����xn��O��wkrzy��m�	#�qb����X�)Ґ�ϋ�;F�.�Z8�aTh�*�#_��˟�;ʴ�\F>��RR�*�i����,�X��<[����:�M�
���fV�:�73w��l��R��	�Me��(˴�?)έу�TG���QX�:Ş��UT	-v�
+]Q�7[\��e�G�$�y��`�Ddh��<�8�a^RHm)7Z��@<\�W���#��zV��������#Wg[(r�L�׎2Q/�j��0��9!k��cZ��V�B�f���j����w�P�f�
�d^Ё��	��3r>ؕi�E[Z)+��������#�\?�����?/~<�a�����0.���X�#)���
|��i{ǚAL-"����%4Ӎ�MD��C��f��#ڌyhF���u�L��m�cL�c�ԗ����tQ_·.[�Z����@�E���jI�j��˅��>[/ɑ��:�X�/7�Ig-�Œx�v�
���Lb�ɡGz�P2Kk���ʗ]���D3x�T*��a@]� ��P.C�������~� �����;�''��]�:E��|�Y�2ϣ�-@���|�Kr���2e"�،>�/�[�\Q��x#r3j����PS\�!Gh��̚z����n�������������猽}w�B�S{��xy��Td��9�*�3�#ڗ�&lz�?Q����ҿ�
+Ҿ��=�	
��^]Q�3���	��"�;��9I�M5�<�l�5��(F}}�hHT��E�]���V-B���A����_��gooB���p��D/����-&J��|d^�Z�ȩcBW�m%a׮����H�pä�{�ă�z������XȪF��*1� L���M�}Yr�����B��v�L�JY��V�Dj�~�~��s,gJ�=<�T���
�t
+�
+bVe�g\�Ա�����^ܕ]�G0uT*�Ү�u��bǧW�m���UB0�<�1����z�u��H��~�QdYO1K��60�g���0��m}E.r��d�<�/g��x��1r���dε�4f��1�y"��5?��!�!S��@Y���g��풱�l~3Z�T�(O��;.k�,Mv�a���*����	��.��P9�-)e�q1f���ĭ�q�u˾/�i�b4Q�Pgݎzw������`�8)vx��BW�D2ysE���V�4�N��K��_y�NR�D竤���e9&�L�mp�?5�`^�=#�:}+ҹ��茮�H*N���θ��^ѥ
Z�����)dr��C��b���G
+�C�E���-���76ϛgO2��IA+g�䔡QZ�&	�����/��rc�h3
+K:���aj�z�
+
�~���*H��Y�r��y
+��`��]D[�7��m	X6�0�ۻ�C�ںdQIʬs�&B�3��(6(�Ez�o�h�<�&��H��T�#ի�Z�z��ش@P��9[�A���UZ�i_b+�I�}���}�ɧ�'[����"e���kg�����F5�]�l�*���\�Ws�Z����\r�U�<�{~#h2�u�������]�.��I�Y��(�Q_����.�I@q��pj4o���ێ�5I��mMՇ\F���D��
Al<�sv�*�P�|èӅ,8�;��HZIc8
c������G�@6֘�Pv)͂Ѥ��=� >rj̎oo>_Y��c�ʾs�Xw������D�쐹�Z*M��<hV/���m��M�;0�b�8-��(��p���xP��J?
+2�i�<(E���¤���\���t
+l}���H�]�f��dq���
+�h'�����LŲj+U��O�����L��pwHnB��m%���)
+:Oc��4a��,��
+����=0� ��N&iQn����S��KV9A�e��YEM%���Y��)E�a�ڠؕ-�tN��6��A���;Y0Ӈx�
+d��@]���.��fF�+���g�di�!��|��E�p<b��O/��`l_�r�c��}I��}���]"���S9�h1H5YV5a��;�QDKBvǨ�X����?�c�r۷r�:��,D�4�LKiu�[�-�`N2@�C� }	o���x<'R�h�
Rў�j�6l��!|�m}0�������+�1zNe@Jfa�>��� "y�R��d�t�DXK�6�L���<���q�)E/��N�8��;0q����}�y��q����U��	�hs�cE��},QM���V��C!��be�#�F怡�;$_���0���@�߅�)�
I��?[!R�S���?�z�؇�)�/����}p�ϼnj�4
+��89v�>�{�ێ��F�'��@zlM�p�m64W�QS�'���%�X�,����z�_���������es�&w/�o�gW�!9�r~S�[�'<�)�Ս��Ns`
��*x�N�x�˭`�.�|��/��Y%u�NPԝ��"�2&ۓ'�GQ���]�?�;�׊���-l*�Ӕ!@<�wG6����aYHO���`Dž񳐐��;�d�0���\�){g��Y�2����U��D�
Nue8�to�2�ńÊ�'�� g�IS��
+Rf�1k�Ah��811�^�����
����p�P_klgl$�w������UT�b��Oj8>E�з�o��4p>�U�����S A��j��`�o!���#��+�%@��Z����A����423w�@诒Ordn�
+ȈQ��2��I�R���pk�v:{r _��~����*	~n���y�KZ��u��-�~�L��I��8�Q
+^��dvA�&O�O�|�q�h3G�v�8���T[�ƾ^���r�o���������o�~_����=9>��e}�Tu������o��<�M�jSfD	ʜ�bʾ!6tp��ھ��
+endstream
+endobj
+
+1445 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1293 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 64
+  /Parent 1 0 R
+  /Contents 1446 0 R
+>>
+endobj
+
+1446 0 obj
+<<
+  /Length 4088
+  /Filter /FlateDecode
+>>
+stream
+x��k���~
8g˗��Mn�8�;Hm}@?��p���}Nh�}��v�3W_l�'iE�Ù�H�����F|��B���O��J��Ç��WO�Yh��gZ�NڨT�+it������W�/���tN���K�	%~��Y<�X(q�~q~��v��zؿ\�]�p��R����Ke��Gq��Ӌ�?O�?Y�c�4�Z�J�C�����~�R)�y5v��Z
+�d
+�h;4�~��W]mg6_)]�el�<�l�R�m�j��|!e�M�(N���X�H��ɯ\�L@�"�6M��t�~�}�`q�W��>@xe�m1�aeLiSp�	�I�o�63�m���ʟ%���Ig�m��r	��R�@��,M�nE2-\���AV��̡
#8:�4�e��t���d
N���Ď����~��z��h	�к��g���_�NI�����Lk��-�I��uP�� ���(����A�
�kDF'��;�f�uF�s]m1`�]/�~���'�e�BN���w�,`�^���3�A�Q8�P��"E><U��9�w ���$��W�X	�
�i��'4K�l;����_�/�V��s����]H^��;����]P1Q{ߋ�N�](�e�T
+zz����¦$SҝP���vZ�N++����|1v�����!(�tp*���T�{�2�uC-�!�'W7}x�G���ʆ�Z�����CO�x��Z�`EFk$k��m�C�hD��m��.XkX|m���5�MV���	����A�=ⰷ���9�?l��U|�I[���U��}3P�b��Z�(�W�:��\F�ĂW��-a�`��6�h�t~�/"(>:�g<�Xc����G��4|�$�L�=�n1R��w����J1I��*e"�'f�i�@`Y���-�l�Al�����P��_�����D�@VFm���}2!?|��cXה[��w�<1�?���uc16ڌ�+���l|X[�7��s�|ger����}�%�Y蝭
�3���)yG{Cl[–/��
�ϐ�r�|���dt?����Vr�iɝH��-f��gI�r8k:��y��1|Hk�ޕ�
Be>�3�a:�عsX��.=�`��:Ӱ@��0�?�A�cO��e��m����I�d���{9���ISl+�y�&�z�*����K6)x�k�Z,N$R����%e ��i�������C4:$�����=�FA���Yĺ60i�y���L�["g��v�]��n0��n�-�ږBO��q2�D|��8�H�N3�Ć���|�X>3�yD3M�K�Yo�T)�[I��>4q�S<�;T���"XD���g2Z��/�Z�$Dc��l9��{���`�^�Gt�;P������
����s��+�(c���J����.�N҄�:h<}2J��ě���f��&.��2h;@[\��h���.��|׿sZ��g e�Tn���v'���>Ůg��ő�;
�z�=.)��8i��uR('�I�E�dZ�de���lb�Ɏ�
?
+yq���uB�V�u���	�[��i؏H^G�nY^��5I3ɫO^,�x����Ϸc+�C�ɋ7�;(�����Nyt�1�D�$`g��ֻ>�l���i�c�Z㸠��4��.�������݋
_���Jaz� 7��&w����P��r��=t$�*1�8L�W�8��(�TT%Z�;j:�Te��i�Vl�/�&}�>d�J�c1��t!4��<�@v�N[T#�9�X���F��I��q���9�	3��A�Φ1!32��,��a��cw�A�iVj
+_�eq0�A���-Tޫ(���XvɁ=�Z�q����W��gO�=S��W�<_�'��R\��F���c9;j���&�8�ۮvW���/��sF��cJ��|��L�i�*�Q`��X����
�+�Y�8p�_�4����5W*@����I��:�r;�P�)`�1��ajuZ�)����khX)���I���Pk�/T�5�>P��^@��IY���q�Ʋ~ u��W���#��i8:��ǀ���H?�j���8����:У^c�ob�S-�t"b�\"�����mff��Év�;)������xO�zw��#n�v�}t��0�^�x:�"�=��(',Yi�����p^�=m2��N�Ν
Tޛs	
(�O(�؜"���)S\��I孍��d�r�HJ�
�䃱8�`��
M�1u�͗F��LЬ�Ґ�c{_L��E%�u��巤��}j�5vʱ���6��uD�v�p��*�&}Y�`��X�:'�Svk'��(�T��
+�RJZ4:�zp2#�љZ�&�PM�\z����ܯ�X!��:��(�V����o}�y��-o"��碍�*����.j����~wy����%Wqg����l�C�'����@�����V�̾Z����t�림�#�mjU`$<ӆ7��C�
^��%����tTj,�n�/�כ��.h=oM�5����d�o��q�v4߁�C��(�bs�L�Ȕ
+,���a�2�6�\:���=�����WX���9�&PP�aPl�k�"3�>e�=���Mŝ:�\�r�Œ�/�/�N��|@§²��6`F�)j�
Vؕ�����ַ6���+J� LgeH���w��73nI굿���d����5JeC|�i�Jz���Yr|���k
+6��������a֎c;iC�q��,��h�z(�O�h�3�Dc�ga��$��WPE�?}=��=?��.S����㲳�Ȋ9�.���K�b��mV&F��� .�8�����h���������{��~�ĔiA�ʸ����Ag��I�W�fه��Vɴ
�ݣa�$�+*�`���4K6��
҄��jS}�d�6Ȳ�������'V���ڇ���]�4��W��ʽh|MԜ����
,�9�K
+J�(���c�gX|����iu@��r�:�l
+��Y4��z�q�h
�
+*�6���)
��Tx�"��s�	�>0w)�q�r"��.e_�Te1�=���hS�v���uvI��B�.������� S��� ��z*�Ynj{Ǟ�6�����W6�Hi�uR���4�<?�W�.��D�ۆ�xԓ��Fk�*�;'rc�e֬�}��#���d���F���֊˜���7��m� ��0�S�x�WP�oe�_
+�}��:��;@�Pi{���\8�ߜ��?ca�O�;A�C��B+���e[�@ɋ�<��㮲1��F��0E�<�LP5�P(���u4(���3n�63ToQ.������#�O3��-���H�?�vXB��%̈́���a�y�V�A/��9��Q}0�x��kW��\Q �i���>ܠ:e�v�����O��	]Nz���7ؑ�5"TQB���xG���ۗ3ѷ�!���͖d��C�����!�[W
U�0t��ɚ>��>a���0@�8�`�Lt]TP����澙$q�]�6�Q�|s�'����ޜ���f��R��-JS=G��d������l���E�2�s��ξ���O;��F�ӱ��k(m�R��Փ����XZ�����������^��p� E[�)��߸'�U��qƹ>��3R;̞��,�{u������;*سVl}���c�(��Q��^M6�ؖ������^ ���a|��9�y
+h����%�+sS|�:���evr������'O�UI�L��S�4���س�o��X��Jy�ǨS�/�5���)����e����/��TT����S�� ����6Rǐ���ӗ�Q�ޔ����x��>6-�|�ZRj��X�����Ɏ�h�[�8����M�D'U!���
+�$
�6f�o�ΰG�p#�`,;�r�S��~�I�}#ơϡ5f�C���X��M�����B/.����^�Q�c/��s]m���
��5S}�?�y�����큚�������?�?~�Z�{�c����V����8������������/���r�˻�<�[�bg�� �ك��엕��BK�|
+endstream
+endobj
+
+1447 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1300 0 R
+      /c1 1299 0 R
+    >>
+    /Font <<
+      /f0 1257 0 R
+      /f1 1269 0 R
+      /f2 1245 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 65
+  /Parent 1 0 R
+  /Contents 1448 0 R
+>>
+endobj
+
+1448 0 obj
+<<
+  /Length 4229
+  /Filter /FlateDecode
+>>
+stream
+x��˒5��_QC���Fo�ր�ـ���9x��]�١���Q���T*���{�;+�Je��%����yh>�|�4�?����F�?o^|�r���jd#�Մ(L��m�7�)雷��n����;����a��n'�����^6J5w��������/W����O�K��Ww�(+���U����9�����+szU���MCq@=�}�������/�?oOh?>�ks�������ݫ^�n1{�^�Z9U�_մ�e<�����)��� �{|=\7ڋ����	�tš���Qx�B��%�����Vn���s�=��HH��.��y�(Dڰ*��������F�5�
M�2T�����AX.�Q�G�
^�Z3�ͲA��X�Qe��6L#̍3�n:X.H�8�8`����-�п�Zi�	E��0M��c����?���0�Fp���9+�&���$Xj��%��q��H��qH���PL0e�--ɷ�����ÿ��w^��pA������RyLj�.�Ԟ@=���jH�ẹ	QX��d4F��F�_C6���I[�s��(NO�}�H��xJ�ؗ��0
�xd�Q�5u�2���VU�b��x㣐^k�O��̈Ej�L���A)/��(8Џ�a�n� TB�BYO�A��t��{>~UXf�nMJ	m�-�Ln�ځe<G�m�͍�B�;M8��ʠ�|$��
+��l�E�/a�dz:'�x�
�����A����z�I��4�Ώٚ����~����Op���#�e#��OA]�~r�ii0{
+�8}��	�}���>�N�[ue��6�@�$%`fJi�ʂ)��8����D���	H$BzC���P��<�vu	��m.�a�7nX:�;���
+�����e����_��%���}�[iL�k����Ձ_�jk�"(�A[o	aD�
+m��������WM���F�0��g�R\P�8�J�BcBKE(ji�b}H�]+��~uD30�ocJ�Μ���tQ��A��9�Ƚ�`@!�!��1ɰ0ʙ���[Yw(*���N����5�G4B����$)��2�	b�0溙��L�J�5����(�͊�| ��-3I�0<C	
j$�m g&X�X���0��N��Y�eI@������Y/ףU�n�p^ʙ�a�.�u����\)�4P�v}`��}Q^���"<����s��VƈE�ND��#t�\/�`����Ϩ)d:�e҃�{�p�4�,�f�\a햊R!l�^?�Oq(�jkTz��,�)*F2��K�`��	�T����%Z~�)ӭ��Pz��v�c�QOJ�^�6��1J�����ocnX�V����z%���sCiuHwq|����j�R��B�V�*�U喕��Gl>%�@�?�3N*�$V��"����f�
%���l���^	�t"��j=q �cw�Lv��թ1��r]�M9ή�(E4j��2+�ދжS���j}F�4,�
+Ee� ��ˤ�*@d��@j�Iqd�f��z]m�דP�f��+?��!��	�Ū9T�m]I���l��'�o7h��F�r6����!������چ�3�?t~b+4F%E�����VX��
+��u���{��@)���yZ뀒[G��op#��-�$�Ѭ�ƙS@M�v0���2c.�ǶA O�"��S���D���q�s�MmQ�"�"��m��(�)�
+��vSm���9ټ��9�˚H�������HY
+t�T]Wh0UL.����צH!�O�*��}%Q�0:���>@=H���F��d���Aɽ93����R���AE�t
+�e�셇�Y��N�Ko]*(�D���nC(�O5����޸���,�kY�-�0:l!H���7dv;%$ݨ�&�s�:���:ﵹ;z�ۆ�~IpnA4.�r��3j��µʛ�j70�B��P%���A[aatz�m恋���t��Yi�
�v<A�>q�*�yMEO��"�.�X�&5�:m����>M�b�������<{�5���
�4o%��l[��H�$@0�2w�RP�;�Ä�@.C�x�MI��H�)Z�n2.g���VJ�6�L���"Ɏ������2�:��	����q٨25�\��9<@�����Cg�U���p�~��aHLS����'�OmWpI��n7뜐�mQ��B��k��b�-�a�cabM���9�hUH�b��N	\E�����1�Y�@�6���X��=M��]�ܭ�#a�j��)4k��Qo�{m��fR���X�A�j��~C���aG��m��t�0u\�
���`o�l�8OU��F�$]2,����6�um�k"ߊ�F�mwn� wY����ĸ��>fؓ��0I��&4�
+>�Y.��WM�${lו�Z��l=�&���ɿ���?(���yE���0���c*TlK�Ֆ���'� �Qv^忯vDb�{"�9�Y�Y��Z]��am�6DU�"y	���z�a��'lHu�y���/{G�i��Zn�A�ۜ����".��RD��>:.G@�G��
R����=���Y31~&	T=�̤��h�|���>�m�	ƶY�z��;��ɖ�-ǘ��vacY	w��ٹA��AE0���`�Nm�(\�ʹ�U�7(�����L�.�}��HTd��W1
+�TU��k��.��:�*�Q�`@znV��ZS�ᡐs�9Y��<�z�rP��;�Ǥ8P�6n.��-W]���
��Z���>��5����ꂴa+�X��
+D�Y�6�j<��O�kk���ϲ�d$3�@01
+�� ��%!N����G���]�83s�X���cV�������@&��@p�������H8�P$k�{*�,�/[~0��S�M�эy�:;\2�7�z�g�i��l`�����ê�e�b���e�
��%��s�f{���t��i*�]�D�E>=�C3������R ��=g�2�
+��%�f�F�u3mH�"�Y=ԌR�+YsM{w�R�8�����/jc��^�M4������H��|�B6��ʵ�#Od�ϰ�y�/$�I@���k~ |��Fq5�9A'Z.�c��ug���}*�V�]�:$�'�)���;���5(�>e��c|T�h�dA���<�����5c��<����0t&�fF����\斂���=����w��D��������f��vm���n�l����-(�0.F窤+�&<D
+e"�\L'a�\�`�&0�<ϘU��bE!�d�$H`�������f3_pI`+�'�…
<�W�,����X�Nm���q��u�������m8�#���4e���.�zJR���g�)&�&�@���g�ѧ�;r�GϮ�k�`w�ZN�G�DM�59���
p�U�b'W��
�Ԓ��?R]��T��� �rm�Tj�A*��
�%���j�PVo�� ��r []��@d���_0L*<�W�_����_<lr�%|�=K��	~�SBP��J���o��r&*���
g_��afؓ&~��83M&u/�ًOd���nqB��Q��q�kQRj6��|&�m�J�~
��iV-��ʐ>�>G&�rP{,�rQ)�L�Tj�Sǿ"�V?wVf�I�>���\zEН��1Mj3���zrk�L���c�Ш�8zdry/����B[�\�rJx巹	丛m�-�ɺA��D�%6@z`�7���R���jfq1X�4 �pI$*�yK��P�I��"v������XJc�H)��8��件e�[7RƉ֙���*9b��lȃ�Գ�D����:�[��8�$�J/�_�i��0�~9�f���$��zIM�AG��D�����H�F������vKl�z�l����3-`���-��f��u��W�aU�o��*�_�����r}�(aI
+�����<
+f�.�-~��r�)���>@Nt�s�K�1Q1�D"�W�rq���i#`���d|�����R�y�����e��W/��7�@G�,��I���jB-
uP�?t�N:�I�m��HN
+JZɖ-Zs�`f�V�(ɨ�u���@�S�����H���j9��7�a`K��J�Ԝ��i��Htb�l��%�|b���L�	�'�"�S	�L����o�7o�~u8�y��w�l~�}��px���ݧ?����?�5�߼x��}tw|��7���������:i�B��]ZCFc��v�1���A

+endstream
+endobj
+
+1449 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1300 0 R
+      /c1 1299 0 R
+    >>
+    /Font <<
+      /f0 1257 0 R
+      /f1 1245 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 66
+  /Parent 1 0 R
+  /Contents 1450 0 R
+>>
+endobj
+
+1450 0 obj
+<<
+  /Length 2663
+  /Filter /FlateDecode
+>>
+stream
+x���r7�}��Rd�.A�1���ؒ�Jj����A�Yo�Rb3U��"�B7���D"��i�ݍ��_��˗c��.��b���`o�.�_+��L2�p텨�w�[)[}��W���T�}Y=To��`����Z0)�r=>��X~�~<�B�e�j�v_o�P�����M����ŭ�z�{��g�g킝5�k�j
�H޹�
+����Q�7�
R�%A�! �n�
+/$��z������v�n��+@�:��������md����N&�ma��{
���L��uH,�Sˏcȁ�{��)k��l��mt	d�[�P�]V��޾��α	ʬ	Ͻ�`��VCXh�ŕ�sg���jXXCkTb���U$�MY�G�j
3Oi�Ґ�
�H=˪�v��3�8��@�I�,�EqR�2��c/�4H`V�V��V-yS�9�:�U(���t�
�H)ƨ����04<���He�lÀ)�4��+	@o�����gvz�eAH�Q�z!�c^J�t�lF2+�.-Qg
+�
+�����(g[ʿ��{���1@:^7�Y�a!�Z�3=��2���L��[!�����#UR�_(�׫�ow���۰ӳA`)�K︪�i��o�l�ߝ>$��7�1��Z���|bϮ_\Z�J@��������%�7���ZsU3I�K�g�]���(ט��M�Ja�A@�4
/�P��S���Tt�9��tbh�ۜ�w����9?�;a�I 1�QI
+��3��5�8�mk�n�}���	��f�0���"�'�3E�ĩ�$#�g�#��1�t�y"y==I���oË#*UD�����Y���o��j�Q?#]���7�:�y#j+j���N��D�E)i
��Q�6{D�lB<����>�2P�$ba�-�� �d��L~ː�ɥG�*L��0�N��l�q�	��2�ZL��:^�%�l�K)���z�x�4\Y/M�zy�����B���@�i��wd���4�u��8���p4�>C��a��.RF�
+E��"��;���Fu|�G�(
͠��B%�Ⱥdb�u��;�D���DZ\NF� �1�9�M|ηEa�E7��
?J�Q��o��T�X�M��L�e�f��r�:����;m㇥t&KF��wc�Ŧ�
r?�^5/<�0L`̂�W���3+Ɂd��Z=-�"���*�r.'me�f6B�(��ױU^�|�
�K��l{�iō��̗*s����c�&��IT�/�	�P6[��E{���F�R¹��%U�mA9!�ld�f�c3Wn�W3�kTdUi)|r�����4U���p�ܟ�uxN���1��f*�E�	���99�Zp�k7���[�q\[��?�%�^3�;�_%�)���ҿlz]�M�r�+�Ɓ�|\��_$�Z���f�z����O�d
��s�
+"]��O�^��q�����T�{�?�λ�J���d��a�Q^����~�4��@͊���ޮ㘅�?o>��V�����[��Vr#���d�7��ު�H�u�V4�V�`7��{Win�`�*]5�c�k�mm�g���p��#�u^m�߶ϯ��z;3����m�@���\膍[v?�v�t�q��'у�6�J�0�X���[{W�Uu�TC��!���H�0��u����?��okd5|��L&�A�[�(t�t�Q|�t���0�6iq��[toT�]��w�:�甾U�Hq ����o�c ����U�l`%G���5֮)�@�Q��Lc9\Ru_�otc�ij��sj�ap��BJ��%�	�3�DԤ�f� .}���O�6��noL�Z�X�w�* ���B�q�;��9X���:mq�th��c�
��a3��tQ-[f��1��da+ݑ��NTa �O��8�W���dgŒQ��b��x�Ϟ�n��¦� Q<�*7pD�o����kS�hGb���Ũ�WW�B��s�L��m��+��d�v��a%6�4:@�G,и�����@�B`�$��'�fk�Fy.k?+�M	�֞/h�v�N����NB�8�[}i�ӽ{~�$O�:�TШ��ȉ7i$��.�ů�`�j84
/Ѫ�{�aKUU�Em^HGl$O�D�}#%7��1Q��'�nҿoM�5H8�����h�D�4Z��9���fh�$zMyJ���nƜze'A��N���D�5�����ّ��f<�7�"�£C��C#��qޕ�q��Ċ�'{�M�Xa�@C���@��a?�@�ҍ��9e1C&�s��fC�r�����7�>�y�L	�k%}��}��'?wR[����*2<�Ud�B�
+�Q�,?
+w' G�W���N��dbJ�ji_s��'�i�U��<�ѩ��f��@N�N<~���zhx�'�����ݓ}��B������;ۆK�8=���2dBjxA`f�m���i�O-�"�X$ɶ��!�Zދ�4ۑ�az��kȀ�LAOc��E�%��ȠX&�Q���d���}�$=V�$2��3�T'sS7�JV�[��\�|չde��bn�t�!%���ȶ|��ί�|š�q�!�<	r@�{e�� )E�e��(�%G�)���n������<n6��~��z������g�׏����۟��������n���z�@��бbFq�h�r�Tn�E�$�wc���
+endstream
+endobj
+
+1451 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 67
+  /Parent 1 0 R
+  /Contents 1452 0 R
+>>
+endobj
+
+1452 0 obj
+<<
+  /Length 4320
+  /Filter /FlateDecode
+>>
+stream
+x��]���~�~.�:�52���)��\��*��#�™��*��S3��Q�Z���"Tq�ӴZ�V��iϾ���Uw��Iם�|��ӎ�<|�=~���_'�����y��}�gRx۽����
���r™3�k!;Τ����/o�N��������%���/g����O'��~�92�xŕ�|ߝq�����g/���a�������C��ю5)�O?��R�g.��������b�$vm��'?q����{�ʂ�Hz��@a�
��KS�OB��B�DH�zg�i�����g2/�Ae���U�����ii��
嘦~�*?����hf�`���?���i�O^�jx��N��6T;��[�TM�z�׋�W?t����)���-��W��Oe+��-,���%�(�"�ުk��4JK���eT"3�0#fӝ������XhÔ�B�p#�2�
LH��(雄5	�j�1�����Q��Ig�\"�� ^q��l:��V��t��j��~(j�G�cs|:�x��(��
o�[�w�{�<�U�߻x����t���Ck�U�y�/
�ۊ� 9(D9��e^?�_�w�G0~����4����I�q<v����JG��6���	75����Y�\�Dgwb'�M�_��J�5h���.��������o���Lr'�	�2i�8���1�{���2i܄�(�����-j���#*4
+MÒW��Ȅ�-I��dE��r����Cj��u"Ÿ����p��M-|� �:r�9ʓ	,i��뀋�hĔc3�z�p�����eI[�\��l��<��Wh�c��װ(5)��x�����j�$"b�D�Y#��dޘg�f�/ ��ŢH�w~l'e�C�8���:�v�@Zf���;�@Ȳc�0����Puנg��T��O�7���~�E,k�9��߶�R�����=>�F6�Hu�y��@h-�l��T'S“�0˕]6��OrW
+��<���L��Lp�/]�v&�����I
0���(�Q��軔�z��R�S�d�3���:r���xE�pX2�+$����^�ᾈ�%��!~k�X:<J�蝸�� ���~��<�N*���9~+sh�^~��p�lJ���^��O�&� �iu����!�)%����+�e����#D���n�9�D�VPrH��l�1�a��v5���5C,�S�G#�S �q�����V\�_*~"�dܛ6�%_EK�0��w-���Ǜ��
m\�n�q��u�Ű�̴�E'E���9���d��F.�
+�p�ӽ���8���_N�xkS,	B�HPQ(7��ڹqR?��EB�k{J��P �YSDuv���LAm:c�����s[�F�}�JPD��Ѥ(�B��F�:�m�;d��2��1:�c����M0#��8�WSD�������~��$6Z���u�F ����|��L)߻9sP�=E�EE��NJ>j�3�%U�����y)t�7W�����A�x�N�NI�a��������HNI`+~Q-`�8�I�C(�{	U�m��[F��i/$B�.���XWx/�U40�3n^Yf_T�>S�-%�9`?_���Zi�;� ��^�a�gJ���1�y?Ĺ�d�>qf��LH�(CI�h��7!�e����5�Sv'B��,���3��|�򏮄��
+o��A���-�Š�vu��ƞ��°1VKWR�%���[��"��/������Ҿh�*�����"��_cN���ż�!WH�
� ��N���^����X���b��:<���zAx�.a)���<���C�bɉv=�%�rŒ��&3y����|�QC}��N��<�z%T��Pg$�Z�~���0�"]�	?WDm�0���J�DQ�~����¤LP�
趡aR�4�f���S�r���b
��9*�]]�}\���ja�e��r��|�`�Xy��f��bò.���
�&%s�
+�x�M
Zӫ�:�Sj��=���t~Z-��??���Ns�>YO�83�安���^��)�fR�Fy�1}��$(Yp4f0����B
+^ٙ�˪���.׃'�D�)dN�I�zgs&���w��
�Y��(GW�k���ײg��$G���G��VLq�ռ*9�'���M�
�=`V��ݪ��k~�:�
�h��Pu�����L�7�
�����V��
+&�����9��[_w�P ��gՆP�1Fh�{�U���>z����?���{�����s�=}:�Y����M��߮h��b��b�w�.Z)!�JM�EnI���e孪o�]�+&�
+�:v}DDQ��L9k\����R9���"���d~��^�#����G��1�I�̵f�H�#����B�&�/E��״@ɤv�j⧌�e
+_����ϭv���t�gFi#��B���AbI�:�蟦���
+��vѼQ!Zm��F*�R5FbE��u��L�$�YG��Ve.3��4NT���{OaQ�l=���S�1e�"�H+7g�N���'��j,�
�[~�mq�u�Ѻ(+�b%��
+�IʹS2�S��p��"@�l'�,�}"
+��\�I�QϬ'��S�X)Z4
��J[�o�'����ȗ7eOge�����	8�<M���t{xt�c����4�W������p%h�g�k�,���)�A�v�wj��~�,�е0x:��D�FU��
�(]��K�Z�y��`�tL�a�>֍�����'��&�<977ё(�;�ū�'o�p�q%�k�b_��S̎�-�P���h��n���<
+����6�j-��uJ�����'�nߝ����l�a���O�.�J���/i�~MR�ؼe�5a�PZU��H%sZ7��a�oT�WJ� ¿�[�p)��f���/��k�Q%�T�\�%���S�K�z�e��sB��Yͼ���(�m*O"2R�Mp(+�DR�ȥ�����k��5���LnTT��?f	Q�{�kI�"���rS�N)�O�U�N+U�l	J&���R&ՎFM�x��YÅY�
.��l=V���˫L.�G�Lf=��,��������$��o� ��a���HZ:�<���HU�
^���n��
�a��I�\�ZbG���|�I�?)��ӹ���V�I��l�
s�s�d3@�}t��J���:c\<���.���ECM��
i%��u6���e��(�<g�[cLa��mV
��h5���G��l�$�6�P�G��d�T-�e��G�W~4Ū��a�d��G�v�&��0�H��PŸ���j��X� �cZ���L/az���[�
+4��R[��3��ض=Hi��E��G���T]��7ܕ�2a��o�S�i��v�9��h�Ȑ��I.z�4����V��u�7��&~ʰj�y%wP��V���2�Hp��8{�+f��K��U�-�١�L�/{_�zK�(pŸ�Z4��]`���x�L�z��-,^	��BwU��	�BoKd���H"|J���Mt����.1/���uȟ��T*��"<8�f�^!�:
+}&�n��H������-Vwo�p~���
%�hb4Y?##\s5����@ah��I�YnAm�p��2n����^5R5{�u�j�S��=%l���Va׏փ%g�X�[���B�ֲʉ��U%��+Q7-t�jv��dp�R��w�Z],�n49�}	�-�]���b��/�������u*a���};��n�5����!���o�ĆWjF����(�CI���#�d����w����-�H���g�����OS�wv���,Y��$��}�
+O���������)������J�b���򽽮/@1���w�|`a�XIS-y�T(�äW~Ѻ%���7��4���+����×x��SZ��h6���pq�z�)E�uN���s⊹��(���~�����?��J�qiU����^?�f=��NzE���Qqh�޶ǀ���kJS�,Y-`��~�ï���ч�7h|�wٛ6����gv��|�=�z:?yl2ቶ��9:s��To��B:א�w�X��O�!�J�*�a�cg��K��;�?�d(��K�n��O��w����N�xA�
w���qXA�y&ʖ����C�m�w��k�_�a&NF��9i��\��o��\{��ҫ�u���֓�NY2|�T�"�2����_qH�'�p���l��[H'��†GP#�0ID��T��pS�B>m�+��.Qʤ��BN}\�h�`re�k�X1�<�s�u��RJ��r(���-+yl�������o�!�ݾ~����;{�n�}����o���b����vg�߽۾}?��||���~�z�����s�<7Fvz��䢽r�5��
/��@
+endstream
+endobj
+
+1453 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1300 0 R
+      /c1 1299 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 68
+  /Parent 1 0 R
+  /Contents 1454 0 R
+>>
+endobj
+
+1454 0 obj
+<<
+  /Length 4575
+  /Filter /FlateDecode
+>>
+stream
+x��]Y��~�_�:��H^.o6m�dm|�Ad/����q�y$�>�f7�X<f��HxZ;ãX$�����~}s�=xp�u��<����z������VG�uO?v��#��w'�s�)�y�9��v����/G�oy���#����������?�^�N���r�=<�9���k�����5�j|��)���U�d�)��K1<m�Xo�Rc-%V?v�_=??zy���gG�xH��P�ia�!2���f՝��@��1U(ϸ�ɑE���ʎt
���j�n�dw1CC�𫒀ݡM����*�7�c/�����e���t)�?R��:p�d����dn�g��~;��p.36?܎��⦾�����Ugl��mZ?ó��4����X|ANv.{?n�5�T�hl�������>ڧ�6��(7tϜ�BogPlf��.����3S�L84���2f/���^�Uw�=Vh�\)9O�
4c�6Rb�6�-�x��m�ڸ_��}g�x�A����;������!�D;j�"�Ez��n�x՝Dt�[ze���O�?AUIN��(<��X;�Nq*��V@�Ć٘[ha���X
+
��Ď��S�'J3ι]F���7thb׏ܼl�
u	j�a�PLm��0l�R[���t"`�E�Hw$��̪����:�LnF9ev���ut2�M`$�=p?l>����d�u����-��ܮ:Ù�އ�}�ڛ��\�&w�/X|g��	P�3g,w��}
$�|TP�l��`��
+8�}�"A}O��aM�x	���"D�ذ�gFX����6΍Y��˟4}ϴ�|����Dۃe0Aō���4.�8$�='z�9�������`g���!u������>��J<�������lp�/3y��~�qw�t����E2"�&EDQ��9�Y|�p�`�h�!�	�v�׉�����F� ��J���V1����7�� ػ�wb��4+�(��P�yn�H?�	]�z����)C����kо݂�:�|��I�.TR9&y#Wv�_[(�����d{��?ў}�,lq��~��,�ڞ�TA����DpǬ���5Х_�v'a�uь9BG��C�,А��F��;yP����=N��mI�Z<	s�b2��$�Qq���j�F��#�����O��I�C+�ч��$ef��1:��0iCD���PM���DQ��ҞmIQЌ��
򰪹��"$�y�?,s:r, �� ���2��x�Oj1��W�7!
+�y���i�d��Erόh����ϪQY
+ɔ�ڵ�S1�������i�7]��E!c*E�>�uP��K��<����xXh���E1�)�B��P
+�2�Y�/�!WW�
+�*Y[���Hje��<d��s�9ä����~�Y�EKY�`�;Ƶ	��=����O����U�KU�V/��n0��ʪ�<,���~B	�C��*L��DZ��V誨D���D4+5H`2jptE0���;�ۢ�M��U,���/N�4��y"2Vs����f+T��׶���ѣÅ1	���):�똱�R�8k{��
+�@W)�#�
��YN�[
�Aj�f�n�e��V���
+7v�!��Ѱ���x&�Y�RKJK#�dVzU�3��;���$J����� �l�i���3�՟S�h���8������8�;�[��;��7ߏ0?c_hWy�S��^2�X`������cC�We��F5�|�����d�׋@2���y/}/(;X쌷��㼭�D��J�m��"�TK=O�l�[&t��  ��:�
�eŚ�=:��|�P������y�&a	M
+m[x��mv�<hQ GӋ�2��
+�G�N>�.-%�AR�h��H��Ys�.����&���a�#��B�p�v
���D#�
+�,��g������^h��J4C<�-3����Ū×��O��a��&e�t����mՏ[���T4B9����i<ul7�7��&��2��*��n��Bd���B������B�ȟ+-{́C�NQ���6w��Ɵ@4G��Ov3�(U�0Z	f�Զi��ar`r��9�|��f�X��4�ɇ�+(U/j���t�d�����-��V;eJ�xCP�C���9�Gz��4��Q�rIo
+�b��.�ⷦ���9(z^hk��x�ϱ����x�A�g^���w��cG$�I0�
@�aą��Z�
��1	ԭ#)C:���
�:|4���p,��e�f�ȿ �C�iJ�D��d�w��/[���'����EѭB���V.d��WIl�C��Sdޠա�š	τthh�n��;��	ޗ��>���f�N)��LK�GP�JǭOhr�L��A��5D�� ��5��@ɫlm�3�1�ڤ��V[t���%�zA�퇡WN�fן愲V���%�d���
dMբ��s!(��EX]{\r�р�w������
�&�]�AK��i���ǕH)�T�I�U��(�:�i�=�p�L���������e��c�G�>}��k�;���E���y�9 ����4
+ޢo�1��ɃY]\T���J,s0^匣��7��^L3:.���+֯!T����}f��E����S*�����`�iVMZ�H�G��ӌ+0BX5SA�2�9q�_g�Y��
+��l�e�Dn���/:l)�Y�u�t��K�m0|�I}�A��������9�#SA��^���^��;!
�{�� J����󳠉���q�������^2�qe�m��&�|ϙ�>\��h/A$�ᦪ�	�^0� ̤I
x���,O�)�.�D4	��Gㆨ���kG�]M���a�pe�f��`�B�h�7�+a��e�R�M��oR�k�/J1��
C�lչ$*(q��ʺ�S����ֱi���
�3nCS�Gu�(��4�	�'i�7y�9��#c�̢�/������'c�H����"�=��bjc�T��(ܙ�⒋%�B
_'�ъx@+0"-W��Q�k�Ś(Ma����t�f���vAۍ�ަ���M�TW�
v0k4��^)u0���I�WK�Ut�Q\�^,�lM�2j���Z{19��6�Ӗ��.yM�����믁�hT�!���ނ�f9N�D��t��0?6�
+��|?�/���@·��x�8 �~��P*0i{&�V�iF�\�8\> H�z�/���DeAA[n
+�L)�uu�E_�z��3*�:�wD��+gI�/v�)ז
+�11B3i���7�����7�<��u������O>�n��}�
���Y��b����)�Z��g6IF���8\�t������TCY�qLs�8����.U(- �y�r�|p��<�U�P�2R��]d�+�o��l�6��—ukdž��� .�
H#�qb�U��'l �)����P����!��S�@V�J#��T�#�{��fp�U�&�=�����:�Jw'TSв~�4ԕ���&5��Ni�v��!Bpӏk�W��[��R?$��2CZ�^ɥ" y]��oԃf'�y���պ������mH5����>4Wow�pE�/�e"4�=Ӻw�2�",�j�e�q�pRs�4�6/uLt�HϨdP�T,�N���m��q����_�j(�Q���y΢Gw�_���R���,����V�#�P[���P"{e�o�f-��bPͿe�Ō�"�Zc{�L�SR0F�{�Ɉ#?@DM��$��]#�?S2�T�v��|l=�����V��j΀_�k�!Voz��ǯ�(�s/ZС��a{P"\&5c�ҶXHb��|�DG)��U<��E2�2�)�Yk��\��(�ڐ@�<���`ҧQ/5w%z��ˏ$ܰ�:�h��t�j���W~'X��2��tvə+�z4&�&^E�3���鼦�c�&�Rl>`b�=#�m��L�Lp^�~֒�{Ӄh��j�{&%� t��#D�����F`r��l �;`����I*δ��6��Ь����S*ͼ񼍞2;�@��.ˮ&�S����ў+�[�N�F���YϨ��L<�qLHˌ�NF3y{��gk)Ό�N�W�$����d��Gn�d��m�I&OZ8g�a�"��X��:�-خ�UJM�I�륎%G)`�@>�4�h���b��օ ���F�_%���4�v��ʣ�[�(�T�	���c�t.I��L�����c,��mf�)�'�e����ײr��}���!���`���$Hx��!:�dK�m�
���TY�掹�� �p:^a�"���z��
+��Ӟ�[����^���,��Dź �j{D�kW�y�z��Y�]1�Fn�=��;�\h��"U_�1�s/l�PŴ���1bH��|9�3�|0����n��S8d!�/��b��?��5�
�HqDR�m�TÃ�=rG�E6�v�#*�.XJI��aJ6m�]2��]��_��A�cҬV=��{~I�8َO���̽����׌{~���B�I2�|X�|���:\���7o���o��O߯���q����/����]wz����݇����y���Wo�?���\��f�#;-�J�Cܫ�XH�
����
+endstream
+endobj
+
+1455 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1257 0 R
+      /f1 1245 0 R
+      /f2 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 69
+  /Parent 1 0 R
+  /Contents 1456 0 R
+>>
+endobj
+
+1456 0 obj
+<<
+  /Length 4521
+  /Filter /FlateDecode
+>>
+stream
+x��]mo���_��eɲ*����5�c��c�A��@��b%)9U.@�����p�r���$�r��>3�w��on�g�������/>o�ދ������gO4��ͩh\˔��7�jf����u��7׿�q����߯o�^_����n��7-3���,��u�����.9痗�O��q#}�v2�yɹ~����o���7�Ssܜ
+�sֵzxy|8�"�,8�����W�:R�~��!��h��\���ϳ����re�:���.isͺ�B�B\�XM�t���§�ᣦ<jx�BG��)=~�^(�4���E�n6�A;���G�f��L�V�z}��2�ކ�k������r�����{o߿�;�Dd
�n�k�U;1 �,{
+P:)iK=��__}���y����lwd˸�v�����h�H��7JN�M��&�q�~�X��n�Kƕ�^�_H�R��!-S�k]���J0�/��i�.)V(Τ��U5Gf5{44j=�wО
���䊇�i=�g�sfZ�E�ϴK��qc8k۶uh-��Q�������ٽ
��Qӱ�?Wи%�
+z7ָ�IXZ	.l3y��S����Jk��P˦�h)�f�2a��C/���p�~���� �3�/��LN�	)7]��)�B'�F�f��PLz�dx�I
w�yJ��
ZS���w�λ\�l�ѭ��Nu
+����*δmy���<� �?��>�7
�PU��Z)�ׇ���������K����z
��͚wR��犍&]���Sd�Q�{7	����h<Ʊ�S�0?��9��m�Sp޶d`���X9�L��f'E�&��M��y�w��~��bm��Kh�U	?�;��p��N������V��#څ-����m5f�V1n��[c��g�֥F�
z��q��������g�H
��y��B9�if�6�JK
�E��3�5��=y�ߙ#7a�'W�����Ҫ��16��d���|s�Hpգ$��wc��s����v��&���pw�q�ۺ��	p7-^�Q�/C5!�:��"Ў��
��I�q�
1+�4����Z['T־iϔ�b�T���#��\����9SZ����2��M�n����c�d�#C������sa��5��[1��$���x�,k6W�ԇ�jV���`S�5��u�6/���T�U(�a�e�=�VtdHԴ�-�F�^�tJFAg��R�xg�.���'�{S1h�tGઠ��(�A����Ɖ�	pK.z��O�3U����ޛzp+$�z��m�m�!�0�@ҟ�n|��W׫?>����Wͣө��ڵb\++�[������Ɔ
�G
v+ǔ�v���=�D��&L��RHB4<�������A���L헜����}���;%��W1�g����p>@�����8E�Ł��Ic
���7^��f�^�ߪM�+�S-���E�}0�wY?�k�B7E4�~�)�0Z?9S�T#�
+,�
+�a�lU�|^�8�Ԏ6g>M�w~Q��W��T��y��
A�S���_6	u��f��a�첀�BݭaR� ���E�pr_&S��(O�@�j(&ڂ������}�T��"0���J��"��|��y`SS9�ʴH��z�b0A'��� ��@G��C.�`�8�!�ʾ��J��a�c��-Fz��>(��Y�3�Q��U����`����Rf��S��ʔ���7���i�	���Z�(IV	�LK�'4��!�	Ԯn��m���E�DG[�L���(a�c�[z��{��ѷ�(�H��g�0�H��c�l��~���:e�
i5�������V����Ie���13N�[ˢvԨ�����g!�1U#�-ks���4���E���P��jxa�(֑����$�1�wb�l�R-�j�����dz����u����e�r+�^��n���\��(�8����0.�n89V�ӯ\��(� o!
22�`�H��DF`QX�%�j0�I/�L��eֲ�ǃU���kA�w���YLT*�9��O��>�h�U�	O�jh8%8F�[(/��ն��d,�d�8�^�jNz���wWt.���*�(�*<�z���-3ʍj\�R�4��K: ǡ[���N2]�b�7�]X�tQe��r�j�]U{j��?�'�hLt����"���-2�����MȗHWH�Y�G��IQ.�`��A�A��2sE�̕>mĢWB�N��;TȚ���#��]e�"Zi���#:w�����(���'���% ���`��!`����o"���臙
��� ����\wFdo[D8g1+(�v9I"�IU��8�h�崖�:8�1�S<9���8�I�_s|c#���
+�7g0��D��\-yGV�2�
Q�I�(��y�az	@���i�h�UX�zƇj[��:�&t�O)��S|01@g#:�p,����㦝�]��Q�~�"��!��1oE�VC�b�_
+��ԕ�)��a��o�)2i�)��b��9�UQ���1x!���y)C7��SV2�$w��>��.����0��賗�v�|���
+@pRc6���hm�A��R%���8+vb�ׄ�(�a	j���JRT�Maq��q�p�"
�jL�!�d�����MG���
+�B��#�9�*�ɳ�� ��.�QᣝN��+��A�\U#���GDgYGʶ�+7�/�Z�z:M��a��6�,�Au�nP��+W?m*EФe6Y�ͦ.����9�&�0�
+}��1�W�
+$�Go��<������r>^�z�.s�]ۿ
L���ClkC�O)�ɺ�F�tB{湬����C�^�0�ik:Xў<�~ꉣ�b�wwD�I[6'�7#xw��B�Ό���z:����}9�0@h|���8�^���=y�)N��"�����Al���rF�b�����A%�����!�W\1.��ͮI�%�a�YE���gS79�	������Sꑤ��0U�ҩ5���K|�Z��ve�,��$�<��S��<��Q�B���!L9#�n�\
+�y���裚zsZ#<��I�fʥ�Lz��Á��Ҭ�������
+��	>h,��=��`�n��S*<��Ծ��՝[:��<��O4\HM�;�-�R�vC��!����eE+�����r��*�9wmU{���0��k�0B��]���4�6l
+�I���3�)e�Z�u�|��7˻��/�xMI*~��J�	�Lo(��D��X��9s�����1p�S�?�T�gB2��u�p%D�CCYm�e�iy�
Y��Yڃ4�9�L���G�5�0�;Cl�EiÅm��M�J�����iA뺳�z>'G�s�̚��
b	o�y�gxM�y�QM��Qa8�B�
�b]UA��
�G5�%-���/��?4��!���Ϭ�2zL��,�V��A��8șX&4&�%#`�F@J&wx�s2I�5��4�?'C5������kq�l=�炙E�Ucį���V��fաf�t^/k;���2���9�jS՞M.�#��
X��{螺2��P�}�
+�Mm�6��+��6	��O
r\$ �d,��YCAIe����Q�t�
ne��t�^��u��t��t=T�dtMsM"x��7���F:�<�HBIKT�nܓ*hk�i�-�bG��(h�;�!��l���D/����R�ƫ����!�0W!��
�mz�\�=����	lQ/y�*���<�Y5=34~P�����y���鐡9�b�1/ҏ�S�â��2;D������͌���g�*�<0I���'��VⳚ#>5��ŗy�ـ�l:�����O��\B��
x|��p[��4���[�/yG��f.Ŝ���>��h͜��U�c)ط��ўi�}GM{j���
�Y+ӯ��nY:@9�7lqcu������8�Y[*� �f��-[��=@;o�`V��Yȴ��
й�/���%7����
K�vG�\�ٛ":_��V�d�"�h�3��nxi�������f�+�
��W7
PC����I
�e�8o��"��E{���G�>T�����	��(�[5�t,��&�z]�d�&�يV�
+�����U�f"q����眿����7iO�=���ʹ;K_R0o�n�+q��KKB�@�����bBgl��d�j���α�OB�;\�dqJ$��5�n
�6�N�w�W_i�Ud"^�q���|!+V���!�\O'��p-��*'b���[�Ty&�S�wzS�K=x��
+<@�'Y��Z��Q!H)-k[�d�z�z<�HW�J0.�1u
ʻ<����Ѡ��dx�2���f[m��F 0=,���a��
�df�;3Q}��B_�<P[�R�J˖�t�nD�R�I.��л�{B_����֕����
�����o��yӅǸ�v2�o�m5��Ϋ��/7�W#�Z�>\�������ןV�O���}��W�����9;��i������?>�����/�n)<�5k�1�ђ�V)���^�(�����
+endstream
+endobj
+
+1457 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1257 0 R
+      /f1 1245 0 R
+      /f2 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 70
+  /Parent 1 0 R
+  /Contents 1458 0 R
+>>
+endobj
+
+1458 0 obj
+<<
+  /Length 4668
+  /Filter /FlateDecode
+>>
+stream
+x��]���~��>�������'e0��(�>�GH�31KU�ߧfv4��Z��Y�U�yw5-��j����{�������������'/�v�������'G�>b�hw�:�0���hI�PLw�9:{M�׿Q";��o�����������Q���L��s����/(���+r���MwJ�6�3c��͍�������S|��H�V��x�^��y�W���񣡷���O���*��j�.G�0�*X�ճ��?��\n̓������zle�d��,�y��d�4x�xz��
�~A���lnx.٭�bU�C��jK;M��Z�=_F��K���J�b�Y4~ֶ*�}4�xsų	��|�6?v�_=;?��ѳ��������t�8��A��Ưy^�%,BY~H�_�I.��[�>�����v�s��f���pa�L�6Ke����������X~A|���ǁǩ���F:`�č�q�NIf�uMVyQV�!JK-��r=�g;��P�Y�x^~���(�jcȘ�Dr�!�ɸ8.5���]�d���Q._���t��'V�����õ�Gaa��8+uHGya��s����럺�o~۔�F;C��}����ey�~�m�e\��d;�A(�B�f�q��(Lt#�R&Z�N�CO�l?:���@h�
z�m�7���=zxϧ���eZ�AM�O���F
pt�q]l0�ئ��	f��z
+�`�H�Јt�̳�%z񼦝2�/��G�N;E�]�'K���#`���M�Y�*®���2��_�?��S3�ѳ�����)���a��MۄxqB���Ӌ�����t�M��A=x�2���yU]YdƜi�5�f�z#k�{�1���Tr�A��P5_Pαg����^�j����jC83��[�p�/#��^ŭ��R-�E@K��"�cw�X�ƨ�˴�y3T�3\1Z��B���<M�)##�|5�a?p>b�{l�憎���8�i��\�h��Ѧ��MA�p�1�2��j��ؐ;*����^�u�M�7��Xs�)&��̉VM�}�z������϶������>z��wRS��l���-���Њil�Q�U`p�#3a���h�0q!�hda.��ԋ��]pu��9]b�dG�b�Iƻ��D��#��\��
+}��DIC�B��P�,(�Ա��7a���1��29D}�c؛w���>npr-C<ET0�㭪��$C��ː/imQ�Vg�gy��-d�A7��ɉ�!&�[�)����>�%��Ni5��$@�:X��@�2�[i�{;�ė�,��CU��D[�ZxS\tӉ���5����h����������{��eO���X�����������3ر>c��>�x����\`�O�H�!G�1�����H:����(�dU~0����P�����m��ŦS��x�"�jn�ߤ�U�������.ǜ[�^�y�����Ũ-1����e&cise�,.�r�_�v���6��S�Ζ�2c�;�ON���B�I��,�=�UDj��N1�(s�R�<�J8r�����@6� ������+�n�Ҁ�8q:,
!i���7��-#)�CD�8�zz��M���UX��\���B��d�5�
��l-@"E<���-n~7��:�Q�󓔡�p���Y؍]��AV�2�΁�k���u��N.܏��f0R]M��_P�g����_�Qhc
+��Ž3uOOz�j���CT֎�?����gp�r��yM��&����D�b�3�i<	5�O{�Rœ1|Vޑ�������n�2��A�Bڮ!�]�C{������r&B�ip�M�(q�9�l1���j�S0nȫ[sKHň�R��(�2D8���y�
+m��
+{�~.
+��TQ
+I8�+x�M
+�J�W;��ktNM
����eH0ȇi>�Q|�������,)�F�g���͓�
�YH�eqF��ln���⧧:BCI�At"������,�x:R�!��!��RC&�	�$a��'X&{!�L��$�aP6:
�tA��"�˭bY��;W�4���g؎$LQ�PI��D�,�˫a�^�8i��!pY}�!�
S��
_}��A�N/,�
G�m�"�
+������6�`�
�G�����P�\&�j�A�>�9�-�M34�<rLk��j�B[�TP��z���<��HF�d���UCG����XJ��Ֆ4Д6����gi��ឺ�<���^�GAk[N*��$�&a+����'�.�d#�8�7�PN�3��7�b��pX~,��R[…\����N�( I���&o0A���]�
+�9?�B�sr���z!i�ye�(F��0�
F��ٓ1�r lz����k��O>˪���E��ZU��h|ޱz��I5p�4�”�1�l�ڢ�i%��0�U�&[�{|,5e��e��Mg�E��QL}13;�����	����1���P�Ձ����%��(mh�n�r�,)���ڤq��1�줰�(s�+�3��e�!rJ
+pg����������t^C��h�YB���n'�~��mʄ��%\3=��a#i�]�AαX:��2@4�c���_����y�,�L9��r�LӚ�`S�����~�f� M��9Eh`yS����;6���S���qQ�x�ba��.zb�����gʬ�횁r¥-JaF�R:Ū�=�������w>B��W8�
+
ɖ��a�<���0x<ꪺ�U�-q���X
���e��')�%��<:��=�zJ�b
a
Q���}K�>�I�9A�Ԃ�Z}A�
����N��e��e�J~ނ�YJ�>�H]��"4:!�\�R�Lj%У�,��D�o��G;(E��9�?/�<m8�����&�-��8c���C��x�Q�|�{���@�c�a�!.B�HR1(�{RADy{p/�^v-�N�Ԃ�#Ŏ\�z\UG���R̮ê�W]T�0��y�V���"`f�8i��1?�L=���t���\����M;������-��-��	�
u4I�!��͝��Ny��)%&�m#��l!��1�)�q����	�bg#f@H|��#� RI�V�1�+x�)뷟�N@e���"��`5y_���l��p�W0�"�`SeSEC͡�/Ԁ���8k�b-��*�8���Aram�����ry��
U�D
�I&��O��XˋD��w0]��d��D����%��0<,{�3*�5��ѽȺ
���� 
,t[��@���On�(0[��R�s
+��-^5���KzQt�d����*V<Tc�a���ȇ*GRϦ]Pt7*I���=�ɋŅ"�v�!r�.V����l8���z�դ/e�7F�!(<8rf.���͛{!�3m��V�Y��oY9A�Q�	o���G��ZιDg�P�s�p�	uV��k�jiF��}����O'gB�
+o���Q���'�K�*ӎ�8Z��P��*Ew7�p��GK
����S���{����@�%VQ?|w8Q\�SU�O_Д[Eu��pU7�:�+�A<�FU��@1�h֧q�Ѵf�mh���U�&���STP�m8��yG\)��e��V��PA}Uh���(i�,
���Q4�u���h\��_�8��D{�H�.�{�}�(Y<�&���h�M싉�.�W
�i��¸�a�i��ƟN@m���(�-f��b�.)1έ��F��#����%�:�H����b���1����E4��_fTa�h|�P�c	/�qԏ.ݞ��9%T	������d>g0��
[4�Ԧ$QR�т��Kc�&�V�&J��o�I"����  ����
�/G;�
�X���P���J77]t�F���A�jM�`L�'
+�_FrF4R5I]I,�Ae�S&%�jN�����k���x&����;��e��Tǎ�s��u�{�R����8�P��5�`|g��fGs�L�0�R�0W[P�$'J*7��?X���kTm�6�~-�ֈE�2?!�ݜK��;���h��jI¨�R�âݟ�'XՂ2�l�'���?ˋ��,ΠLY����\����j��X�I,��&�72�aY���_|��b����ƹk;�&6��/�-����4��B	H⎎�8="|r��i�9Q��rd�"�T~S`9X���ͻUIB�t�r�8��B��#G���f�h4�ʰ������0a�v°�Y6aF�	�ƫ�=!cm3$��BY�ԇY�z��)���i�ʈ
�C&J]�*WoL�9��2B���^W�*)���'�+
���-�X����^V�E<�\Oe�j7�S1��k{U!�\!y��R��s�X�G����MHj�%Q����A0~����7CT�8�r�Q\0.��� Y�_�9u�'�p���~"'�:{��))�������6K����馁w�������f���"5
tc���Jq�!tĔ���}4��U6ͬ\���D�T���^�7*f�uc�	?�Zb�9LV����/Hi�q%5I��2P1�
+��*��<QY9��gf�j/
U���M2)(j���l�pj�o@��I
ڒT�
���ѻ��W�^o���������{�������_~�?�����}ӝ=�v��]�����ϯ~������טI%qT)�on�����������
+endstream
+endobj
+
+1459 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1300 0 R
+      /c1 1299 0 R
+    >>
+    /Font <<
+      /f0 1245 0 R
+      /f1 1257 0 R
+      /f2 1251 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 71
+  /Parent 1 0 R
+  /Contents 1460 0 R
+>>
+endobj
+
+1460 0 obj
+<<
+  /Length 4848
+  /Filter /FlateDecode
+>>
+stream
+x��]y�[��?���Y����H7��4i��
�m�u�m�f�:
+�~��I�{����- �������7u���?>����}�|�㿾zu����͏��'O�Ϟ?;�����Gbp�)ǹ�UL;g�������|x��	~|}{���	.ߝ\��A���f~z|�����+�����z5(9l���
+|�b���w���'���WnG@��0\~y����7'/�~vr��(�i	�7���y�¼v��f5\��2��Ђ��fYe��ݠ$b�_ܢ�7���\��)[�T��g:�N@T'p�%�N&�nB*~��f�Il����~���IV\)@���� ���n�\��g�╮n)�SA�-�ĆUL�̴�L�k��-�T�y�MP�nS	O"7����C�\�V�9��)���1	�Z
����,���`P��R2���g�<��}�Lla
+)c�<���% E<>/SI����#���zI��iBw�)X4�
+�֖dR�`컾���X'��5Q�NgeЦ��?W��q�,zR.!y�#�i�S�#J3ι=�A,�*�)�K-��-�I�h���%`%�B[zg���NK��5Ր�i�����X�?-��YEG��2����V�>�p�}��~l)�F[}0��,���d��К�#��`܋:v{)t�M�4sZwr��[`òOW��$^񴁘[Eq���	f��H�ͺp�WAo>D�	M� i��w2h?hz�.ٔ������v>��t'���Ǎ�in�5,83��(H�y�Km�� xM	���V�#�jǼpN��3�Ё=�s�%�j�sT���v�q5<��L�(���w�ԧhkeX|�]�;§'��t0p��-��>lv˃�τ��&y/�xT�A�n�0z&�`�)=��H
-�Α�,]���)��.[�ETՋW�9g�qP2��1�b��<>-�9ت� �a����y��r/�����T1�G��t�ī���8�<H'4���Ot �}p5x��YŸv!��OU���7�U;l;2�&L�m&��~��7���}���~[:��hX:��n�-�o��0��=_���n-���m��N�?�1b朽I����f#�5!8pD"�E�P���:��x>�R�߸��3zH�q�"�P�7���Q�#�A�%▛͡ت�юI��ǩ�<�u'<�<������YF�����~��>8\u
�dFqa���y�=3�rI��geԜ=���}(u�fM���	�6g9�G���
+�iڇ��Sȏ�uK;�+C��M\[���~�@z!{��
����@���u�mF���|6/f%g����~ۓ��FK����O��v	������
�V�<D�p;�ܯ�8���t��R�1LH>
!�C����B՜�-����
���XP�h��7
+��k�
+�'c�6�Lp�k-�l�3h3M�7���6
�B�����g�6��<�B�u�眈v҉g����9�Z���
�"�b���:�lg�W�h@�	G�x*���b��29�q�h�
+�ҽM�L�ڂv;d;�KQ�̶eٷ�y���Y�����)��]e�8͸ժ�?����k-�I1E
+�Po%�
�v��5��i+י�6�t��}��p,j����e��$�IL��
)*h,�����g��QB�վ�C�4�C�Zz��g�cf�o,�T�����
[Yt��J����D��Q�J�"���5_K�Xڝ/��� �.i�,����}SWd5��(Ô������@�3���,���9|ٍ9)g���ii�à	�q�<�}i;i�)��]#�)E�b���5��

+�NG���L����:��Q*l��\@�o	3�[e�Z�����c���&����ƛ
�hK�/�QY����bc�im
+e������m��h�����E"��>}�����/��k=�?��Y�Uy&�N8qoˆ�ˏhɸ��h�6e��k�X�:D9xJ��i�.%�I���H�`}|�L��	��(�\�Vc�v6��
m�-�����ܔ��企�[m�	��~�v�rqh���R�%����
��g�-��\�3n��ܜR5�D���tV<�� i�r�c�Mst��} `�]�{1��U�x͂?R-˒�q�mf��v=L��n��7��-�T�>�c�j��e�I�}�{���,n�=@`wC������Y����Ʋ��$(\Jr#/1�h#;W�f�@V�Y+SN)u�d5��4B3�EP������|��%�O�?AਨQ6��dF�qW��q&
���K���jד�#����?h��Oi��(SȦD�p�G��
+&&��	T	�$�|��'Y�!�4�+�Z0���t[�Zf����I�(��S#~fE�����qњs$]�2�w]�Zj9��Q.F�FK�x�L�.[�l��`�G/�9�fR���p��y&��"�+��+b �t3S��&�
�ͼ_$L�z�#��)��C=n?�HJ�CT�UHB���Ax�I;�by���0{U�굱��ڻ�9��X�Ϣ�&֟��<�T�42s�Օ����.�4*.���@)(�Q;rQ/�
�
+�vLo�*���� !�|q��	L{R$6�-�+�ֺ��Z��%g\:+��d�Qf4�qRZ�w:�?Xi�@�*�}M�!�����
+Hyw���<�=aT���ѕ��S<�}oL3{kP�^D
=ݣE���Hp��^�e�
+&�{g��k�s�֪s<uOի ;��1
+�@�m^!����Zd� ��T�����i��գ#R��h�-Y��S|�OQ���(���J��y�F�s�A��t|;挗�
�`+�\�����*�|�Xy�����8_�X�[ނC���d�1���m��A�`�>p��t���? ��@�9��QP�[��M��D�P��<�16��
+�2�.dv�ܨ����%�՞}x��peۆtdC$Kج9��2�v0Ǧ�
��*c�X�N)
q�vM�O`R#|0~
h�(0���/Ey��
+�:�}_?^

��$5���M��^�s��bM���,8;�V-F���()F�{�\�::F�9��|�1�dإ��1��yLS�xsY�P�0n]'������/���cY��x��fJrq�ÍDVv����ݮh�̡�9�]�Di�]��ְ!�D�Xۦ��O��Y�b�H���9V��'��_	��3bk4���%���?�7{������y�r������&��j�6�K���[ꨚ�D`��^�"��@Jz��	��d8(��-�������=��:���ooZ<<��T�8�R!&���I�N0΃���>L�\jt ���X[jJ�GKb��ir R�ތO`rz�l�M���î2H�5�i�M
�<�JIƝ�}B�a�f��R�i1�K�3��ʱW�#k�rZ���!��=w�E⥧�W����ߦ���N��6=Іr~h혐^�E�G&�[����')�@C��騥~��G���f�8���	I�wr28k8o0���gG��io���޽"�gJ黺sV����*���\C5�Ds
+P�&����,��׺[ᥟM�h���nY�9q]h�[t���Ιr�渚�=&E6(���H�(Rh���� �
����
+�y
��[[Th��l�^��[Z���&�b&�c���Pׄς�v*j�����:���y�ݱp�v�<A�o_I%��!�M��"��$/��w5�1������}dds��o_��\.tͫ�ƽ��\��͗In��"�$�����M�r��>(W���df��#LU{ad��eH	�e�oN�;�j����	Ρ�:����j�q�u}
29n̋��r�ٶcz����z[�I��v�"���qԲ(�#�`H�C�UQwA�Э]�WBI.���H�|�!�i)�"�ѼnQ�c�vH�Ai'����mM&�P}4.F,�����BuD\T�1�C
+èe� /nDZ�d^y����jM�kL���8������_
+���8Y�G.�$]��DZI�wzkkve��"y2U(0
+U�6�-�%+K�A3�oTE#}�9~ǐ�����t,��V-�K5M�A�勳x�O��Z�T&��M�~�l�o��7�Z@�Wɴ2(w�H�Sv�!�J$������D,�%3�s��Þd���u}���X��R�0Z���
+�7UQ|*�!oQ�><��,��F2!�q��3q�pgs�cR
+5��B�W�dU6@Y��ݵ3]`���S���%V��#�a�<ܺtuyy�ӏ�R�t�#��Q]��}�PR>/S{Re�{��}V�.�X-�m��i��q,�R��c�L$َ3������K��b��
+)5׬�P���[y8��DW��gwkr�^�-w�G��wɊ�1x��Zz�4eõK�aZo�hπ�0*
+��ӿ�-�(e���2}�X�N��I
+:����i��KG�:ͼK��^/��`Yڲ?���\��ߩV�t��[�u�j����.�,�a���
hl�^ӱl�h�����M�f~
ꠙ�� �l�T�����=0�0��`	�(���
�=��H�j[L�0�s\�I������\[R�%}Iy�>1acK1T��eB���x�2qQK��7&�zg���}�RN#�H��@WF7QIZ����W3rJ%�����g�{���
f#�BB3奴��d����Tm6!S��}G��%�
�A�C�e��,�eҨы�G�h!Ÿz�m?�7&I��p�W~�#�~�n��ͫ��xO�z���_��y��⳷����0~�������Ǜ���۷�7�Ư.7���/�߾Z�����k�1rs!{PJ.�]�L
m����
+endstream
+endobj
+
+1461 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 1299 0 R
+      /c1 1300 0 R
+    >>
+    /Font <<
+      /f0 1257 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 72
+  /Parent 1 0 R
+  /Contents 1462 0 R
+>>
+endobj
+
+1462 0 obj
+<<
+  /Length 440
+  /Filter /FlateDecode
+>>
+stream
+x��T=o�0��+�-0���8!�⣕Z	�-ـ!��e�Pp�����NHC���������.�B�G���iH�}�G�P@@�P�|�{�]�\�\P�
�#��@�r@�hK�
A�Ď|K@���S�!��ߎ��e3��m�C���V�UQ�F-=�	Hߟ��Q�&�p�6�nv`�T�q�O�9ra���W&�hM�*�,�J"��l���X%l�AEʍ	���4�h��t����VE��"�f���Y�9�4nU���V�$�Z�X�_�i�[�8+�Q��ktj���ɻ�uE��^�jn}%ڢl����<������Ɉ��W�־�oy����x%4ϦiE�1�)o���e���Y0�,q���8���9�2�>W�0�����fq�N���g��1I�j��%�XoC�N��t,�`�0}Ι���֜Bi�,���lmN1
+endstream
+endobj
+
+1463 0 obj
+<<
+  /Title <FEFF004F006E0074006F006C006F006700ED0061002000660069006C006F007300F30066006900630061002000790020006D006F00640065006C006F0020006F007000650072006100630069006F006E0061006C00200065006E00200073006F006600740077006100720065>
+  /Author <FEFF004A0065007300FA00730020005000E900720065007A0020004C006F00720065006E007A006F>
+  /Creator (Typst 0.14.2)
+  /ModDate (D:20260424234756+01'00)
+  /CreationDate (D:20260424234756+01'00)
+>>
+endobj
+
+1464 0 obj
+<<
+  /Length 1212
+  /Type /Metadata
+  /Subtype /XML
+>>
+stream
+<?xpacket begin="" id="W5M0MpCehiHzreSzNTczkc9d"?><x:xmpmeta xmlns:x="adobe:ns:meta/" x:xmptk="xmp-writer"><rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"><rdf:Description rdf:about="" xmlns:dc="http://purl.org/dc/elements/1.1/"  xmlns:xmp="http://ns.adobe.com/xap/1.0/"  xmlns:xmpMM="http://ns.adobe.com/xap/1.0/mm/"  xmlns:xmpTPg="http://ns.adobe.com/xap/1.0/t/pg/"  xmlns:pdf="http://ns.adobe.com/pdf/1.3/" ><dc:title><rdf:Alt><rdf:li xml:lang="x-default">Ontología filosófica y modelo operacional en software</rdf:li></rdf:Alt></dc:title><dc:creator><rdf:Seq><rdf:li>Jesús Pérez Lorenzo</rdf:li></rdf:Seq></dc:creator><xmp:CreatorTool>Typst 0.14.2</xmp:CreatorTool><dc:language><rdf:Bag><rdf:li>es</rdf:li></rdf:Bag></dc:language><xmp:ModifyDate>2026-04-24T23:47:56+01:00</xmp:ModifyDate><xmp:CreateDate>2026-04-24T23:47:56+01:00</xmp:CreateDate><xmpTPg:NPages>28</xmpTPg:NPages><dc:format>application/pdf</dc:format><xmpMM:InstanceID>ZKwEWq8nPEeKyHz6JNDJwg==</xmpMM:InstanceID><xmpMM:DocumentID>ew9rg43v5afrI59+6KrAKA==</xmpMM:DocumentID><xmpMM:RenditionClass>proof</xmpMM:RenditionClass><pdf:PDFVersion>1.7</pdf:PDFVersion></rdf:Description></rdf:RDF></x:xmpmeta><?xpacket end="r"?>
+endstream
+endobj
+
+1465 0 obj
+<<
+  /Type /Catalog
+  /Pages 1 0 R
+  /Metadata 1464 0 R
+  /PageLabels 62 0 R
+  /Lang (es)
+  /StructTreeRoot 63 0 R
+  /MarkInfo <<
+    /Marked true
+    /Suspects false
+  >>
+  /ViewerPreferences <<
+    /Direction /L2R
+  >>
+  /Outlines 2 0 R
+>>
+endobj
+
+xref
+0 1466
+0000000000 65535 f
+0000000016 00000 n
+0000000327 00000 n
+0000000408 00000 n
+0000000810 00000 n
+0000001071 00000 n
+0000001229 00000 n
+0000001520 00000 n
+0000001805 00000 n
+0000002134 00000 n
+0000002237 00000 n
+0000002657 00000 n
+0000002901 00000 n
+0000003062 00000 n
+0000003342 00000 n
+0000003709 00000 n
+0000004008 00000 n
+0000004411 00000 n
+0000004762 00000 n
+0000005158 00000 n
+0000005381 00000 n
+0000005538 00000 n
+0000005794 00000 n
+0000005936 00000 n
+0000006065 00000 n
+0000006204 00000 n
+0000006484 00000 n
+0000006632 00000 n
+0000006953 00000 n
+0000007209 00000 n
+0000007520 00000 n
+0000007666 00000 n
+0000007807 00000 n
+0000008107 00000 n
+0000008472 00000 n
+0000008592 00000 n
+0000008975 00000 n
+0000009166 00000 n
+0000009501 00000 n
+0000009773 00000 n
+0000010174 00000 n
+0000010590 00000 n
+0000010765 00000 n
+0000011165 00000 n
+0000011488 00000 n
+0000011872 00000 n
+0000012243 00000 n
+0000012582 00000 n
+0000012930 00000 n
+0000013331 00000 n
+0000013623 00000 n
+0000013962 00000 n
+0000014245 00000 n
+0000014616 00000 n
+0000014767 00000 n
+0000014919 00000 n
+0000015083 00000 n
+0000015307 00000 n
+0000015658 00000 n
+0000015905 00000 n
+0000016033 00000 n
+0000016272 00000 n
+0000016540 00000 n
+0000016899 00000 n
+0000018267 00000 n
+0000018961 00000 n
+0000019583 00000 n
+0000020223 00000 n
+0000020370 00000 n
+0000021661 00000 n
+0000022488 00000 n
+0000022963 00000 n
+0000023470 00000 n
+0000024025 00000 n
+0000024548 00000 n
+0000025407 00000 n
+0000026362 00000 n
+0000026533 00000 n
+0000027144 00000 n
+0000027755 00000 n
+0000028054 00000 n
+0000028593 00000 n
+0000029148 00000 n
+0000029511 00000 n
+0000029946 00000 n
+0000030405 00000 n
+0000030640 00000 n
+0000031283 00000 n
+0000031918 00000 n
+0000032529 00000 n
+0000033148 00000 n
+0000033781 00000 n
+0000033814 00000 n
+0000035810 00000 n
+0000036031 00000 n
+0000036120 00000 n
+0000036213 00000 n
+0000036324 00000 n
+0000036413 00000 n
+0000036506 00000 n
+0000036611 00000 n
+0000036705 00000 n
+0000036822 00000 n
+0000036913 00000 n
+0000037008 00000 n
+0000037122 00000 n
+0000037213 00000 n
+0000037308 00000 n
+0000037422 00000 n
+0000037513 00000 n
+0000037608 00000 n
+0000037722 00000 n
+0000037813 00000 n
+0000037908 00000 n
+0000038028 00000 n
+0000038119 00000 n
+0000038214 00000 n
+0000038479 00000 n
+0000038593 00000 n
+0000038684 00000 n
+0000038779 00000 n
+0000038896 00000 n
+0000038987 00000 n
+0000039082 00000 n
+0000039194 00000 n
+0000039284 00000 n
+0000039378 00000 n
+0000039488 00000 n
+0000039578 00000 n
+0000039672 00000 n
+0000039786 00000 n
+0000039883 00000 n
+0000039978 00000 n
+0000040092 00000 n
+0000040186 00000 n
+0000040281 00000 n
+0000040401 00000 n
+0000040492 00000 n
+0000040587 00000 n
+0000040704 00000 n
+0000040795 00000 n
+0000040890 00000 n
+0000041007 00000 n
+0000041098 00000 n
+0000041193 00000 n
+0000041318 00000 n
+0000041409 00000 n
+0000041506 00000 n
+0000041601 00000 n
+0000041718 00000 n
+0000041809 00000 n
+0000041904 00000 n
+0000042027 00000 n
+0000042118 00000 n
+0000042213 00000 n
+0000042329 00000 n
+0000042419 00000 n
+0000042513 00000 n
+0000042732 00000 n
+0000042947 00000 n
+0000043038 00000 n
+0000043133 00000 n
+0000043250 00000 n
+0000043341 00000 n
+0000043436 00000 n
+0000043550 00000 n
+0000043641 00000 n
+0000043736 00000 n
+0000043853 00000 n
+0000043944 00000 n
+0000044039 00000 n
+0000044153 00000 n
+0000044244 00000 n
+0000044339 00000 n
+0000044464 00000 n
+0000044578 00000 n
+0000044669 00000 n
+0000044764 00000 n
+0000044878 00000 n
+0000044969 00000 n
+0000045064 00000 n
+0000045170 00000 n
+0000045265 00000 n
+0000045494 00000 n
+0000045611 00000 n
+0000045702 00000 n
+0000045797 00000 n
+0000045911 00000 n
+0000046002 00000 n
+0000046096 00000 n
+0000046208 00000 n
+0000046298 00000 n
+0000046392 00000 n
+0000046607 00000 n
+0000046698 00000 n
+0000046793 00000 n
+0000046904 00000 n
+0000046998 00000 n
+0000047093 00000 n
+0000047207 00000 n
+0000047301 00000 n
+0000047396 00000 n
+0000047516 00000 n
+0000047607 00000 n
+0000047702 00000 n
+0000047816 00000 n
+0000047910 00000 n
+0000048005 00000 n
+0000048108 00000 n
+0000048203 00000 n
+0000048536 00000 n
+0000048650 00000 n
+0000048744 00000 n
+0000048839 00000 n
+0000048953 00000 n
+0000049044 00000 n
+0000049139 00000 n
+0000049256 00000 n
+0000049347 00000 n
+0000049442 00000 n
+0000049556 00000 n
+0000049647 00000 n
+0000049742 00000 n
+0000049854 00000 n
+0000049944 00000 n
+0000050038 00000 n
+0000050148 00000 n
+0000050238 00000 n
+0000050332 00000 n
+0000050446 00000 n
+0000050537 00000 n
+0000050632 00000 n
+0000050746 00000 n
+0000050837 00000 n
+0000050932 00000 n
+0000051052 00000 n
+0000051143 00000 n
+0000051238 00000 n
+0000051352 00000 n
+0000051443 00000 n
+0000051538 00000 n
+0000051652 00000 n
+0000051743 00000 n
+0000051838 00000 n
+0000051958 00000 n
+0000052049 00000 n
+0000052144 00000 n
+0000052261 00000 n
+0000052352 00000 n
+0000052447 00000 n
+0000052564 00000 n
+0000052655 00000 n
+0000052750 00000 n
+0000052867 00000 n
+0000052958 00000 n
+0000053053 00000 n
+0000053167 00000 n
+0000053258 00000 n
+0000053353 00000 n
+0000053466 00000 n
+0000053556 00000 n
+0000053650 00000 n
+0000053869 00000 n
+0000054003 00000 n
+0000054100 00000 n
+0000054201 00000 n
+0000054299 00000 n
+0000054388 00000 n
+0000054538 00000 n
+0000054630 00000 n
+0000054779 00000 n
+0000054888 00000 n
+0000054982 00000 n
+0000055241 00000 n
+0000055336 00000 n
+0000055445 00000 n
+0000055540 00000 n
+0000055635 00000 n
+0000055783 00000 n
+0000055889 00000 n
+0000055984 00000 n
+0000056090 00000 n
+0000056185 00000 n
+0000056291 00000 n
+0000056386 00000 n
+0000056500 00000 n
+0000056595 00000 n
+0000056690 00000 n
+0000057043 00000 n
+0000057146 00000 n
+0000057241 00000 n
+0000057344 00000 n
+0000057438 00000 n
+0000057540 00000 n
+0000057634 00000 n
+0000057898 00000 n
+0000058003 00000 n
+0000058098 00000 n
+0000058190 00000 n
+0000058455 00000 n
+0000058564 00000 n
+0000058659 00000 n
+0000058762 00000 n
+0000058857 00000 n
+0000058957 00000 n
+0000059052 00000 n
+0000059221 00000 n
+0000059345 00000 n
+0000059434 00000 n
+0000059660 00000 n
+0000059880 00000 n
+0000059969 00000 n
+0000060214 00000 n
+0000060459 00000 n
+0000060548 00000 n
+0000060768 00000 n
+0000060988 00000 n
+0000061077 00000 n
+0000061322 00000 n
+0000061567 00000 n
+0000061656 00000 n
+0000061876 00000 n
+0000062096 00000 n
+0000062185 00000 n
+0000062436 00000 n
+0000062681 00000 n
+0000062765 00000 n
+0000062854 00000 n
+0000063116 00000 n
+0000063378 00000 n
+0000063699 00000 n
+0000063793 00000 n
+0000063942 00000 n
+0000064030 00000 n
+0000064317 00000 n
+0000064664 00000 n
+0000064773 00000 n
+0000064868 00000 n
+0000064977 00000 n
+0000065072 00000 n
+0000065167 00000 n
+0000065512 00000 n
+0000065681 00000 n
+0000065805 00000 n
+0000065902 00000 n
+0000066131 00000 n
+0000066360 00000 n
+0000066580 00000 n
+0000066677 00000 n
+0000066928 00000 n
+0000067176 00000 n
+0000067426 00000 n
+0000067523 00000 n
+0000067748 00000 n
+0000067969 00000 n
+0000068192 00000 n
+0000068289 00000 n
+0000068537 00000 n
+0000068788 00000 n
+0000069033 00000 n
+0000069130 00000 n
+0000069353 00000 n
+0000069579 00000 n
+0000069799 00000 n
+0000069896 00000 n
+0000070144 00000 n
+0000070392 00000 n
+0000070637 00000 n
+0000070721 00000 n
+0000070818 00000 n
+0000071086 00000 n
+0000071348 00000 n
+0000071610 00000 n
+0000071931 00000 n
+0000072032 00000 n
+0000072127 00000 n
+0000072244 00000 n
+0000072342 00000 n
+0000072433 00000 n
+0000072786 00000 n
+0000072881 00000 n
+0000072984 00000 n
+0000073075 00000 n
+0000073189 00000 n
+0000073284 00000 n
+0000073375 00000 n
+0000073752 00000 n
+0000073846 00000 n
+0000073996 00000 n
+0000074085 00000 n
+0000074188 00000 n
+0000074282 00000 n
+0000074373 00000 n
+0000074462 00000 n
+0000074564 00000 n
+0000074659 00000 n
+0000074751 00000 n
+0000074840 00000 n
+0000074942 00000 n
+0000075037 00000 n
+0000075129 00000 n
+0000075221 00000 n
+0000075338 00000 n
+0000075433 00000 n
+0000075531 00000 n
+0000075834 00000 n
+0000075932 00000 n
+0000076035 00000 n
+0000076129 00000 n
+0000076238 00000 n
+0000076333 00000 n
+0000076734 00000 n
+0000076863 00000 n
+0000076986 00000 n
+0000077081 00000 n
+0000077179 00000 n
+0000077282 00000 n
+0000077377 00000 n
+0000077790 00000 n
+0000078141 00000 n
+0000078278 00000 n
+0000078376 00000 n
+0000078471 00000 n
+0000078566 00000 n
+0000078708 00000 n
+0000078797 00000 n
+0000078905 00000 n
+0000079000 00000 n
+0000079092 00000 n
+0000079181 00000 n
+0000079294 00000 n
+0000079393 00000 n
+0000079488 00000 n
+0000079580 00000 n
+0000079669 00000 n
+0000079769 00000 n
+0000079863 00000 n
+0000079963 00000 n
+0000080057 00000 n
+0000080324 00000 n
+0000080427 00000 n
+0000080525 00000 n
+0000080640 00000 n
+0000080738 00000 n
+0000080844 00000 n
+0000080942 00000 n
+0000081051 00000 n
+0000081149 00000 n
+0000081466 00000 n
+0000081635 00000 n
+0000081815 00000 n
+0000081912 00000 n
+0000082163 00000 n
+0000082408 00000 n
+0000082653 00000 n
+0000082750 00000 n
+0000082979 00000 n
+0000083199 00000 n
+0000083419 00000 n
+0000083516 00000 n
+0000083767 00000 n
+0000084020 00000 n
+0000084111 00000 n
+0000084356 00000 n
+0000084453 00000 n
+0000084676 00000 n
+0000084904 00000 n
+0000085001 00000 n
+0000085221 00000 n
+0000085318 00000 n
+0000085566 00000 n
+0000085811 00000 n
+0000086056 00000 n
+0000086153 00000 n
+0000086379 00000 n
+0000086599 00000 n
+0000086819 00000 n
+0000086916 00000 n
+0000087164 00000 n
+0000087412 00000 n
+0000087656 00000 n
+0000087753 00000 n
+0000087976 00000 n
+0000088197 00000 n
+0000088416 00000 n
+0000088513 00000 n
+0000088764 00000 n
+0000089012 00000 n
+0000089257 00000 n
+0000089354 00000 n
+0000089583 00000 n
+0000089803 00000 n
+0000090023 00000 n
+0000090120 00000 n
+0000090371 00000 n
+0000090624 00000 n
+0000090718 00000 n
+0000090963 00000 n
+0000091060 00000 n
+0000091289 00000 n
+0000091517 00000 n
+0000091608 00000 n
+0000091828 00000 n
+0000091925 00000 n
+0000092176 00000 n
+0000092429 00000 n
+0000092520 00000 n
+0000092765 00000 n
+0000092849 00000 n
+0000092946 00000 n
+0000093208 00000 n
+0000093470 00000 n
+0000093732 00000 n
+0000093905 00000 n
+0000094014 00000 n
+0000094109 00000 n
+0000094237 00000 n
+0000094335 00000 n
+0000094426 00000 n
+0000094521 00000 n
+0000094641 00000 n
+0000094732 00000 n
+0000094827 00000 n
+0000094916 00000 n
+0000095281 00000 n
+0000095371 00000 n
+0000095465 00000 n
+0000095595 00000 n
+0000095906 00000 n
+0000096018 00000 n
+0000096113 00000 n
+0000096209 00000 n
+0000096319 00000 n
+0000096413 00000 n
+0000096503 00000 n
+0000096798 00000 n
+0000096908 00000 n
+0000097020 00000 n
+0000097134 00000 n
+0000097238 00000 n
+0000097350 00000 n
+0000097451 00000 n
+0000097588 00000 n
+0000097683 00000 n
+0000097778 00000 n
+0000097879 00000 n
+0000098017 00000 n
+0000098186 00000 n
+0000098310 00000 n
+0000098415 00000 n
+0000098647 00000 n
+0000098873 00000 n
+0000099096 00000 n
+0000099306 00000 n
+0000099407 00000 n
+0000099512 00000 n
+0000099766 00000 n
+0000100014 00000 n
+0000100270 00000 n
+0000100361 00000 n
+0000100620 00000 n
+0000100715 00000 n
+0000100820 00000 n
+0000101046 00000 n
+0000101269 00000 n
+0000101500 00000 n
+0000101591 00000 n
+0000101840 00000 n
+0000101935 00000 n
+0000102040 00000 n
+0000102297 00000 n
+0000102545 00000 n
+0000102798 00000 n
+0000102892 00000 n
+0000103154 00000 n
+0000103249 00000 n
+0000103354 00000 n
+0000103580 00000 n
+0000103806 00000 n
+0000104037 00000 n
+0000104130 00000 n
+0000104365 00000 n
+0000104459 00000 n
+0000104564 00000 n
+0000104822 00000 n
+0000105071 00000 n
+0000105327 00000 n
+0000105421 00000 n
+0000105692 00000 n
+0000105787 00000 n
+0000105871 00000 n
+0000105976 00000 n
+0000106244 00000 n
+0000106509 00000 n
+0000106771 00000 n
+0000107033 00000 n
+0000107176 00000 n
+0000107271 00000 n
+0000107391 00000 n
+0000107486 00000 n
+0000107581 00000 n
+0000107739 00000 n
+0000107828 00000 n
+0000107936 00000 n
+0000108031 00000 n
+0000108123 00000 n
+0000108212 00000 n
+0000108317 00000 n
+0000108412 00000 n
+0000108504 00000 n
+0000108593 00000 n
+0000108707 00000 n
+0000108802 00000 n
+0000108894 00000 n
+0000108983 00000 n
+0000109091 00000 n
+0000109186 00000 n
+0000109278 00000 n
+0000109389 00000 n
+0000109487 00000 n
+0000109582 00000 n
+0000109875 00000 n
+0000109990 00000 n
+0000110085 00000 n
+0000110180 00000 n
+0000110292 00000 n
+0000110383 00000 n
+0000110544 00000 n
+0000110639 00000 n
+0000110734 00000 n
+0000110829 00000 n
+0000110923 00000 n
+0000111017 00000 n
+0000111111 00000 n
+0000111362 00000 n
+0000111629 00000 n
+0000111721 00000 n
+0000111825 00000 n
+0000111970 00000 n
+0000112128 00000 n
+0000112217 00000 n
+0000112322 00000 n
+0000112417 00000 n
+0000112509 00000 n
+0000112598 00000 n
+0000112700 00000 n
+0000112795 00000 n
+0000112887 00000 n
+0000112976 00000 n
+0000113078 00000 n
+0000113173 00000 n
+0000113265 00000 n
+0000113354 00000 n
+0000113456 00000 n
+0000113551 00000 n
+0000113643 00000 n
+0000113732 00000 n
+0000113863 00000 n
+0000113954 00000 n
+0000114045 00000 n
+0000114140 00000 n
+0000114235 00000 n
+0000114341 00000 n
+0000114432 00000 n
+0000114701 00000 n
+0000114851 00000 n
+0000114940 00000 n
+0000115048 00000 n
+0000115143 00000 n
+0000115235 00000 n
+0000115324 00000 n
+0000115429 00000 n
+0000115524 00000 n
+0000115616 00000 n
+0000115705 00000 n
+0000115808 00000 n
+0000115902 00000 n
+0000115993 00000 n
+0000116125 00000 n
+0000116275 00000 n
+0000116364 00000 n
+0000116467 00000 n
+0000116561 00000 n
+0000116652 00000 n
+0000116741 00000 n
+0000116846 00000 n
+0000116941 00000 n
+0000117033 00000 n
+0000117122 00000 n
+0000117227 00000 n
+0000117322 00000 n
+0000117414 00000 n
+0000117538 00000 n
+0000117688 00000 n
+0000117777 00000 n
+0000117885 00000 n
+0000117980 00000 n
+0000118072 00000 n
+0000118161 00000 n
+0000118272 00000 n
+0000118367 00000 n
+0000118459 00000 n
+0000118548 00000 n
+0000118653 00000 n
+0000118748 00000 n
+0000118840 00000 n
+0000118977 00000 n
+0000119135 00000 n
+0000119224 00000 n
+0000119329 00000 n
+0000119424 00000 n
+0000119516 00000 n
+0000119605 00000 n
+0000119710 00000 n
+0000119805 00000 n
+0000119897 00000 n
+0000119986 00000 n
+0000120091 00000 n
+0000120186 00000 n
+0000120278 00000 n
+0000120367 00000 n
+0000120475 00000 n
+0000120570 00000 n
+0000120662 00000 n
+0000120751 00000 n
+0000120996 00000 n
+0000121107 00000 n
+0000121276 00000 n
+0000121448 00000 n
+0000121537 00000 n
+0000121760 00000 n
+0000121980 00000 n
+0000122069 00000 n
+0000122317 00000 n
+0000122562 00000 n
+0000122651 00000 n
+0000122871 00000 n
+0000123091 00000 n
+0000123180 00000 n
+0000123424 00000 n
+0000123668 00000 n
+0000123757 00000 n
+0000123976 00000 n
+0000124195 00000 n
+0000124284 00000 n
+0000124528 00000 n
+0000124772 00000 n
+0000124861 00000 n
+0000125084 00000 n
+0000125303 00000 n
+0000125392 00000 n
+0000125640 00000 n
+0000125885 00000 n
+0000125974 00000 n
+0000126194 00000 n
+0000126414 00000 n
+0000126503 00000 n
+0000126751 00000 n
+0000126996 00000 n
+0000127085 00000 n
+0000127305 00000 n
+0000127525 00000 n
+0000127614 00000 n
+0000127859 00000 n
+0000128104 00000 n
+0000128188 00000 n
+0000128277 00000 n
+0000128539 00000 n
+0000128801 00000 n
+0000129006 00000 n
+0000129104 00000 n
+0000129210 00000 n
+0000129301 00000 n
+0000129689 00000 n
+0000129784 00000 n
+0000129893 00000 n
+0000129988 00000 n
+0000130313 00000 n
+0000130447 00000 n
+0000130542 00000 n
+0000130637 00000 n
+0000130728 00000 n
+0000130841 00000 n
+0000130935 00000 n
+0000131029 00000 n
+0000131404 00000 n
+0000131494 00000 n
+0000131600 00000 n
+0000131691 00000 n
+0000131805 00000 n
+0000131899 00000 n
+0000131990 00000 n
+0000132263 00000 n
+0000132355 00000 n
+0000132478 00000 n
+0000132573 00000 n
+0000132664 00000 n
+0000133005 00000 n
+0000133100 00000 n
+0000133220 00000 n
+0000133315 00000 n
+0000133409 00000 n
+0000133678 00000 n
+0000133793 00000 n
+0000133885 00000 n
+0000133993 00000 n
+0000134087 00000 n
+0000134326 00000 n
+0000134695 00000 n
+0000134819 00000 n
+0000134910 00000 n
+0000135005 00000 n
+0000135114 00000 n
+0000135209 00000 n
+0000135340 00000 n
+0000135431 00000 n
+0000135522 00000 n
+0000135617 00000 n
+0000135745 00000 n
+0000135851 00000 n
+0000135957 00000 n
+0000136052 00000 n
+0000136184 00000 n
+0000136353 00000 n
+0000136573 00000 n
+0000136678 00000 n
+0000136901 00000 n
+0000136997 00000 n
+0000137217 00000 n
+0000137445 00000 n
+0000137536 00000 n
+0000137746 00000 n
+0000137841 00000 n
+0000137946 00000 n
+0000138203 00000 n
+0000138451 00000 n
+0000138707 00000 n
+0000138801 00000 n
+0000139036 00000 n
+0000139131 00000 n
+0000139236 00000 n
+0000139462 00000 n
+0000139688 00000 n
+0000139916 00000 n
+0000140007 00000 n
+0000140217 00000 n
+0000140312 00000 n
+0000140417 00000 n
+0000140687 00000 n
+0000140793 00000 n
+0000140899 00000 n
+0000141150 00000 n
+0000141403 00000 n
+0000141494 00000 n
+0000141729 00000 n
+0000141827 00000 n
+0000141932 00000 n
+0000142161 00000 n
+0000142384 00000 n
+0000142612 00000 n
+0000142706 00000 n
+0000142916 00000 n
+0000143011 00000 n
+0000143116 00000 n
+0000143381 00000 n
+0000143486 00000 n
+0000143732 00000 n
+0000143984 00000 n
+0000144078 00000 n
+0000144313 00000 n
+0000144411 00000 n
+0000144516 00000 n
+0000144749 00000 n
+0000144970 00000 n
+0000145199 00000 n
+0000145299 00000 n
+0000145509 00000 n
+0000145605 00000 n
+0000145710 00000 n
+0000145968 00000 n
+0000146218 00000 n
+0000146472 00000 n
+0000146568 00000 n
+0000146803 00000 n
+0000146903 00000 n
+0000147008 00000 n
+0000147241 00000 n
+0000147470 00000 n
+0000147699 00000 n
+0000147799 00000 n
+0000148009 00000 n
+0000148109 00000 n
+0000148214 00000 n
+0000148472 00000 n
+0000148726 00000 n
+0000148992 00000 n
+0000149092 00000 n
+0000149192 00000 n
+0000149427 00000 n
+0000149523 00000 n
+0000149628 00000 n
+0000149861 00000 n
+0000150090 00000 n
+0000150315 00000 n
+0000150525 00000 n
+0000150623 00000 n
+0000150728 00000 n
+0000150998 00000 n
+0000151104 00000 n
+0000151210 00000 n
+0000151455 00000 n
+0000151708 00000 n
+0000151802 00000 n
+0000152037 00000 n
+0000152132 00000 n
+0000152237 00000 n
+0000152488 00000 n
+0000152594 00000 n
+0000152700 00000 n
+0000152926 00000 n
+0000153165 00000 n
+0000153256 00000 n
+0000153347 00000 n
+0000153557 00000 n
+0000153655 00000 n
+0000153760 00000 n
+0000154008 00000 n
+0000154253 00000 n
+0000154506 00000 n
+0000154597 00000 n
+0000154832 00000 n
+0000154930 00000 n
+0000155035 00000 n
+0000155304 00000 n
+0000155410 00000 n
+0000155516 00000 n
+0000155622 00000 n
+0000155728 00000 n
+0000155834 00000 n
+0000156054 00000 n
+0000156282 00000 n
+0000156373 00000 n
+0000156607 00000 n
+0000156702 00000 n
+0000156807 00000 n
+0000157079 00000 n
+0000157185 00000 n
+0000157291 00000 n
+0000157397 00000 n
+0000157645 00000 n
+0000157901 00000 n
+0000157992 00000 n
+0000158227 00000 n
+0000158322 00000 n
+0000158427 00000 n
+0000158677 00000 n
+0000158783 00000 n
+0000158889 00000 n
+0000158995 00000 n
+0000159221 00000 n
+0000159452 00000 n
+0000159543 00000 n
+0000159753 00000 n
+0000159848 00000 n
+0000159953 00000 n
+0000160217 00000 n
+0000160326 00000 n
+0000160432 00000 n
+0000160677 00000 n
+0000160941 00000 n
+0000161038 00000 n
+0000161129 00000 n
+0000161364 00000 n
+0000161459 00000 n
+0000161543 00000 n
+0000161648 00000 n
+0000161909 00000 n
+0000162174 00000 n
+0000162435 00000 n
+0000162696 00000 n
+0000162786 00000 n
+0000163065 00000 n
+0000163157 00000 n
+0000163281 00000 n
+0000163373 00000 n
+0000163467 00000 n
+0000163750 00000 n
+0000163859 00000 n
+0000163950 00000 n
+0000164227 00000 n
+0000164373 00000 n
+0000164464 00000 n
+0000164559 00000 n
+0000164650 00000 n
+0000164756 00000 n
+0000164854 00000 n
+0000164957 00000 n
+0000165051 00000 n
+0000165210 00000 n
+0000165371 00000 n
+0000165462 00000 n
+0000165569 00000 n
+0000165666 00000 n
+0000165759 00000 n
+0000165851 00000 n
+0000165971 00000 n
+0000166064 00000 n
+0000166161 00000 n
+0000166255 00000 n
+0000166347 00000 n
+0000166467 00000 n
+0000166560 00000 n
+0000166657 00000 n
+0000166751 00000 n
+0000166843 00000 n
+0000166951 00000 n
+0000167048 00000 n
+0000167142 00000 n
+0000167264 00000 n
+0000167360 00000 n
+0000167456 00000 n
+0000167716 00000 n
+0000167807 00000 n
+0000168179 00000 n
+0000168397 00000 n
+0000168527 00000 n
+0000168613 00000 n
+0000168756 00000 n
+0000168922 00000 n
+0000169016 00000 n
+0000169102 00000 n
+0000169245 00000 n
+0000169411 00000 n
+0000169505 00000 n
+0000169591 00000 n
+0000169734 00000 n
+0000169900 00000 n
+0000169994 00000 n
+0000170080 00000 n
+0000170223 00000 n
+0000170389 00000 n
+0000170483 00000 n
+0000170569 00000 n
+0000170712 00000 n
+0000170878 00000 n
+0000170972 00000 n
+0000171058 00000 n
+0000171201 00000 n
+0000171367 00000 n
+0000171461 00000 n
+0000171547 00000 n
+0000171690 00000 n
+0000171856 00000 n
+0000171950 00000 n
+0000172080 00000 n
+0000172166 00000 n
+0000172309 00000 n
+0000172475 00000 n
+0000172569 00000 n
+0000172655 00000 n
+0000172798 00000 n
+0000172964 00000 n
+0000173058 00000 n
+0000173144 00000 n
+0000173287 00000 n
+0000173453 00000 n
+0000173547 00000 n
+0000173633 00000 n
+0000173776 00000 n
+0000173942 00000 n
+0000174036 00000 n
+0000174122 00000 n
+0000174265 00000 n
+0000174431 00000 n
+0000174525 00000 n
+0000174611 00000 n
+0000174754 00000 n
+0000174920 00000 n
+0000175014 00000 n
+0000175100 00000 n
+0000175243 00000 n
+0000175409 00000 n
+0000175503 00000 n
+0000175624 00000 n
+0000175710 00000 n
+0000175853 00000 n
+0000176019 00000 n
+0000176113 00000 n
+0000176199 00000 n
+0000176342 00000 n
+0000176508 00000 n
+0000176602 00000 n
+0000176688 00000 n
+0000176831 00000 n
+0000176997 00000 n
+0000177091 00000 n
+0000177177 00000 n
+0000177320 00000 n
+0000177486 00000 n
+0000177580 00000 n
+0000177666 00000 n
+0000177809 00000 n
+0000177978 00000 n
+0000178071 00000 n
+0000178157 00000 n
+0000178300 00000 n
+0000178464 00000 n
+0000178557 00000 n
+0000178678 00000 n
+0000178764 00000 n
+0000178907 00000 n
+0000179071 00000 n
+0000179164 00000 n
+0000179250 00000 n
+0000179393 00000 n
+0000179557 00000 n
+0000179650 00000 n
+0000179736 00000 n
+0000179879 00000 n
+0000180045 00000 n
+0000180139 00000 n
+0000180225 00000 n
+0000180368 00000 n
+0000180534 00000 n
+0000180628 00000 n
+0000180714 00000 n
+0000180857 00000 n
+0000181023 00000 n
+0000181117 00000 n
+0000181203 00000 n
+0000181346 00000 n
+0000181512 00000 n
+0000181606 00000 n
+0000181727 00000 n
+0000181813 00000 n
+0000181956 00000 n
+0000182122 00000 n
+0000182216 00000 n
+0000182302 00000 n
+0000182445 00000 n
+0000182611 00000 n
+0000182705 00000 n
+0000182791 00000 n
+0000182934 00000 n
+0000183100 00000 n
+0000183194 00000 n
+0000183280 00000 n
+0000183423 00000 n
+0000183589 00000 n
+0000183683 00000 n
+0000183769 00000 n
+0000183912 00000 n
+0000184078 00000 n
+0000184172 00000 n
+0000184258 00000 n
+0000184401 00000 n
+0000184567 00000 n
+0000184661 00000 n
+0000184782 00000 n
+0000184868 00000 n
+0000185011 00000 n
+0000185177 00000 n
+0000185271 00000 n
+0000185357 00000 n
+0000185500 00000 n
+0000185666 00000 n
+0000185760 00000 n
+0000185846 00000 n
+0000185989 00000 n
+0000186155 00000 n
+0000186249 00000 n
+0000186335 00000 n
+0000186478 00000 n
+0000186644 00000 n
+0000186738 00000 n
+0000186824 00000 n
+0000186967 00000 n
+0000187133 00000 n
+0000187227 00000 n
+0000187313 00000 n
+0000187456 00000 n
+0000187622 00000 n
+0000187716 00000 n
+0000187801 00000 n
+0000187887 00000 n
+0000188030 00000 n
+0000188196 00000 n
+0000188290 00000 n
+0000188376 00000 n
+0000188519 00000 n
+0000188685 00000 n
+0000188779 00000 n
+0000188891 00000 n
+0000188977 00000 n
+0000189120 00000 n
+0000189286 00000 n
+0000189380 00000 n
+0000189466 00000 n
+0000189609 00000 n
+0000189775 00000 n
+0000189869 00000 n
+0000189955 00000 n
+0000190098 00000 n
+0000190264 00000 n
+0000190358 00000 n
+0000190444 00000 n
+0000190587 00000 n
+0000190753 00000 n
+0000190846 00000 n
+0000190932 00000 n
+0000191075 00000 n
+0000191239 00000 n
+0000191332 00000 n
+0000191440 00000 n
+0000191529 00000 n
+0000191618 00000 n
+0000191707 00000 n
+0000191798 00000 n
+0000191858 00000 n
+0000191918 00000 n
+0000191978 00000 n
+0000192038 00000 n
+0000192098 00000 n
+0000192158 00000 n
+0000192218 00000 n
+0000192278 00000 n
+0000192338 00000 n
+0000192399 00000 n
+0000192460 00000 n
+0000192521 00000 n
+0000192582 00000 n
+0000192643 00000 n
+0000192704 00000 n
+0000192765 00000 n
+0000192826 00000 n
+0000192887 00000 n
+0000192948 00000 n
+0000193009 00000 n
+0000193070 00000 n
+0000193131 00000 n
+0000193192 00000 n
+0000193253 00000 n
+0000193314 00000 n
+0000193375 00000 n
+0000193436 00000 n
+0000193497 00000 n
+0000193674 00000 n
+0000194709 00000 n
+0000194800 00000 n
+0000195050 00000 n
+0000196983 00000 n
+0000205078 00000 n
+0000205257 00000 n
+0000206160 00000 n
+0000206252 00000 n
+0000206512 00000 n
+0000208165 00000 n
+0000215971 00000 n
+0000216151 00000 n
+0000217231 00000 n
+0000217323 00000 n
+0000217575 00000 n
+0000219733 00000 n
+0000229012 00000 n
+0000229200 00000 n
+0000229456 00000 n
+0000229543 00000 n
+0000229800 00000 n
+0000230501 00000 n
+0000231041 00000 n
+0000231209 00000 n
+0000231479 00000 n
+0000231571 00000 n
+0000231851 00000 n
+0000233072 00000 n
+0000241141 00000 n
+0000241301 00000 n
+0000241559 00000 n
+0000241648 00000 n
+0000241887 00000 n
+0000242757 00000 n
+0000246313 00000 n
+0000246498 00000 n
+0000246768 00000 n
+0000246855 00000 n
+0000247110 00000 n
+0000247867 00000 n
+0000249053 00000 n
+0000249232 00000 n
+0000249487 00000 n
+0000249574 00000 n
+0000249822 00000 n
+0000250541 00000 n
+0000251319 00000 n
+0000251487 00000 n
+0000251758 00000 n
+0000251849 00000 n
+0000252138 00000 n
+0000253204 00000 n
+0000266364 00000 n
+0000266404 00000 n
+0000266444 00000 n
+0000266868 00000 n
+0000267228 00000 n
+0000267283 00000 n
+0000267338 00000 n
+0000267394 00000 n
+0000267449 00000 n
+0000267504 00000 n
+0000267559 00000 n
+0000267614 00000 n
+0000267669 00000 n
+0000267724 00000 n
+0000267780 00000 n
+0000267835 00000 n
+0000267890 00000 n
+0000267946 00000 n
+0000268002 00000 n
+0000268058 00000 n
+0000268114 00000 n
+0000268170 00000 n
+0000268226 00000 n
+0000268281 00000 n
+0000268337 00000 n
+0000268393 00000 n
+0000268449 00000 n
+0000268505 00000 n
+0000268560 00000 n
+0000268615 00000 n
+0000268671 00000 n
+0000268727 00000 n
+0000268783 00000 n
+0000268838 00000 n
+0000268893 00000 n
+0000268948 00000 n
+0000269004 00000 n
+0000269060 00000 n
+0000269116 00000 n
+0000269171 00000 n
+0000269226 00000 n
+0000269281 00000 n
+0000269337 00000 n
+0000269393 00000 n
+0000269449 00000 n
+0000269504 00000 n
+0000269560 00000 n
+0000269615 00000 n
+0000269671 00000 n
+0000269726 00000 n
+0000269781 00000 n
+0000269836 00000 n
+0000269892 00000 n
+0000269948 00000 n
+0000270004 00000 n
+0000270060 00000 n
+0000270115 00000 n
+0000270170 00000 n
+0000270226 00000 n
+0000270281 00000 n
+0000270337 00000 n
+0000270393 00000 n
+0000270448 00000 n
+0000270503 00000 n
+0000271007 00000 n
+0000271411 00000 n
+0000271891 00000 n
+0000272311 00000 n
+0000272739 00000 n
+0000273155 00000 n
+0000273527 00000 n
+0000274031 00000 n
+0000274415 00000 n
+0000274719 00000 n
+0000275068 00000 n
+0000275357 00000 n
+0000275650 00000 n
+0000276059 00000 n
+0000276460 00000 n
+0000276901 00000 n
+0000277322 00000 n
+0000277723 00000 n
+0000278168 00000 n
+0000278622 00000 n
+0000279000 00000 n
+0000279511 00000 n
+0000279830 00000 n
+0000280396 00000 n
+0000289344 00000 n
+0000289807 00000 n
+0000290222 00000 n
+0000290709 00000 n
+0000291268 00000 n
+0000291631 00000 n
+0000292130 00000 n
+0000292597 00000 n
+0000293088 00000 n
+0000293575 00000 n
+0000294010 00000 n
+0000294477 00000 n
+0000294888 00000 n
+0000295387 00000 n
+0000295826 00000 n
+0000296121 00000 n
+0000296505 00000 n
+0000296874 00000 n
+0000297355 00000 n
+0000297732 00000 n
+0000298081 00000 n
+0000298450 00000 n
+0000298863 00000 n
+0000299363 00000 n
+0000306972 00000 n
+0000307327 00000 n
+0000311629 00000 n
+0000311965 00000 n
+0000313520 00000 n
+0000313875 00000 n
+0000319208 00000 n
+0000319563 00000 n
+0000325043 00000 n
+0000325379 00000 n
+0000329360 00000 n
+0000329696 00000 n
+0000333979 00000 n
+0000334296 00000 n
+0000338349 00000 n
+0000338685 00000 n
+0000342434 00000 n
+0000342789 00000 n
+0000347632 00000 n
+0000348006 00000 n
+0000353748 00000 n
+0000354084 00000 n
+0000356005 00000 n
+0000356341 00000 n
+0000360671 00000 n
+0000361026 00000 n
+0000365876 00000 n
+0000366212 00000 n
+0000368599 00000 n
+0000368954 00000 n
+0000372893 00000 n
+0000373248 00000 n
+0000377620 00000 n
+0000377937 00000 n
+0000381000 00000 n
+0000381336 00000 n
+0000385505 00000 n
+0000385841 00000 n
+0000390151 00000 n
+0000390468 00000 n
+0000393212 00000 n
+0000393548 00000 n
+0000397949 00000 n
+0000398285 00000 n
+0000402941 00000 n
+0000403277 00000 n
+0000407879 00000 n
+0000408215 00000 n
+0000412964 00000 n
+0000413300 00000 n
+0000418229 00000 n
+0000418527 00000 n
+0000419047 00000 n
+0000419496 00000 n
+0000420800 00000 n
+trailer
+<<
+  /Size 1466
+  /Root 1465 0 R
+  /Info 1463 0 R
+  /ID [(ew9rg43v5afrI59+6KrAKA==) (ZKwEWq8nPEeKyHz6JNDJwg==)]
+>>
+startxref
+421061
+%%EOF
\ No newline at end of file
diff --git a/assets/presentation/docs/ontologia-filosofia-software.typ b/assets/presentation/docs/ontologia-filosofia-software.typ
new file mode 100644
index 0000000..161f249
--- /dev/null
+++ b/assets/presentation/docs/ontologia-filosofia-software.typ
@@ -0,0 +1,963 @@
+#set document(
+  title: "Ontología filosófica y modelo operacional en software",
+  author: "Jesús Pérez Lorenzo",
+)
+
+#set page(
+  paper: "a4",
+  margin: (top: 2.5cm, bottom: 2.5cm, left: 2.8cm, right: 2.8cm),
+  numbering: "1",
+  number-align: right,
+)
+
+#set text(
+  font: "New Computer Modern",
+  size: 11pt,
+  lang: "es",
+)
+
+#set heading(numbering: "1.1.")
+
+#set par(
+  justify: true,
+  leading: 0.75em,
+  first-line-indent: 1.2em,
+)
+
+#show heading: it => {
+  set text(weight: "bold")
+  set par(first-line-indent: 0em)
+  v(0.8em)
+  it
+  v(0.4em)
+}
+
+#show heading.where(level: 1): it => {
+  set text(size: 14pt, fill: rgb("#c04000"))
+  it
+}
+
+#show heading.where(level: 2): it => {
+  set text(size: 12pt, fill: rgb("#444444"))
+  it
+}
+
+#show table: it => {
+  set text(size: 9pt)
+  it
+}
+
+// ─── PORTADA ───────────────────────────────────────────────────────────────
+
+#align(center)[
+  #v(2cm)
+  #text(size: 20pt, weight: "bold", fill: rgb("#c04000"))[
+    Ontología filosófica \
+    y modelo operacional en software
+  ]
+  #v(0.6cm)
+  #text(size: 12pt, fill: rgb("#666666"), style: "italic")[
+    Genealogía conceptual, autores y uso contemporáneo
+  ]
+  #v(0.4cm)
+  #line(length: 60%, stroke: rgb("#c04000") + 0.5pt)
+  #v(0.4cm)
+  #text(size: 11pt)[Jesús Pérez Lorenzo]
+  #v(0.2cm)
+  #text(size: 10pt, fill: rgb("#888888"))[OntoRef · 2026]
+  #v(2cm)
+]
+
+#outline(
+  title: "Contenidos",
+  indent: 1.5em,
+  depth: 2,
+)
+
+#pagebreak()
+
+// ─── 1. UNIVERSALES E INVARIANTES ──────────────────────────────────────────
+
+= ¿Son los invariantes el equivalente a los universales filosóficos?
+
+La pregunta es legítima pero la correspondencia es más fina de lo que parece a primera vista.
+
+== Los universales en ontología filosófica
+
+Los *universales* responden a una pregunta específica: _¿qué comparten múltiples particulares que los hace del mismo tipo?_ Son propiedades, relaciones o cualidades que pueden instanciarse en varios individuos distintos — "rojez", "humanidad", "ser-mayor-que". El debate clásico gira en torno a su estatus ontológico:
+
+- *Realismo platónico*: los universales existen independientemente de sus instancias — las Formas habitan un reino separado del mundo sensible.
+- *Realismo moderado* (Aristóteles): los universales existen _en_ las cosas, no aparte de ellas. No hay humanidad sin seres humanos concretos.
+- *Nominalismo* (Ockham, siglo XIV): solo existen los particulares. Los universales son nombres (_nomina_), convenciones lingüísticas sin referente ontológico propio.
+- *Conceptualismo* (posición media): los universales existen en la mente como conceptos, no en el mundo ni como meros nombres.
+
+== Los invariantes como propiedades esenciales de un singular
+
+Los invariantes en un modelo operacional de software no son universales en ninguno de los sentidos anteriores. No preguntan qué comparten múltiples proyectos — preguntan _qué no puede cambiar en este proyecto concreto sin que deje de ser este proyecto_.
+
+La analogía filosófica precisa es la distinción aristotélica entre *propiedades esenciales y accidentales*: la esencia es aquello sin lo cual una cosa no sería lo que es; el accidente es lo que puede cambiar sin alterar su naturaleza. Un invariante operacional declara: "esta propiedad es esencial para la identidad de este proyecto."
+
+Kripke formalizó esto de manera más rigurosa en _Naming and Necessity_ (1980): ciertas propiedades son *rígidas* — verdaderas en todos los mundos posibles accesibles al individuo. El agua es necesariamente H₂O aunque alguien la llame "phlogiston". Un invariante operacional tiene este carácter modal: no puede ser falso en ningún estado posible del proyecto sin que el proyecto haya dejado de ser ese proyecto. Los ADRs que lo modifican no _violan_ el invariante — constituyen un nuevo individuo.
+
+== Dónde sí hay resonancia con los universales
+
+La familia semántica es compartida: ambos capturan _lo que no varía_. El universal no varía a través de la instanciación (la rojez es la misma en esta manzana y en esa). El invariante no varía a través de la evolución temporal del proyecto. El eje difiere — instanciación vs. tiempo — pero el gesto filosófico es análogo.
+
+== El estatus ontológico: hechos institucionales
+
+Hay una diferencia de fondo que ninguna de las categorías anteriores captura. Los universales platónicos existen independientemente de que alguien los declare. Los invariantes operacionales son *hechos institucionales* en el sentido de John Searle (_La construcción de la realidad social_, 1995): son verdad porque un colectivo los declaró y formalizó. Como el dinero, los contratos o las reglas del ajedrez — su existencia depende de actos declarativos sostenidos por una práctica social. Un invariante que el equipo no reconoce como tal no es un invariante real; es teatro.
+
+Esto los hace más frágiles que los universales metafísicos, pero también más honestos sobre su naturaleza.
+
+#pagebreak()
+
+// ─── 2. TABLA COMPARATIVA ──────────────────────────────────────────────────
+
+= Genealogía filosófica del modelo operacional
+
+La siguiente tabla mapea los elementos básicos de la ontología filosófica a sus equivalentes en modelos operacionales de software, con autores y origen temporal.
+
+#v(0.5em)
+
+#table(
+  columns: (1.8fr, 1.5fr, 1.8fr, 1.8fr),
+  inset: 7pt,
+  align: (left, left, left, left),
+  fill: (col, row) => if row == 0 { rgb("#c04000") } else if calc.odd(row) { rgb("#fdf6f0") } else { white },
+  table.header(
+    text(fill: white, weight: "bold")[Concepto filosófico],
+    text(fill: white, weight: "bold")[Autor / Origen],
+    text(fill: white, weight: "bold")[Equivalente operacional],
+    text(fill: white, weight: "bold")[Implementación],
+  ),
+  [*Universales*],
+  [Platón (428–348 aC) \ _República_, _Parménides_],
+  [Tipos / Schemas],
+  [Contratos Nickel: `String`, `[| 'Accepted, 'Proposed |]`],
+
+  [*Esencia / Accidente*],
+  [Aristóteles (384–322 aC) \ _Metafísica_],
+  [Invariante / Estado mutable],
+  [`invariante = true` en `core.ncl` vs. dimensiones en `state.ncl`],
+
+  [*Potencia / Acto*],
+  [Aristóteles \ _Metafísica_ IX],
+  [Estado actual → deseado + Gates],
+  [`current_state` / `desired_state` / `gate.ncl`],
+
+  [*Categorías* (10 predicamentos)],
+  [Aristóteles \ _Categorías_],
+  [Tipos de nodo ontológico],
+  [`invariants`, `tensions`, `practices`, `dimensions`, `gates`],
+
+  [*Sustancia primera / segunda*],
+  [Aristóteles \ _Categorías_],
+  [Instancia / Tipo],
+  [Proyecto concreto vs. schema de proyecto],
+
+  [*Identidad a través del tiempo*],
+  [Locke (1689) _Ensayo_ \ Parfit (1984) _Reasons_],
+  [Versioning + ADR lifecycle],
+  [`status = Superseded`, `superseded_by` — identidad que evoluciona formalmente],
+
+  [*Dialéctica*],
+  [Hegel (1807) \ _Fenomenología del Espíritu_],
+  [Tensiones activas],
+  [`tensions` en `core.ncl` — sin síntesis, gestionadas permanentemente],
+
+  [*Logos como tensión de opuestos*],
+  [Heráclito (535–475 aC)],
+  [Tensión constitutiva irreducible],
+  [Dualidad ontología / reflection como tensión de diseño],
+
+  [*Mundos posibles*],
+  [Leibniz (1710) _Teodicea_ \ Kripke (1980) _Naming and Necessity_],
+  [FSM — transiciones permitidas],
+  [ADRs como restricciones de accesibilidad entre estados],
+
+  [*Propiedades rígidas / esenciales*],
+  [Kripke (1980) \ _Naming and Necessity_],
+  [Invariantes como necesidad, no contingencia],
+  [Un invariante es verdad en todos los estados posibles del proyecto],
+
+  [*Límites del lenguaje = límites del mundo*],
+  [Wittgenstein (1921) \ _Tractatus Logico-Philosophicus_],
+  [Schema como frontera de lo expresable],
+  [Lo que no tipea en Nickel no puede existir en el sistema],
+
+  [*Juegos de lenguaje*],
+  [Wittgenstein (1953) \ _Investigaciones filosóficas_],
+  [Modos operacionales],
+  [Un modo define las reglas del juego para un contexto operacional],
+
+  [*Verdad como correspondencia*],
+  [Tarski (1933) \ _El concepto de verdad_],
+  [Typecheck / validación formal],
+  [`nickel export` como verificación de correspondencia con el schema],
+
+  [*Hechos institucionales*],
+  [Searle (1995) \ _La construcción de la realidad social_],
+  [ADRs como decisiones declaradas],
+  [Una constraint existe porque el equipo la declaró y comprometió],
+
+  [*Ontología formal vs. regional*],
+  [Husserl (1913) \ _Ideas I_],
+  [Schema formal vs. ontología del proyecto],
+  [`adrs/schema.ncl` (formal) vs. `.ontology/core.ncl` (regional)],
+
+  [*Filosofía del proceso*],
+  [Whitehead (1929) \ _Process and Reality_],
+  [Ontología (ser) vs. Reflection (devenir)],
+  [La dualidad estructural del protocolo — capas con ritmos distintos],
+
+  [*Realismo estructural*],
+  [Ladyman & Ross (2007) \ _Every Thing Must Go_],
+  [Grafo como lo real, no las entidades],
+  [El DAG captura relaciones; los nodos son secundarios],
+
+  [*Lógica de predicados*],
+  [Frege (1879) \ _Begriffsschrift_],
+  [Queries sobre el grafo],
+  [`where severity == "Hard" | select id claim`],
+)
+
+#v(0.8em)
+
+== Matices que la tabla no captura
+
+*Aristóteles es el arquitecto base.* La esencia/accidente es exactamente invariante/estado-mutable. La potencia/acto es exactamente `current_state`/`desired_state` con la gate como umbral de actualización. Que esto haya tardado 2.300 años en aparecer explícitamente en software es indicativo de cuánto tiempo estuvo implícito en las decisiones de diseño.
+
+*Whitehead contra Aristóteles es la tensión interna del modelo.* Aristóteles ve la realidad como sustancias con propiedades. Whitehead la ve como procesos — no hay cosas, hay eventos. Un modelo operacional de software vive en esa grieta: la ontología es aristotélica (qué _es_ el proyecto), la capa de reflection es whiteheadiana (cómo _deviene_). Mantenerlas separadas es un reconocimiento implícito de que ninguna basta sola.
+
+*Kripke resuelve lo que Aristóteles dejó ambiguo.* Aristóteles distingue esencial de accidental pero no formaliza la modalidad. Kripke lo hace: una propiedad esencial es rígida — verdadera en todos los mundos posibles accesibles desde el individuo. Los invariantes operacionales tienen exactamente este carácter modal. Los ADRs que los modifican no violan los invariantes — constituyen un nuevo individuo.
+
+*La ausencia notable: Heidegger.* El _ser-en-el-mundo_ como condición de posibilidad de cualquier ontología explícita no aparece en ningún modelo operacional de software. Lo que Heidegger llamaría el trasfondo pre-temático — el contexto que no se puede formalizar porque es la condición de posibilidad de la formalización — es exactamente lo que los READMEs intentan capturar y fracasan. Un modelo operacional no lo resuelve; lo que hace es reducir la superficie de ese trasfondo haciendo explícito lo que puede serlo.
+
+#pagebreak()
+
+// ─── 3. QUIÉN CREÓ LA ONTOLOGÍA OPERACIONAL ────────────────────────────────
+
+= La ontología operacional en software: origen, autores y uso actual
+
+== El problema que había que resolver
+
+A finales de los años 1970, los sistemas de inteligencia artificial enfrentaban un problema conocido como *el cuello de botella del conocimiento*: los sistemas expertos eran eficaces en dominios muy acotados pero no podían compartir conocimiento entre sí ni razonar fuera de su dominio específico. Cada sistema construía su representación del mundo de forma ad hoc, incompatible con el resto.
+
+La pregunta que emergió fue: ¿es posible especificar formalmente el conocimiento de un dominio de manera que múltiples sistemas puedan compartirlo y razonar sobre él?
+
+== Los fundadores
+
+=== John McCarthy — el precursor (1958–1980s)
+
+John McCarthy, fundador del término "inteligencia artificial" y creador de LISP, fue el primero en proponer que los sistemas de IA necesitaban _representaciones explícitas del mundo_ para razonar sobre él. Su noción de *situational calculus* (1969, con Patrick Hayes) — un lenguaje formal para describir el mundo y sus cambios — es el antecedente directo de la ontología operacional.
+
+McCarthy adoptó el término "ontología" de la filosofía para referirse a los compromisos sobre qué tipos de cosas existen en el dominio modelado. No era una metáfora — era un préstamo deliberado de la tradición aristotélica.
+
+=== Marvin Minsky — frames y estructuras de conocimiento (1974)
+
+En su influyente artículo _A Framework for Representing Knowledge_ (MIT AI Lab, 1974), Minsky propuso los *frames* como estructuras para organizar el conocimiento: cada frame representa un concepto con slots (atributos) y valores por defecto. Es la primera implementación computacional de algo cercano a la sustancia aristotélica con propiedades.
+
+Los frames son el antecedente directo de los esquemas orientados a objetos y, más tarde, de las ontologías formales.
+
+=== Tom Gruber — la definición canónica (1993)
+
+Tom Gruber, en el Knowledge Systems Laboratory de Stanford, formuló la definición que se convirtió en estándar: _"An ontology is a specification of a conceptualization"_ (_A translation approach to portable ontology specifications_, Knowledge Acquisition, 1993).
+
+La definición es deliberadamente minimalista: una ontología no captura la realidad — captura un _punto de vista compartido_ sobre un dominio. Esto la aleja del realismo metafísico y la sitúa en el territorio de los hechos institucionales de Searle.
+
+Gruber también desarrolló ONTOLINGUA, el primer sistema formal para escribir y compartir ontologías entre sistemas de IA heterogéneos.
+
+=== Nicola Guarino — ontología formal en sistemas de información (1990s)
+
+Nicola Guarino, del Instituto de Ciencias y Tecnologías Cognitivas (ISTC-CNR) de Italia, desarrolló la distinción entre *ontología de alto nivel* (conceptos universales como tiempo, espacio, proceso, objeto) y *ontología de dominio* (conceptos específicos de un campo).
+
+Fundó la conferencia _Formal Ontology in Information Systems_ (FOIS, 1998), aún activa, que establece el puente entre la filosofía analítica y la ingeniería del conocimiento. Su trabajo sobre la distinción entre ontologías *descriptivas* (cómo los agentes conceptualizan el mundo) y *prescriptivas* (cómo deben conceptualizarlo para interoperar) es directamente aplicable a los modelos operacionales de software.
+
+=== Doug Lenat — Cyc: el proyecto más ambicioso (1984–2017)
+
+Doug Lenat inició el proyecto *Cyc* en 1984 en MCC (Microelectronics and Computer Technology Corporation), con el objetivo de codificar todo el conocimiento de sentido común humano como una ontología operacional. Fue el proyecto de representación de conocimiento más grande y longevo de la historia.
+
+Cyc alcanzó más de 600.000 conceptos y 5 millones de aserciones. Demostró tanto el potencial como los límites del enfoque: la escala del conocimiento implícito humano es prácticamente inabarcable mediante codificación manual.
+
+=== Tim Berners-Lee, James Hendler y Ora Lassila — la Web Semántica (2001)
+
+En su artículo seminal en _Scientific American_ (2001), propusieron extender la Web con significado semántico: páginas que no solo humanos sino también máquinas pudieran "entender". La ontología era el mecanismo de interoperabilidad semántica.
+
+Esto produjo las tecnologías RDF (Resource Description Framework), OWL (Web Ontology Language) y SPARQL — el ecosistema de la Web Semántica — que popularizaron el uso de ontologías en ingeniería de software más allá de la IA.
+
+== Línea temporal condensada
+
+#table(
+  columns: (1fr, 3fr),
+  inset: 7pt,
+  fill: (col, row) => if row == 0 { rgb("#c04000") } else if calc.odd(row) { rgb("#fdf6f0") } else { white },
+  table.header(
+    text(fill: white, weight: "bold")[Año],
+    text(fill: white, weight: "bold")[Hito],
+  ),
+  [1958], [McCarthy propone representación explícita del conocimiento para IA],
+  [1969], [McCarthy & Hayes — situational calculus: lógica de cambios en el mundo],
+  [1974], [Minsky — frames: primera estructura computacional análoga a la sustancia aristotélica],
+  [1984], [Lenat inicia Cyc — codificación masiva de conocimiento de sentido común],
+  [1991], [Primeras ontologías de alto nivel: DOLCE, SUMO — conceptos universales compartidos],
+  [1993], [Gruber — definición canónica: "ontología = especificación de una conceptualización"],
+  [1998], [Primera conferencia FOIS (Formal Ontology in Information Systems)],
+  [2001], [Berners-Lee et al. — Web Semántica: RDF, OWL, SPARQL],
+  [2004], [OWL se convierte en recomendación W3C — estándar industrial],
+  [2006–2010], [Wikidata, DBpedia — ontologías abiertas a escala masiva],
+  [2012], [Google Knowledge Graph — ontología operacional en producción para miles de millones de usuarios],
+  [2020s], [LLMs entrenados sobre grafos de conocimiento ontológicos; ontologías como contexto estructurado para agentes de IA],
+)
+
+== Uso actual
+
+=== Biomedicina — el dominio más maduro
+
+La ontología operacional tiene su mayor penetración en ciencias biomédicas:
+
+- *Gene Ontology (GO)*: iniciada en 1998, describe funciones génicas de forma estandarizada entre organismos. Usada en todos los laboratorios de biología molecular del mundo.
+- *SNOMED CT*: más de 350.000 conceptos médicos con relaciones formales. Estándar en sistemas de historia clínica electrónica en Europa, EEUU, Australia.
+- *ICD-11* (OMS): clasificación internacional de enfermedades en formato ontológico desde 2018.
+- *Human Phenotype Ontology (HPO)*: diagnóstico de enfermedades raras mediante matching ontológico entre síntomas del paciente y fenotipos conocidos.
+
+=== Grafos de conocimiento empresariales
+
+- *Google Knowledge Graph* (2012): más de 500.000 millones de hechos estructurados que alimentan los resultados de búsqueda y el panel de conocimiento.
+- *Wikidata*: ontología abierta con más de 100 millones de elementos, mantenida colaborativamente, usada como fuente de conocimiento por Wikimedia, investigación académica y sistemas de IA.
+- *schema.org*: vocabulario ontológico estándar para marcado semántico de páginas web, creado por Google, Microsoft, Yahoo y Yandex. Presente en la mayoría de sitios web modernos.
+
+=== Industria y manufactura
+
+- *ISO 15926* (petróleo y gas): ontología para integración de datos a lo largo del ciclo de vida de plantas industriales. Usada por Shell, BP, Statoil.
+- *IFC* (Industry Foundation Classes): ontología para información de construcción. Base del estándar BIM (Building Information Modeling) en arquitectura e ingeniería civil.
+- *Industry 4.0 ontologies*: RAMI 4.0, Asset Administration Shell — representación formal de activos industriales en fábricas conectadas.
+
+=== Inteligencia artificial y agentes
+
+- *ConceptNet*: grafo de conocimiento de sentido común con 21 millones de aserciones, usado en entrenamiento de modelos de lenguaje.
+- *Entrenamiento de LLMs*: Wikidata y DBpedia son fuentes estructuradas en los corpus de entrenamiento de GPT, LLaMA y modelos similares.
+- *Contexto para agentes*: el uso emergente más relevante — ontologías como contexto estructurado que se inyecta en agentes de IA para que operen con comprensión del dominio en lugar de inferirlo del lenguaje natural.
+
+=== Software de proyecto — la frontera actual
+
+El uso de ontologías operacionales _dentro_ de proyectos de software — no para describir dominios externos sino para describir el propio proyecto — es el desarrollo más reciente y menos formalizado.
+
+Los precursores son los ADRs (Architecture Decision Records, propuestos por Michael Nygard en 2011) y los modelos C4 (Simon Brown). Pero estos son documentación, no grafos consultables.
+
+La propuesta de *ontoref* pertenece a esta frontera: aplicar las estructuras de la ontología formal — invariantes tipados, tensiones explícitas, estados formalizados, gates verificables — al proyecto de software como entidad que se describe a sí misma. El proyecto _es_ el dominio, y la ontología _es_ el modelo operacional de ese dominio.
+
+Lo que distingue este enfoque de la documentación tradicional es que el modelo es:
+
+- *Consultable*: mediante pipelines estructuradas, no búsqueda de texto
+- *Verificable*: el typecheck valida la coherencia en cada commit
+- *Ejecutable*: los modos operacionales son grafos acíclicos que el sistema puede ejecutar
+- *Machine-readable*: los agentes de IA pueden consumirlo directamente como contexto estructurado
+
+== Conclusión
+
+La ontología operacional en software no es una importación reciente de la filosofía — es una convergencia que lleva setenta años ocurriendo, desde los frames de Minsky hasta los grafos de conocimiento de Google. Lo que ha cambiado con la proliferación de agentes de IA es la urgencia: cuando un agente puede modificar código, infraestructura o configuración de forma autónoma, la ausencia de un modelo explícito del proyecto no es una deuda técnica. Es una pérdida de control.
+
+La ontología no es la solución a ese problema. Es el nombre del problema que hay que resolver.
+
+#pagebreak()
+
+// ─── 4. TAOÍSMO Y ONTOLOGÍA OPERACIONAL ────────────────────────────────────
+
+= Taoísmo, Yin-Yang y ontología operacional
+
+== Lo que la tradición occidental no vio
+
+La ontología occidental desde Parménides privilegia el *ser* sobre el *no-ser*, la *presencia* sobre la *ausencia*, la *identidad* sobre la *diferencia*. Toda la tradición aristotélica-escolástica-analítica trabaja desde ese punto de partida.
+
+El Taoísmo invierte la pregunta. El Tao (道) es previo a la distinción ser/no-ser. El primer capítulo del _Tao Te Ching_ (Laozi, ~siglo VI-IV aC):
+
+#align(center)[
+  #text(style: "italic")[
+    道可道，非常道。名可名，非常名。\
+    El Tao que puede nombrarse no es el Tao eterno.\
+    El nombre que puede pronunciarse no es el nombre eterno.
+  ]
+]
+
+Lo que no puede formalizarse no es una limitación del modelo — es la condición de posibilidad de todo lo que sí puede formalizarse. La ontología operacional que puede expresarse completamente en un schema no captura la realidad operacional completa. Heidegger llegó al mismo lugar por camino diferente: el *trasfondo pre-temático* que hace posible cualquier acto de comprensión explícita es precisamente lo que escapa a la formalización.
+
+== Yin-Yang — polaridad constitutiva, no oposición
+
+La confusión más frecuente: yin y yang no son *contrarios* sino *complementarios que se definen mutuamente*. Las propiedades estructurales del par son:
+
+- Cada polo *contiene una semilla del opuesto* — representada por el punto de color contrario en el símbolo
+- El exceso de un polo *genera su contrario* (極則反 — la extremidad se invierte)
+- La tensión entre ellos *nunca se resuelve* en síntesis — solo se gestiona dinámicamente
+- Uno *no existe sin el otro* — yin sin yang es vacío inerte, yang sin yin es movimiento sin forma
+
+Esto es estructuralmente distinto de la dialéctica hegeliana, que busca síntesis — resolver la contradicción en una unidad superior. El yin-yang no resuelve. La tensión es constitutiva, permanente. Hegel usa la contradicción como *palanca*; el Taoísmo la habita como *condición*.
+
+Heráclito (535-475 aC) es el occidental más cercano: su logos como tensión dinámica de opuestos — "el camino hacia arriba y hacia abajo son el mismo" — es estructuralmente idéntico al yin-yang. La diferencia: Heráclito vive en la tensión, Hegel la usa para superarla.
+
+== Correspondencias con el modelo operacional
+
+#table(
+  columns: (1.4fr, 1.3fr, 1.8fr, 1.8fr),
+  inset: 7pt,
+  fill: (col, row) => if row == 0 { rgb("#c04000") } else if calc.odd(row) { rgb("#fdf6f0") } else { white },
+  table.header(
+    text(fill: white, weight: "bold")[Concepto taoísta],
+    text(fill: white, weight: "bold")[Texto / Época],
+    text(fill: white, weight: "bold")[Equivalente occidental más cercano],
+    text(fill: white, weight: "bold")[Correspondencia operacional],
+  ),
+  [*Tao* (道) — principio previo a toda distinción],
+  [Laozi, _Tao Te Ching_ \ (~6-4 aC)],
+  [Ser de Heidegger — previo a los entes],
+  [Trasfondo pre-temático que ningún schema captura completamente],
+
+  [*Yin / Yang* — polaridad que no se resuelve],
+  [Tradición Zhou (~1100 aC) \ sistematizada en _I Ching_],
+  [Heráclito — logos como tensión de opuestos],
+  [Ontología (ser) / Reflection (devenir) — capas que se necesitan mutuamente],
+
+  [*Wu wei* (無為) — acción sin forzar],
+  [Laozi, _Tao Te Ching_],
+  [Sin equivalente exacto en Occidente],
+  ["No enforcement" — coherencia que emerge de adopción justificada, no impuesta],
+
+  [*Ji* (機) — el momento exacto de transformación],
+  [Zhuangzi, _Zhuangzi_ \ (~4 aC)],
+  [Kripke — mundos posibles accesibles],
+  [Gates — umbral cualitativo entre estados, no condición binaria],
+
+  [*Wu ji → Tai ji* — de lo indiferenciado a la polaridad],
+  [Zhou Dunyi (1017-1073) \ _Taijitu shuo_],
+  [Husserl — ontología formal vs. regional],
+  [schema.ncl (vacío universal) → core.ncl (polaridad concreta del proyecto)],
+
+  [*Relatividad perspectival*],
+  [Zhuangzi — la rana del pozo],
+  [Wittgenstein tardío — juegos de lenguaje],
+  [Toda ontología es una conceptualización, no una captura de la realidad (Gruber)],
+)
+
+== Las cuatro correspondencias profundas
+
+*Yin/Yang ↔ Ontología/Reflection.* La dualidad ya está en el diseño del protocolo. La ontología (core.ncl) captura lo que ES — invariantes, estructura, ser; capa yang: formal, dura, definida. La reflection captura lo que DEVIENE — operaciones, modos, movimiento; capa yin: fluida, procesual. Como en el símbolo: cada una contiene una semilla de la otra. La ontología describe estados *futuros* (desired_state) — contiene movimiento. La reflection puede *actualizar* la ontología vía ADRs — contiene forma. Separarlas es necesario porque tienen ritmos de cambio distintos; fundirlas colapsaría el yin en el yang.
+
+*Wu wei ↔ "No enforcement".* Uno de los invariantes del protocolo: coherencia voluntaria que emerge de la adopción justificada, sin mecanismo de enforcement. Esto es wu wei aplicado a diseño de sistemas. No impones — creas las condiciones para que el alineamiento emerja naturalmente. Un sistema de tipos fuertes fuerza; una ontología operacional orienta. La diferencia es exactamente la diferencia entre imponer el Tao y fluir con él.
+
+*Tensiones ↔ Yin-Yang permanente.* Las tensiones en un modelo operacional ("formalización vs. fricción de adopción", "ontología vs. reflection") son explícitamente irresolubles. No se busca síntesis — se gestionan. Si eliminas la tensión entre formalización y adopción maximizando uno de los polos, el sistema muere: o es un schema perfecto que nadie adopta, o es caos operacional que todos adoptan pero nadie puede razonar sobre él.
+
+*Gates ↔ Ji (機).* El concepto taoísta de ji es el punto de inflexión donde el potencial se actualiza — no un estado sino el umbral entre estados. Dinámico, cualitativo, contextual. Las gates en el modelo operacional son exactamente esto: no checkboxes — son condiciones que marcan el momento en que un estado puede transformarse en otro.
+
+== La crítica taoísta a la ontología clasificatoria
+
+Los sistemas como OWL intentan eliminar la ambigüedad mediante clasificación precisa. El _Tao Te Ching_ advierte: cuanto más preciso el nombre, más excluyes. La realidad vive en el espacio *entre* las categorías, no en las categorías mismas.
+
+Esto explica por qué las ontologías biomédicas — las más maduras del campo — sufren continuamente con casos límite, dependencia del contexto y la necesidad de múltiples ejes de clasificación simultáneos. La vista taoísta diría: estáis intentando resolver una tensión constitutiva. La precisión y la cobertura son yin y yang — no se resuelven, se equilibran.
+
+Zhuangzi va más lejos que Laozi. Toda distinción ontológica es perspectival. Esto no lleva al nihilismo sino a la *apropiada contextualidad*: la ontología correcta para un proyecto no es la más formalmente precisa sino la que el equipo puede sostener y evolucionar. Gruber llegó al mismo lugar en 1993 sin saberlo: una ontología es "una especificación de una conceptualización" — un punto de vista compartido, no una verdad metafísica.
+
+#pagebreak()
+
+// ─── 5. DAGs ────────────────────────────────────────────────────────────────
+
+= DAGs — genealogía y función en ontología operacional
+
+== El problema que fuerza el grafo
+
+Antes del grafo hay un problema: ¿cómo representar conocimiento estructurado de forma que se pueda razonar sobre él? La prosa no es consultable ni verificable. Las listas planas pierden las relaciones. Las jerarquías de árbol pierden los nodos con múltiples padres. El grafo emerge como la estructura mínima que captura relaciones arbitrarias entre conceptos.
+
+Pero el grafo general tiene un problema: puede tener ciclos. Y los ciclos en ontología son epistemológicamente problemáticos.
+
+== Por qué la aciclicidad — el problema de la justificación circular
+
+Dos tradiciones independientes llegan al mismo resultado:
+
+*Aristóteles* (_Analíticos Posteriores_): una definición válida procede de género más diferencia específica. El género es más general que lo definido — la jerarquía es estrictamente descendente. Si A se define en términos de B y B en términos de A, ninguno está definido. La justificación circular es un defecto lógico.
+
+*Russell* (1901): la paradoja del conjunto de todos los conjuntos que no se contienen a sí mismos. La solución en _Principia Mathematica_ (1910, con Whitehead) es la *teoría de tipos*: los tipos forman una jerarquía estricta — un tipo de nivel N solo puede referirse a tipos de nivel N-1 o inferior. Esto es estructuralmente un DAG. La aciclicidad no es una restricción arbitraria — es la solución formal al colapso por autorreferencia.
+
+La convergencia: en lógica, ontología formal y teoría de tipos, la aciclicidad es el requisito de que las definiciones sean *bien fundadas*. No puede haber descenso infinito ni circularidad.
+
+== Línea genealógica
+
+#table(
+  columns: (0.8fr, 2fr, 2.5fr),
+  inset: 7pt,
+  fill: (col, row) => if row == 0 { rgb("#c04000") } else if calc.odd(row) { rgb("#fdf6f0") } else { white },
+  table.header(
+    text(fill: white, weight: "bold")[Año],
+    text(fill: white, weight: "bold")[Autor / Hito],
+    text(fill: white, weight: "bold")[Contribución al DAG en ontología],
+  ),
+  [384-322 aC],
+  [Aristóteles — _Analíticos Posteriores_],
+  [Definición bien fundada: jerarquía sin circularidad. Género → diferencia → especie.],
+
+  [1879],
+  [Frege — _Begriffsschrift_],
+  [Lógica de predicados: relaciones formales entre conceptos. Base para representar grafos de conocimiento.],
+
+  [1901-1910],
+  [Russell & Whitehead — _Principia Mathematica_],
+  [Teoría de tipos como DAG: aciclicidad como solución al colapso por autorreferencia.],
+
+  [1933],
+  [Tarski — "El concepto de verdad"],
+  [Verdad como correspondencia entre proposiciones y estructuras (grafos). Base de la validación formal.],
+
+  [1958],
+  [US Navy — PERT charts (programa Polaris)],
+  [DAG para planificación de proyectos: camino crítico como recorrido del grafo de dependencias temporales.],
+
+  [1969],
+  [Collins & Quillian — "Retrieval time from semantic memory"],
+  [Primera red semántica computacional: is-a como DAG. La topología del grafo predice tiempos de reacción humanos.],
+
+  [1974],
+  [Minsky — "A Framework for Representing Knowledge"],
+  [Frames con herencia múltiple: DAG sobre árbol. Primer sistema con nodos de múltiples padres.],
+
+  [1976],
+  [Feldman — Make],
+  [Build system como DAG de dependencias. Orden topológico para compilación incremental. Ciclos = error explícito.],
+
+  [1985],
+  [Brachman & Schmolze — KL-ONE],
+  [Description Logic: subsunción formal como DAG. Antecedente directo de OWL.],
+
+  [1988],
+  [Pearl — _Probabilistic Reasoning in Intelligent Systems_],
+  [Causalidad como DAG: la asimetría temporal hace la causalidad inherentemente acíclica.],
+
+  [2000],
+  [Pearl — _Causality_],
+  [Do-calculus: intervención como operación sobre el DAG causal. Separación entre correlación y causa.],
+
+  [2001],
+  [Berners-Lee et al. — Web Semántica],
+  [RDF/OWL: grafos de conocimiento como estándar industrial. DAG de subsunción en jerarquías de clases.],
+
+  [2005],
+  [Torvalds — Git],
+  [Historia de commits como DAG: irreversibilidad del tiempo en control de versiones. Merges = nodos con múltiples padres.],
+)
+
+== Las cuatro funciones del DAG en ontología operacional
+
+*① Traversal determinista para análisis de impacto.* Si cambias un nodo, el DAG define exactamente qué nodos son alcanzables desde él (descendientes) y qué nodos lo alcanzan (ancestros). Esto es análisis de impacto: "si modifico este invariante, ¿qué tensiones, prácticas y gates se ven afectadas?" En un grafo cíclico esto es indecidible — el cierre transitivo de un nodo puede ser todo el grafo.
+
+*② Orden topológico para ejecución.* Los modos operacionales son DAGs de pasos con dependencias. El ordenamiento topológico da el orden de ejecución válido sin necesidad de un scheduler externo. Si el paso B depende del paso A, A siempre precede a B en cualquier topological sort.
+
+*③ Justificación bien fundada — sin circularidad epistémica.* El debate clásico en epistemología enfrenta fundacionalismo (el conocimiento descansa en fundamentos no circulares — DAG con nodos raíz) contra coherentismo (red de creencias mutuamente sustentadas — puede ser cíclica). La ontología operacional es fundacionalista por diseño: los invariantes son nodos raíz justificados por ADR explícito, no por otros nodos del grafo. Un ciclo significaría: "este invariante existe porque esta práctica existe, y esta práctica existe porque este invariante existe" — dogma circular, no arquitectura.
+
+*④ Rollback con orden definido.* Deshacer cambios requiere un orden inverso bien definido. En un grafo cíclico no existe ese orden. El DAG garantiza que el rollback siempre tiene una secuencia válida.
+
+== La tensión con el Taoísmo — DAG vs. ciclo
+
+El DAG es *yang* en su estructura: direccional, jerárquico, asimétrico, irreversible. Codifica la flecha del tiempo — las causas preceden a los efectos, las definiciones preceden a lo definido.
+
+El yin-yang taoísta es *cíclico*: yin genera yang, yang genera yin. Las estaciones, el día y la noche, vida y muerte — todo retorna.
+
+La resolución está en separar dos temporalidades:
+
+- *Sincrónica* (snapshot): el DAG es válido dentro de un momento dado. La ontología en $t_0$ es un grafo acíclico.
+- *Diacrónica* (evolución): la secuencia de snapshots puede tener patrones cíclicos. Lo que era tensión puede convertirse en invariante; lo que era práctica puede ser abandonado y redescubierto.
+
+El DAG captura la *estructura del ser en un momento*. El ciclo taoísta captura el *proceso de transformación a través del tiempo*. Son complementarios, no contradictorios — exactamente como ontología y reflection son complementarios en el protocolo. El *I Ching* — el libro de los cambios — es un sistema para razonar sobre las transformaciones de un estado a otro: 64 hexagramas como estados (nodos), transformaciones entre ellos como aristas. El sistema completo es un grafo. No un DAG — un grafo con ciclos, porque la realidad vuelve.
+
+#pagebreak()
+
+// ─── 6. FÍSICA CUÁNTICA ─────────────────────────────────────────────────────
+
+= Física cuántica, computación cuántica y ontología operacional
+
+== El golpe más profundo — Aristóteles estaba equivocado en el nivel fundamental
+
+La ontología aristotélica asume que las sustancias tienen propiedades definidas. La mecánica cuántica demuestra que, en el nivel fundamental, las propiedades no son definidas hasta que se miden. Un electrón no *tiene* posición — la posición emerge del acto de medición.
+
+Esto no es ignorancia del observador sobre una propiedad preexistente. El *teorema de Kochen-Specker* (1967) lo prueba formalmente: es imposible asignar valores definidos a todos los observables cuánticos simultáneamente, incluso antes de la medición. La indeterminación es *ontológica*, no epistémica. El suelo sobre el que se construyeron 2.300 años de ontología occidental se mueve.
+
+== Las tres conexiones genuinas
+
+=== Complementariedad de Bohr ↔ Yin-Yang y la dualidad Ontología/Reflection
+
+Niels Bohr (1885-1962) formuló el *principio de complementariedad*: onda y partícula son dos descripciones mutuamente excluyentes pero igualmente necesarias del mismo fenómeno. No puedes observar ambas simultáneamente — el aparato de medición determina qué aspecto revelas.
+
+Bohr lo conectó explícitamente con el Taoísmo: cuando fue ennoblecido en 1947 adoptó el símbolo yin-yang como parte de su escudo de armas con la inscripción _Contraria sunt complementa_ — los contrarios son complementarios.
+
+La dualidad ontología/reflection del modelo operacional tiene la misma estructura. El "aparato de medición" — qué preguntas haces al sistema — determina qué aspecto revelas. La capa que uses (core.ncl o los modos de reflection) determina qué ves. Ninguna es más verdadera — son complementarias.
+
+=== Contextualidad cuántica ↔ Ontología perspectival
+
+El teorema de Kochen-Specker prueba algo más radical que la incertidumbre de Heisenberg: el resultado de medir un observable cuántico depende de *qué otros observables se miden simultáneamente*. Las propiedades son *contextuales* — no existen independientemente del contexto de medición.
+
+Esto convierte en hecho físico lo que tres tradiciones de esta discusión habían sostenido filosóficamente:
+
+- *Zhuangzi*: toda distinción ontológica es perspectival
+- *Wittgenstein tardío*: el significado depende del juego de lenguaje
+- *Gruber*: una ontología es una especificación de una conceptualización, no una captura de la realidad
+
+Para la ontología operacional: el "estado" de un proyecto es siempre relativo al actor y al contexto de consulta. Un agente de IA y un desarrollador humano consultando el mismo core.ncl están haciendo mediciones con aparatos distintos. Lo que ven es complementario, no idéntico.
+
+=== El problema del observador ↔ El trasfondo que no puede formalizarse
+
+En mecánica cuántica, la medición colapsa la función de onda — el observador es parte del sistema. No hay _view from nowhere_ (Nagel, 1986). Todo conocimiento es conocimiento *desde* una posición.
+
+Esto es exactamente lo que Heidegger llamó el trasfondo pre-temático — la condición de posibilidad de cualquier formalización que ella misma no puede ser formalizada. Y lo que el _Tao Te Ching_ afirma: el Tao que puede nombrarse no es el Tao eterno.
+
+La ontología operacional que crees que describes objetivamente está siendo constituida por el acto de escribirla. Esto no es un defecto del modelo — es un límite estructural de toda representación.
+
+== La interpretación más relevante — Mecánica Cuántica Relacional
+
+Carlo Rovelli propuso en 1996 (_Relational Quantum Mechanics_, International Journal of Theoretical Physics) que no existen estados absolutos — solo estados *relativos a otros sistemas*. El estado de una partícula no es una propiedad intrínseca — es una propiedad relacional respecto al sistema con el que interactúa.
+
+Esto es la ontología más radical posible y conecta directamente con Ladyman y Ross (realismo estructural): no hay propiedades, solo relaciones. Lo real son las estructuras relacionales, no las entidades.
+
+Para el modelo operacional: el "estado" de un proyecto (current_state en state.ncl) no es una propiedad intrínseca del proyecto — es relacional respecto al actor que consulta, los objetivos perseguidos, el momento del ciclo de vida. Dos actores con marcos de referencia distintos pueden tener mediciones del estado coherentes pero distintas. No hay contradicción — hay contextualidad relacional.
+
+== La computación cuántica — honestidad sobre sus límites
+
+#table(
+  columns: (1.5fr, 2fr, 2fr),
+  inset: 7pt,
+  fill: (col, row) => if row == 0 { rgb("#c04000") } else if calc.odd(row) { rgb("#fdf6f0") } else { white },
+  table.header(
+    text(fill: white, weight: "bold")[Aspecto],
+    text(fill: white, weight: "bold")[Situación real (2026)],
+    text(fill: white, weight: "bold")[Relevancia para ontología operacional],
+  ),
+  [Tamaño de grafos],
+  [La ventaja cuántica aparece a escala masiva],
+  [Los grafos de ontología de proyecto son pequeños — sin ventaja cuántica],
+
+  [Algoritmos relevantes],
+  [Grover: O(√N) búsqueda. Shor: factorización polinómica],
+  [No mapean a consultas típicas sobre ontologías de proyecto],
+
+  [Decoherencia],
+  [Cualquier interacción colapsa la superposición],
+  [El uso práctico requiere aislamiento que contradice operación continua],
+
+  [Redes semánticas cuánticas],
+  [Investigación activa — relaciones con amplitudes de probabilidad],
+  [Captura gradualidad que OWL binario pierde — relevante a mediano plazo],
+
+  [Bayesian networks cuánticas],
+  [Extensión de Pearl con amplitudes complejas],
+  [Modelar dependencias con comportamiento emergente no reducible],
+
+  [LLMs y espacios de Hilbert],
+  [Los embeddings son productos internos en espacio vectorial de alta dimensión],
+  [Directamente relevante: la misma matemática cuántica, aplicada a semántica],
+)
+
+== La conexión no obvia — LLMs como geometría cuántica aplicada
+
+Los Large Language Models operan en espacios vectoriales de alta dimensión donde el significado es relacional y contextual. No es metáfora cuántica — es la misma matemática de espacios de Hilbert que usa la mecánica cuántica, aplicada a representación semántica.
+
+La ontología operacional (core.ncl, invariantes, tensiones) es, en este contexto, una *base* en ese espacio — un conjunto de vectores que fijan el marco de referencia en el que el agente opera. Inyectar el contexto ontológico al inicio de una sesión es, en la geometría de estos modelos, fijar el marco de referencia antes de que el agente empiece a computar.
+
+La física cuántica importa aquí no porque vayamos a correr ontologías en computadores cuánticos — sino porque demuestra que las asunciones ontológicas clásicas (propiedades definidas, estados absolutos, observador separado del sistema) son físicamente incorrectas en el nivel fundamental. La ontología operacional es contextual, relacional y perspectival *por diseño correcto*, no por limitación.
+
+#pagebreak()
+
+// ─── 7. RELEVANCIA EN SOFTWARE E INFRA ──────────────────────────────────────
+
+= Por qué es relevante en software e infraestructuras digitales
+
+== El problema real no es técnico — es epistémico
+
+El software moderno tiene un déficit estructural que no se resuelve con mejores herramientas:
+
+#align(center)[
+  #block(
+    fill: rgb("#1a1a1a"),
+    radius: 4pt,
+    inset: 12pt,
+    width: 90%,
+  )[
+    #set text(font: "Courier New", size: 9pt, fill: rgb("#cccccc"))
+    Lo que el sistema hace           ≠  Lo que el equipo cree que hace \
+    Lo que la infra tiene desplegado ≠  Lo que Terraform declara \
+    Lo que el ADR decidió            ≠  Lo que el código implementa \
+    Lo que el agente asume           ≠  Lo que el contexto real requiere
+  ]
+]
+
+Cada una de esas brechas es silenciosa, acumulativa, y solo se descubre cuando falla en producción. No son fallos de las personas — son consecuencias inevitables de operar un sistema complejo sin un modelo del propio sistema.
+
+== Lo que la ontología operacional hace que nada más hace
+
+#table(
+  columns: (1.5fr, 1.5fr),
+  inset: 7pt,
+  fill: (col, row) => if row == 0 { rgb("#c04000") } else if calc.odd(row) { rgb("#fdf6f0") } else { white },
+  table.header(
+    text(fill: white, weight: "bold")[Documentación tradicional],
+    text(fill: white, weight: "bold")[Ontología operacional],
+  ),
+  [Prosa — no consultable],
+  [Grafo tipado — consultable con queries estructuradas],
+
+  [Estática — se desactualiza sin señal],
+  [Verificada en cada commit mediante typecheck],
+
+  [Describe lo que hubo],
+  [Formaliza lo que ES y lo que debe ser],
+
+  [Solo legible por humanos],
+  [Legible por humanos, agentes de IA y pipelines de CI],
+
+  [Captura el qué],
+  [Captura el qué, el por qué, y lo que no puede cambiar],
+
+  [Nadie la actualiza bajo presión],
+  [El proceso la actualiza — es el path, no documentación paralela],
+)
+
+*Formaliza lo que no puede romperse.* Los invariantes no son comentarios — son constraints tipadas que el sistema verifica antes del runtime, no después.
+
+*Captura las tensiones activas.* "Coste vs. disponibilidad", "velocidad de despliegue vs. estabilidad" viven en la ontología, no en la cabeza del miembro más veterano del equipo. Cuando ese alguien se va, los trade-offs permanecen.
+
+*Da al agente de IA el contexto que no puede inferir del código.* Un agente puede leer todo el código de un sistema y no saber que hay un acuerdo que impide escalar horizontalmente cierto servicio, o que la decisión de usar gRPC interno fue consecuencia de una tensión específica de latencia que no está en ningún archivo. La ontología operacional hace eso explícito y machine-readable.
+
+== Por qué los DAGs — especificidad técnica
+
+Los DAGs no son una elección estética. Son la única estructura que garantiza cuatro propiedades simultáneamente en sistemas operacionales:
+
+*Análisis de impacto decidible.* Antes de cambiar un invariante o un nodo de infra — recorre el DAG y sabes exactamente qué se ve afectado. Concreto: en Terraform, el grafo de recursos es un DAG. `terraform plan` recorre ese grafo para determinar qué debe crearse, modificarse o destruirse en qué orden. Si hubiera ciclos, el plan sería indecidible.
+
+*Orden topológico para ejecución fiable.* Cualquier operación con dependencias — despliegue, migración, rollback, onboarding — tiene un orden correcto. El DAG lo da sin necesidad de un coordinador central. Este es el principio detrás de Make (1976), Bazel, Gradle, y cualquier build system no trivial.
+
+*Justificación sin circularidad.* Si el invariante A existe porque la práctica B lo requiere, y la práctica B existe porque el invariante A lo define — tienes dogma, no arquitectura. El DAG hace esa circularidad visible e inaceptable.
+
+*Rollback determinista.* El inverso topológico del DAG es siempre computable. Sin DAG, el rollback es un proceso manual que depende de que alguien recuerde el orden correcto.
+
+== Por qué la reflection — la ontología sin operación es un museo
+
+La ontología captura el *ser*. Pero los sistemas de software no solo son — *operan*. La reflection es la capa que hace la ontología ejecutable:
+
+*Modos como DAGs operacionales.* Un procedimiento sin DAG es una lista de pasos que alguien puede ejecutar en cualquier orden, saltar, o ignorar. Un modo como DAG tiene dependencias explícitas — el sistema puede verificar que se ejecutó correctamente antes de avanzar.
+
+*ADRs con constraints tipadas.* Una decisión sin constraints es prosa. Con constraints tipadas es un contrato que el CI puede verificar en cada commit. La pregunta "¿estamos violando alguna decisión arquitectural?" pasa de ser reflexión manual a ser comprobación automática.
+
+*Detección de drift.* La diferencia entre el estado sellado (lo que declaraste que era verdad) y el estado actual (lo que realmente hay desplegado) es medible si y solo si formalizaste el estado sellado. Sin eso, la pregunta "¿ha derivado esto?" no tiene respuesta objetiva.
+
+== Las tres presiones que hacen esto urgente ahora
+
+Estos conceptos no son nuevos. Lo nuevo es que tres presiones convergen simultáneamente haciendo que la ausencia del modelo operacional sea costosa de una forma que antes era tolerable:
+
+*Presión 1: Los sistemas superan la capacidad de comprensión individual.* Un monolito de 2005 podía vivir en la cabeza de dos personas. Un sistema cloud-native de 2026 con decenas de microservicios, múltiples regiones, pipelines automatizados y SLOs por servicio no cabe en ninguna cabeza. El modelo mental individual ha dejado de ser suficiente como mecanismo de coordinación.
+
+*Presión 2: Los agentes de IA tienen acceso de escritura.* Hasta hace poco, una herramienta de IA que generaba código incorrecto producía texto que un humano revisaba antes de ejecutar. Hoy los agentes pueden abrir PRs, modificar configuración, ejecutar Terraform, y desplegar. El ciclo humano-en-el-loop se comprime. Un agente sin contexto ontológico en ese entorno no es un asistente — es un actor autónomo con comprensión incompleta del sistema que modifica.
+
+*Presión 3: Los equipos son multi-actor y distribuidos.* Humanos en zonas horarias distintas, agentes de IA con distintos contextos de sesión, pipelines de CI ejecutando en paralelo — todos escribiendo sobre el mismo sistema sin coordinador central. Sin un protocolo compartido y machine-readable para "qué puede cambiarse, bajo qué condiciones, con qué justificación", la coordinación depende de que todo el mundo lea el mismo Confluence actualizado. Eso nunca funcionó.
+
+== La síntesis
+
+Lo que la ontología operacional, reflection y DAGs aportan en conjunto es algo que ninguna de sus partes aporta sola:
+
+#align(center)[
+  #block(
+    stroke: rgb("#c04000") + 1pt,
+    radius: 4pt,
+    inset: 14pt,
+    width: 88%,
+  )[
+    #set text(size: 11pt)
+    Un sistema que sabe lo que es, puede verificar que sigue siendo \
+    lo que dice ser, y puede comunicarlo a cualquier actor — humano, \
+    agente o CI — de forma estructurada y sin ambigüedad.
+  ]
+]
+
+Eso no es documentación. Es un modelo operacional vivo.
+
+La ausencia de ese modelo no es un problema de disciplina del equipo. Es una propiedad emergente de sistemas complejos operados sin representación explícita. El drift, la amnesia de decisiones, la ilusión de control en infraestructura declarativa — no son fallos de las personas. Son consecuencias inevitables.
+
+La ontología operacional es la respuesta a la pregunta: *¿qué estructura mínima necesita un sistema para conocerse a sí mismo de forma que ese conocimiento pueda ser usado, verificado y compartido?*
+
+#pagebreak()
+
+// ─── 8. REFERENCIAS ─────────────────────────────────────────────────────────
+
+= Referencias y lecturas recomendadas
+
+== Filosofía clásica y ontología
+
+#set par(first-line-indent: 0em)
+
+*Aristóteles.* _Metafísica_. Traducción de Tomás Calvo Martínez. Madrid: Gredos, 1994. [~330 aC] \
+#text(fill: rgb("#666666"), size: 9.5pt)[El libro fundamental para la distinción esencia/accidente, potencia/acto y las categorías. El libro IX sobre potencia es la base del modelo estado-actual/estado-deseado.]
+
+*Aristóteles.* _Analíticos Posteriores_. Traducción de Miguel Candel Sanmartín. Madrid: Gredos, 1988. [~350 aC] \
+#text(fill: rgb("#666666"), size: 9.5pt)[Formaliza el requisito de definiciones bien fundadas — antecedente directo de la aciclicidad en DAGs.]
+
+*Laozi.* _Tao Te Ching_. Traducción y comentarios de Stephen Mitchell. Nueva York: Harper Perennial, 1992. [~siglo VI-IV aC] \
+#text(fill: rgb("#666666"), size: 9.5pt)[Texto fundacional del Taoísmo. Los primeros y últimos capítulos son los más relevantes para la relación con ontología operacional.]
+
+*Zhuangzi.* _Zhuangzi: The Complete Writings_. Traducción de Brook Ziporyn. Indianapolis: Hackett, 2020. [~siglo IV aC] \
+#text(fill: rgb("#666666"), size: 9.5pt)[Especialmente los capítulos internos (1-7). La relatividad perspectival de Zhuangzi como crítica de toda ontología clasificatoria.]
+
+*Frege, G.* _Begriffsschrift_. Halle: Nebert, 1879. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Primera formulación de la lógica de predicados. El antecedente formal de toda representación de conocimiento estructurado.]
+
+*Russell, B. & Whitehead, A.N.* _Principia Mathematica_. Cambridge: Cambridge University Press, 1910-1913. \
+#text(fill: rgb("#666666"), size: 9.5pt)[La teoría de tipos como DAG para evitar la autorreferencia circular. El capítulo introductorio explica el principio del círculo vicioso.]
+
+*Wittgenstein, L.* _Tractatus Logico-Philosophicus_. Londres: Kegan Paul, 1922. \
+#text(fill: rgb("#666666"), size: 9.5pt)["Los límites de mi lenguaje son los límites de mi mundo." El fundamento de por qué el schema Nickel define la frontera de lo representable en el sistema.]
+
+*Wittgenstein, L.* _Investigaciones filosóficas_. Oxford: Blackwell, 1953. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Juegos de lenguaje como práctica. El significado como uso — los modos operacionales como juegos de lenguaje de un proyecto.]
+
+*Hegel, G.W.F.* _Fenomenología del Espíritu_. Traducción de Manuel Jiménez Redondo. Valencia: Pre-Textos, 2006. [1807] \
+#text(fill: rgb("#666666"), size: 9.5pt)[Dialéctica tesis-antítesis-síntesis. Contrástese con el yin-yang: la diferencia entre síntesis y tensión gestionada permanentemente.]
+
+*Husserl, E.* _Ideas relativas a una fenomenología pura y una filosofía fenomenológica_. México: FCE, 1993. [1913] \
+#text(fill: rgb("#666666"), size: 9.5pt)[La distinción ontología formal/regional — estructura vacía universal vs. conceptos de dominio específico.]
+
+*Whitehead, A.N.* _Process and Reality_. Nueva York: Free Press, 1978. [1929] \
+#text(fill: rgb("#666666"), size: 9.5pt)[La realidad como proceso, no sustancia. El fundamento filosófico de la capa reflection como devenir distinto del ser ontológico.]
+
+*Kripke, S.* _Naming and Necessity_. Cambridge: Harvard University Press, 1980. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Propiedades esenciales rígidas y semántica de mundos posibles. El fundamento modal de los invariantes operacionales.]
+
+*Searle, J.* _The Construction of Social Reality_. Nueva York: Free Press, 1995. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Hechos institucionales como fundamento de las decisiones declaradas. El ADR como acto de habla colectivo que crea realidad operacional.]
+
+*Parfit, D.* _Reasons and Persons_. Oxford: Oxford University Press, 1984. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Identidad a través del tiempo. La pregunta "¿es el mismo proyecto?" cuando sus invariantes cambian vía ADR.]
+
+*Ladyman, J. & Ross, D.* _Every Thing Must Go: Metaphysics Naturalized_. Oxford: Oxford University Press, 2007. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Realismo estructural: las relaciones son más reales que las entidades. El DAG como estructura fundamental, los nodos como secundarios.]
+
+*Quine, W.V.O.* "On What There Is." _Review of Metaphysics_, 2(5), 1948. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El concepto de compromiso ontológico: toda teoría asume la existencia de ciertas entidades. Todo schema asume una ontología aunque no lo declare.]
+
+*Tarski, A.* "The Concept of Truth in Formalized Languages." En _Logic, Semantics, Metamathematics_. Oxford: Clarendon Press, 1956. [1933] \
+#text(fill: rgb("#666666"), size: 9.5pt)[Verdad como correspondencia. El fundamento de la validación formal — nickel typecheck como verificación tarsquiana.]
+
+== Inteligencia artificial y representación del conocimiento
+
+*Minsky, M.* "A Framework for Representing Knowledge." MIT AI Lab Memo 306, 1974. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Los frames como primera estructura computacional análoga a la sustancia aristotélica con propiedades heredables.]
+
+*Collins, A.M. & Quillian, M.R.* "Retrieval time from semantic memory." _Journal of Verbal Learning and Verbal Behavior_, 8(2), 240-247, 1969. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Primera red semántica computacional. Demuestra que la topología del grafo de conocimiento predice el comportamiento cognitivo.]
+
+*Brachman, R.J. & Schmolze, J.G.* "An overview of the KL-ONE knowledge representation system." _Cognitive Science_, 9(2), 171-216, 1985. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Description Logic: la subsunción como relación formal que genera el DAG de conceptos. Antecedente teórico de OWL.]
+
+*Gruber, T.* "A translation approach to portable ontology specifications." _Knowledge Acquisition_, 5(2), 199-220, 1993. \
+#text(fill: rgb("#666666"), size: 9.5pt)[La definición canónica: ontología como especificación de una conceptualización. Referencia obligatoria para cualquier trabajo en ontología de software.]
+
+*Lenat, D.B. & Guha, R.V.* _Building Large Knowledge-Based Systems: Representation and Inference in the Cyc Project_. Reading: Addison-Wesley, 1990. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El proyecto más ambicioso de ontología operacional. Sus fracasos son tan informativos como sus logros.]
+
+*Pearl, J.* _Probabilistic Reasoning in Intelligent Systems_. San Mateo: Morgan Kaufmann, 1988. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Bayesian networks como DAGs causales. La formalización de por qué la causalidad es inherentemente acíclica.]
+
+*Pearl, J.* _Causality: Models, Reasoning, and Inference_. Cambridge: Cambridge University Press, 2000. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El do-calculus. Distinción entre correlación y causa como operaciones sobre el DAG. Segunda edición (2009) incluye el modelo causal completo.]
+
+*Guarino, N. (ed.)* _Formal Ontology in Information Systems_. Amsterdam: IOS Press, 1998. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Actas de FOIS'98. La conferencia fundacional del campo. Especialmente el artículo introductorio de Guarino sobre ontología formal vs. de dominio.]
+
+*Berners-Lee, T., Hendler, J. & Lassila, O.* "The Semantic Web." _Scientific American_, 284(5), 34-43, 2001. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El artículo que popularizó las ontologías en ingeniería de software. La visión de la web como grafo de conocimiento estructurado.]
+
+== Ingeniería de software y praxis
+
+*Nygard, M.* "Documenting Architecture Decisions." Cognitect Blog, 2011. Disponible en: cognitect.com/blog \
+#text(fill: rgb("#666666"), size: 9.5pt)[La propuesta original de ADRs como documentación ligera de decisiones. El precursor directo de los ADRs con constraints tipadas.]
+
+*Brown, S.* _Software Architecture for Developers_. Leanpub, 2014. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El modelo C4 como alternativa a los diagramas UML. Introduce la idea de múltiples niveles de abstracción para describir un sistema.]
+
+*Hohpe, G. & Woolf, B.* _Enterprise Integration Patterns_. Boston: Addison-Wesley, 2003. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Patrones como vocabulario compartido — la idea de que nombrar los patrones operacionales es un acto ontológico.]
+
+== Perspectivas adicionales
+
+*Zhou Dunyi.* _Taijitu shuo_ (Explicación del Diagrama del Supremo Último). [1017-1073] \
+#text(fill: rgb("#666666"), size: 9.5pt)[La cosmología taoísta sistematizada: wu ji → tai ji → yin-yang. La transición de lo indiferenciado a la polaridad como modelo de emergencia estructural.]
+
+*Heidegger, M.* _Ser y Tiempo_. Traducción de Jorge Eduardo Rivera. Santiago: Editorial Universitaria, 1997. [1927] \
+#text(fill: rgb("#666666"), size: 9.5pt)[El trasfondo pre-temático que ninguna ontología captura completamente. La ausencia notable en todos los modelos operacionales de software.]
+
+*Heráclito.* _Fragmentos_. Edición de Rodolfo Mondolfo. Buenos Aires: Losada, 2007. [~500 aC] \
+#text(fill: rgb("#666666"), size: 9.5pt)[El único pensador presocrático estructuralmente equivalente al yin-yang. El fragmento DK22B60 — "el camino hacia arriba y hacia abajo son el mismo" — es el punto de partida.]
+
+*Bateson, G.* _Steps to an Ecology of Mind_. Nueva York: Ballantine, 1972. \
+#text(fill: rgb("#666666"), size: 9.5pt)[La diferencia que hace una diferencia. Epistemología sistémica que conecta Taoísmo, cibernética y representación del conocimiento. Perspectiva poco citada en CS pero fundamentalmente relevante.]
+
+*Winograd, T. & Flores, F.* _Understanding Computers and Cognition_. Norwood: Ablex, 1986. \
+#text(fill: rgb("#666666"), size: 9.5pt)[La crítica más influyente a la IA simbólica desde Heidegger. Argumenta que el trasfondo de comprensión no puede formalizarse — límite estructural de toda ontología operacional.]
+
+== Física cuántica y fundamentos
+
+*Bohr, N.* "Can Quantum-Mechanical Description of Physical Reality Be Considered Complete?" _Physical Review_, 48, 696-702, 1935. \
+#text(fill: rgb("#666666"), size: 9.5pt)[La respuesta de Bohr al artículo EPR de Einstein-Podolsky-Rosen. Formula el principio de complementariedad como posición filosófica irrenunciable.]
+
+*Bohr, N.* "Discussion with Einstein on Epistemological Problems in Atomic Physics." En Schilpp (ed.), _Albert Einstein: Philosopher-Scientist_. Evanston: Library of Living Philosophers, 1949. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El debate más importante de la física del siglo XX sobre interpretación cuántica. Bohr y Einstein representan dos ontologías irreconciliables.]
+
+*Heisenberg, W.* _Physics and Philosophy: The Revolution in Modern Science_. Nueva York: Harper & Row, 1958. \
+#text(fill: rgb("#666666"), size: 9.5pt)[La reflexión filosófica más accesible de uno de los fundadores. El capítulo sobre el papel del lenguaje en física cuántica conecta directamente con Wittgenstein.]
+
+*Schrödinger, E.* "Die gegenwärtige Situation in der Quantenmechanik." _Naturwissenschaften_, 23(48-49-50), 1935. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El artículo del gato. Introduce el concepto de entrelazamiento (_Verschränkung_) y demuestra que la superposición cuántica no puede interpretarse clásicamente.]
+
+*Kochen, S. & Specker, E.P.* "The Problem of Hidden Variables in Quantum Mechanics." _Journal of Mathematics and Mechanics_, 17(1), 59-87, 1967. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Prueba formal de que es imposible asignar valores definidos a todos los observables cuánticos simultáneamente. La contextualidad como hecho matemático, no filosófico.]
+
+*Bell, J.S.* _Speakable and Unspeakable in Quantum Mechanics_. Cambridge: Cambridge University Press, 1987. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Colección de ensayos del autor del teorema de Bell (1964). El capítulo "Against measurement" es una crítica filosófica fundamental al vocabulario de la física cuántica.]
+
+*Aspect, A., Grangier, P. & Roger, G.* "Experimental Realization of Einstein-Podolsky-Rosen-Bohm Gedankenexperiment." _Physical Review Letters_, 49(2), 91-94, 1982. \
+#text(fill: rgb("#666666"), size: 9.5pt)[La confirmación experimental de que las desigualdades de Bell se violan. Prueba que la naturaleza es no-local — no hay variables ocultas locales.]
+
+*Everett, H.* "Relative State Formulation of Quantum Mechanics." _Reviews of Modern Physics_, 29(3), 454-462, 1957. \
+#text(fill: rgb("#666666"), size: 9.5pt)[La interpretación de muchos mundos. Relevante porque ofrece una ontología donde todos los estados posibles son reales — el antípoda de la ontología operacional que colapsa posibilidades en decisiones.]
+
+*Rovelli, C.* "Relational Quantum Mechanics." _International Journal of Theoretical Physics_, 35(8), 1637-1678, 1996. \
+#text(fill: rgb("#666666"), size: 9.5pt)[No existen estados absolutos — solo estados relativos a otros sistemas. La ontología más radical de la lista y la más relevante para entender por qué el estado de un proyecto es siempre relacional.]
+
+*Rovelli, C.* _Helgoland_. Milano: Adelphi, 2020. \
+#text(fill: rgb("#666666"), size: 9.5pt)[La versión divulgativa de la mecánica cuántica relacional. Accesible y filosóficamente riguroso. El capítulo sobre relaciones y el budismo es relevante para la conexión con el Taoísmo.]
+
+*Nagel, T.* _The View from Nowhere_. Oxford: Oxford University Press, 1986. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El argumento de que toda objetividad requiere un punto de vista. En tensión directa con el problema del observador cuántico y la ontología perspectival de Zhuangzi.]
+
+*Nielsen, M.A. & Chuang, I.L.* _Quantum Computation and Quantum Information_. Cambridge: Cambridge University Press, 2000. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El texto estándar de computación cuántica. Los capítulos 2-3 sobre el formalismo matemático (espacios de Hilbert, operadores, entrelazamiento) son la base para entender la conexión con LLMs.]
+
+*Deutsch, D.* _The Fabric of Reality_. Londres: Allen Lane, 1997. \
+#text(fill: rgb("#666666"), size: 9.5pt)[La defensa más ambiciosa de la interpretación de muchos mundos. El capítulo sobre epistemología cuántica conecta con la cuestión de qué puede conocerse sobre un sistema complejo.]
+
+== Ingeniería de software e infraestructura
+
+*Evans, E.* _Domain-Driven Design: Tackling Complexity in the Heart of Software_. Boston: Addison-Wesley, 2003. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El lenguaje ubicuo como ontología operacional informal. El bounded context es estructuralmente equivalente a una ontología regional de Husserl aplicada a dominios de negocio.]
+
+*Kleppmann, M.* _Designing Data-Intensive Applications_. Sebastopol: O'Reilly, 2017. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El capítulo sobre modelos de datos es el más relevante: la elección de modelo (relacional, documental, grafo) es una decisión ontológica — define qué relaciones pueden existir.]
+
+*Morris, K.* _Infrastructure as Code_. Sebastopol: O'Reilly, 2016. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Formaliza la declaratividad en infraestructura. El capítulo sobre principios hace explícita la distinción entre estado deseado y estado actual — el núcleo del modelo state.ncl.]
+
+*Humble, J. & Farley, D.* _Continuous Delivery_. Boston: Addison-Wesley, 2010. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El pipeline de despliegue como DAG de verificaciones. La idea de que cada etapa del pipeline es un gate que el sistema debe cruzar formalmente.]
+
+*Kim, G., Humble, J., Debois, P. & Willis, J.* _The DevOps Handbook_. Portland: IT Revolution Press, 2016. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El flujo de trabajo DevOps como protocolo operacional. La sección sobre feedback loops es estructuralmente equivalente al ciclo ontología/reflection.]
+
+*Dehghani, Z.* "How to Move Beyond a Monolithic Data Lake to a Distributed Data Mesh." Martin Fowler Blog, 2019. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Data mesh como ontología operacional distribuida para datos. Cada dominio es responsable de su propia ontología de datos — la aplicación más cercana al modelo ontoref en infraestructura de datos.]
+
+*Fowler, M.* "Who Needs an Architect?" _IEEE Software_, 20(5), 11-13, 2003. \
+#text(fill: rgb("#666666"), size: 9.5pt)[Define arquitectura como "las decisiones que son difíciles de cambiar" — equivalente exacto de los invariantes en ontología operacional.]
+
+*Richardson, C.* _Microservices Patterns_. Shelter Island: Manning, 2018. \
+#text(fill: rgb("#666666"), size: 9.5pt)[El patrón Saga como DAG de transacciones distribuidas. Cada paso del saga tiene depends_on implícito — el DAG operacional aplicado a consistencia eventual.]
+#text(fill: rgb("#666666"), size: 9.5pt)[La crítica más influyente a la IA simbólica desde Heidegger. Argumenta que el trasfondo de comprensión no puede formalizarse — límite de toda ontología operacional.]
diff --git a/assets/presentation/docs/ontologia_motor.md b/assets/presentation/docs/ontologia_motor.md
new file mode 100644
index 0000000..10ddeef
--- /dev/null
+++ b/assets/presentation/docs/ontologia_motor.md
@@ -0,0 +1,53 @@
+No pude acceder directamente al artículo de Sainath Palla en *AI in Plain English* (la página bloquea el acceso automatizado), pero pude reconstruir su contenido a partir de las extensas citas textuales que el artículo anterior de Axius-SDC reprodujo, además de información complementaria del autor sobre AI-FDE y la ontología de Palantir.
+
+# Resumen: "Palantir AI FDE: cuando la ontología se convierte en motor de estrategia"
+
+**Autor:** Sainath Palla (consultor senior de Palantir Foundry, fundador de Analyticorex). Publicado el 25 de marzo de 2026.
+
+**Tesis central**
+
+Palla argumenta que el AI-FDE de Palantir (la versión asistida por IA del clásico *Forward Deployed Engineer*) representa un cambio cualitativo, no incremental, en cómo se construyen los sistemas empresariales. Lo decisivo no es que la IA escriba código más rápido, sino que opera *sobre la ontología*: el modelo de datos vivo, los workflows existentes y el contexto operacional de la empresa. El AI-FDE traduce peticiones en lenguaje natural a operaciones de Foundry, encargándose de tareas como crear pipelines de transformación de datos, gestión de repositorios y construcción/mantenimiento de la ontología.
+
+**El argumento clave**
+
+> "AI-FDE no opera sobre código aislado. Trabaja sobre la ontología, el modelo de datos existente y los workflows que ya representan al negocio. Ese contexto compartido es lo que le permite generar algo significativo. Sin él, solo obtienes andamiaje más rápido."
+
+Esa frase final —*sin la ontología solo obtienes andamiaje más rápido*— es la idea que vertebra el ensayo. Para Palla, los enfoques de IA centrados en el modelo (más parámetros, más velocidad) están atacando el problema equivocado: el cuello de botella empresarial no es generar código, sino dotar a la IA de un sustrato semántico sobre el cual razonar.
+
+**El cambio en el rol del practicante**
+
+Palla describe un desplazamiento del trabajo humano:
+
+> "Menos esfuerzo va a escribir transformaciones, conectar datasets y unir workflows básicos, y más a decidir qué debe construirse, revisar cómo está estructurado y dar forma a cómo se comporta el sistema una vez en marcha."
+
+El cuello de botella se mueve de la *implementación* al *juicio*. El practicante deja de ser un ejecutor para convertirse en arquitecto y revisor.
+
+**El gobierno mediante branching**
+
+Palla destaca un aspecto que evita el caos: cuando el AI-FDE realiza cambios, siempre crea una "Branch Proposal" y la presenta para revisión humana. Los humanos quedan liberados de la tarea tediosa de "conectar datos" y pasan al rol más avanzado de revisar y aprobar las ontologías construidas y las acciones propuestas por la IA para garantizar que se "alinean con la intención del negocio". Es, en esencia, *Git para operaciones empresariales*: cada cambio propuesto por la IA pasa por un pull request antes de tocar la realidad.
+
+**La metáfora del ferrocarril**
+
+Aquí Palla introduce su advertencia, que el artículo de Axius-SDC retomó críticamente:
+
+> "Un cambio en un lugar empieza a influir en otros. Nada está realmente aislado. Los cambios no se quedan locales. Se propagan por todo el sistema."
+
+A medida que las vías se tienden más rápido, el reto deja de ser construir una ruta aislada y pasa a ser cómo se cruzan, cómo se comparte la capacidad y cómo los cambios en una parte de la red afectan a todo lo demás. La velocidad de construcción aumenta la complejidad de las intersecciones.
+
+**Conclusión del autor**
+
+> "Los practicantes que piensan en términos de sistemas, no solo de implementaciones, se vuelven más importantes. A medida que la capacidad aumenta, la complejidad no desaparece. Se desplaza. Y alguien tiene que darle forma."
+
+Para Palla, el AI-FDE no elimina al ingeniero: lo eleva. La IA escala la *construcción*, pero el *diseño sistémico*, las decisiones de gobernanza y la coherencia entre dominios siguen siendo trabajo humano —y se vuelven más críticos, no menos.
+
+---
+
+**Diferencia esencial entre los dos artículos**
+
+| | Palla (este artículo) | Axius-SDC (artículo anterior) |
+|---|---|---|
+| Postura | Celebratoria: la ontología como "motor de estrategia" valida el enfoque de Palantir | Crítica: la ontología centralizada genera dependencia del proveedor |
+| Foco | El cambio del rol del practicante hacia el juicio sistémico | Quién es el dueño de ese juicio (los FDEs de Palantir vs. los tuyos) |
+| Metáfora | Red ferroviaria que crece y exige nueva disciplina | Estándar de ancho de vía distribuido en cada segmento |
+
+Axius-SDC tomó las observaciones de Palla como punto de partida —admitiendo que tiene razón en el diagnóstico— para luego argumentar que la solución debería ser una gobernanza distribuida y de código abierto, en lugar de una ontología propietaria gestionada por los ingenieros desplegados de Palantir.
diff --git a/assets/presentation/docs/ontologia_motor_resumen.md b/assets/presentation/docs/ontologia_motor_resumen.md
new file mode 100644
index 0000000..c61e766
--- /dev/null
+++ b/assets/presentation/docs/ontologia_motor_resumen.md
@@ -0,0 +1,89 @@
+# Resumen: "Palantir AI FDE: Cuando la ontología se convierte en motor de estrategia"
+
+**Autor:** Sainath Palla — 25 de marzo de 2026
+
+---
+
+**Punto de partida: el caso de uso de la cadena de suministro**
+
+Palla abre con un ejemplo concreto: un sistema de resolución de escasez de inventario. Datos de ERP, proveedores y logística fluyen hacia una ontología modelada en torno a materiales, proveedores y pedidos. Pipelines calculan la escasez, una aplicación Workshop muestra opciones al planificador, y AIP Logic sugiere si conviene acelerar, sustituir o transferir inventario.
+
+El sistema funciona, pero la observación clave es esta: **la mayor parte del esfuerzo no está en la recomendación**, sino en todo lo que la rodea — diseñar la ontología, alinear los pipelines, cablear los workflows y mantener consistentes los writebacks y la auditoría. Esa es la dificultad real del primer caso de uso.
+
+---
+
+**Lo que ha cambiado: la IA cruza al otro lado**
+
+Antes, la IA vivía *dentro* del caso de uso (asistiendo decisiones ya construidas). El AI-FDE rompe esa frontera: ahora la IA participa en *cómo se construye el propio sistema*. Escala pipelines, sugiere extensiones de la ontología, conecta workflows y ensambla la aplicación. El trabajo no desaparece, se desplaza: menos esfuerzo en transformaciones y plumbing, más en decidir qué construir y cómo debe comportarse.
+
+---
+
+**Lo que se desbloquea**
+
+Aparentemente es solo velocidad. Pero la velocidad solo importa porque el AI-FDE opera sobre un sustrato con contexto previo:
+
+> "Sin él, solo obtienes andamiaje más rápido. Con él, empiezas a obtener sistemas usables mucho más rápido."
+
+El verdadero desbloqueo es la **capacidad de extender un sistema operacional existente**, donde cada pieza nueva se construye sobre contexto compartido.
+
+---
+
+**Donde empieza a romperse**
+
+Lo mismo que acelera la expansión la vuelve más difícil de gestionar. Cuando se añaden casos —rendimiento de proveedores, optimización de inventario, planificación de producción— todos operan sobre la misma ontología, los mismos datos, los mismos workflows. **Nada queda realmente aislado**. Lo nuevo no es la propagación de cambios, sino la *velocidad y frecuencia* con que se forman esas conexiones.
+
+Metáfora: una **red ferroviaria**. Tender vías rápido es fácil; lo difícil es cómo se cruzan, cómo se comparte la capacidad y cómo un cambio en un punto afecta al resto.
+
+---
+
+**Consecuencia: la velocidad amplifica el comportamiento**
+
+Cuando el sistema está alineado, el valor se compone rápido. Cuando hay desalineación, los pequeños fallos también se propagan —y son más difíciles de rastrear. La pregunta deja de ser "¿cómo construyo este caso?" para convertirse en "¿qué le pasa al sistema cuando este caso se añade?".
+
+---
+
+**El cuello de botella se ha movido**
+
+La capa inferior (integración, pipelines, plantillas de workflow) se comprime. Pero la capa superior **no se comprime**: juicio de dominio, disciplina ontológica, diseño de gobernanza, límites de decisión y dónde dejar al humano en el loop. El AI-FDE cambia *qué tan rápido puedes construir*; no elimina la necesidad de decidir *qué vale la pena construir*.
+
+---
+
+**El sistema empieza a entenderse a sí mismo**
+
+Cuando la ontología se estabiliza, el sistema deja de ser solo ejecutable y se vuelve **legible** —tanto para humanos como para la propia IA. El AI-FDE accede a la estructura de la ontología, las relaciones entre objetos, los workflows, las evaluaciones y las decisiones históricas. Eso le permite detectar señales: relaciones faltantes, workflows que terminan demasiado pronto, bucles de feedback incompletos, patrones duplicados.
+
+> "El siguiente conjunto de casos de uso ya no viene solo desde fuera. Empieza a emerger desde dentro del propio sistema."
+
+En ese momento, **la ontología deja de ser solo un modelo operacional y se comporta como un motor de estrategia**, mostrando para qué está listo el sistema a continuación. Esta es la idea que da título al artículo.
+
+---
+
+**Vuelta a los fundamentos**
+
+A medida que construir se vuelve más fácil, los fundamentos pesan más. Una ontología mal definida o relaciones inconsistentes ya no solo generan fricción: *moldean el comportamiento de múltiples workflows simultáneos*. La diferencia entre construir para entregar y construir para crecer solo se ve cuando el sistema se expande.
+
+---
+
+**Las guardarraíles tienen que llegar antes**
+
+Cuando muchos workflows dependen del mismo contexto, corregir deja de ser un acto local. Las definiciones claras, las relaciones consistentes y los límites de decisión explícitos ya no son tareas de limpieza posterior: son **parte del diseño desde el inicio**. Una vez que el sistema se expande, no estás arreglando un caso de uso —estás remodelando una capa operacional compartida.
+
+---
+
+**Reflexión final para practicantes**
+
+> "Los practicantes que piensan en términos de sistemas, no solo de implementaciones, se vuelven más importantes. A medida que la capacidad aumenta, la complejidad no desaparece. Se desplaza. Y alguien tiene que darle forma."
+
+El valor ya no está en qué tan rápido cableas un workflow, sino en qué tan bien entiendes el sistema como un todo: cómo se estructuran las decisiones, cómo se ensambla el contexto, cómo se hacen visibles los trade-offs y cómo interactúan las partes a lo largo del tiempo.
+
+---
+
+## Idea-fuerza
+
+El artículo describe una **transición en tres tiempos**:
+
+1. **De ejecutor a constructor:** la IA pasa de asistir dentro del caso de uso a participar en construir el sistema.
+2. **De constructor a intérprete:** una vez la ontología es rica, la IA puede leer el estado del sistema y proponer qué falta.
+3. **De intérprete a estratega:** la ontología deja de ser un modelo de datos y empieza a generar las próximas preguntas del negocio.
+
+Por eso la conclusión no es "la IA reemplaza al ingeniero", sino que el cuello de botella se mueve hacia arriba: hacia el juicio sistémico, la disciplina ontológica y el diseño de cómo el sistema se comporta cuando crece.
diff --git a/assets/presentation/docs/ontoref-dags-diferenciacion.md b/assets/presentation/docs/ontoref-dags-diferenciacion.md
new file mode 100644
index 0000000..50918b8
--- /dev/null
+++ b/assets/presentation/docs/ontoref-dags-diferenciacion.md
@@ -0,0 +1,107 @@
+# Ontoref — Por qué los DAGs de ontoref son diferentes
+
+Referencia para posts, slides, presentaciones.
+Contexto: análisis de por qué DAGs son ubicuos pero nadie tiene la perspectiva de ontoref.
+
+---
+
+## DAGs son ubicuos — y todos hacen lo mismo
+
+En CI/CD, compiladores, runbooks, Airflow, dbt, Bazel: el DAG es un **modelo de ejecución**.
+Las aristas significan "esto antes que aquello". El grafo describe *cómo se computa*,
+no *qué es lo que se computa*. Es ordenación topológica, no conocimiento.
+
+```
+CI/CD:      test → build → deploy
+Compiler:   parse → typecheck → codegen → link
+Runbook:    check_health → drain → restart → verify
+```
+
+Todos inerts. El grafo no sabe qué es el proyecto. No tiene semántica sobre sus propias aristas.
+
+---
+
+## Ontoref usa DAGs en dos modos distintos
+
+**DAG ontológico — aristas con tipo semántico:**
+
+```
+Practice "ncl-schemas" --implements--> Principle "type-safety"
+Practice "ncl-schemas" --enforces-->   Constraint "zero-runtime-deps"
+Practice "ncl-schemas" --enables-->    Capability "agent-queryable-state"
+Practice "ncl-schemas" --tension-->    Practice "adoption-friction"
+```
+
+No es "esto antes que aquello". Es conocimiento — compromisos formales sobre qué ES
+el proyecto y cómo se relacionan sus conceptos. El equivalente sería un compilador
+que sabe por qué existe y qué trade-offs tomó.
+
+**DAG de reflexión — aristas con contrato ejecutable + restricción de actor:**
+
+```
+step "validate-state" (actor: any)    --> step "run-mode" (actor: developer)
+step "run-mode" (actor: developer)    --> step "report" (actor: agent|developer)
+```
+
+No es solo ejecución. El DAG es validado como contrato antes de ejecutarse,
+registra su propio progreso, y sabe quién puede hacer qué.
+
+---
+
+## El gap que nadie ha cerrado
+
+```
+Knowledge Graph world:  semántica sobre el dominio del negocio
+Software tooling world: DAGs sobre la ejecución del software
+ontoref:                semántica sobre el proyecto de software en sí mismo
+```
+
+El mundo de knowledge graphs (W3C, RDF/OWL, Talisman) aplica semántica a dominios
+empresariales — productos, clientes, transacciones. Nunca al proyecto de software mismo.
+
+El mundo de herramientas de software (GitHub, Jira, CI/CD) trata el conocimiento del
+proyecto como documentos no tipados. Usa DAGs para ejecución, nunca para representación
+de conocimiento.
+
+---
+
+## La pieza que lo hace no-trivial
+
+Los dos DAGs se hablan:
+
+- El DAG ontológico dice *qué es* el proyecto (prácticas, principios, constraints activos)
+- El DAG de reflexión *opera sobre ese estado* — detecta deriva, ejecuta modos, registra transiciones
+- Las migraciones propagan cambios del DAG ontológico a proyectos consumidores
+
+Sin esa conexión: ontología bonita que nadie mantiene (buenas intenciones)
+o runbooks sin semántica (ejecución ciega).
+
+---
+
+## Por qué no había emergido antes
+
+Tres condicionantes que ahora se alinean:
+
+1. **Agentes AI** — que necesitan consumir conocimiento estructurado del proyecto,
+   no texto libre. Antes no había consumidor que lo requiriera con urgencia.
+2. **NCL/Nickel** — contratos y tipos que permiten expresar ontología sin triplestore externo.
+   RDF/OWL requería infraestructura pesada y expertise especializado.
+3. **La escala correcta** — Knowledge Graphs empresariales son proyectos de años.
+   Ontoref opera a escala de repositorio, con adopción incremental y sin enforcement.
+
+---
+
+## Para slides
+
+**Una línea:** Todos usan DAGs para HACER cosas. Ontoref usa DAGs para CONOCER cosas
+sobre el proyecto, Y para HACER cosas, Y hace que las dos capas se hablen.
+
+**Contraste:**
+
+| DAGs tradicionales | DAGs en ontoref |
+|---|---|
+| Modelo de ejecución | Modelo de conocimiento + ejecución |
+| Aristas: "depende de" | Aristas: `implements`, `enforces`, `tension`, `enables` |
+| Inert respecto al proyecto | Semánticamente cargado con el estado del proyecto |
+| Evalúa y descarta | Evalúa, registra, propaga |
+| Grafo describe proceso externo | Grafo describe el proyecto mismo |
diff --git a/assets/presentation/docs/ontoref-scope-proyecto-infra-personal.md b/assets/presentation/docs/ontoref-scope-proyecto-infra-personal.md
new file mode 100644
index 0000000..64cbe4f
--- /dev/null
+++ b/assets/presentation/docs/ontoref-scope-proyecto-infra-personal.md
@@ -0,0 +1,157 @@
+# Ontoref — Scope: Proyecto, Infraestructura, Personal
+
+Referencia para posts, slides, presentaciones.
+Contexto: ontoref como protocolo de auto-conocimiento para cualquier sistema con identidad e intención,
+más allá de gestión de proyectos de software.
+
+---
+
+## La afirmación central
+
+Ontoref no es una herramienta de gestión de proyectos.
+Es un **protocolo de auto-conocimiento para cualquier sistema con identidad e intención**.
+
+Los tres dominios comparten la misma estructura profunda:
+
+```
+              Ontology (Yin)              Reflection (Yang)
+              ──────────────────────────  ──────────────────────────────
+Proyecto      Qué ES el software          Qué DEVIENE — ADRs, modos,
+              principios, prácticas       migraciones, deriva de docs
+
+Infra         Qué ES el sistema           Qué DEVIENE — provisioning,
+              servicios, deps,            drift detection, remediation,
+              constraints, topología      runbooks como DAGs ejecutables
+
+Personal      Quién ERES                  Quién DEVIENS — decisiones con
+              valores, roles,             consecuencias duraderas, hábitos,
+              principios vitales          revisiones, deriva de intención
+```
+
+---
+
+## Infraestructura
+
+Infra ya usa DAGs (Terraform, Pulumi, Ansible) pero todos inerts: describen
+*cómo se provisiona*, no *por qué existe así*. Ontoref sobre infra captura
+las tensiones reales:
+
+```
+Servicio "auth" --tension--> Principio "stateless"
+    porque necesita session state pero el principio dice no
+
+Constraint "zero-downtime" --enforces--> Práctica "blue-green"
+    y eso está en el ADR-infra-003, no en el Terraform
+```
+
+jj + Radicle + NATS federation = GitOps soberano donde cada entorno
+(dev, staging, prod) se describe a sí mismo y puede ser consultado
+desde el proyecto que lo consume.
+
+**El problema que resuelve:** infraestructura que deriva sin registro de por qué
+se tomaron las decisiones que la definen. El ADR de infra captura el "por qué"
+con el mismo peso que un ADR arquitectónico de software.
+
+---
+
+## Personal
+
+El dominio más radical. La misma pregunta que ontoref hace a un proyecto —
+*¿qué eres, qué tensiones activas tienes, en qué estado estás respecto a donde
+quieres estar?* — es exactamente la pregunta que el conocimiento personal
+estructurado necesita responder.
+
+No un segundo cerebro de notas. Un grafo de compromisos.
+
+```
+Valor "profundidad" --tension--> Rol "padre"
+    porque el tiempo es finito y los dos lo demandan
+
+Decisión "dejar empresa X" --constraint--> Futuro laboral
+    con el mismo peso que un ADR arquitectónico
+
+Estado "aprendiendo Rust" --blocker--> "contribuir a ontoref-core"
+    FSM personal, no lista de tareas
+```
+
+**El problema que resuelve:** la distancia entre quién intentas ser y quién
+estás siendo crece sin mecanismo de detección. Reflection personal es ese mecanismo.
+
+---
+
+## Lo que todos los dominios comparten
+
+Los tres tienen:
+
+- **Identidad** — qué son, qué principios los definen
+- **Tensiones activas** — trade-offs que no se resuelven, se gestionan
+- **Decisiones con consecuencias duraderas** — el equivalente de ADRs
+- **Estado observable** — dónde están vs. dónde quieren estar
+- **Deriva** — la distancia entre intención y realidad que crece con el tiempo
+- **Operaciones** — los modos que reducen esa deriva
+
+Y todos se benefician del mismo stack:
+- Almacenamiento soberano (NCL versionado junto al sujeto)
+- Federación (NATS, modos federados)
+- Multi-superficie (CLI + UI + MCP + GraphQL)
+- VCS soberano (jj + Radicle)
+
+---
+
+## El stack completo — lo que añade cada capa
+
+**Almacenamiento soberano:**
+El enterprise knowledge graph asume un triplestore central. Ontoref invierte eso:
+el conocimiento vive *con el sujeto*, en NCL versionado. Nadie puede quitarte
+tu ontología porque está en tu repo/entorno/diario.
+
+**Modos federados:**
+Proyectos que se describen a sí mismos Y pueden ser descubiertos y consultados
+por otros proyectos — sin servidor central.
+
+```
+proyecto-A consulta: ¿qué proyectos implementan "auth-pattern-X"?
+proyecto-B, proyecto-C responden desde sus propias ontologías
+sin intermediario que lo sepa todo
+```
+
+**CLI + UI + MCP + GraphQL:**
+
+| Superficie | Consumidor |
+|---|---|
+| CLI (Nushell) | Desarrollador en terminal, scripts, CI |
+| UI (axum) | Managers, revisión visual, onboarding |
+| MCP | Agentes AI |
+| GraphQL | Herramientas externas, dashboards, queries ad-hoc |
+
+GraphQL sobre el DAG ontológico reduce la barrera de consumo sin sacrificar estructura.
+SPARQL es el estándar semántico pero requiere expertise. GraphQL es lo que todo el mundo escribe.
+
+**jj + Radicle:**
+- jj: cada working copy es siempre un commit — alinea con ontoref, no hay "estado sucio"
+- Radicle: P2P, local-first, sin GitHub — si el código es soberano, el conocimiento sobre ese código también
+
+---
+
+## Comparativa de posicionamiento
+
+| Enterprise KG (Talisman) | Ontoref |
+|---|---|
+| Triplestore central | NCL distribuido, por sujeto |
+| Equipo dedicado de ontólogos | El sujeto se mantiene a sí mismo |
+| Dominio: datos organizacionales | Dominio: proyecto + infra + personal |
+| GitHub/GitLab | jj + Radicle (soberano, P2P) |
+| SPARQL sobre HTTP | CLI + UI + MCP + GraphQL |
+| Un grafo global | Federación de sujetos auto-descritos |
+| Escala: organización | Escala: individuo → proyecto → ecosistema |
+
+---
+
+## Para slides — una línea por dominio
+
+- **Proyecto:** El código sabe compilar. Ontoref hace que sepa qué es y por qué.
+- **Infra:** Terraform sabe provisionar. Ontoref hace que sepa por qué existe así.
+- **Personal:** Tu agenda sabe qué haces. Ontoref hace que sepas quién eres y hacia dónde derivas.
+
+**Cierre:** El sujeto más pequeño que puede adoptar ontoref es una persona.
+El más grande es un ecosistema de proyectos federados.
diff --git a/assets/presentation/docs/ontoref-yin-yang-diferenciacion.md b/assets/presentation/docs/ontoref-yin-yang-diferenciacion.md
new file mode 100644
index 0000000..983e10c
--- /dev/null
+++ b/assets/presentation/docs/ontoref-yin-yang-diferenciacion.md
@@ -0,0 +1,67 @@
+# Ontoref — Yin/Yang: Lo que Talisman no resuelve
+
+Referencia para posts, slides, presentaciones.
+Contexto: extracción del análisis comparativo con "Steal This Deck" (Talisman, KGC 2026).
+
+---
+
+## El gap del argumento de Talisman
+
+Talisman describe infraestructura de conocimiento como artefacto de construcción —
+vocabularios, ontologías, knowledge graphs. El problema que identifica (governance
+"forever, not a project") lo *nombra* pero no lo *resuelve*. Su stack termina en
+"Govern" como sexto paso, sin mecanismo interno de mantenimiento.
+
+**La pregunta que no responde:**
+- ¿Quién vigila que la ontología no derive respecto al sistema real?
+- ¿Quién detecta cuando una decisión arquitectónica invalida un nodo del grafo?
+- ¿Quién propaga los cambios a los consumidores?
+
+Sin ese cierre operacional, la ontología más rigurosa acaba siendo buenas intenciones
+formalizadas. Archaeology, no infrastructure.
+
+---
+
+## Ontoref es las dos mitades
+
+```
+Ontology (Yin — lo que ES)          Reflection (Yang — lo que DEVIENE)
+───────────────────────────────     ────────────────────────────────────
+.ontology/core.ncl                  reflection/modes/*.ncl
+Nodos, edges, constraints           DAGs ejecutables, pasos reportables
+Compromisos formales                Operaciones que cambian estado
+state.ncl (fotografía)              state.ncl FSM (transición activa)
+ADRs (decisión tomada)              migrations/ (propagación al ecosistema)
+describe project                    sync diff --docs (detección de deriva)
+```
+
+Sin Reflection, la ontología cristaliza — describe lo que el proyecto *quería ser*
+en el momento en que se escribió.
+
+Sin Ontology, Reflection es ejecución sin verdad — sabe *qué hacer* pero no *qué es*.
+
+---
+
+## La tensión nombrada en core.ncl
+
+> "Ontology vs Reflection: Ontology captures what IS (invariants, structure, being).
+> Reflection captures what BECOMES (operations, drift, memory).
+> Both must coexist without one dominating. This tension is onref's core identity."
+
+Talisman articula bien el Yin. Ontoref es la respuesta al Yang que su framework no provee.
+
+---
+
+## Para slides
+
+**Una línea:** Talisman dice "construye ontologías, goviernalas". Ontoref es el mecanismo
+de governance interno que hace que esa instrucción sea ejecutable.
+
+**Contraste visual:**
+
+| Talisman (KGC 2026) | Ontoref |
+|---|---|
+| Construye knowledge infrastructure | Construye + mantiene desde adentro |
+| Governance como paso final | Reflection como loop continuo |
+| Ontología como artefacto | Ontología + operaciones coexistentes |
+| "Forever, not a project" (aspiración) | FSM + modes + migrations (mecanismo) |
diff --git a/assets/presentation/docs/posts/en/ai-knowledge-tool-who-keeps-it-alive.md b/assets/presentation/docs/posts/en/ai-knowledge-tool-who-keeps-it-alive.md
new file mode 100644
index 0000000..eb79641
--- /dev/null
+++ b/assets/presentation/docs/posts/en/ai-knowledge-tool-who-keeps-it-alive.md
@@ -0,0 +1,97 @@
+---
+# Post metadata
+id: "ai-knowledge-tool-who-keeps-it-alive"
+title: "AI is a Knowledge Tool. But Who Keeps the Knowledge Alive?"
+slug: "ai-knowledge-tool-who-keeps-it-alive"
+subtitle: "Talisman is right — and stops exactly where the hard problem begins"
+excerpt: "Jessica Talisman's KGC 2026 talk is the clearest articulation of why AI strategies fail: organizations invest in data infrastructure and expect reasoning to emerge. She's right. But her solution — build knowledge infrastructure — stops exactly where the hard problem begins: who keeps the knowledge from drifting?"
+
+# Publication info
+author: "Jesús Pérez"
+date: "2026-05-10"
+published: false
+featured: false
+
+# Categorization
+category: "ontoref"
+tags: ["ontoref", "knowledge-graphs", "ontology", "ai-infrastructure", "reflection"]
+
+# Display
+read_time: "7 min read"
+sort_order: 1
+css_class: "category-ontoref"
+category_description: "Ontoref — protocol and tooling for structured self-knowledge in software projects"
+category_published: true
+---
+
+# AI is a Knowledge Tool. But Who Keeps the Knowledge Alive?
+
+Jessica Talisman delivered a talk at KGC 2026 called "Stop Betting, Start Building." The data she opens with is brutal:
+
+- **89%** of firms report zero productivity impact from AI after three years (NBER, Feb 2026, n=5,937 executives)
+- Experienced developers are **19% slower** with AI tools, not faster (METR RCT, 2025)
+- The average AI-using worker gains **−14 minutes** per week in net productivity (Foxit/Sapio, March 2026)
+
+The market is bullish. The evidence is not.
+
+Her diagnosis is correct: *AI is a knowledge tool, not a data tool.* Organizations are pouring money into data lakes, vector stores, and ETL pipelines and expecting context and reasoning to emerge from raw tokens. It won't. Models were trained on linked data, RDF triples, and controlled vocabularies — the top fifteen C4 training sources are knowledge-graph-heavy. They're then deployed against environments stripped of all that structure, and we wonder why they hallucinate.
+
+Her prescription is also correct: build knowledge infrastructure. Controlled vocabularies first. Taxonomies. Thesauri. Ontologies — formal commitments, classes, properties, constraints. Knowledge graphs on top. And govern them. Forever, not as a project.
+
+She's right. And she stops exactly where the hard problem begins.
+
+## The Governance Gap
+
+The sixth step in her stack is "Govern — this is infrastructure, stewardship, versioning, growth, forever." She names the problem. But naming it is not a mechanism.
+
+The real question is: *who ensures the ontology doesn't drift from the system it describes?*
+
+Software evolves. An architectural decision made in March invalidates a node in the graph by October. A new team joins with different mental models. A library gets replaced and a constraint becomes fictional. Knowledge graphs accumulate debt the same way codebases do — silently, without an alarm.
+
+In the enterprise knowledge graph world, the answer to governance is headcount: hire ontologists, librarians, knowledge engineers. That works at organizational scale with dedicated budgets. It is not a viable answer for a software project, an infrastructure environment, or an individual trying to maintain structured self-knowledge.
+
+## What's Actually Missing: the Operational Loop
+
+Every serious ontology without an operational closure layer becomes archaeology within eighteen months. Beautiful, formally correct, and describing a system that no longer exists.
+
+The missing piece is not more knowledge. It's a feedback loop that:
+
+1. **Observes** the current state of the system against its declared intent
+2. **Detects** drift before it becomes permanent
+3. **Executes** operations that reduce that drift
+4. **Records** decisions with lasting architectural weight
+5. **Propagates** changes to everything that depends on this knowledge
+
+This is the Yang to the ontology's Yin. Talisman describes the Yin in precise detail. The Yang is what makes it not an artifact.
+
+## Ontoref: Both Halves
+
+Ontoref is a protocol for structured self-knowledge in software projects. It operates as two coexisting layers that cannot function without each other:
+
+**Ontology (what IS):** typed nodes and edges representing practices, principles, tensions, capabilities. Not documentation — formal commitments. A `Practice` node that `implements` a `Principle`, `enforces` a `Constraint`, `enables` a `Capability`, and is in active `tension` with another `Practice`. The project as a knowledge graph.
+
+**Reflection (what BECOMES):** executable DAGs — operational modes that run against the ontological state, detect drift, report transitions, and propagate changes. The `sync diff --docs` command that catches when a crate's documentation has drifted from its ontology node. The FSM in `state.ncl` that tracks where each project dimension is versus where it intends to go. The migrations system that ensures protocol changes reach every consumer project.
+
+The tension between these two layers is not a design flaw — it's named explicitly in the project's own ontology as its core identity:
+
+> *"Ontology captures what IS. Reflection captures what BECOMES. Both must coexist without one dominating."*
+
+Without Reflection, the Ontology crystallizes. It becomes a snapshot of what the project wanted to be at the moment it was written. Without Ontology, Reflection is execution without truth — it knows how to operate but not what it's operating on.
+
+## The Accuracy Numbers
+
+Talisman's data on what structured knowledge actually does to AI accuracy is worth repeating:
+
+- Question-answering on enterprise SQL: **16% → 72%** when an ontology checks and repairs LLM-generated queries (Allemang & Sequeda, data.world AI Lab, 2024)
+- GraphRAG vs. vector RAG: **3.4× accuracy** across 43 enterprise queries (Diffbot KG-LM Benchmark, 2023)
+- Vector RAG collapses to **0% accuracy past five entities per query**. KG-grounded retrieval sustains performance well beyond that
+
+The accuracy gap is not closed by a bigger model. It is closed by a defined schema, an ontology, and a validated query.
+
+But only if the ontology is kept alive. Only if someone — or something — is running the operational loop that keeps it from drifting into fiction.
+
+That's the part Talisman's framework doesn't provide. That's what Reflection is for.
+
+---
+
+*Ontoref is open source. The protocol specification, Nushell automation, and Rust crates are at [github.com/jesusperezlorenzo/ontoref](https://github.com/jesusperezlorenzo/ontoref).*
diff --git a/assets/presentation/docs/posts/en/dags-everywhere-none-know-what-they-are.md b/assets/presentation/docs/posts/en/dags-everywhere-none-know-what-they-are.md
new file mode 100644
index 0000000..e52d686
--- /dev/null
+++ b/assets/presentation/docs/posts/en/dags-everywhere-none-know-what-they-are.md
@@ -0,0 +1,106 @@
+---
+# Post metadata
+id: "dags-everywhere-none-know-what-they-are"
+title: "DAGs Are Everywhere. None of Them Know What They Are."
+slug: "dags-everywhere-none-know-what-they-are"
+subtitle: "Every build system, CI pipeline, and runbook uses DAGs for execution. Ontoref uses them for knowledge."
+excerpt: "CI/CD pipelines, compilers, runbooks, data orchestrators — they all use directed acyclic graphs. Every single one of them uses DAGs as execution models: this before that, topological ordering, dependency resolution. None of them use DAGs to represent what the system is, why it exists, or what trade-offs define it. That's the gap ontoref fills."
+
+# Publication info
+author: "Jesús Pérez"
+date: "2026-05-10"
+published: false
+featured: false
+
+# Categorization
+category: "ontoref"
+tags: ["ontoref", "dag", "ontology", "knowledge-graphs", "software-architecture"]
+
+# Display
+read_time: "6 min read"
+sort_order: 2
+css_class: "category-ontoref"
+category_description: "Ontoref — protocol and tooling for structured self-knowledge in software projects"
+category_published: true
+---
+
+# DAGs Are Everywhere. None of Them Know What They Are.
+
+Every build system uses DAGs. Every CI/CD pipeline uses DAGs. Compilers, data orchestrators, runbooks, package managers, Kubernetes operators — all DAGs. Directed acyclic graphs are so ubiquitous in software infrastructure that the question "why DAGs?" barely registers as a question anymore.
+
+But there's a second question nobody asks: *what do those DAGs represent?*
+
+The answer is always the same: **execution order**. This before that. Topological sorting. Dependency resolution. The graph describes *how* something computes, not *what* it is.
+
+```
+CI/CD:      test → build → deploy
+Compiler:   parse → typecheck → codegen → link
+Runbook:    check_health → drain → restart → verify
+```
+
+These graphs are inert with respect to the system they operate on. A CI pipeline doesn't know what the project is, why it was built this way, or what trade-offs define its architecture. It knows that tests must pass before deployment. That's all it knows.
+
+## The Semantic Gap
+
+This inertness is not a flaw — it's appropriate. Build systems should be fast and mechanical. Runbooks should be executable without requiring philosophical knowledge about the system. Execution graphs and knowledge graphs are different tools for different purposes.
+
+The problem is that the software industry has reached for DAGs for *everything except knowledge representation*. The knowledge of what a system is — its principles, its architectural decisions, its active tensions, its capability surface — lives in documents. Wikis. READMEs. Tickets. Verbal tradition.
+
+These are all unstructured, un-typed, un-queryable, and guaranteed to drift from the actual system within months. They describe what the project was, not what it is.
+
+## Ontoref's Dual DAG
+
+Ontoref uses DAGs in two fundamentally different modes, and the distinction matters:
+
+**Ontological DAG — semantically typed edges:**
+
+```
+Practice "ncl-schemas"  --implements-->  Principle "type-safety"
+Practice "ncl-schemas"  --enforces-->    Constraint "zero-runtime-deps"
+Practice "ncl-schemas"  --enables-->     Capability "agent-queryable-state"
+Practice "ncl-schemas"  --tension-->     Practice "adoption-friction"
+```
+
+These edges are not "depends on." They are typed relationships: `implements`, `enforces`, `enables`, `tension`. The graph is the knowledge — formal commitments about what the project is and how its concepts relate. The equivalent would be a compiler that knows why it exists and what architectural trade-offs it made.
+
+**Reflection DAG — executable contracts with actor restrictions:**
+
+```
+step "validate-state"  (actor: any)        -->  step "run-mode"  (actor: developer)
+step "run-mode"        (actor: developer)  -->  step "report"    (actor: agent|developer)
+```
+
+This is not just execution ordering. The DAG is validated as a contract before it runs, records its own progress through state transitions, and encodes who can execute each step. An agent trying to run a developer-restricted step receives a typed rejection, not a runtime error.
+
+## The Missing Connection
+
+What makes this non-trivial is that the two DAGs communicate:
+
+- The ontological DAG declares what the project IS — active constraints, practices in tension, current state dimensions
+- The reflection DAG OPERATES on that state — detecting drift, executing modes, recording transitions
+- Migrations propagate changes in the ontological DAG to downstream consumer projects
+
+In every other system that uses DAGs, the graph is evaluated and discarded. In ontoref, evaluation modifies the state that the next execution reads. The graph has memory.
+
+## Why This Didn't Exist Before
+
+Three things had to align:
+
+**Agents as consumers.** Before agentic AI, there was no automated consumer that needed structured knowledge about a project. Humans could read the wiki. Agents cannot — they need typed, queryable, machine-readable project knowledge to work accurately rather than hallucinate context. Ontoref's knowledge DAG is what agents consume via MCP.
+
+**Configuration languages with contracts.** Expressing an ontology without an external triplestore required a configuration language with types and contracts. Nickel (NCL) provides exactly that: a typed, lazy configuration language where schema violations are caught at evaluation time, not at runtime. RDF/OWL would have required dedicated infrastructure and specialist expertise.
+
+**Repository scale, not enterprise scale.** Enterprise knowledge graphs are multi-year initiatives. Ontoref operates at the scale of a single repository — incremental adoption, no enforcement, no dedicated team. The smallest unit that can adopt it is a project of one.
+
+## The Consequence
+
+When a project has a knowledge DAG that its agents can consume, the accuracy arithmetic changes entirely. The numbers from KGC 2026 research:
+
+- Ontology-grounded retrieval: **3.4× more accurate** than vector RAG on enterprise queries
+- Ontology-validated queries: accuracy jumps from **16% to 72%** on SQL question-answering
+
+The gap is closed not by a bigger model but by structured knowledge the model can reason against. A DAG where the edges mean something.
+
+---
+
+*Ontoref is open source. The protocol specification, Nushell automation, and Rust crates are at [github.com/jesusperezlorenzo/ontoref](https://github.com/jesusperezlorenzo/ontoref).*
diff --git a/assets/presentation/docs/posts/en/one-protocol-three-subjects.md b/assets/presentation/docs/posts/en/one-protocol-three-subjects.md
new file mode 100644
index 0000000..254b81a
--- /dev/null
+++ b/assets/presentation/docs/posts/en/one-protocol-three-subjects.md
@@ -0,0 +1,122 @@
+---
+# Post metadata
+id: "one-protocol-three-subjects"
+title: "One Protocol, Three Subjects"
+slug: "one-protocol-three-subjects"
+subtitle: "The same question applies to your project, your infrastructure, and yourself"
+excerpt: "Ontoref started as a protocol for software project self-knowledge. The same architecture — ontological DAG for what IS, reflection DAG for what BECOMES — applies without modification to infrastructure environments and to individuals. The subject changes. The question is identical: what are you, what tensions define you, where are you versus where you intend to be?"
+
+# Publication info
+author: "Jesús Pérez"
+date: "2026-05-10"
+published: false
+featured: false
+
+# Categorization
+category: "ontoref"
+tags: ["ontoref", "ontology", "infrastructure", "personal-knowledge", "reflection"]
+
+# Display
+read_time: "7 min read"
+sort_order: 3
+css_class: "category-ontoref"
+category_description: "Ontoref — protocol and tooling for structured self-knowledge in software projects"
+category_published: true
+---
+
+# One Protocol, Three Subjects
+
+The question ontoref asks of a software project is precise: *What are you? What principles define you? What tensions are you actively managing? What decisions have you made with lasting consequences? Where are you versus where you intend to be?*
+
+This turns out to be the same question that infrastructure environments need answered. And individuals.
+
+The subject changes. The question is identical.
+
+## What All Three Have in Common
+
+Any system with identity and intent shares the same structural problem:
+
+- **Identity** — what it is, what principles define it
+- **Active tensions** — trade-offs that don't resolve, they're managed
+- **Durable decisions** — choices with lasting consequences that constrain the future
+- **Observable state** — where it is versus where it intends to be
+- **Drift** — the distance between intention and reality, growing silently with time
+- **Operations** — the actions that reduce that drift
+
+Software projects have all of these. So do infrastructure environments. So do people.
+
+And they all benefit from the same architecture: an ontological layer that captures what IS, and a reflection layer that captures what BECOMES and operates to keep those two aligned.
+
+## Infrastructure
+
+Infrastructure already uses DAGs — Terraform, Pulumi, Ansible. All of them describe *how* the system is provisioned. None of them capture *why it exists as it does*.
+
+The `why` is where knowledge lives:
+
+```
+Service "auth"       --tension-->    Principle "stateless"
+    reason: needs session state, but the principle forbids it
+
+Constraint "zero-downtime"  --enforces-->  Practice "blue-green"
+    reason: ADR-infra-003, after the 2025 deploy incident
+```
+
+This information is not in the Terraform plan. It's in the memory of the engineer who was there, in a Slack thread from eight months ago, or nowhere at all. When that engineer leaves, the infrastructure loses its own history — it knows what it is, not why.
+
+Infrastructure ontoref captures the architectural decisions of the infrastructure itself, with the same ADR formalism as software projects. It tracks state dimensions — which environments are stable, which are in active migration, what's blocked and why. The FSM doesn't track deployment status; it tracks *epistemic status* — how well the infrastructure understands itself.
+
+Federation via NATS means each environment (dev, staging, prod) describes itself and can be queried from the projects that depend on it. Sovereign GitOps: the infrastructure knowledge lives with the infrastructure code, not in a vendor's database.
+
+## Personal
+
+This is the most radical application. But it follows directly from the same logic.
+
+A person with values, roles, and long-term intentions faces exactly the same structural problem as a software project:
+
+```
+Value "depth"       --tension-->  Role "parent"
+    reason: time is finite and both demand it fully
+
+Decision "left company X"  --constraint-->  Future career choices
+    with the same architectural weight as ADR-001
+
+State "learning Rust"      --blocker-->     "contribute to ontoref-core"
+    a personal FSM dimension, not a task list item
+```
+
+The knowledge graph here is not a productivity system. It's not a second brain of notes. It's a graph of *commitments* — typed, provenanced, with active tensions named rather than suppressed.
+
+The drift problem is identical: who you are becoming versus who you intended to be. Reflection modes for a person might be weekly reviews, annual state assessments, or triggered reviews when a major decision introduces new constraints. The operational loop is the same — observe, detect drift, execute, record.
+
+## The Same Stack, Different Subject
+
+All three domains run on the same infrastructure:
+
+**Sovereign storage:** Knowledge lives alongside its subject — in the project repo, in the infrastructure code, in a personal vault. No vendor holds it. No platform dependency. NCL files versioned with the same discipline as code.
+
+**Multi-surface access:**
+
+| Surface | Consumer |
+|---|---|
+| CLI (Nushell) | Developer, scripts, CI |
+| UI (axum) | Visual review, onboarding |
+| MCP | AI agents |
+| GraphQL | External tools, dashboards |
+
+**VCS alignment:** jj's model — where the working copy is always a commit — aligns naturally with ontoref's approach to continuous state capture. There is no "dirty state"; every moment is capturable. Radicle as transport makes the whole stack sovereign and P2P — code collaboration without a central platform.
+
+**Federation:** Projects querying each other's knowledge graphs without a central broker. Infrastructure environments exposing their self-knowledge to the projects that depend on them. Individuals whose personal knowledge graph informs their contributions to the projects they work on.
+
+## Why Scale Matters
+
+The smallest unit that can adopt ontoref is a person. One developer, one project, one infrastructure environment. No team required. No budget conversation.
+
+The largest unit is an ecosystem of federated, self-describing projects that can query each other's knowledge without any of them being the authoritative central source.
+
+That range — individual to ecosystem — is possible because the protocol is the same at every scale. The ontological DAG and the reflection DAG don't change shape based on the size of the subject. They change in density and depth, not in structure.
+
+The organizations that Talisman argues should invest in knowledge infrastructure are trying to build this top-down, with dedicated teams and multi-year programs. Ontoref makes it available bottom-up — starting with a single project, a single environment, a single person, each describing themselves in the same language.
+
+---
+
+*Ontoref is open source. The protocol specification, Nushell automation, and Rust crates are at [github.com/jesusperezlorenzo/ontoref](https://github.com/jesusperezlorenzo/ontoref).*
diff --git a/assets/presentation/docs/posts/en/your-ontology-should-live-with-your-code.md b/assets/presentation/docs/posts/en/your-ontology-should-live-with-your-code.md
new file mode 100644
index 0000000..59f3fa6
--- /dev/null
+++ b/assets/presentation/docs/posts/en/your-ontology-should-live-with-your-code.md
@@ -0,0 +1,106 @@
+---
+# Post metadata
+id: "your-ontology-should-live-with-your-code"
+title: "Your Ontology Should Live With Your Code"
+slug: "your-ontology-should-live-with-your-code"
+subtitle: "Sovereignty, federation, and why knowledge infrastructure without a home isn't infrastructure at all"
+excerpt: "Enterprise knowledge graphs live in a triplestore maintained by a dedicated team on a vendor's platform. When the team changes, the budget gets cut, or the vendor pivots, the knowledge disappears. Ontoref takes the opposite approach: sovereign, local-first knowledge that lives alongside its subject, versioned, queryable across four surfaces, and federated without a central broker."
+
+# Publication info
+author: "Jesús Pérez"
+date: "2026-05-10"
+published: false
+featured: false
+
+# Categorization
+category: "ontoref"
+tags: ["ontoref", "sovereignty", "federation", "radicle", "jujutsu", "knowledge-infrastructure"]
+
+# Display
+read_time: "6 min read"
+sort_order: 4
+css_class: "category-ontoref"
+category_description: "Ontoref — protocol and tooling for structured self-knowledge in software projects"
+category_published: true
+---
+
+# Your Ontology Should Live With Your Code
+
+Jessica Talisman, in her KGC 2026 talk, says something worth highlighting: *"Your ontology is your moat — your IP."*
+
+She's right. And she's describing the enterprise knowledge graph model, where the moat is held in a vendor's triplestore, maintained by a dedicated team, on a platform with a subscription. When the team changes, the budget gets cut, or the vendor pivots, the moat drains.
+
+Ontoref takes the opposite approach.
+
+## What Sovereign Knowledge Means
+
+In ontoref, the knowledge of what a project is — its principles, practices, tensions, architectural decisions, operational state — lives as NCL files alongside the code. In the repository. Versioned with the same discipline as the source.
+
+This is not a documentation approach. It's a storage decision with architectural consequences:
+
+- **No vendor can take it from you.** The ontology is in your repo, not their database.
+- **No platform migration.** GitHub, Gitea, Radicle — the knowledge moves with the code.
+- **No dedicated infrastructure.** The daemon runs locally; the files are the source of truth.
+- **Offline by default.** Knowledge that requires a network connection to exist is not infrastructure — it's a service.
+
+The closest analogy is SQLite vs. PostgreSQL: local-first, embedded, always available, no connection required. When you need distribution, you add it. But the baseline is sovereignty.
+
+## Four Surfaces, One Source
+
+Sovereign storage doesn't mean isolated. Ontoref exposes the same ontological knowledge across four surfaces simultaneously:
+
+**CLI (Nushell):** The developer's native interface. `ontoref describe project`, `ontoref graph ontology`, `ontoref sync diff --docs`. Fast, scriptable, CI-composable. The surface for humans who live in the terminal and for automated checks that run in pipelines.
+
+**UI (axum):** A web interface for visual inspection, onboarding, and management. Not a dashboard over a remote API — a local server that renders the same NCL-backed knowledge. The surface for project managers, new contributors, or anyone who prefers navigating a graph visually.
+
+**MCP:** The agent surface. Ontoref's MCP endpoint is the semantic layer that sits above the transport protocol. As Talisman's framework correctly identifies, MCP moves bytes — it doesn't establish shared meaning. Ontoref provides that meaning through its ontological layer, exposed via MCP to agents that need structured knowledge about the project to work accurately rather than hallucinate context.
+
+**GraphQL:** The integration surface. SPARQL is the semantic web standard, but GraphQL is what most developers know and every tooling ecosystem supports. A GraphQL API over the ontological DAG removes the barrier of expertise without sacrificing structure.
+
+Four surfaces, one canonical source: the NCL files in the repository.
+
+## jj and Radicle
+
+The storage model pairs naturally with a specific class of VCS tooling.
+
+**Jujutsu (jj):** A Git-compatible VCS where the working copy is always a commit — there is no dirty state. This aligns directly with ontoref's approach to continuous state capture: every project state is capturable, every transition is a first-class event. jj's operation log is the VCS equivalent of ontoref's reflection session history.
+
+**Radicle:** P2P, local-first, sovereign code collaboration. If the code is sovereign, the knowledge about that code should be too. Radicle as transport means the federation of ontoref-enabled projects doesn't depend on any centralized platform — no GitHub, no GitLab, no hosted Gitea. Nodes discover each other and exchange knowledge directly.
+
+Together: sovereign code and sovereign knowledge, distributed without a central broker.
+
+## Federation Without a Central Broker
+
+The enterprise model for knowledge federation is a central knowledge graph that all projects point to. One team maintains it. One platform hosts it. Every query routes through it.
+
+Ontoref's federation model is different:
+
+```
+project-A queries: which projects implement "zero-trust-auth"?
+project-B, project-C respond from their own ontologies
+no central server that knows everything
+```
+
+This is possible because each project is self-describing. The ontological DAG in `.ontology/core.ncl` is queryable without any external dependency. Federation via NATS connects the daemons — each project's daemon registers itself, publishes its capabilities, and responds to queries from other daemons in the federation.
+
+The result: ecosystem-level visibility without ecosystem-level centralization. You can discover which projects share your architectural constraints, which implement compatible practices, which are in conflicting states — without any of them requiring a shared platform.
+
+## Why This Architecture Now
+
+Two shifts made this viable:
+
+**Agentic AI as the consumer.** Before agents, knowledge infrastructure was built for human consumption — ontologists maintaining graphs for analysts to query. Agents are different: they need machine-readable, structured, typed knowledge to operate accurately. Ontoref's multi-surface model is designed for a world where the primary consumer of project knowledge is not a human reading a wiki but an agent making decisions in a context window.
+
+**Local-first as the default.** The decade-long trend toward cloud-hosted everything is reversing for knowledge infrastructure. The risks are clear: vendor lock-in, platform discontinuity, sovereignty loss, and the compounding cost of external dependencies for what should be internal knowledge. Local-first with optional federation is the architecture that survives platform changes.
+
+## The Moat Argument
+
+Talisman's framing — "your ontology is your moat" — is more accurate than she might intend it to be. A moat works because it's attached to the thing it protects. A moat stored in someone else's database is not a moat. It's a service agreement.
+
+When the knowledge of what your project is, why it exists as it does, and what decisions have lasting consequences lives alongside the code that implements those decisions — versioned, queryable, sovereign, federated — it is genuinely yours. It accumulates. It survives team changes. It informs agents. It can be queried by projects that depend on you.
+
+That's the kind of moat worth building.
+
+---
+
+*Ontoref is open source. The protocol specification, Nushell automation, and Rust crates are at [github.com/jesusperezlorenzo/ontoref](https://github.com/jesusperezlorenzo/ontoref).*
diff --git a/assets/presentation/docs/posts/es/dags-en-todos-lados-ninguno-sabe-lo-que-es.md b/assets/presentation/docs/posts/es/dags-en-todos-lados-ninguno-sabe-lo-que-es.md
new file mode 100644
index 0000000..222316d
--- /dev/null
+++ b/assets/presentation/docs/posts/es/dags-en-todos-lados-ninguno-sabe-lo-que-es.md
@@ -0,0 +1,106 @@
+---
+# Post metadata
+id: "dags-en-todos-lados-ninguno-sabe-lo-que-es"
+title: "Los DAGs están en todos lados. Ninguno sabe lo que es."
+slug: "dags-en-todos-lados-ninguno-sabe-lo-que-es"
+subtitle: "Cada build system, pipeline CI y runbook usa DAGs para ejecución. Ontoref los usa para conocimiento."
+excerpt: "Pipelines CI/CD, compiladores, runbooks, orquestadores de datos — todos usan grafos acíclicos dirigidos. Todos sin excepción los usan como modelos de ejecución: esto antes que aquello, ordenación topológica, resolución de dependencias. Ninguno los usa para representar qué es el sistema, por qué existe, o qué trade-offs lo definen. Ese es el gap que llena ontoref."
+
+# Publication info
+author: "Jesús Pérez"
+date: "2026-05-10"
+published: false
+featured: false
+
+# Categorization
+category: "ontoref"
+tags: ["ontoref", "dag", "ontology", "knowledge-graphs", "arquitectura-software"]
+
+# Display
+read_time: "6 min read"
+sort_order: 2
+css_class: "category-ontoref"
+category_description: "Ontoref — protocolo y herramientas para el auto-conocimiento estructurado en proyectos de software"
+category_published: true
+---
+
+# Los DAGs están en todos lados. Ninguno sabe lo que es.
+
+Cada build system usa DAGs. Cada pipeline CI/CD usa DAGs. Compiladores, orquestadores de datos, runbooks, gestores de paquetes, operadores de Kubernetes — todos DAGs. Los grafos acíclicos dirigidos son tan ubicuos en la infraestructura de software que la pregunta "¿por qué DAGs?" ya ni registra como pregunta.
+
+Pero hay una segunda pregunta que nadie hace: *¿qué representan esos DAGs?*
+
+La respuesta es siempre la misma: **orden de ejecución**. Esto antes que aquello. Ordenación topológica. Resolución de dependencias. El grafo describe *cómo* se computa algo, no *qué* es.
+
+```
+CI/CD:       test → build → deploy
+Compilador:  parse → typecheck → codegen → link
+Runbook:     check_health → drain → restart → verify
+```
+
+Estos grafos son inertes respecto al sistema sobre el que operan. Un pipeline CI no sabe qué es el proyecto, por qué se construyó así, o qué trade-offs definen su arquitectura. Sabe que los tests deben pasar antes del deploy. Eso es todo lo que sabe.
+
+## El gap semántico
+
+Esta inercia no es un defecto — es apropiada. Los build systems deben ser rápidos y mecánicos. Los runbooks deben ser ejecutables sin requerir conocimiento filosófico sobre el sistema. Los grafos de ejecución y los grafos de conocimiento son herramientas distintas para propósitos distintos.
+
+El problema es que la industria del software ha recurrido a DAGs para *todo excepto la representación de conocimiento*. El conocimiento sobre qué es un sistema — sus principios, sus decisiones arquitectónicas, sus tensiones activas, su superficie de capacidades — vive en documentos. Wikis. READMEs. Tickets. Tradición verbal.
+
+Todo esto es no-estructurado, no-tipado, no-consultable, y garantizado a derivar del sistema real en meses. Describe lo que el proyecto era, no lo que es.
+
+## El doble DAG de ontoref
+
+Ontoref usa DAGs en dos modos fundamentalmente distintos, y la distinción importa:
+
+**DAG ontológico — aristas con tipo semántico:**
+
+```
+Practice "ncl-schemas"  --implements-->  Principle "type-safety"
+Practice "ncl-schemas"  --enforces-->    Constraint "zero-runtime-deps"
+Practice "ncl-schemas"  --enables-->     Capability "agent-queryable-state"
+Practice "ncl-schemas"  --tension-->     Practice "adoption-friction"
+```
+
+Estas aristas no son "depende de". Son relaciones tipadas: `implements`, `enforces`, `enables`, `tension`. El grafo es el conocimiento — compromisos formales sobre qué es el proyecto y cómo se relacionan sus conceptos. El equivalente sería un compilador que sabe por qué existe y qué trade-offs arquitectónicos tomó.
+
+**DAG de reflexión — contratos ejecutables con restricción de actor:**
+
+```
+step "validate-state"  (actor: any)        -->  step "run-mode"  (actor: developer)
+step "run-mode"        (actor: developer)  -->  step "report"    (actor: agent|developer)
+```
+
+Esto no es solo ordenación de ejecución. El DAG se valida como contrato antes de ejecutarse, registra su propio progreso a través de transiciones de estado, y codifica quién puede ejecutar cada paso. Un agente que intenta ejecutar un paso restringido a developer recibe un rechazo tipado, no un error de runtime.
+
+## La conexión faltante
+
+Lo que hace esto no-trivial es que los dos DAGs se comunican:
+
+- El DAG ontológico declara lo que el proyecto ES — constraints activos, prácticas en tensión, dimensiones de estado actuales
+- El DAG de reflexión OPERA sobre ese estado — detectando deriva, ejecutando modos, registrando transiciones
+- Las migraciones propagan cambios en el DAG ontológico a proyectos consumidores descendentes
+
+En todos los demás sistemas que usan DAGs, el grafo se evalúa y se descarta. En ontoref, la evaluación modifica el estado que la siguiente ejecución lee. El grafo tiene memoria.
+
+## Por qué esto no existía antes
+
+Tres cosas tuvieron que alinearse:
+
+**Los agentes como consumidores.** Antes de la IA agéntica, no había ningún consumidor automatizado que necesitara conocimiento estructurado sobre un proyecto. Los humanos podían leer el wiki. Los agentes no pueden — necesitan conocimiento del proyecto tipado, consultable y machine-readable para trabajar con precisión en lugar de alucinar contexto. El DAG de conocimiento de ontoref es lo que los agentes consumen vía MCP.
+
+**Lenguajes de configuración con contratos.** Expresar una ontología sin un triplestore externo requería un lenguaje de configuración con tipos y contratos. Nickel (NCL) provee exactamente eso: un lenguaje de configuración tipado y lazy donde las violaciones de schema se detectan en tiempo de evaluación, no de runtime. RDF/OWL habría requerido infraestructura dedicada y expertise especializado.
+
+**Escala de repositorio, no de empresa.** Los knowledge graphs enterprise son iniciativas de varios años. Ontoref opera a la escala de un repositorio único — adopción incremental, sin enforcement, sin equipo dedicado. La unidad más pequeña que puede adoptarlo es un proyecto de uno.
+
+## La consecuencia
+
+Cuando un proyecto tiene un DAG de conocimiento que sus agentes pueden consumir, la aritmética de precisión cambia por completo. Los números de la investigación presentada en KGC 2026:
+
+- Retrieval anclado en ontología: **3.4× más preciso** que vector RAG en queries enterprise
+- Queries validadas por ontología: la precisión salta del **16% al 72%** en question-answering sobre SQL
+
+La brecha se cierra no con un modelo más grande sino con conocimiento estructurado contra el que el modelo puede razonar. Un DAG donde las aristas significan algo.
+
+---
+
+*Ontoref es open source. La especificación del protocolo, la automatización en Nushell, y los crates de Rust están en [github.com/jesusperezlorenzo/ontoref](https://github.com/jesusperezlorenzo/ontoref).*
diff --git a/assets/presentation/docs/posts/es/la-ia-herramienta-conocimiento-quien-mantiene-vivo.md b/assets/presentation/docs/posts/es/la-ia-herramienta-conocimiento-quien-mantiene-vivo.md
new file mode 100644
index 0000000..0e72a01
--- /dev/null
+++ b/assets/presentation/docs/posts/es/la-ia-herramienta-conocimiento-quien-mantiene-vivo.md
@@ -0,0 +1,97 @@
+---
+# Post metadata
+id: "la-ia-herramienta-conocimiento-quien-mantiene-vivo"
+title: "La IA es una herramienta de conocimiento. ¿Pero quién mantiene vivo el conocimiento?"
+slug: "la-ia-herramienta-conocimiento-quien-mantiene-vivo"
+subtitle: "Talisman tiene razón — y se detiene exactamente donde empieza el problema difícil"
+excerpt: "La charla de Jessica Talisman en KGC 2026 es la articulación más clara de por qué fallan las estrategias de AI: las organizaciones invierten en infraestructura de datos y esperan que emerja el razonamiento. Tiene razón. Pero su solución — construir infraestructura de conocimiento — se detiene exactamente donde empieza el problema difícil: ¿quién evita que el conocimiento derive?"
+
+# Publication info
+author: "Jesús Pérez"
+date: "2026-05-10"
+published: false
+featured: false
+
+# Categorization
+category: "ontoref"
+tags: ["ontoref", "knowledge-graphs", "ontology", "ai-infrastructure", "reflection"]
+
+# Display
+read_time: "7 min read"
+sort_order: 1
+css_class: "category-ontoref"
+category_description: "Ontoref — protocolo y herramientas para el auto-conocimiento estructurado en proyectos de software"
+category_published: true
+---
+
+# La IA es una herramienta de conocimiento. ¿Pero quién mantiene vivo el conocimiento?
+
+Jessica Talisman dio una charla en KGC 2026 llamada "Stop Betting, Start Building." Los datos con los que abre son contundentes:
+
+- **89%** de las empresas no reportan impacto de productividad de la IA después de tres años (NBER, Feb 2026, n=5.937 ejecutivos)
+- Los developers experimentados son **19% más lentos** con herramientas de IA, no más rápidos (METR RCT, 2025)
+- El trabajador promedio que usa IA gana **−14 minutos** netos por semana en productividad (Foxit/Sapio, marzo 2026)
+
+El mercado es alcista. La evidencia, no.
+
+Su diagnóstico es correcto: *la IA es una herramienta de conocimiento, no de datos.* Las organizaciones invierten en data lakes, vector stores y pipelines ETL esperando que el contexto y el razonamiento emerjan de tokens crudos. No va a ocurrir. Los modelos fueron entrenados sobre linked data, triples RDF y vocabularios controlados — las quince principales fuentes de entrenamiento de C4 son intensivas en knowledge graphs. Luego se despliegan en entornos despojados de toda esa estructura, y nos preguntamos por qué alucinan.
+
+Su prescripción también es correcta: construye infraestructura de conocimiento. Primero vocabularios controlados. Taxonomías. Tesauros. Ontologías — compromisos formales, clases, propiedades, constraints. Knowledge graphs encima. Y goviernalos. Para siempre, no como un proyecto.
+
+Tiene razón. Y se detiene exactamente donde empieza el problema difícil.
+
+## El gap de governance
+
+El sexto paso en su stack es "Govern — esto es infraestructura, stewardship, versionado, crecimiento, para siempre." Nombra el problema. Pero nombrarlo no es un mecanismo.
+
+La pregunta real es: *¿quién asegura que la ontología no derive respecto al sistema que describe?*
+
+El software evoluciona. Una decisión arquitectónica tomada en marzo invalida un nodo del grafo en octubre. Se incorpora un nuevo equipo con modelos mentales distintos. Una librería se reemplaza y un constraint se convierte en ficción. Los knowledge graphs acumulan deuda igual que los codebases — silenciosamente, sin alarmas.
+
+En el mundo enterprise de knowledge graphs, la respuesta al governance es headcount: contratar ontólogos, bibliotecarios, knowledge engineers. Funciona a escala organizacional con presupuestos dedicados. No es una respuesta viable para un proyecto de software, un entorno de infraestructura, o un individuo que intenta mantener auto-conocimiento estructurado.
+
+## Lo que realmente falta: el loop operacional
+
+Toda ontología seria sin un cierre operacional se convierte en arqueología en dieciocho meses. Hermosa, formalmente correcta, y describiendo un sistema que ya no existe.
+
+La pieza que falta no es más conocimiento. Es un loop de retroalimentación que:
+
+1. **Observa** el estado actual del sistema contra su intención declarada
+2. **Detecta** la deriva antes de que se vuelva permanente
+3. **Ejecuta** operaciones que reducen esa deriva
+4. **Registra** decisiones con peso arquitectónico duradero
+5. **Propaga** los cambios a todo lo que depende de este conocimiento
+
+Este es el Yang al Yin de la ontología. Talisman describe el Yin con precisión. El Yang es lo que evita que sea un artefacto.
+
+## Ontoref: las dos mitades
+
+Ontoref es un protocolo para el auto-conocimiento estructurado en proyectos de software. Opera como dos capas coexistentes que no pueden funcionar la una sin la otra:
+
+**Ontología (lo que ES):** nodos y aristas tipadas que representan prácticas, principios, tensiones, capacidades. No documentación — compromisos formales. Un nodo `Practice` que `implements` un `Principle`, `enforces` un `Constraint`, `enables` una `Capability`, y está en `tension` activa con otra `Practice`. El proyecto como knowledge graph.
+
+**Reflection (lo que DEVIENE):** DAGs ejecutables — modos operacionales que corren contra el estado ontológico, detectan deriva, registran transiciones y propagan cambios. El comando `sync diff --docs` que detecta cuándo la documentación de un crate ha derivado de su nodo ontológico. El FSM en `state.ncl` que registra dónde está cada dimensión del proyecto versus adónde pretende ir. El sistema de migraciones que asegura que los cambios de protocolo lleguen a cada proyecto consumidor.
+
+La tensión entre estas dos capas no es un defecto de diseño — está nombrada explícitamente en la ontología del propio proyecto como su identidad central:
+
+> *"Ontology captures what IS. Reflection captures what BECOMES. Both must coexist without one dominating."*
+
+Sin Reflection, la Ontología cristaliza. Se convierte en una fotografía de lo que el proyecto quería ser en el momento en que se escribió. Sin Ontología, Reflection es ejecución sin verdad — sabe cómo operar pero no sobre qué opera.
+
+## Los números de precisión
+
+Los datos de Talisman sobre lo que el conocimiento estructurado hace a la precisión de la IA merecen repetirse:
+
+- Question-answering sobre SQL enterprise: **16% → 72%** cuando una ontología verifica y repara queries generadas por LLM (Allemang & Sequeda, data.world AI Lab, 2024)
+- GraphRAG vs. vector RAG: **3.4× precisión** en 43 queries enterprise (Diffbot KG-LM Benchmark, 2023)
+- Vector RAG colapsa a **0% precisión más allá de cinco entidades por query**. El retrieval anclado en KG sostiene el rendimiento mucho más allá
+
+La brecha de precisión no se cierra con un modelo más grande. Se cierra con un schema definido, una ontología, y una query validada.
+
+Pero solo si la ontología se mantiene viva. Solo si algo o alguien ejecuta el loop operacional que evita que derive hacia la ficción.
+
+Esa es la parte que el framework de Talisman no provee. Para eso existe Reflection.
+
+---
+
+*Ontoref es open source. La especificación del protocolo, la automatización en Nushell, y los crates de Rust están en [github.com/jesusperezlorenzo/ontoref](https://github.com/jesusperezlorenzo/ontoref).*
diff --git a/assets/presentation/docs/posts/es/tu-ontologia-deberia-vivir-con-tu-codigo.md b/assets/presentation/docs/posts/es/tu-ontologia-deberia-vivir-con-tu-codigo.md
new file mode 100644
index 0000000..295152b
--- /dev/null
+++ b/assets/presentation/docs/posts/es/tu-ontologia-deberia-vivir-con-tu-codigo.md
@@ -0,0 +1,106 @@
+---
+# Post metadata
+id: "tu-ontologia-deberia-vivir-con-tu-codigo"
+title: "Tu ontología debería vivir con tu código"
+slug: "tu-ontologia-deberia-vivir-con-tu-codigo"
+subtitle: "Soberanía, federación, y por qué la infraestructura de conocimiento sin hogar no es infraestructura"
+excerpt: "Los knowledge graphs enterprise viven en un triplestore mantenido por un equipo dedicado en la plataforma de un vendor. Cuando el equipo cambia, el presupuesto se recorta, o el vendor pivota, el conocimiento desaparece. Ontoref toma el enfoque opuesto: conocimiento soberano y local-first que vive junto a su sujeto, versionado, consultable a través de cuatro superficies, y federado sin un broker central."
+
+# Publication info
+author: "Jesús Pérez"
+date: "2026-05-10"
+published: false
+featured: false
+
+# Categorization
+category: "ontoref"
+tags: ["ontoref", "soberania", "federacion", "radicle", "jujutsu", "infraestructura-conocimiento"]
+
+# Display
+read_time: "6 min read"
+sort_order: 4
+css_class: "category-ontoref"
+category_description: "Ontoref — protocolo y herramientas para el auto-conocimiento estructurado en proyectos de software"
+category_published: true
+---
+
+# Tu ontología debería vivir con tu código
+
+Jessica Talisman, en su charla de KGC 2026, dice algo que merece destacarse: *"Tu ontología es tu moat — tu IP."*
+
+Tiene razón. Y está describiendo el modelo enterprise de knowledge graphs, donde el moat está retenido en el triplestore de un vendor, mantenido por un equipo dedicado, en una plataforma con suscripción. Cuando el equipo cambia, el presupuesto se recorta, o el vendor pivota, el moat se drena.
+
+Ontoref toma el enfoque opuesto.
+
+## Qué significa conocimiento soberano
+
+En ontoref, el conocimiento de qué es un proyecto — sus principios, prácticas, tensiones, decisiones arquitectónicas, estado operacional — vive como archivos NCL junto al código. En el repositorio. Versionado con la misma disciplina que el source.
+
+Esto no es un enfoque de documentación. Es una decisión de almacenamiento con consecuencias arquitectónicas:
+
+- **Ningún vendor puede quitártelo.** La ontología está en tu repo, no en su base de datos.
+- **Sin migración de plataforma.** GitHub, Gitea, Radicle — el conocimiento se mueve con el código.
+- **Sin infraestructura dedicada.** El daemon corre localmente; los archivos son la fuente de verdad.
+- **Offline por defecto.** El conocimiento que requiere conexión de red para existir no es infraestructura — es un servicio.
+
+La analogía más cercana es SQLite vs. PostgreSQL: local-first, embebido, siempre disponible, sin conexión requerida. Cuando necesitas distribución, la añades. Pero el baseline es soberanía.
+
+## Cuatro superficies, una fuente
+
+Almacenamiento soberano no significa aislado. Ontoref expone el mismo conocimiento ontológico a través de cuatro superficies simultáneamente:
+
+**CLI (Nushell):** La interfaz nativa del developer. `ontoref describe project`, `ontoref graph ontology`, `ontoref sync diff --docs`. Rápido, scriptable, composable en CI. La superficie para humanos que viven en la terminal y para checks automatizados que corren en pipelines.
+
+**UI (axum):** Una interfaz web para inspección visual, onboarding y gestión. No un dashboard sobre una API remota — un servidor local que renderiza el mismo conocimiento respaldado por NCL. La superficie para project managers, nuevos contribuidores, o cualquiera que prefiera navegar un grafo visualmente.
+
+**MCP:** La superficie de agentes. El endpoint MCP de ontoref es la capa semántica que se sienta encima del protocolo de transporte. Como identifica correctamente el framework de Talisman, MCP mueve bytes — no establece significado compartido. Ontoref provee ese significado a través de su capa ontológica, expuesta vía MCP a agentes que necesitan conocimiento estructurado del proyecto para trabajar con precisión en lugar de alucinar contexto.
+
+**GraphQL:** La superficie de integración. SPARQL es el estándar del semantic web, pero GraphQL es lo que la mayoría de developers conocen y cada ecosistema de herramientas soporta. Una API GraphQL sobre el DAG ontológico elimina la barrera de expertise sin sacrificar estructura.
+
+Cuatro superficies, una fuente canónica: los archivos NCL en el repositorio.
+
+## jj y Radicle
+
+El modelo de almacenamiento se combina naturalmente con una clase específica de herramientas VCS.
+
+**Jujutsu (jj):** Un VCS compatible con Git donde el working copy es siempre un commit — no hay dirty state. Esto se alinea directamente con el enfoque de ontoref de captura continua de estado: cada estado del proyecto es capturable, cada transición es un evento de primera clase. El operation log de jj es el equivalente VCS del historial de sesiones de reflexión de ontoref.
+
+**Radicle:** Colaboración de código P2P, local-first, soberana. Si el código es soberano, el conocimiento sobre ese código también debería serlo. Radicle como transporte significa que la federación de proyectos con ontoref no depende de ninguna plataforma centralizada — sin GitHub, sin GitLab, sin Gitea hosteado. Los nodos se descubren entre sí e intercambian conocimiento directamente.
+
+Juntos: código soberano y conocimiento soberano, distribuidos sin un broker central.
+
+## Federación sin broker central
+
+El modelo enterprise para federación de conocimiento es un knowledge graph central al que todos los proyectos apuntan. Un equipo lo mantiene. Una plataforma lo aloja. Cada query lo atraviesa.
+
+El modelo de federación de ontoref es distinto:
+
+```
+proyecto-A consulta: ¿qué proyectos implementan "zero-trust-auth"?
+proyecto-B, proyecto-C responden desde sus propias ontologías
+sin servidor central que lo sepa todo
+```
+
+Esto es posible porque cada proyecto se describe a sí mismo. El DAG ontológico en `.ontology/core.ncl` es consultable sin ninguna dependencia externa. La federación vía NATS conecta los daemons — el daemon de cada proyecto se registra, publica sus capacidades, y responde a queries de otros daemons en la federación.
+
+El resultado: visibilidad a nivel de ecosistema sin centralización a nivel de ecosistema. Puedes descubrir qué proyectos comparten tus constraints arquitectónicos, cuáles implementan prácticas compatibles, cuáles están en estados conflictivos — sin que ninguno requiera una plataforma compartida.
+
+## Por qué esta arquitectura ahora
+
+Dos cambios hicieron esto viable:
+
+**La IA agéntica como consumidor.** Antes de los agentes, la infraestructura de conocimiento se construía para consumo humano — ontólogos manteniendo grafos para que los analistas los consultaran. Los agentes son diferentes: necesitan conocimiento machine-readable, estructurado y tipado para operar con precisión. El modelo multi-superficie de ontoref está diseñado para un mundo donde el consumidor principal del conocimiento del proyecto no es un humano leyendo un wiki sino un agente tomando decisiones en una ventana de contexto.
+
+**Local-first como el default.** La tendencia de una década hacia todo en la nube está revirtiendo para la infraestructura de conocimiento. Los riesgos son claros: vendor lock-in, discontinuidad de plataforma, pérdida de soberanía, y el coste compuesto de dependencias externas para lo que debería ser conocimiento interno. Local-first con federación opcional es la arquitectura que sobrevive a los cambios de plataforma.
+
+## El argumento del moat
+
+El framing de Talisman — "tu ontología es tu moat" — es más acertado de lo que ella puede pretender. Un moat funciona porque está unido a lo que protege. Un moat almacenado en la base de datos de otra persona no es un moat. Es un acuerdo de servicio.
+
+Cuando el conocimiento de qué es tu proyecto, por qué existe como existe, y qué decisiones tienen consecuencias duraderas vive junto al código que implementa esas decisiones — versionado, consultable, soberano, federado — es genuinamente tuyo. Se acumula. Sobrevive a cambios de equipo. Informa a agentes. Puede ser consultado por proyectos que dependen de ti.
+
+Ese es el tipo de moat que vale la pena construir.
+
+---
+
+*Ontoref es open source. La especificación del protocolo, la automatización en Nushell, y los crates de Rust están en [github.com/jesusperezlorenzo/ontoref](https://github.com/jesusperezlorenzo/ontoref).*
diff --git a/assets/presentation/docs/posts/es/un-protocolo-tres-sujetos.md b/assets/presentation/docs/posts/es/un-protocolo-tres-sujetos.md
new file mode 100644
index 0000000..bd77b57
--- /dev/null
+++ b/assets/presentation/docs/posts/es/un-protocolo-tres-sujetos.md
@@ -0,0 +1,122 @@
+---
+# Post metadata
+id: "un-protocolo-tres-sujetos"
+title: "Un protocolo, tres sujetos"
+slug: "un-protocolo-tres-sujetos"
+subtitle: "La misma pregunta aplica a tu proyecto, tu infraestructura, y a ti mismo"
+excerpt: "Ontoref comenzó como un protocolo para el auto-conocimiento de proyectos de software. La misma arquitectura — DAG ontológico para lo que ES, DAG de reflexión para lo que DEVIENE — aplica sin modificación a entornos de infraestructura y a individuos. El sujeto cambia. La pregunta es idéntica: ¿qué eres, qué tensiones te definen, dónde estás versus dónde pretendes estar?"
+
+# Publication info
+author: "Jesús Pérez"
+date: "2026-05-10"
+published: false
+featured: false
+
+# Categorization
+category: "ontoref"
+tags: ["ontoref", "ontology", "infraestructura", "conocimiento-personal", "reflection"]
+
+# Display
+read_time: "7 min read"
+sort_order: 3
+css_class: "category-ontoref"
+category_description: "Ontoref — protocolo y herramientas para el auto-conocimiento estructurado en proyectos de software"
+category_published: true
+---
+
+# Un protocolo, tres sujetos
+
+La pregunta que ontoref hace a un proyecto de software es precisa: *¿Qué eres? ¿Qué principios te definen? ¿Qué tensiones estás gestionando activamente? ¿Qué decisiones has tomado con consecuencias duraderas? ¿Dónde estás versus dónde pretendes estar?*
+
+Resulta que es la misma pregunta que los entornos de infraestructura necesitan responder. Y los individuos.
+
+El sujeto cambia. La pregunta es idéntica.
+
+## Lo que los tres tienen en común
+
+Cualquier sistema con identidad e intención comparte el mismo problema estructural:
+
+- **Identidad** — qué es, qué principios lo definen
+- **Tensiones activas** — trade-offs que no se resuelven, se gestionan
+- **Decisiones duraderas** — elecciones con consecuencias que restringen el futuro
+- **Estado observable** — dónde está versus dónde pretende estar
+- **Deriva** — la distancia entre intención y realidad, creciendo silenciosamente con el tiempo
+- **Operaciones** — las acciones que reducen esa deriva
+
+Los proyectos de software tienen todo esto. Los entornos de infraestructura también. Las personas también.
+
+Y todos se benefician de la misma arquitectura: una capa ontológica que captura lo que ES, y una capa de reflexión que captura lo que DEVIENE y opera para mantener ambas alineadas.
+
+## Infraestructura
+
+La infraestructura ya usa DAGs — Terraform, Pulumi, Ansible. Todos describen *cómo* se provisiona el sistema. Ninguno captura *por qué existe como existe*.
+
+El "por qué" es donde vive el conocimiento:
+
+```
+Servicio "auth"            --tension-->    Principio "stateless"
+    razón: necesita session state, pero el principio lo prohíbe
+
+Constraint "zero-downtime" --enforces-->   Práctica "blue-green"
+    razón: ADR-infra-003, después del incidente de deploy de 2025
+```
+
+Esta información no está en el plan de Terraform. Está en la memoria del engineer que estuvo allí, en un hilo de Slack de hace ocho meses, o en ningún lado. Cuando ese engineer se va, la infraestructura pierde su propia historia — sabe lo que es, no por qué.
+
+El ontoref de infraestructura captura las decisiones arquitectónicas de la propia infraestructura, con el mismo formalismo de ADR que los proyectos de software. Registra dimensiones de estado — qué entornos están estables, cuáles en migración activa, qué está bloqueado y por qué. El FSM no registra el estado de deploy; registra el *estado epistémico* — qué tan bien se entiende a sí misma la infraestructura.
+
+La federación vía NATS significa que cada entorno (dev, staging, prod) se describe a sí mismo y puede ser consultado desde los proyectos que dependen de él. GitOps soberano: el conocimiento de infraestructura vive con el código de infraestructura, no en la base de datos de un vendor.
+
+## Personal
+
+Esta es la aplicación más radical. Pero se sigue directamente de la misma lógica.
+
+Una persona con valores, roles e intenciones a largo plazo enfrenta exactamente el mismo problema estructural que un proyecto de software:
+
+```
+Valor "profundidad"   --tension-->   Rol "padre"
+    razón: el tiempo es finito y los dos lo demandan completamente
+
+Decisión "dejé empresa X"  --constraint-->   Elecciones de carrera futuras
+    con el mismo peso arquitectónico que un ADR-001
+
+Estado "aprendiendo Rust"  --blocker-->   "contribuir a ontoref-core"
+    una dimensión FSM personal, no un ítem de lista de tareas
+```
+
+El knowledge graph aquí no es un sistema de productividad. No es un segundo cerebro de notas. Es un grafo de *compromisos* — tipados, provenados, con tensiones activas nombradas en lugar de suprimidas.
+
+El problema de deriva es idéntico: quién estás llegando a ser versus quién pretendías ser. Los modos de reflexión para una persona pueden ser revisiones semanales, evaluaciones anuales de estado, o revisiones disparadas cuando una decisión mayor introduce nuevos constraints. El loop operacional es el mismo — observa, detecta deriva, ejecuta, registra.
+
+## El mismo stack, sujeto distinto
+
+Los tres dominios corren en la misma infraestructura:
+
+**Almacenamiento soberano:** El conocimiento vive junto a su sujeto — en el repo del proyecto, en el código de infraestructura, en una bóveda personal. Ningún vendor lo retiene. Sin dependencia de plataforma. Archivos NCL versionados con la misma disciplina que el código.
+
+**Acceso multi-superficie:**
+
+| Superficie | Consumidor |
+|---|---|
+| CLI (Nushell) | Developer, scripts, CI |
+| UI (axum) | Revisión visual, onboarding |
+| MCP | Agentes AI |
+| GraphQL | Herramientas externas, dashboards |
+
+**Alineación VCS:** El modelo de jj — donde el working copy es siempre un commit — se alinea naturalmente con el enfoque de ontoref de captura continua de estado. No hay "dirty state"; cada momento es capturable. Radicle como transporte hace el stack soberano y P2P — colaboración de código sin plataforma central.
+
+**Federación:** Proyectos consultando los knowledge graphs de otros sin un broker central. Entornos de infraestructura exponiendo su auto-conocimiento a los proyectos que dependen de ellos. Individuos cuyo knowledge graph personal informa sus contribuciones a los proyectos en los que trabajan.
+
+## Por qué la escala importa
+
+La unidad más pequeña que puede adoptar ontoref es una persona. Un developer, un proyecto, un entorno de infraestructura. Sin equipo requerido. Sin conversación de presupuesto.
+
+La unidad más grande es un ecosistema de proyectos federados y auto-descritos que pueden consultarse mutuamente sin que ninguno sea la fuente central autoritativa.
+
+Ese rango — de individuo a ecosistema — es posible porque el protocolo es el mismo a cada escala. El DAG ontológico y el DAG de reflexión no cambian de forma según el tamaño del sujeto. Cambian en densidad y profundidad, no en estructura.
+
+Las organizaciones que Talisman argumenta deberían invertir en infraestructura de conocimiento intentan construirla top-down, con equipos dedicados y programas de varios años. Ontoref lo hace disponible bottom-up — empezando con un solo proyecto, un solo entorno, una sola persona, cada uno describiéndose a sí mismo en el mismo lenguaje.
+
+---
+
+*Ontoref es open source. La especificación del protocolo, la automatización en Nushell, y los crates de Rust están en [github.com/jesusperezlorenzo/ontoref](https://github.com/jesusperezlorenzo/ontoref).*
diff --git a/assets/presentation/docs/talisman-article/a-two-week-sprint-for-knowledge.md b/assets/presentation/docs/talisman-article/a-two-week-sprint-for-knowledge.md
new file mode 100644
index 0000000..14c9e4b
--- /dev/null
+++ b/assets/presentation/docs/talisman-article/a-two-week-sprint-for-knowledge.md
@@ -0,0 +1,101 @@
+# A Two-Week Sprint for Knowledge
+## From Agile to Design Thinking
+
+**Jessica Talisman, MLS**
+*Apr 21, 2026*
+
+---
+
+In 1986, Hirotaka Takeuchi and Ikujiro Nonaka published "The New New Product Development Game" in the *Harvard Business Review*, describing how Honda, Canon, and Fuji-Xerox organized teams to build copiers and automobiles. Jeff Sutherland read that article, combined it with lean manufacturing principles and object oriented programming, and in 1993 adapted it for software development at Easel Corporation. Ken Schwaber formalized the method at the OOPSLA conference in 1995. By 2001, seventeen software developers gathered at a ski lodge in Snowbird, Utah, and signed the Agile Manifesto, declaring that they were "uncovering better ways of developing software." Not knowledge or meaning. A manufacturing framework for software.
+
+The Agile Manifesto's own language is explicit about its scope, yet the industry has applied Agile methodology to virtually everything in a technological environment — including knowledge work, semantic systems and now artificial intelligence. And within the same breath, organizations and experts regale the industry with talk of innovation and the emergence of a new paradigm in ways of working. Yet those ways of working are locked inside of Agile methodologies and Scrum, designed for the factory floor and the codebase — not for the messy, recursive, deeply human work of making meaning out of complexity.
+
+## Scrum Is a Manufacturing Schedule
+
+Let's return to the Takeuchi and Nonaka paper, where the title — "The New New Product Development Game" — describes how product teams at Japanese firms developed physical goods such as the FX-3500 copier and the Honda City car. They draw on the rugby metaphor because the team "tries to go the distance as a unit, passing the ball back and forth." The unit of delivery is a working prototype of a material artifact. The feedback loop is a machine that works or fails.
+
+Sutherland's 2014 book, *Scrum: The Art of Doing Twice the Work in Half the Time*, reinforces this with military and manufacturing case studies such as the F-86 fighter production and Toyota. Sutherland's book rebuilds the argument with the same assumption — that the thing being built has discrete, inspectable, testable parts that, when finished, are finished.
+
+The 2020 Scrum Guide defines the Increment as "a concrete stepping stone toward the Product Goal," one that is "additive to all prior Increments and thoroughly verified, ensuring that all Increments work together." These words are sensible if you are building a mobile application, with a login screen and a settings page. But for knowledge systems, these words are destructive, because the components of a semantic architecture are not additive in the way Scrum dictates. A controlled vocabulary is not a subcomponent of a taxonomy the way a login screen is a subcomponent of an app. A taxonomy without disciplined, vocabulary control as its underpinnings does not function well, if at all. You cannot ship half a taxonomy and call it an Increment.
+
+You see, Scrum assumes tame problems that can be time-boxed, estimated with planning poker, decomposed into user stories, tracked on a velocity chart and declared done against a checklist. If only knowledge were this tame. Knowledge organization is anything but tame. More than fifty years old, in 1973 Horst Rittel and Melvin Webber warned that applying the engineering paradigm uniformly to all problems was "bound to fail." And yet here we are, applying a manufacturing schedule and methodologies to knowledge, expecting that knowledge will fit into the confines of old ways of working — an artifact of the Third Industrial Revolution imposed on the demands of the Fourth.
+
+In 2016, Klaus Schwab wrote about this historical pattern. The First Industrial Revolution, beginning around 1760, gave us steam and mechanical production. The Second, in the late nineteenth century, gave us electricity and the assembly line — the world that produced Toyota, lean manufacturing, and the production logic that Scrum still carries as its core ethos. The Third, beginning in the 1960s, gave us semiconductors, mainframe computing, and eventually the internet — the world in which Agile was born, by software developers, for software development. The Fourth Industrial Revolution, Schwab argued, is fundamentally different, in kind, as it is defined by the fusion of artificial intelligence, cyber-physical systems and knowledge technologies that blur the boundaries between the physical, digital and biological worlds.
+
+The problems of this era are not necessarily engineering problems with stable requirements. We are wrestling with problems such as how to ground machine intelligence in meaning, how to organize evolving knowledge and how to build systems that can explain themselves. These are wicked problems that demand new epistemologies and frameworks for problem solving. Agile may have been the right methodology for the Third Industrial Revolution but we are no longer in the Third Industrial Revolution. The frameworks have not caught up, and the cost of that lag is exacerbated every time an organization forces Third Industrial Revolution principles applied to Fourth Industrial Revolution work.
+
+In other words, this is not a product problem or a platform problem — it is learning to embrace new ways of working. This is an infrastructure problem that is governed by epistemology — a fundamental philosophy pertaining to knowledge.
+
+## Why Knowledge Cannot Be Sprinted
+
+An ontology is, per Thomas Gruber's foundational definition, "a specification of a conceptualization." A conceptualization is a whole, not parts of a whole. Specifying half of it produces logical incoherence, not a minimally viable product. For example, ANSI/NISO Z39.19, the standard for constructing controlled vocabularies, requires resolving synonymy, homonymy and scope before terms enter production, because any inconsistency propagates downstream into every subsequent use. SKOS, the W3C standard for encoding knowledge organization systems, requires that hierarchical relationships form coherent structures with transitive closure and OWL 2 requires that class axioms be satisfiable. These are preconditions, necessary for logical systems to reason and express meaning.
+
+In practice, I have watched organizations attempt Agile delivery of knowledge systems in two patterns and neither ends well. In the first, the team holds stand-ups, ceremonies and retrospectives. The taxonomist brings tickets and adds fifteen terms to the product vocabulary. Tickets close and a velocity chart goes up. Nothing is integrated, because integration is the hard part and integration does not fit into two weeks so that goes to the backlog. Six months later, the vocabulary is a collection of inconsistently scoped terms, the ontology has orphan classes and the knowledge graph — if it exists — cannot reliably answer queries. But the burndown chart is immaculate and leadership is blindly satisfied because boxes were checked.
+
+In the second pattern, the team borrows from other methodologies while still calling it Agile. The taxonomist conducts multi-week domain analysis and stakeholder interviews — that is ethnography. She does card sorting and concept modeling — that is design research. She iterates with subject matter experts — that is participatory design. She ships, eventually, a coherent model. The sprint cadence was irrelevant to the work. Martin Fowler named this phenomenon in his 2018 keynote: "faux-agile" — the Agile Industrial Complex.
+
+Ron Jeffries, another original signatory of the Agile Manifesto, told developers the same year, 2018, to abandon the word entirely. Dave Thomas declared in 2014 that "Agile" had become "a magnet for anyone with points to espouse, hours to bill or products to sell." When three of the seventeen authors of a framework say it has been hollowed out, one has to wonder if Agile is appropriate for all technology work and if it is time to revisit ways of working.
+
+## Innovation Does Not Fit in a Backlog
+
+The deeper issue is that Agile and Scrum were never designed to inspire or foster innovation. They were designed to deliver known requirements with increasing efficiency. The Scrum Guide's entire apparatus — product backlog, sprint planning, Definition of Done — presupposes that the problem has been formulated by a product owner who can write it in a user story. But knowledge work, like design, is wicked. Rittel and Webber's first property of a wicked problem is that "There is no definitive formulation." The second property states that there is no stopping rule. And the fifth property illuminates that every solution is a "one-shot operation" because there is no opportunity to learn by trial-and-error without consequence.
+
+Given the nature of wicked problems, it would seem that Agile is ill fitted for wicked problems. And for that matter, ill fitted for the dynamic nature of knowledge, where knowledge infrastructures are never one-shot operations, necessitate stopping rules and without a doubt defy definitive formulations.
+
+Therefore, knowledge organization in an enterprise domain is wicked in exactly this sense. The problem of how to represent "customer" in a telecommunications company's semantic model has no pre-existing correct answer. Its formulation depends on who is doing what with the representation, and that changes as you build. The scope of the vocabulary shifts as you interview stakeholders. The taxonomy reorganizes as you encounter edge cases. The ontology's axioms tighten or loosen as the reasoner surfaces contradictions. Don't be mistaken — this does not bend to the Agile principle of welcoming changing requirements, as the requirements are constituted *through* the work. The problem is under construction, often evolving, during the solving.
+
+Organizations that manage all work through Scrum are, in effect, optimizing for predictability in domains that demand exploration. The results resist innovation and take on the appearance of productivity. Velocity charts rise while the underlying intellectual architecture is flattened into bite-sized deployments, uncoupled from the continuum that is knowledge.
+
+## Design Thinking Belongs in the Swamp
+
+Donald Schön drew the distinction that "In the varied topography of professional practice, there is a high, hard ground where practitioners can make effective use of research-based theory and technique, and there is a swampy lowland where situations are confusing 'messes' incapable of technical solution." Agile lives on the high ground while knowledge work lives in the swamp. Schön's term for how competent practitioners navigate the swamp was "reflection-in-action" — the iterative, tacit recalibration that happens while the work is happening.
+
+Design thinking, in the lineage of Herbert Simon, Richard Buchanan and Schön, is the methodology of the swamp. It assumes the problem is under construction. Its five phases — empathize, define, ideate, prototype, test — are not a linear pipeline but a recursive, iterative process that expects the definition of the problem to change as understanding deepens.
+
+Research in the *Handbook of Human-Centered Artificial Intelligence* confirms this alignment — integrating design thinking into AI development makes systems more user-focused, iterative and impactful, emphasizing data governance, transparency, explainability and bias mitigation. Carnegie Mellon University's Tepper School of Business now teaches a dedicated program on design thinking with AI, arguing that the combination addresses "complex, systemic problems" that linear frameworks cannot manage successfully. IBM's enterprise design thinking practice, developed over two decades, has been applied across client engagements precisely because it keeps human experience at the center while technology evolves underneath — design thinking is "timeless," while technology is "timestamped."
+
+A 2026 study in *Scientific Reports* examined the relationship between higher order thinking, generative AI chatbot use and engineering creativity, finding that higher order thinking had significantly greater importance for creative outcomes than AI tool use alone — reinforcing that cognitive depth, not tool velocity, drives innovation. Research in the *Proceedings of the Design Society* further argues that generative AI is shifting the designer's role from executor to paradigm-setter, making design thinking more essential as AI handles production tasks while humans must guide framing, ethics and meaning.
+
+To be clear, these frameworks are not different flavors of the same methodology. Scrum and design thinking represent totally different epistemologies. Scrum assumes the problem was handed to you while design thinking assumes the problem must be found and wrestled with.
+
+## Is AI a Knowledge Tool?
+
+This brings us to the question the industry avoids and I, for one, have been itching to ask. Is AI a knowledge tool? How do you classify it?
+
+A large language model generates plausible sequences of text. Returning to the Bender, Gebru, McMillan-Major, and Mitchell paper mentioned earlier, the authors defined a language model as a system for stitching together linguistic forms according to probabilistic patterns, "without any reference to meaning." Without reference to meaning, as LLMs out-of-the-box are absent of external grounding. They optimize for plausibility, not truth and therefore as a standalone, can be best categorized as an information-processing tool.
+
+It becomes a knowledge tool only when grounded in a semantic layer — controlled vocabularies, taxonomies, ontologies, knowledge graphs — that provides reference and provenance. The industry's own corrective measures and associated marketplace confirms this as fact. Retrieval-augmented generation, knowledge graphs for LLMs and what recent discourse calls "context engineering" are, underneath the packaging, the semantic infrastructure that should have been built first. But it was bolted on after the fact to prevent hallucination that was pretty much guaranteed the moment the model was deployed without one.
+
+The RAG paper from 2020 called out this problem, identifying that parametric memory is opaque and cannot be updated and that retrieval from structured sources is how provenance is captured and encoded. But it seems the industry has a memory problem or it was too early for most to hear the message.
+
+Ironically, provenance is a core element of knowledge management and ontology work, as provenance is a principled tenant of the discipline and problem space. Without provenance, knowledge cannot exist. Organizations are only now discovering they need it, because their generative systems cannot explain themselves. Call it lineage, call it traces. But if you are building for knowledge, call it provenance, as provenance includes traces, lineage, citations, the temporal and the spatial.
+
+So if generative AI is an information processing tool, I posture that for AI to become a knowledge tool, there must be provenance and descriptive logic. And to be frank, nobody sprints their way into provenance — it is architected and sustained.
+
+## Knowledge Is a Spiral, Not a Sprint
+
+The irony does not escape the moment. The same Nonaka who co-authored the 1986 paper that inspired Scrum went on to develop the SECI model of knowledge creation in 1995, describing knowledge as a spiral between tacit and explicit forms, moving through socialization, externalization, combination and internalization.
+
+The spiral does not have an end, as is true with spirals, emblematic of life. Each rotation creates new knowledge that becomes the substrate for the next rotation. There is no Definition of Done because the definition of what is being done is itself being constructed. Scrum took the team structure from Nonaka's first paper and discarded his next decade of work because it did not fit neatly with Scrum. The mechanisms of knowledge work, as Nonaka spent a decade arguing, are metaphor, tacit skill, apprenticeship, shared space and the patient conversion of hunches into models and models back into practice.
+
+The Ontology Pipeline™ mirrors this spiral. You start with a controlled vocabulary — a disciplined process involving agreements as to what terms mean. You mature it into a taxonomy by introducing hierarchy, structuring terms into parent-child relations. Thereafter, the taxonomy is matured into a thesaurus by adding associative relationships and synonymy, encoded in SKOS. The thesaurus is baseline work from which an ontology is developed, adding formal semantics using OWL and RDF. You architect a knowledge graph, assembling components to include the knowledge assets developed through the Ontology Pipeline™ iterative framework.
+
+Each stage prepares the ground for the next and stages cannot be skipped or bypassed. You cannot ship the knowledge graph without an ontology, and the ontology must be logically consistent and coherent to make sense of data. A knowledge graph without vocabulary control risks becoming a confident liar or failing to work with natural language. Because knowledge assets are dimensionally interconnected, knowledge is indeed a spiral, unable to be contained by manufacturing logic.
+
+## The Pattern and the Cure
+
+Organizations budget for platforms — Confluence, SharePoint, LLM subscriptions, vector databases, a RAG vendor. But they do not budget for descriptive logic and knowledge, perhaps because the problem is not solved by a single tool and is seemingly too amorphous in an Agile, data-centric world. The controlled vocabulary is cut unless it delivers value for the database. Often, the taxonomy is cut because it does not demo well and its value sits in the plumbing, delivering its punch over time. The result is an ever-growing pile of documents with a chatbot stapled to the front, confidently answering questions in the voice of a system that has no reference for any of the terms sprinkled throughout the documents.
+
+Then the system hallucinates. It contradicts itself across sessions. It surfaces outdated policy as current and surfaces private Outlook messages from deactivated employees. I have seen this in real life more times than I can count. And the optics are worse when a vendor tool is responsible for revealing private chats and email transactions. Usually the executive sponsor asks why the investment has not worked, ignoring the evidence of faux pas such as the surfacing of private communications. All in a day's work, when burndown charts and velocity set the precedent for ways of working and reward systems.
+
+Executives ignore the fact that the system was never grounded. It was never grounded because grounding requires the slow, expensive, disciplined work of knowledge organization — work that does not fit into a sprint, does not demo well and fails to produce a velocity chart. The organization chooses to bypass investment in the descriptive knowledge apparatus because the lines are too blurry to clearly define according to Agile speak. Because knowledge is infrastructure, not a product or a platform. Knowledge is the gas in the tank, the fuel for meaning, the material that justifies logical reasoning.
+
+Design thinking gives that knowledge work a home. It is comfortable with ambiguity. It treats problem discovery as an equal to problem solving. It certainly does not demand a shippable increment every fourteen days. Instead, design thinking demands that you understand what you are building before you build it — and that you keep revising that understanding as the work proceeds. Library and information science, with over a century of standards for knowledge organization, provides the methods while design thinking provides the epistemology. Together, they offer what Agile never could — a framework suited to the nature of knowledge itself.
+
+A two-week sprint to deliver what? If you are building knowledge systems, two weeks will get you nothing coherent. Knowledge is not built fast and broken into shape. It is built carefully, with provenance, governance and patience. Knowledge is not a product to be manufactured and no — you cannot buy off-the-shelf knowledge that perfectly matches a domain or organization. Knowledge is nuanced, specific to each organization, domain and person.
+
+In the spirit of VC-speak, knowledge is the moat, and that moat is not Agile.
+
+---
+
+*Jessica Talisman, MLS — Intentional Arrangement*
diff --git a/assets/presentation/docs/talisman-article/a-two-week-sprint-for-knowledge.pdf b/assets/presentation/docs/talisman-article/a-two-week-sprint-for-knowledge.pdf
new file mode 100644
index 0000000..cb5a1d8
--- /dev/null
+++ b/assets/presentation/docs/talisman-article/a-two-week-sprint-for-knowledge.pdf
@@ -0,0 +1,2032 @@
+%PDF-1.7
+%����
+
+1 0 obj
+<<
+  /Type /Pages
+  /Count 7
+  /Kids [105 0 R 107 0 R 109 0 R 111 0 R 113 0 R 115 0 R 117 0 R]
+>>
+endobj
+
+2 0 obj
+<<
+  /Type /Outlines
+  /First 3 0 R
+  /Last 9 0 R
+  /Count 7
+>>
+endobj
+
+3 0 obj
+<<
+  /Parent 2 0 R
+  /Next 4 0 R
+  /Title (Scrum Is a Manufacturing Schedule)
+  /Dest 98 0 R
+>>
+endobj
+
+4 0 obj
+<<
+  /Parent 2 0 R
+  /Next 5 0 R
+  /Prev 3 0 R
+  /Title (Why Knowledge Cannot Be Sprinted)
+  /Dest 99 0 R
+>>
+endobj
+
+5 0 obj
+<<
+  /Parent 2 0 R
+  /Next 6 0 R
+  /Prev 4 0 R
+  /Title (Innovation Does Not Fit in a Backlog)
+  /Dest 100 0 R
+>>
+endobj
+
+6 0 obj
+<<
+  /Parent 2 0 R
+  /Next 7 0 R
+  /Prev 5 0 R
+  /Title (Design Thinking Belongs in the Swamp)
+  /Dest 101 0 R
+>>
+endobj
+
+7 0 obj
+<<
+  /Parent 2 0 R
+  /Next 8 0 R
+  /Prev 6 0 R
+  /Title (Is AI a Knowledge Tool?)
+  /Dest 102 0 R
+>>
+endobj
+
+8 0 obj
+<<
+  /Parent 2 0 R
+  /Next 9 0 R
+  /Prev 7 0 R
+  /Title (Knowledge Is a Spiral, Not a Sprint)
+  /Dest 103 0 R
+>>
+endobj
+
+9 0 obj
+<<
+  /Parent 2 0 R
+  /Prev 8 0 R
+  /Title (The Pattern and the Cure)
+  /Dest 104 0 R
+>>
+endobj
+
+10 0 obj
+<<
+  /Type /StructTreeRoot
+  /RoleMap <<
+    /Datetime /Span
+    /Terms /Part
+    /Title /P
+    /Strong /Span
+    /Em /Span
+  >>
+  /K [18 0 R]
+  /ParentTree <<
+    /Nums [0 11 0 R 1 12 0 R 2 13 0 R 3 14 0 R 4 15 0 R 5 16 0 R 6 17 0 R]
+  >>
+  /ParentTreeNextKey 7
+>>
+endobj
+
+11 0 obj
+[77 0 R 76 0 R 75 0 R 73 0 R 71 0 R 71 0 R 72 0 R 71 0 R 71 0 R 71 0 R 71 0 R 71 0 R 71 0 R 71 0 R 71 0 R 70 0 R 70 0 R 70 0 R 70 0 R 70 0 R 70 0 R 70 0 R 69 0 R 68 0 R 68 0 R 68 0 R 68 0 R 68 0 R 68 0 R 68 0 R 66 0 R 67 0 R 66 0 R 66 0 R 66 0 R 66 0 R 66 0 R 66 0 R]
+endobj
+
+12 0 obj
+[65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 64 0 R 64 0 R 64 0 R 64 0 R 64 0 R 64 0 R 64 0 R 64 0 R 63 0 R 63 0 R 63 0 R 63 0 R 63 0 R 63 0 R 63 0 R 63 0 R 63 0 R 63 0 R 63 0 R 62 0 R 62 0 R 62 0 R 62 0 R 62 0 R 62 0 R 62 0 R 62 0 R 62 0 R 62 0 R 61 0 R 61 0 R 61 0 R]
+endobj
+
+13 0 obj
+[60 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 58 0 R 58 0 R 58 0 R 58 0 R 58 0 R 58 0 R 58 0 R 58 0 R 58 0 R 57 0 R 57 0 R 57 0 R 57 0 R 57 0 R 57 0 R 57 0 R 56 0 R 56 0 R 56 0 R 56 0 R 56 0 R 56 0 R 55 0 R 54 0 R 54 0 R 54 0 R 54 0 R 54 0 R 54 0 R 54 0 R]
+endobj
+
+14 0 obj
+[54 0 R 54 0 R 54 0 R 53 0 R 53 0 R 53 0 R 53 0 R 51 0 R 51 0 R 51 0 R 51 0 R 51 0 R 51 0 R 51 0 R 51 0 R 51 0 R 52 0 R 51 0 R 51 0 R 50 0 R 50 0 R 50 0 R 50 0 R 50 0 R 50 0 R 49 0 R 48 0 R 48 0 R 48 0 R 48 0 R 48 0 R 48 0 R 48 0 R 47 0 R 47 0 R 47 0 R 47 0 R 47 0 R 45 0 R 46 0 R 45 0 R 45 0 R 45 0 R 45 0 R 45 0 R 45 0 R 45 0 R]
+endobj
+
+15 0 obj
+[45 0 R 45 0 R 45 0 R 45 0 R 45 0 R 42 0 R 44 0 R 42 0 R 42 0 R 42 0 R 42 0 R 42 0 R 43 0 R 42 0 R 42 0 R 42 0 R 41 0 R 41 0 R 41 0 R 41 0 R 41 0 R 40 0 R 39 0 R 39 0 R 38 0 R 38 0 R 38 0 R 38 0 R 38 0 R 38 0 R 38 0 R 38 0 R 37 0 R 37 0 R 37 0 R 37 0 R 37 0 R 37 0 R 37 0 R 37 0 R 36 0 R 36 0 R 36 0 R 36 0 R]
+endobj
+
+16 0 obj
+[35 0 R 35 0 R 35 0 R 35 0 R 35 0 R 35 0 R 35 0 R 34 0 R 34 0 R 34 0 R 33 0 R 32 0 R 32 0 R 32 0 R 32 0 R 31 0 R 31 0 R 31 0 R 31 0 R 31 0 R 31 0 R 31 0 R 30 0 R 30 0 R 30 0 R 30 0 R 30 0 R 30 0 R 30 0 R 30 0 R 30 0 R 29 0 R 29 0 R 29 0 R 29 0 R 29 0 R 29 0 R 29 0 R 28 0 R 27 0 R 27 0 R]
+endobj
+
+17 0 obj
+[27 0 R 27 0 R 27 0 R 27 0 R 27 0 R 27 0 R 27 0 R 27 0 R 26 0 R 26 0 R 26 0 R 26 0 R 26 0 R 26 0 R 26 0 R 25 0 R 25 0 R 25 0 R 25 0 R 25 0 R 25 0 R 25 0 R 25 0 R 25 0 R 24 0 R 24 0 R 24 0 R 24 0 R 24 0 R 24 0 R 24 0 R 24 0 R 23 0 R 23 0 R 23 0 R 23 0 R 23 0 R 23 0 R 22 0 R 21 0 R 19 0 R 19 0 R 19 0 R 20 0 R 19 0 R]
+endobj
+
+18 0 obj
+<<
+  /Type /StructElem
+  /S /Document
+  /P 10 0 R
+  /K [77 0 R 76 0 R 74 0 R 73 0 R 71 0 R 70 0 R 69 0 R 68 0 R 66 0 R 65 0 R 64 0 R 63 0 R 62 0 R 61 0 R 60 0 R 59 0 R 58 0 R 57 0 R 56 0 R 55 0 R 54 0 R 53 0 R 51 0 R 50 0 R 49 0 R 48 0 R 47 0 R 45 0 R 42 0 R 41 0 R 40 0 R 39 0 R 38 0 R 37 0 R 36 0 R 35 0 R 34 0 R 33 0 R 32 0 R 31 0 R 30 0 R 29 0 R 28 0 R 27 0 R 26 0 R 25 0 R 24 0 R 23 0 R 22 0 R 19 0 R]
+>>
+endobj
+
+19 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [21 0 R 40 41 42 20 0 R 44]
+  /Pg 117 0 R
+>>
+endobj
+
+20 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 19 0 R
+  /K [43]
+  /Pg 117 0 R
+>>
+endobj
+
+21 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 19 0 R
+  /K [39]
+  /Pg 117 0 R
+>>
+endobj
+
+22 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [38]
+  /Pg 117 0 R
+>>
+endobj
+
+23 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [32 33 34 35 36 37]
+  /Pg 117 0 R
+>>
+endobj
+
+24 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [24 25 26 27 28 29 30 31]
+  /Pg 117 0 R
+>>
+endobj
+
+25 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [15 16 17 18 19 20 21 22 23]
+  /Pg 117 0 R
+>>
+endobj
+
+26 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [8 9 10 11 12 13 14]
+  /Pg 117 0 R
+>>
+endobj
+
+27 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [39 40 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 117 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 117 0 R
+  >> <<
+    /Type /MCR
+    /MCID 2
+    /Pg 117 0 R
+  >> <<
+    /Type /MCR
+    /MCID 3
+    /Pg 117 0 R
+  >> <<
+    /Type /MCR
+    /MCID 4
+    /Pg 117 0 R
+  >> <<
+    /Type /MCR
+    /MCID 5
+    /Pg 117 0 R
+  >> <<
+    /Type /MCR
+    /MCID 6
+    /Pg 117 0 R
+  >> <<
+    /Type /MCR
+    /MCID 7
+    /Pg 117 0 R
+  >>]
+  /Pg 115 0 R
+>>
+endobj
+
+28 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 18 0 R
+  /T (The Pattern and the Cure)
+  /K [38]
+  /Pg 115 0 R
+>>
+endobj
+
+29 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [31 32 33 34 35 36 37]
+  /Pg 115 0 R
+>>
+endobj
+
+30 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [22 23 24 25 26 27 28 29 30]
+  /Pg 115 0 R
+>>
+endobj
+
+31 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [15 16 17 18 19 20 21]
+  /Pg 115 0 R
+>>
+endobj
+
+32 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [11 12 13 14]
+  /Pg 115 0 R
+>>
+endobj
+
+33 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 18 0 R
+  /T (Knowledge Is a Spiral, Not a Sprint)
+  /K [10]
+  /Pg 115 0 R
+>>
+endobj
+
+34 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [7 8 9]
+  /Pg 115 0 R
+>>
+endobj
+
+35 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [0 1 2 3 4 5 6]
+  /Pg 115 0 R
+>>
+endobj
+
+36 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [40 41 42 43]
+  /Pg 113 0 R
+>>
+endobj
+
+37 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [32 33 34 35 36 37 38 39]
+  /Pg 113 0 R
+>>
+endobj
+
+38 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [24 25 26 27 28 29 30 31]
+  /Pg 113 0 R
+>>
+endobj
+
+39 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [22 23]
+  /Pg 113 0 R
+>>
+endobj
+
+40 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 18 0 R
+  /T (Is AI a Knowledge Tool?)
+  /K [21]
+  /Pg 113 0 R
+>>
+endobj
+
+41 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [16 17 18 19 20]
+  /Pg 113 0 R
+>>
+endobj
+
+42 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [5 44 0 R 7 8 9 10 11 43 0 R 13 14 15]
+  /Pg 113 0 R
+>>
+endobj
+
+43 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 42 0 R
+  /K [12]
+  /Pg 113 0 R
+>>
+endobj
+
+44 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 42 0 R
+  /K [6]
+  /Pg 113 0 R
+>>
+endobj
+
+45 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [38 46 0 R 40 41 42 43 44 45 46 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 113 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 113 0 R
+  >> <<
+    /Type /MCR
+    /MCID 2
+    /Pg 113 0 R
+  >> <<
+    /Type /MCR
+    /MCID 3
+    /Pg 113 0 R
+  >> <<
+    /Type /MCR
+    /MCID 4
+    /Pg 113 0 R
+  >>]
+  /Pg 111 0 R
+>>
+endobj
+
+46 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 45 0 R
+  /K [39]
+  /Pg 111 0 R
+>>
+endobj
+
+47 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [33 34 35 36 37]
+  /Pg 111 0 R
+>>
+endobj
+
+48 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [26 27 28 29 30 31 32]
+  /Pg 111 0 R
+>>
+endobj
+
+49 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 18 0 R
+  /T (Design Thinking Belongs in the Swamp)
+  /K [25]
+  /Pg 111 0 R
+>>
+endobj
+
+50 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [19 20 21 22 23 24]
+  /Pg 111 0 R
+>>
+endobj
+
+51 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [7 8 9 10 11 12 13 14 15 52 0 R 17 18]
+  /Pg 111 0 R
+>>
+endobj
+
+52 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 51 0 R
+  /K [16]
+  /Pg 111 0 R
+>>
+endobj
+
+53 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [3 4 5 6]
+  /Pg 111 0 R
+>>
+endobj
+
+54 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [35 36 37 38 39 40 41 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 111 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 111 0 R
+  >> <<
+    /Type /MCR
+    /MCID 2
+    /Pg 111 0 R
+  >>]
+  /Pg 109 0 R
+>>
+endobj
+
+55 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 18 0 R
+  /T (Innovation Does Not Fit in a Backlog)
+  /K [34]
+  /Pg 109 0 R
+>>
+endobj
+
+56 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [28 29 30 31 32 33]
+  /Pg 109 0 R
+>>
+endobj
+
+57 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [21 22 23 24 25 26 27]
+  /Pg 109 0 R
+>>
+endobj
+
+58 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [12 13 14 15 16 17 18 19 20]
+  /Pg 109 0 R
+>>
+endobj
+
+59 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [1 2 3 4 5 6 7 8 9 10 11]
+  /Pg 109 0 R
+>>
+endobj
+
+60 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 18 0 R
+  /T (Why Knowledge Cannot Be Sprinted)
+  /K [0]
+  /Pg 109 0 R
+>>
+endobj
+
+61 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [38 39 40]
+  /Pg 107 0 R
+>>
+endobj
+
+62 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [28 29 30 31 32 33 34 35 36 37]
+  /Pg 107 0 R
+>>
+endobj
+
+63 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [17 18 19 20 21 22 23 24 25 26 27]
+  /Pg 107 0 R
+>>
+endobj
+
+64 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [9 10 11 12 13 14 15 16]
+  /Pg 107 0 R
+>>
+endobj
+
+65 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [0 1 2 3 4 5 6 7 8]
+  /Pg 107 0 R
+>>
+endobj
+
+66 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [30 67 0 R 32 33 34 35 36 37]
+  /Pg 105 0 R
+>>
+endobj
+
+67 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 66 0 R
+  /K [31]
+  /Pg 105 0 R
+>>
+endobj
+
+68 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [23 24 25 26 27 28 29]
+  /Pg 105 0 R
+>>
+endobj
+
+69 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 18 0 R
+  /T (Scrum Is a Manufacturing Schedule)
+  /K [22]
+  /Pg 105 0 R
+>>
+endobj
+
+70 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [15 16 17 18 19 20 21]
+  /Pg 105 0 R
+>>
+endobj
+
+71 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [4 5 72 0 R 7 8 9 10 11 12 13 14]
+  /Pg 105 0 R
+>>
+endobj
+
+72 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 71 0 R
+  /K [6]
+  /Pg 105 0 R
+>>
+endobj
+
+73 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [3]
+  /Pg 105 0 R
+>>
+endobj
+
+74 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [75 0 R]
+>>
+endobj
+
+75 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 74 0 R
+  /K [2]
+  /Pg 105 0 R
+>>
+endobj
+
+76 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [1]
+  /Pg 105 0 R
+>>
+endobj
+
+77 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 18 0 R
+  /K [0]
+  /Pg 105 0 R
+>>
+endobj
+
+78 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /MFKKFX+Georgia-Bold
+  /Encoding /Identity-H
+  /DescendantFonts [79 0 R]
+  /ToUnicode 82 0 R
+>>
+endobj
+
+79 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType2
+  /BaseFont /MFKKFX+Georgia-Bold
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 81 0 R
+  /DW 0
+  /CIDToGIDMap /Identity
+  /W [0 0 1000 1 1 758.3008 2 2 253.90625 3 3 684.08203 4 4 863.28125 5 5 635.7422 6 6 378.90625 7 7 1126.4648 8 8 571.77734 9 9 631.83594 10 10 648.9258 11 11 657.71484 12 12 520.01953 13 13 353.51562 14 14 689.9414 15 15 397.46094 16 16 393.0664 17 17 816.89453 18 18 344.23828 19 19 663.08594 20 20 576.66016 21 21 595.21484 22 22 512.6953 23 23 531.25 24 24 595.7031 25 25 1015.625 26 26 328.125 27 27 1023.4375 28 28 685.5469 29 29 676.7578 30 30 445.80078 31 31 679.6875 32 32 562.0117 33 33 715.33203 34 34 757.3242 35 35 566.89453 36 36 833.9844 37 37 839.35547 38 38 671.3867 39 39 548.33984 40 40 701.1719]
+>>
+endobj
+
+80 0 obj
+<<
+  /Length 14
+  /Filter /FlateDecode
+>>
+stream
+x�������
r|
+endstream
+endobj
+
+81 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /MFKKFX+Georgia-Bold
+  /Flags 131076
+  /FontBBox [-27.34375 -216.79688 1145.0195 756.34766]
+  /ItalicAngle 0
+  /Ascent 756.34766
+  /Descent -216.79688
+  /CapHeight 692.8711
+  /StemV 168.6
+  /CIDSet 80 0 R
+  /FontFile2 83 0 R
+>>
+endobj
+
+82 0 obj
+<<
+  /Length 1166
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+40 beginbfchar
+<0001> <0041>
+<0002> <0020>
+<0003> <0054>
+<0004> <0077>
+<0005> <006F>
+<0006> <002D>
+<0007> <0057>
+<0008> <0065>
+<0009> <006B>
+<000A> <0053>
+<000B> <0070>
+<000C> <0072>
+<000D> <0069>
+<000E> <006E>
+<000F> <0074>
+<0010> <0066>
+<0011> <004B>
+<0012> <006C>
+<0013> <0064>
+<0014> <0067>
+<0015> <004A>
+<0016> <0073>
+<0017> <0063>
+<0018> <0061>
+<0019> <006D>
+<001A> <002C>
+<001B> <004D>
+<001C> <004C>
+<001D> <0075>
+<001E> <0049>
+<001F> <0068>
+<0020> <0079>
+<0021> <0043>
+<0022> <0042>
+<0023> <0076>
+<0024> <0044>
+<0025> <004E>
+<0026> <0046>
+<0027> <003F>
+<0028> <0050>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+83 0 obj
+<<
+  /Length 11819
+  /Filter /FlateDecode
+>>
+stream
+x��|	xו�U��{w��^����E��Z�RK-�T�
!!	��%�Xl@2`�	�e�06^�ǖ��
i���e2�$^��$��yvg�(��s�����nuK�`�ͼ��T}�~����s�
j8$Ķ��Vln���~�up�.�����8
@o��[�7��@� tz{o߿���a�2�۷���i+
ʖ
�w����P�?(�	�wl�}����P��w��i�V�
����ݻ�?zk���w�+������K0<��p��_�`I��Ikd� @}��rL��@!N��Lz�3����Z}03��?����������	tmAM(
7�](�����R���@3�
zaR“�z��M8��^4�Ï�
�����Z�/�8M� ���#-Bd�&h���&��E��1���� B3��>������6jD��(�28�#8
�~�H	���U�wf�&����c�Vh�,�;��J�y�1�D����ނ���U����D�4&�D#�P�8��yy��c�
+]N���f��MF��i5j�R!�eI ��LA]�Y�<�t���h>��d:C���0~���F��҅W����m0g��z��Yh�MLd��Vd��o��w=߰#c�����4��<�e��H}�U����~U4gU�:�N�@��<�k�D�
Ug	Ph���1�!|
����x�'C��z���dLWJ�f/ܳ�2��e�Q(C�e�Ҹ܎�؛�c��ȅ�{������}��2do:9��a{'�c����e(_/=���5l��x,���=��ק��O͏F2ʺ�#�Ό���HC��4�����/:ɱ��N���2+����3�ND#�X߰��>i�~i4�
+�h$7���z��c^ߋ�l��;�/�z�ăT�a;߰���Tkl���o�����d_]F�>�s=א{����|��]J*�O�U�FP˪�:��[��-�|NO>�eUW�\!�9hΈ=n��U]|��U�G%�m���ǝF�HKǕV��ṱ?C����dNo>��1L6�=cc�<�8�3�;5{p3�1��ٖ����.]�;5��1g��t��َ��k@㪮Z�ې�Kv�%!#���X�[Va)d(_s����@g��������3#vva��+��}bE���P�N��b�2ꯜO]�t��v�as4��\ٕKs���,�B8�!zpɅ��\rp�d�y�F�
>$-��W�XM
۫2��w��s�S]�$�›�'�)U�g��-�d��cϽ�g�pF�u�Y��C�x�V�-+�wq
c�Z�����A���n�o%��g	Xz�GGW������3��ήg	D��,M��S�&�2ʎ��ݛ�B��SP_x�@^�):k#װ�>�z�S�.2kC��tE�����in�k�����}�'}���������];r���y�?���NA�CI����F��|�K=���lt
+�GZ����Zٕ9X�̈�i���5d.tte.�;��D6�s�q
��(��12B�)�����+#:3�˥xw��ؘs�Oϥ���U���S �H������ ��
+���yw��N�G�`S�e���+�h�`�G���	�����(��
n��mp#���2��?�L:�����Ͳ' ���Y��i=[�T8)�)�����Jw���R��c�"�cK�������z�2��_�vq[�[\ü���E!���f/�z���D(A%�Z[a�y�Z���l���G��XX�De���0mK�SB�t		�;�6$��і‰�#L�…G�l�����;��kPy�?@�uH�C��fq�נ�T�@��ي�_�Ft��K�H���!�7�{o̾�^]L������U����Z^�����6��)���W�U\T��F��1d'�P�LQ����R+��}���`8Z�N�1�a���sn������'
��`&#v���aB�c�#v�n�y�}�'���p�
����L��� d��xB�N��_�%(�f+���,�
���된7�e;W�wWq�<��o��\�}�S�!x6qUw�+dב�𺂢/=�`�&���T+����BS��7�G����?"߁�/~9��k��u�&�:n�{����6�
����[ǺX�Z}T��Xn!��<G<~�J��%�$�qa��LmQ��Iu��Y]�ݠ\����P�Wޤ�
+�Usn%)71h�wq&>'b�ˠ��X�̘	3������UZk���6���P;mK	�T]g׳�jԝFBwB��6$����H���Qde�TLQ�}��P7�TH���*,�@�ےӪ"d3%L��e�\�QI������j�s�����5�+!�z6�h��_}/[U���k��yR.�_y�{�%��]Ȁ��ts3鰪�涿�&�v}��=��X�T��f �i��526 (���&:��:�/��J�K{J��a��1��n�i����.��"d1c��=���dE
Jĭdn������&��'���m+�����ͯ��v�X��gw���sV��\�fI����9mk�V�%�o�08�O�-��u�t�M/67�X��UN\�x��vY[M��ջ�[V���A_'�N�\"�^!Fd �Dԁ������u����gO�����:�����A�1H��}~"fo�*�Jk�|;T���y�T>�O�VߡV�}���^?
+��G�*Uq�ZI�4�C�RZ�:D��<��+��L3]Ϥ���z+��ޣ�M{�1��Z}Ey9�U{JZ:��*�J��"������U�f<>�8��i�eI���62��F[
+�0ɀuc
_����� �i!M��rN�1�dV�	%�C��s���5E%
��[0s�̙<��bS��\\L��n�ɨ�S܀��B�Wo�nW�M,�-"-��&2��<s�b/��~Ch�M�ڍY[�B�P�:��m�u����OB��M=��H�T��y"��.2�`�X���;�[�{�2�x(Dƍ�
r�>L[��
܇�/E��o��D�����0�Q6�U8!`-Nݵ��֏n�4�T�V�
+��lMēh^��]��ں��������������[�*�n|���)�yU�7,���H�G���t65}�ƭ/67u�W�>�f�+?�֎�o/b�}Ir���������{��x�e˅
��MɨG�w�F��$��(�'c�<�')��28y&̸��(M%"��:1�	��PDW'ٸ3)Ė&��r
+�X<)��a�*�����*����
+�iN6ZxFw�r>�|t�b6~+W��m����i�;>?����$S(�9YL�4>(��u��FqAq!K]@���AT�r�էJ^�|6C.��?���܉*[<\wB��cه�qg+t�ZY�����+���Zt�oGMpE|��؍%��xC����BC47rZ]�6����hv0����L��S�eS��4�\���#G��A��9���,,�������m%{ɛ�{Kh�u$O�}���yI �I�����,4Cw�=� -�i��~0�^�)��3T���J����}v�9�z%����N�d�S��F4����>2�8mL�c*�5�RB8��X��Ɯ���d0�%���d_|o\ޝ��;X�2�? �Ʉ$�D<�߸�"a��ίǜ'��x��a�vm��E�O����C��k�;��a����h�:�ڢ��҃kw޽��埫�����7�b��@ͧĺ����7�%���~tqj�w���m2Y�p��5;��������T��"@=SO]&/A1,�
b�2�X��v���"�9)/g��9O
+���IҪ3�3��axi����ń��;�e�0gl�s����|8
��8>��)����nd���k`M���l�Q6��
s2�.���_{n
+(X�s�;[bo5q6��4/��o���O����̣V�3-.i+�=��߲��͙7����e��X��P��r9뎛��١ �����ܝL�Qo��Z����F�"�/�>�K�A���� GH�$�k�0Ï{��1&븅�SB���@Aj�|&���e9
.����,����p�֘Ja(���=~�#:$7|�&�$$��6�h�WQ%[V�u�5�Ş0r�c����>��}#lZ����x�|�V�Gs��h���lv����0�
�gi
��ub���Рg�00vbq_p8�i�9�ND��H��%�x1{o�ܬ	l�%�~���M���j"n%��uč��+06�?�\�ぽ+��]�r���嚨�h���s{�7|�F�#e��nhmݹ3�V���%��mb#Yn(������a&��<\��kA�Ɲ���TO)FkNW� �'>!�&����V�����"���[�Ӟ�y���f>���Ι�ic^��.�����嶾%_I�9��)��[��܅��lcb�섴��#�������Ik#aQ�CA���m6s����� ��([mʯ���Z�9�����u���0�鑩��b���
+����Q�m6�äU���
��"ؑBˌ����G�x6.ta|��-�N�R	�4�IGv��/O���m�֤G|`�c'�M�T0�%��S�+:~�غ	3�cg��{7+d��W �Hk!�bm������٠i��0Z�#^��1ngul	�N(��ܣ�a�^f�c?
n��/&�S)�b"!���v(��@#��hy��øB��*?�<1���-ee���qd��������v;JM��?�YR��[�ʶث-v�{�о3��z�.9�}a�
?ho�����ˎ�RxLܪ9�ED�,��!�Tu�p ����/()%KKTU��ɄBvp)Z��^��'�	�~�H�n��\]���L��da �cc��,$5K�����6kK%��5�P^�%S�!��?��0�6
"?�)�z���!q-`�B����}u�GO�9�&>�sn1{]�l��T��؀#��z��e�rb��x�
�v� �]��r�h���7��7(���P(tq?�'>f�I�Ȫ���O����v��H��iRu=�w���'�-6�I�5 >/ٲ
+�-1�Fl΀�N�t�0�#gðy���`�z���֟u�W��/zG�9"�NM�Gu����|�Ͼ�o>	7˚h
����TJ�\�V�o)���ް�ȃ'��{�@��V����`F�󅤐�/�z��V�X����3V���Å�O(`͆�2�0��}�O��D<)k�1��Ϭ�������U'�Ev�[ \z�˲�1C��Vs]��p����R��H���ȟ�D%>;j�-1ZM���2�Z�v�J�����+��W8z嚩�>G^�E�6�qq���^\�	��6�v�nq)�#U��K�*�B��j�+�Z2V����9g��-��B-w9�
գ�RU��Q�&�&aS�&��.�o-{���0���hKMc+0-LSa,��Ĝ���$��s�+��#�y�\>�r�؍挿E�x�rY�n���ŧ{&5��7�P\�'TR��������f���c�9W���ҿ^?�Fw΄��T
���e+��_�}�/�����%#���e����A7������D����[ev[}��jr*����ؔ3?�{��{6�v�s{6쩳�TѪ*{�����?��� l=�v�jKA`��4�
+��hM��pZCE�{���M�tx0��`E����`����dV�#�v�;!�f�X�y�1g|s�?��%���uoq%��k]����۱qݶtJY��t�5T��E|$[,S)Օ���<��z�/t߼�E�U�5��z�t{cF��j�a���_�?#/AL�7�
+U	k�=���b�Bcl��6�M��5��M>%|%�)|)ay��l�`}#��$G�	��"L����f�i�h������?���b��\t��dOQO�d�������ƐJ1�X�X
���O���r���a~[���=�(j�6���
�b>�]_�[MN_��h���i⿼yEס�?=3�Iq�ho��h��iyq���2�������g|�E3_�~RL���<q��s9b�1��(t�l�"UD��q
3@���a�X0��ܩ�J�(�
+�^h��a���Ci3���)�r?҄�E�y� l���MRb�74ԯnj��᩟�v��s���.�X�3��-����;���M�޳����u��-�ZVU�c7�B��/�{��[6�-�C@@���Y2�����:��HgH�R�cח���r���u�Ɉ���j�h�i��N���گEPDT2�7^��=BZ��Z�����EL���>Zo-za�}��j,��;�W1MKp2���|�LKV	���GR<<�\��hH��`�3�-��-��X�D���w\N����Mu�[hO�WX�6|��.j�_�Rhuե�EF�e4�R�:צ��Ӎ����4+�.Csь�r7�5�U49-&���3X-w߸dυ��&W�}���ص�U~CBo��\R�.,rۼ��
7��Ki�L� �S�ӵnQ[cԝ���,y	6Ó� t�4�����4]7^�FӴ�\�p���I؝�͈�'�����Q����R���p<��4Z��s�����ʘ2����h�[姫�������G�J]��ӷ��$�p��s�����\c ����
��7|�5ހ<���5�X5���:W!�6ؘ�-�O��֢��3Jy_���?ϵ��n��(�,��NNZ�m&)�J�7�ms�-!}���(�"�sNĕ^)�`��[�w��,ܖU�f������/���%�p�N0
+�w��)U�E���:�;��@�N��|fZ��������"��<|PV�u{_����g�49��C�j7��=���
��&;����S�jR6�)�+e�P�8t�ReV*U��`���^m���jUq0$ӣ	�Hɓ�d�������������o��-V
�%R��Ѡ�@�O��jDP�ɂ<J�K
6�Q�Ղ�P�JmE�;��a�x�^�^4�RGJŠۘ�ѕHD6ϭ�t�s��|ՠ�y�ׁ�<��b�q3�+��3��}�C������.�+���&л��k�v{����F���b�{����"��O�z����H��p��B|S\cg�d��u��B��6B�
�e��ԨdH@� �Uv��,�M̤�`�Ar���V%S
P�^<PL'#�1�|.��J>�1-�v|�#�3��9d�5Bv�����Mz��b�9H5��8�G������
��T��u��{�`���,�����ai�Ѥ6q��+$����Z���#0(�j�E갆t��L�
+���o��
+#`��|Xe%������F���l)*U�
+����'JgK��Q�[��1棋���=R�;1���_�1UK�������3���V��zڳ$�ذ��Ev��@���nt8��:��w9M��,��������ѷ�&��hv�۶m^"���A���/��Rr{���dǀ���ը�dH��DF��
+���(�Wԥ5T�	��u�"�P�((��͹q��tu�3������xEW)1���s���t�6yHS�z6�p�-�n���K��<e��reM��4��{���!.[Á���Q��(Ι��QõQ������bXє����R[vBr����"��zʿ8���pj�Zd�p�Q�m[�Q_X��S�eF�;��O����¸�Qhr��ً3��i�X���@ԭ�[��Y���GI���9���$-szS��r�{��A�X
+�8�EN=M���
+,��",�/XL��RC"�ݣ<V�%���B�!��.�9��If�e�v[}�:[gx��?r�rOD����*�:띝�[��"Z/��
+=W�"��$6eW�|���X�����l�q`>�M�n~?
�jP����x>���JNG׸�>?P3��m�]�Wka�B�[+ܽ��S{���I�
+�m�k�U$_`/�K7����3��q�D��{�RW�ae�܇>F�������eU���{��W��eF�J��1����q�Y(���eW��
DծX�Z[�vnW�'(��uFQ�T�Fd4H���a�����x攁7�N�B�!d���� ����-!Lݹ�Z�	K����eƊ�w��A\��90NF�y�˞��[Q��o�լSv��q���%����׆w/k����g�<��o�Hy+�O�j湙�?�[��?W�u,^Q8F�ϡ�s�j�&q���o�oH�oI�q~)�����+����$�Ѓ�v��A���7
+p����BC�@�T�o#��qP$JR#�	ՄṼ5��c�6��U;��
�ө����a�+�.k��7���e���Ď
k���Y��d�9�!vK��GDU��h[�g.���[�N}�1�V��T��a�Xdt�·�%�6R�З��aG� k�xlٮb_�hR��C��$Q����}�qe��KL�l�{��ΐ�@.�mB3�jt�u�.���/g��G�Z���bI%�2�b�
z�V�$�.Vn���ͤ�Z�1c�q�H������[9S�M�h5ҙ\~�>��Tmo����T~t�T��cA"�x`�����.��–���W��a5^���D׸�,1'SC�,�b�g���-�o�S~b?0��c�;�H�z�Ez���t��u�׏���
+��7�t��c���c��84�1%�t�|�wc��/�k�+w��\N�TYGYO�`)�PO�K`��[T��R��j��	9M]��q�oz�oSY����i�g��\�u9��56�@���huU�?���V��V�����53�rn��k֬nC*�>�%��04D[ͥ�w�}���V�ZO�K?�VX,�N�3
+g�JOʎA�����'x��RaV*N���{2�G~��ޤ��/��T��~�����(�R��MzHc�4b�!��1%]�2�d�~eBY��){�����$��hL�����)	�ϽQa1�
��<EO���|쳗�q������~O��u��+���
+��D�l1ɰ��#�UTsv뭇�������#&�i�Pw\	��ۊln:w¶L�����w�U*㼈CӊK'��;ɛ�S���`�жI@�\_.��'�Α�	~"�\����66W�*�07Fk��;�G�	�g�M6'������E}���Wwy사��J�
+��w��+�Y����u����%�Y�����o:y��P�+�C��	a�Ͻf&AuO 1�X8��o�_e_�B�M�J��f��k�
+t�W(��4��H�N���?;)��TW�d���֢ozql�륾&�(*]f��02~C,@8�#��;$�#�/���b,�KJ�x��U�'�&����+�����C��?=��T�
+[�!��/�~f|hE��Pۊ!J����GѨ扻w6��p㊞ơ����jݵ��u`I�ˎ�°Q�{<���
����\�,�dI�a��o_끐�ߢ2VD;�H?Q,a'�	�s����
���sѢ܅��Ya����s:!ۿe��ߞv����˳�=������m#�Ȅ���=�:t�+T�M�������-g�kS$�'����N|>{��ܽ"�"?�ý������[���s?ş?�_l�����qŏ����������0?����9���Y�T��A���	��h���~ �B��5Jr��ԃ�$�
)+QVj*�-p�F!B��>�黡���,�����r@%B�]hWp!&�nEx٭� ����P+���(;M���gߣ�
+u�?B��:h��@��1������G�}Dt��2�y�fY\�2��������)S\T���	�R��Nm��v]\��ޡ�Ǩ��>���sS�����_YVI�—@A �u����
��
���`�"����0MA�$�RA$OS���D�@-��i
+��V��@�v�i
+
+`X�@�����`\�����4*��(�?O��
+��i
+
+��<-�"4�
��<-��ѫyZ
��yZ	.��yZK�3yZ��G$Z�� �C���H�W�V
Fښ�)�����|���4�*��

}}����o�hF��sy��U�6ay�9�MX��[m���9~̘:Ǐ0�s�X��<$�V\_ޖ�)(�o�h�T?ǃ]��D;��ʟ��8�/It!�G�ӍB̏���J�?�Ӹ�e����U�����Ux$:��+j�4�2L+�Y�@Ί�+�YP_���f��5s�o�~�{�����`l���U�ðia7���zWJ�B?�]S��m����4�O%�A~��&�
���u>Q�%� 1�A8X
;`�
�
�
+��A��vPzb�w���`	석��*�
�`;Ð��!�=�O�������y��/�J��cqP%�7Aʤ��R(�(�B����������V���o�4�~�
ð]ꑓ���:C0,��9�U���~���nN�-���K���Vs��9n8X�%�r��k^<�n�!����l�
��%�vH�K�N�)7�yvC?���̵�/�_���z��JZ��/'��R/xF�RzPj�K�z���R۹95�h�%��{�J2ꃽ�E�1'���X[`�g��K�[`'�V �Z�A�T>;�3�I-7֎|X^�ꗞx\=o\�S��a%����G�4�v_��]FWz���my������2�1�>�+Z�I�-��In.��]���͵`�4�
I�?}�99�~B�9-�V3�T�aPz���a�4���~p͝0�X!l�p?U��F�/�+�$��}V"���
i��Z�� l� ��b�y[���Z�\i+����`@��{M�O��G���\~#�v�
���ÞO9�.ϵ�;Kc�3{�R[�kϗ\�2腝�O:�����Ҏ�$�����ŧ�Y(�k��D)�R5�"��JR��H-�Z���->�N�g��WJr'꧟�-x��v_S~��Eڍ�pç�baY�d�p���1W��=f[��O��|�/@B��U�@,�6H�jɎ��u�%�M����:@1�"��>HCJ!HL���s���ػs��d��#�;���/я^W����`_}��N�^��ה��k9Z���:&��}Tx��G�xT�����S��}����d��x�����3�Sό+ؓ���G��#�
+v|2�2��41���ey辇�x�b�<D>�����
+��'ŗ��dϋ����|��
+/ �()��b�'�G���Ώ�'�)�.���H|^�Nƞ��x��~���h�2j?q�	��S<��S<{�)�}s��>9Q����	�3�����ސ>@� ��l��L����f�7���5�¿�?�����e���
셗
�7_6��x���ϱ��s_;G��s�����+�W�^�{�=NJ�c#���c���H1{tT�
�_��w���;{X�a�0qx�g<�`�E�#v��H�$��ړG��#{�������#o��>�������{x���:��P[`wh-V��h2k�t���X�%Ě�!�"�,��zƠ�hu�J���
+
I�4�
[$,���n�I�^�����T-�N��IP�0�s���W)X2�`�R�v$P��-�K3&�-��f�)����-edž�7t3�p�8:��3C�"�3c�[��k
+�q��e��M���Ǐ;�Ry*�f�ZVweә8&N�!��������0�i?gߦ�w�{3o��Cg����̯���J�Xߪ�(�v8�;�l(�'����sEx����0n1
+�
��mr-��jx(��
VÇh
+endstream
+endobj
+
+84 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /CDDLLL+Georgia-Italic
+  /Encoding /Identity-H
+  /DescendantFonts [85 0 R]
+  /ToUnicode 88 0 R
+>>
+endobj
+
+85 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType2
+  /BaseFont /CDDLLL+Georgia-Italic
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 87 0 R
+  /DW 0
+  /CIDToGIDMap /Identity
+  /W [0 0 1000 1 1 599.1211 2 2 461.42578 3 3 537.1094 4 4 879.39453 5 5 241.21094 6 6 670.89844 7 7 572.7539 8 8 297.36328 9 9 285.15625 10 10 471.6797 11 11 347.16797 12 12 749.02344 13 13 431.15234 14 14 589.84375 15 15 618.65234 16 16 562.5 17 17 527.83203 18 18 814.9414 19 19 572.7539 20 20 538.08594 21 21 575.1953 22 22 653.8086 23 23 575.1953 24 24 701.66016 25 25 822.2656 26 26 561.03516 27 27 453.61328 28 28 383.78906 29 29 328.61328 30 30 975.58594 31 31 553.71094 32 32 374.02344 33 33 642.08984 34 34 389.64844 35 35 578.125 36 36 609.8633 37 37 559.5703 38 38 517.5781 39 39 269.53125 40 40 927.2461 41 41 603.5156]
+>>
+endobj
+
+86 0 obj
+<<
+  /Length 14
+  /Filter /FlateDecode
+>>
+stream
+x���������
+endstream
+endobj
+
+87 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /CDDLLL+Georgia-Italic
+  /Flags 131140
+  /FontBBox [-176.26953 -216.79688 1089.8438 756.34766]
+  /ItalicAngle -13
+  /Ascent 756.34766
+  /Descent -216.79688
+  /CapHeight 692.8711
+  /StemV 95.4
+  /CIDSet 86 0 R
+  /FontFile2 89 0 R
+>>
+endobj
+
+88 0 obj
+<<
+  /Length 1180
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+41 beginbfchar
+<0001> <0046>
+<0002> <0072>
+<0003> <006F>
+<0004> <006D>
+<0005> <0020>
+<0006> <0041>
+<0007> <0067>
+<0008> <0069>
+<0009> <006C>
+<000A> <0065>
+<000B> <0074>
+<000C> <0044>
+<000D> <0073>
+<000E> <006E>
+<000F> <0054>
+<0010> <0068>
+<0011> <006B>
+<0012> <0048>
+<0013> <0061>
+<0014> <0076>
+<0015> <0064>
+<0016> <0042>
+<0017> <0075>
+<0018> <0052>
+<0019> <0077>
+<001A> <0053>
+<001B> <0063>
+<001C> <003A>
+<001D> <0066>
+<001E> <0057>
+<001F> <0062>
+<0020> <002D>
+<0021> <0043>
+<0022> <0049>
+<0023> <0070>
+<0024> <0050>
+<0025> <0079>
+<0026> <004A>
+<0027> <002C>
+<0028> <004D>
+<0029> <004C>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+89 0 obj
+<<
+  /Length 13367
+  /Filter /FlateDecode
+>>
+stream
+x��|	X׹�?�VF�Ѿ�f��Hb$`	�0�;`[6��l�4N��%Nq�4�I蒥q�$�Y�8v��ݾ�5u����67i���&4I�&�5
�wF��M�}���9�?���1  ���>���,�s؟�޺��q�y< ��֝�]���ߣP}�278����
��
�w
@���@�1i�o
��}����;�s��Ρ-}�_��
��vW�熉�(O
4�vnw߮�_z��
/���G�B>��=80�w`����Sm��;r
+@v
+��߀��A0�oP�Js�J��J��f��Q�����e�����)�����S�/�V�1<
�����_���
�a:p���!��v�zX
+O�	�U`ZX���R�5��
�
+K������]x�oc$��X��A?�Vh�&lv��av�a��84�r�^8��<o��70C<¾IX[��D����o0c�ށ�a7��pn�{�
xw��7�|	�'�"��o�������x�V����VF�BE((/+������XgI��n�Z�&�A�ch��*R��
+��$p�X���}Φ8\.WO�P�_Y�^�Ϯ����Ġ�O�K>Qv.�W���kᛚ�����r�1�-�aE�A���|z[����ۛk�y�˵�+�"�}�H��7
�CA8�.j⛊BAȵ��簖%��-��9��P0��po�n�%O��p/��r�B���r����-�� ��C�<���M9��^n[.ٗ�Sܹ���-�l�
P�|߆���
+�›�BxL���A.Gz�f���^.=ȍ������{B��և�9US�Iה#�o�>�����|sn�7�Xں�Cű��\���ŭ.���鱆��X�O�6�����P�
+�`~O��nG��އ֙�΍���z���kz�Oo��G�����|����&�{�r�.���!tp�\����PU谮;GJ-��=�CA�}UwZ�����BMo��}Uwz��C+h�%{s�.����E��Z�R+���;.��ɼύ�rX/?���5}����+ ��o�kṖ�ޱ�ɹc�y���ε��
�{�tt簾ɹ����[zrL� 9D
-����g��1_���)'G$ܾ
+a!Gz�
+}\��]\S�t�8rɮnwu�8򟈐&�j{zz
+hC8�]@OS
t�u��L��PЕ;�ٝ/s���$�@O�E-S�-�5���|���^�
+>	He�rJ���1҃�f���|{���M8pD���n�A H��>g	������s/�9&��uO9�{8F�=:��|{�n.=�@���N��4�78V`%D���z�\����9TђK�)Aj�
r�kX(x��9���\�y���wsW�8�7��zzB��&�>	���svk�$67:	�%@Ħ��IX�������k�������� ג'������Ǹn��?Gz��U���X���`u�����=�p��'��4)�3�õp�3l�f�fB��.���_Gwgw�X�#�l�q�\\:7�ѝ�jv����_X)ǥo�ּ!8	��C���ϲ�;�t�gl,_�]�ccc�1�g�<	S������B�$H3��$v�Cj:ƻ)�y��q�z�C��1ؾz�Q
G$�9

+0>#�I@��O~��	�t.�ץsa ����G����F��od7�NA=\J�;�u7��dyXˈ097��hQ�dU*�=�4F"���
I#��A/��a�mz���c��Q��
�pK��Xe*Ұ�!y�!c�I�N4�Q�J'�
�3zZ-�3�MjR��/�0ՙ���D����r����QA��	��L��0ƼP0o����F�´���
+��ڊ��0�=E}}��W`�
+,���q	&F	~	V]��Wj1��lqb�5.�D�|�_�ݰ��\]Ϯ�����xo��-?�#�E��ug����lt23w2�f=�Q���ռ������{��*��y�Vj�X=���[�e�YZ�S�]�O�p�B�I�ڠ6�P��H
`�
 .�AIV�0��0j4�xS��l�2]k1�U���a�6'�b��Qq�qr���:�hDȵ�(�xT���^��y�1�~Is���1���N�?�N�	�0&Q	�B�s:{�����@3�Z�w���jD7��G�+�%X��l�J'F\dh}�����P�߿����|1s��^��@G����M\��l"V7�%5.v�Oc0,O�T�7[{-�>Z����\�uc0��C��oI��1�U)����}i��h�4A3�	�Y��z�j��8�z��|0`>���tb:	_�
i������S��X��W�}�g���s������է��|[�{��
+|���=�����7��fǯ�n��ڑk^Yw���]���l��ψ�`�%�VW����1�vr���*����{1Y�bĸ�8j�����_4��o�r�Q����(��w���#�\m���~�
�I˸�M��29��Ϩh�b��g�Fhn���3B&�ʐ��1>�^�Y��I7�.�[�D5%��\�H��ॲ��K�%%�"������n��̛��XL"
+KL"�=�4�|PY)t1!��������b�f���Y�e�^&�Ť�i��D��rE���5��v�;:���E�x��x�c���_W�>ќΓT.?0��gHA�xp�wkեb�1C랤�du��ڃ����V&��8�#T�ؖ7~3�Q�2��g;����0�N�!4���+S#�:� ������d'i
�jːm��(q�y4��0���9��@St��U���{<|���=�\�̳�[���o�����x���d*�a/���KE�^o�P�el�M{���qOXI���@���H�M��<��٨��Ų'g����QNE�,F�$	�`.-���ˢMb���$qG��}3��j�ZͳZ�2�޵����ھ��o�>��}w����bc��Ɠx�̿�L��"��lQ>y��.�q��R)����0=~��E�Q�ր��n��h���!ރ,�'�j%N�Z�R�4j���;��[F��¡��q&<97u^ň�������WĈ�JX:j6�[c!��
+sּ�|�,7�L�m�q�3�&4�Q�&S�&�e�z�H��"��v/9�~�
+��v�a��T���b^op��
�騐E"CBmc>@�DBwqB6 !z
+<"y����R�"Z��JV]U������1HJeA�X��ǿ�0��Vl��~�������?C=�;���ۇVos:M5����qJ�x��z��uss��/�Ѥ�dJ�-cKN�L��p������{��&��}���e	��Tm
5l�V}em������f-�v.�宺-^oYgB:�����ݲ��Q8��;]�jJ]SQ6�EP��x�z����.eK{K�K�RIi�$2�PJ����ʤ�DE�@�j���2p�;N�
	��7�vA�e�%*
!;�
(��l�993��?-Fy�[�Ҍ��E�v%퍙���<��)�;�t4fX�n�]�,k�%IK�{��(��j�L闰���ۭ�r�L[�Y%�Z��y��[x���&�@��FW��\f�5o5�f�q���#it�]�a7F;15ھ�B��Ө&8]X����8P_R�k�H�������h������u�IDH�����*&3f�������*�_2�ٔti���G�}��>I�S�)Y�Z��Z%',�Cz�
��=p@9Kj�z����p�gX!/C|G1�=�,���	�Q9!�����a(��)Ju'c֠NN
-j4ţt�
+�0� �	b^�	ó
惙J�<�U6�\�
7	#�LRʊ��-�R���
+Il�������=�_^�x���g3˗���/��4�������ђ���$\�'B�r�=�o)?tǿ<}����{����"J>���W{����lwνE��PCɥ��0,\Ĉ��э��"�3VV�Z�S���,�;Yg�W�k��Z�Rk'�`�$QT�hw�i��2�l+�B����.���TAnde�S�Ũ!J�.b�c¦n��O��ny���f�:���V�F���ӊVK���z�;�?���ʷ�&F5�6�_�U���%;,T�@2�@ƻ�mw{K������>����6j��0�0LV%����h�Ni�UѸ��(�z�F��(�*
+vwc�������y�:�e�W�I�%��E�꒤^�ƺ����{���^8?3N�7=^�ۄ��ؗi�g�������ʺm�ٿw����W�?�}jמ})�R��DevQ:"]���/�����y����ܼg?fts�#ڈ���O2*�L�Yb�R%�*����h���hQ5_¹"Zt�R���k�����~x���Hp2Ȏ��`V
���by�cћ1�y���ihH�(���y=-����[b���H�̓CAIHꠀ�E��$I�|%����(Ƕ��~��4��-J��r���Vo������u5BJy�c�N����n���%�֔�xm�V��d��N7)��=��3�ʲ&��X�Tʷ��,�E�d'^��]K&ރ8<x��"4�U���F�~�C����Q���[��fK��ZˈE�w%35Rq����q�.fI�����\؜4�悉jF}mjJ4�fw�d��n<���Zd�B�6-�+W�L������-z��y�f!�����3�h�hz���"&ӌ�w�wm�Ndg�*c�����b;�c92.�������k���
d��NA�|��iяP�����k�C����AY��n�)��8�Y�Ÿ����Xۄ�"��a�8��/�L�@�hƲ!,��3A�l;ka�21S�a,�F#�,d�SB�<_gcC`�g�<d��r��A�y���4�ct�ϙ-M��'��m�m.ff3��k/-Ub�j�:��,o6���N��w*Rr���Ҡ���҂�R8���<䍄�\���q�+q��EN���"�Ë�!
+�)0RVq]�/*�PAV�<*-6��Rc�X��ˀ�E%c�D���1:��1!��eb�b2fbC�b�{'��j���j�9�щ>�~���w��i~�,7s�.���1��K�T��>��X�}I�/��{�.��H��"Uo!�F�-WPgr���X�La5R�P�Gn��D�J�Ոsղ�La�lzf��O�g
炜�̘����Lɩ˭1kU�]�Vg"=֮�uX9�>\q ̔t�����:�QiT*MI��X~�S��D���(�rJ��y�6�%�n�;��2�aհ�6�5wk�(�r�J���P�$(����DD�b�/c���c���g�d��>��٤j8t[��ز�/ �"��H$��H��JYRlըT�*F����:�?*�%�Lz},}F�.>2]4 �-1!;f'g�ȷRNa'��T$��$�S�K�����j�8�����>q:�)�\Se�����5���eIv�ڋX�ecN�>���@kW8��8l�����g��oO�ʔϾ4�����#2�"����C糩H,����8w
��^<�潘�2�
Y����+������nO7��R%%Dž�Q�i!����������n�Lvw��pJ���Ou����&����~%`d�l���F��li�i���
+~�����)$n�(F'��?���J����`A�����\��s�u-���Ew��n�p\o�3���yڶ����fFp�`���י���4��0�Z����S)�Qhj��>T&N��#�Tj*�8�)�J��hQ%�V�Tf���Ta�l#���*��2��M�^��;�;TJ%�ڔ�Ƥ-��y��=������a��
+���:�
+��1s�T�P�w�Mq�2�:�6��&9K70&�dj2
�
D�	���`�������k�f&C��+�R4+y�
a���z����^��/wRan(�o����
�+^^z��5�ԅc���;��uJk6�T���U)���*�^���z�f���
����Yz�z��+�Iׅ_p��m����j����'2�C��oIZ�
�_�ܟ�*-*�*Z'�I��s�R�&\�JxX
+W?�t�h��n��*�hNjѬ�T��Q�
+��2_2��_�1��r	o�ϼ�,���;d�����$bFV���-�i�f��Io��|��_+¯Q����UO���ӱB��U�]6k��|{M]��ht-/�k"j'f����n?d�������֧�%y߲c�={�ע�j��k�����6�2V�)�A�����0��d�DSF�C%�ZIR�֮4Pm ]�nO�O�H����i�����<g�|\5���p����r�|i�K�K[����|=�?�ct������I�g���[��Lko+�pB��Ғ��XC�8�0����7-"�y�:O��� ��`uv��(t�m���m�7��$�o+�-�pg����hi$}sp�\Vb�v��ظz��y���-���"n7Z�*W|���^k4����/,N��O����];��p�=>UAPw�֘ce�"�s��?.�6WoNJ_�j�X�
�V
!�sz�Fl�x4Y�
+S�H�Ռ؁�E��`F̾ăU�3�LH_km=�d�N'�c���*�q�ȵGJ�q�'Ī�*�3Ʀ���:�Χ:ժ�3t\����L��2b��YK��d�$�\�:v�%�M�d �&9(H�$
�/�Wוfb�$YA�ʚ|��#y���5^7�<~CeuӶ����ig��&ޫ��	�������.��n���ru]��Go��~���:��ؾj`\�ڹ%�~L'h����-��hM\L��6“r�Amb�O�f���+���M:��咙Ѻ[�w"�Y��ZTU�Z�3�s �o�x"�5��Dy��)��|u`[�RV1"���Ҋڊ�
+��w����k�z-���%bt�J$���
+Fg�$S!$���do�$1�e,J�,�U���@}(�\2�;��K�����^�)�Vp��H!q�
D���!V��мz�|f�M4|Z��ߕWt��W&z�+�j�7�#�t�+�=jM<�<���cԸ��H��zH����]b����YS�XC����B��4
+��Z�fhZ�Ԍ�G�37���I<�voD�7
d�d� )���@���m�4f����mT�G��!ÏTG��%
+�q��h��=Ou��'�U�-�_}]d_�*~8�T���!��|esS̞�557�mO5E<�Ӝ<)�@Q0��3�Y��@��N�7��GR��pcX��A֫p�p?���y��P>Z�İ*_>�)[��~y�]]g��ά]�dfbl����`̦;�Y{k����>q_���)w���1���{�����_��1����Yfq��d����j���&��S)<�vr�K7~U�����o��Z�s�ޙ���ul��o�k���^L)��c��S�b���k�h14YP�z�v;���1b�b�j���FkĚM"6&b�b�8(IQS[�6I��u`-�VX{t-�)d7R�n��T�.���%*J4	5X���v_b�}/վ��i�����Ӊ7
+a��ZZ�_��|Fy����)��7>�}_P���F��2�Uu�|ͺ������{��H�3�|��ت����������_���{������Z{�m�u͗U�ݳ?�㭃��z�s>��JC܄Շ�Y�`�ۼ+
�+�)�C����yM���i���gw.���܇s�#��4��@=�Z�n
#���,^w�[GT˫�uު����Κ]�e�����7�
G�+�P�q���/۾,�W���w�m�֪ܞ�(��!�Z�@B�4K�^����e��``�p���d�T틵�Z�H�m؆�|�{�9{����6��N����h��$���_���0�\N���²�2BVH�gu�,:`� �7PX��p�i�%����~��}~o�^-�6o�FQ����Sb��_��:6wʼ��7mr���/6�۽���ۛz��;:	��u`��ꖤn��߰��_KJ۰7��hH��m��\l���@��u���x�lيo��5�^��sG5��O.I�Msow��{J��0�q��19�ZҪa�$�ѴuB�0.΅�\�;�,��4�iܦQ�+�F��
+�(�|	_Ʒ�w�'w� �����$��G�BV
+�"��t_PR��@�-�=���V��/�m����&�)Z^v�����+KO���ÿ���s݆�`��r;l��}��7��٠�ν���BX�1�F�v�xME�&���^�!�fD8/�S�Σ+$O�(Q�aY!0�
+�
��R�?�繏X�2�שּׂ���:@�%��vt ��7k3 7{�l3�\a�H�� �AE04j�X
�z�rC�a���*��I��H̐�F*#�^װ7)�b�Uo]f�`%��<C����)��	Pa��Bw�{��jB�f�"򐱘���K��� �����d�rX�"Q�Ix1�W*g�h[��XT���f/�׽�*�
+z-�Y��t�ok�����w��T�z�U��0d"|�/�ξ�
��M��I��������.6&�P��l:��O������rRv
+�X y�����:O�g]`�p�m�pP}��,z��
���"��C��y����|���=��/����1��h(��W~L�V�<1/�=�F��uZ]PQ���5Em�tQ�bM��@����E�Z��EE�֟V#R�!�^L%���D���J�Mӣ�5˵�;����V:F׫֑:t/X�c˕V�p��!�1��r�J�iM]��9晜�zBE��lO��0)�ͧ����
+Q�>&Y^'JBQ�O2��잽'Q<	��G;�@�ƫ��|,	�ހE
5��F�����-�-�:{�Ş}3��-�7���4^G��q¢û6<|�-ȯ_+�,L:K}�3���"[>����u��tx*��ڍ+�ҳ���ޏ��6rf�F�r�ٹߒ�R.4���2�Ě�C�jd6{�9�2�C��D5͚.�V�
��u���2u���xL���u񸙣��I���Z.Y������{�Ӆ��Bhe��#x^�'�O�(-��`dio���{מ���0K�	r�'���l�㗸JD�(���gNU�5���_��m-��Kɝ��r�tkG�Ϟ�����F}�8�1��X���/�|	�2�|'`Oj��� ��xT�e��R�S�D���G<�Ή	ɞ���i�#P7%#e��.����UU�1}�r���{�e�WC��G��^�#�)���O�A797���}�K���?�����3zL��3��u�N(�c�a�3�Y��.$�y����+�H	|I��ϼ:���q�Q/I}9����ֶ�džwb���o��ع�������7�<w�Õ�Gf�Z޶����Z��pg<�<+/]5^O�ص}��_e��������6�2��)�B_rIKz:<;0R�)X�0��_̑����8gO���G������~֟��笗��
+��f�����@�X�϶ȉ��l��ܣ|Ǎv�Z=s�.4�K(X�pd��q�(�
+G)膟R�|K�J�2:)ʄ�}ȟ�v��\G�Qh���
u��R^e�4W1��ʣ��-�a#J6�ԣ��%U%�Kp�<�J�bt���ۚk%}�"F�-����Q���.�,�����d��?^��*$o�3�4O���&�Q#V�K�q[ ��2�!7~1��]�|*������.� �n��k~�c��@\���lKlڙ_i�Z�U�\�����]~(�ҪR6u��ó�#*L��K*�:�)v�Ȍ�#7��ݷ�����-�qp
+�O$p�B��ŵ>�,��.M$V�l��U?�_���jW��kl���?�8�)(�K����C�x��t�pT$U��
+f ����""C� �S�U�A�)��Ba�����
+n�b'0;���H���Ʃk������>z-�3Vz������m0y�yw�ku������T���tn\z��{�=D�����e*���V+��N�ZeV��*��S_��Y7�X��37���~��\B�	K�$c,5���u���
�m~����t�3��Mň�I�eR���l4,
�D�+�T����Z���~J��R*t[+��b@�F�҉�@[�
�_�j#�(�a�Uk��W���,//�)򎓯�Mν��(FL�auu��ʂ��b!�!��6R`6pE��_���[���{���|�7��+�E�e�n�R�^�mL��j���#���Sׯ�/)1VE�U�ީ�ƎV��߾�;�
5��FZ[�;wZ���&�)���y�=�b5+m�T���l��K�0����ܫ��)��Iu�����M��=TM!�C���?�P
+eeL�RTF+�`>EFA*Q.�Y拊�Rb)��0�c�F#����9#n4rg���Ŋ�l�*v��ܧd��v!�T@��(\�^���d-�w���^ݘ~d䖇�t�W��8C��l���Qӷ0Bo��[%	�Qk�<?�4^�Xy�
[����2�m�8�M����eF��M1:Ic@�ރ�
��I��m��Y��Mh	��
��k��u�5aųB`fz���t$��/۟���`w�l���*6r�"#44t4��6�9]�4��yMC��{M���P�Rv
+�a6Y۰Lˈ�c#�Q�	.��t���ņao������d"��H��b���dbrn6٧����%N� �%K:1���5��{�`K�cee���JhA����ޒ�U�*�-�xY�`%^�T��&��Ngkn8� �QN/t6Xy�PN����gR�'�
�Hw��7�G������6O2fj
+�ndON��1�_\�I?�33`K0Q�K�dJؽ��z�v�������4�
�_�>�?����F����|�6��֖|K&c4�/���U���"�H��M�B����P�!]E�7?���g]mr�R)���8t���s
�Ċ�>z�8�I��a�ge���p
���4��G��	�i!�����1T�S��%%�Uj�J�.�$�ݹ���1uJU(���*7�	�qG��� ���tk@�^̂
�Lj�ʜ.�/����
+�������\��{��*U=R��_�aߜ]7�`p�AetQ��H�0Z��s7��:���=�D�(���	�&��������c*�:��#���~7s�('�d>�c��X;��>���f-t@	د����r%`@�h�7�9ȐˡI���'�ƹ^r��>h�[���H�?���ٹ�#�#'!C�n$��H�@������r%ԑ_�Y���ȯ@���7�g��<<�!�uh"��(��ٻ?���ݐ�g!#㡟pνL�7w-����C�'=s?�K�R�{�|�p���w�G�G��$G~Kf�M�q�M�
+�/��U�j���E7Snꏚ�{�8�d�����7†1��8h�wӗ�>��r��U�Om详n�ᗠ�� ph�M�_�g ��z��D��e�$�'h%��B�����

+�0	��/�
+��X
F��H�R����'%X��b
�@�U`(,U�	X��,�$c�(�2pbSX�ثX
��V��`��/�*(&��R�� ���dߑ`�"<��p0����<?�y~m���7J�pPʏ`��$�A���)�$8�H�A��a
F���`�"|��$�����T�lF�
+w
F�a	��y�u��<�u��(����%�d�{K���湿�y�`�4O~�i���CR��:C��G�r����\�~��S��S��S��O��	��^�mp�`)�셭�
���D��t�������F^�g���3�-j#�#�%!��s��wE���A%�!5��
+�[`/����O�;,=�`l�!�
�A#석��*�[a���T���p�*>��6�}�Sz�'w���8X+�3Rx'�P!��s�*�Z{"��!�P���~��W�}�Z�I��m��Y9�~/H}F`�T�p��>	{�0�������c��z��DЊ�3���nim����v�Jh�}��A(�`h�����Ҋ�I�+�IZS~7�
++�
�B�3?��T�g�4'��:�`�D���qw�4�Q�T���V=������=�a
,���ؽ�Z�%��~�"͘��A�][`��7_F}��N�/�@���}�j���䱖׶�_B�>�O�O��R�e���o�ڪvj��:�.ώֿ�h�#Z������/Sѕ�[�
���^�v-�"�?�W����I�}�����wN�T�i�DX����?��,̃z��pBHN�y� ��F��
+i�.`\'�p���a��@X��WH�!�l��t^�sY���0��+�ތu�N���~��y�0��Ok�ŭ�u�\���ί�:8xMve+Z�����5\�J��&2I6��d�'{^������V,�����yE[a}���:�lM�>8$Q��߾�
���\�^��r�
���*�<��CZ���Z��a%�z`)x��a�@�@�!
� �!
+ 4�A��5�$ۥ�b�������PM�m
]#���oe�¥�k[
�k�׉_g~M����M���{�d�I��=�5?����ٯ~-�~�t����swϗ�!�{�G�������;��]~v�N;~����.�M������������e,w&���[s��ށ�~�΢���^_M�9]�yvӳC�������c޹��YrR��$'U�4��"��~�}-��S4#������y��w�{�e_~��~�J��җ���>{�>�%|k�[lf6wl&�������~og��X��G��/$^�3�oz����#���"���|��K�����5���ʾ{�ɾ�G���v�����)�~�}�|	���s�<"ן�05�8���И�=5&�_ر#{�=9��>~7v����i��~�D�U%NdN��'F��+�vv�H9{�1�>v�~��2��7{�HP��sGxo��>�%G~v�?�p�=2|������D�S�����`�.hK�h�_���n^˹h'�u�h�6��d�h��F!/gm�r�l*g
�r�$�X�.giFGQ-�RQr��"HN�N���r��^FNд@'�M&�
+��	����:%KǕ,S�P�d;�XN��]��
k��թ\4�>��U��@{Nձ>�e�h ��<�AW��y����i���Ĭ�G�o�9�a�ر�/~�q�,@==��\����pIO�
_*鑢�#�)&q��s�!����~�7�{�7җ��^�s��咧rD�7�!�|N��ٿ*�����`�6�o�Ƒ�h�}葇�/���w�������6��V� Լ/ }��AG�߷_?�����I���~��
+endstream
+endobj
+
+90 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /NILTCE+Georgia
+  /Encoding /Identity-H
+  /DescendantFonts [91 0 R]
+  /ToUnicode 94 0 R
+>>
+endobj
+
+91 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType2
+  /BaseFont /NILTCE+Georgia
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 93 0 R
+  /DW 0
+  /CIDToGIDMap /Identity
+  /W [0 0 1000 1 1 670.89844 2 2 571.28906 3 3 409.66797 4 4 241.21094 5 5 558.59375 6 6 429.6875 7 7 269.53125 8 8 613.76953 9 9 565.91797 10 10 279.29688 11 11 389.64844 12 12 590.8203 13 13 345.21484 14 14 483.39844 15 15 292.96875 16 16 539.0625 17 17 503.90625 18 18 286.1328 19 19 509.27734 20 20 880.8594 21 21 565.91797 22 22 596.1914 23 23 814.9414 24 24 535.64453 25 25 618.65234 26 26 575.1953 27 27 454.10156 28 28 582.03125 29 29 574.21875 30 30 291.9922 31 31 767.08984 32 32 560.0586 33 33 432.1289 34 34 410.15625 35 35 737.3047 36 36 609.8633 37 37 749.02344 38 38 496.58203 39 39 725.09766 40 40 410.15625 41 41 642.08984 42 42 599.1211 43 43 374.02344 44 44 710.4492 45 45 504.8828 46 46 443.84766 47 47 269.53125 48 48 517.5781 49 49 325.1953 50 50 561.03516 51 51 551.7578 52 52 653.3203 53 53 694.33594 54 54 744.1406 55 55 603.5156 56 56 528.3203 57 57 653.8086 58 58 492.1875 59 59 756.34766 60 60 927.2461 61 61 226.5625 62 62 856.9336 63 63 615.2344 64 64 564.9414 65 65 502.4414 66 66 701.66016 67 67 975.58594 68 68 559.5703 69 69 468.75 70 70 601.5625 71 71 312.5 72 72 666.5039 73 73 539.0625 74 74 226.5625 75 75 478.51562 76 76 942.3828]
+>>
+endobj
+
+92 0 obj
+<<
+  /Length 13
+  /Filter /FlateDecode
+>>
+stream
+x���
+�6�	�
+endstream
+endobj
+
+93 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /NILTCE+Georgia
+  /Flags 131076
+  /FontBBox [-73.73047 -216.79688 990.2344 756.34766]
+  /ItalicAngle 0
+  /Ascent 756.34766
+  /Descent -216.79688
+  /CapHeight 692.8711
+  /StemV 95.4
+  /CIDSet 92 0 R
+  /FontFile2 95 0 R
+>>
+endobj
+
+94 0 obj
+<<
+  /Length 1670
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+76 beginbfchar
+<0001> <0041>
+<0002> <0070>
+<0003> <0072>
+<0004> <0020>
+<0005> <0032>
+<0006> <0031>
+<0007> <002C>
+<0008> <0030>
+<0009> <0036>
+<000A> <00B7>
+<000B> <0049>
+<000C> <006E>
+<000D> <0074>
+<000E> <0065>
+<000F> <0069>
+<0010> <006F>
+<0011> <0061>
+<0012> <006C>
+<0013> <0067>
+<0014> <006D>
+<0015> <0039>
+<0016> <0038>
+<0017> <0048>
+<0018> <006B>
+<0019> <0054>
+<001A> <0075>
+<001B> <0063>
+<001C> <0068>
+<001D> <0064>
+<001E> <006A>
+<001F> <004E>
+<0020> <0062>
+<0021> <0073>
+<0022> <201C>
+<0023> <0077>
+<0024> <0050>
+<0025> <0044>
+<0026> <0076>
+<0027> <0047>
+<0028> <201D>
+<0029> <0043>
+<002A> <0046>
+<002B> <002D>
+<002C> <0058>
+<002D> <0078>
+<002E> <007A>
+<002F> <002E>
+<0030> <004A>
+<0031> <0066>
+<0032> <0053>
+<0033> <0033>
+<0034> <0045>
+<0035> <004B>
+<0036> <004F>
+<0037> <004C>
+<0038> <0035>
+<0039> <0042>
+<003A> <0079>
+<003B> <0055>
+<003C> <004D>
+<003D> <2019>
+<003E> <2014>
+<003F> <0059>
+<0040> <0034>
+<0041> <0037>
+<0042> <0052>
+<0043> <0057>
+<0044> <0071>
+<0045> <002F>
+<0046> <005A>
+<0047> <003A>
+<0048> <0056>
+<0049> <00F6>
+<004A> <2018>
+<004B> <003F>
+<004C> <2122>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+95 0 obj
+<<
+  /Length 21869
+  /Filter /FlateDecode
+>>
+stream
+x���	xչ7~��h_f���H���Hm�,۱�{'��$��D���=NB��,@CBB�����8�-84P����-�����RB{[�Bi�?�R-�"�{Α�8,�~������f���w����= @��um�=�6�]�m`���*�7��z݆�w��VW�Aܸ~�����c;��7n\�Z�3�o
����q������1ذu���^x]	�ι�����v���|��o
�����������n۱�j�\�{\w'��c�֏M��
��>Ȼ�j�U%ǀ�c����@c�W��ޕ^��?��;�UP��?�����p���e�_�)���?�
+���]�5���ց�C�p
�F�����@-��=�<�Z
��nbE�~Z�2�l��3��,�}�7`
+t�.�}�Py��n�`����3�*!C8�U�~	��?��������^�G����d���]`�h&��q�D3� 6-mj��ץS�D\��E�H�&�9����]N��f��MF�^��Z�Z�T�eR	EDᄵ}�M�;�^�P���~��	2����+Nr|�"�޻>��=�~�0Ntq����@כ�0
�}
+4,���E��6s��&l��FG'���a'��"T�
+��9���k_��E�9���kWŢ`��;�Z ]��� �Ģz~�t���'�c�D���z��������m��	=_�#8!m����e7M��'�1�\t��m�X3ʫ�q�V�� WŢ��ܸɱ���NP
�od�����8:7�N�
�c(����脢}�w�1�o<�9��'������/:ȣ��M,z{��av�����G�h;44d�E٣�\��юX�ss[,
+�B,Z�M
�݌>s�j�=;7�G�����6����M������\�u�[�}�h����t���WwUvUN���ю!4˱(��lG_�[��(O�̞�ʞ�����A}�y��������h@��
����<�!���]�jB`8�����߹r���i��@�.�k���.��::zt�d���e���zz��u���op��,}�c�붡	ft#l�EY�
]��9�W7T}�W}&d�R��=�H
+T`^�e5;
�z��	�lp�1!.D㥃C��+R���T���PElHF�f��^z�H;�M�`M,�8�x���k�Q��&�Qtd�zĴ9X=2s�(�E�M���?�1:76N@�sx}����}�tH�
탄�D#%�1l󄅏E'j��Ǿ�M0��dp��<�2�	�G����Y<<�v�т��/Ez��ur�7��R�	�}������
��q���s"<�dx�y����gH���

�&��h�$P�
�����IX:4	:\� GV�&�@�e;7uL���$X�olFٮ�sC�Q��uG�.v��uT��r�	�X2��<��cf�~h�16	��}(|��Cl��r���G��BlG{�	2�7�xp�`�cB�@��휘����px�Y1�MY�s�&k�;��N���$ȗ�dpBtL���G��8����GG����I0��|z�X�1	��@�$<؇�Hs;9/�fb�#6	VE{�T�Hd@
+	d��m)A
�_x��&�꼺�W�@
+>>(
��WpP�c3Q�O$�%���S� ����{،w��K���1NGj���7Qjor:=^�B�qN��}
+mF�Q�&��
z¬��a��)**��B�&Km��ɤ���"1yHC�y��iK6��
+��8��.%�zK6������ʧ0�u-�.�^-�i��h���u-0�3z`HA7������x�ܢ�(jo�vrZ�ro�m�"[MԴ���o��2�Vx�?/tK�f�]	[�6F�z]M���&��)
U�����!�h
�`�U@o���G�
Ԃ�5~F�M
�a����Ҕ(�h$�Th��MA!-4։���T��YMg�g/M���(�vRa6ZΚR&o&'Ko��i���,}p^�ƃD�R�n��~$Q�{]t����5�E�
+5k�_��\Х6�&�d�"���z+<@_�5Y��ytK�d�������M]�w2�忝/�ɋ�ʎ�<����ܴ>��}0���y��@74�2-�|��3��LTg�N��ja�.`1YZ`*i&�3Y9�T���J֓/6�.�Ϭ��;]�!W�7�F����E7�����/Xy�֚ ��m_8cP�y�6�뻞�f�::��8����U
+lo'�2e�G��
:��h5�w�}�.�������g��
+�;M6O�`d�>��sb
��†�_E�$$E�
Rq�:d����c0�5Io���o)��,�!���Q���
+i�ϘuVE����iFs�#�;��wN�.��i)�����|��<�i��#�ুPH�>��������<4J9_�.]�A��b���oJ��d��?�p�z]��,	����n8W+i9����+R���ٯ_�U�!U)�D[AI1?��ӬW��V�������Pj1���b�D�f)���;����¬Z���U����v��
NC�A��V�[T#9�I!	�d.ٛ$B>_�
+��
+�Y Lgs�)!���|�- P��� R)�����\VN��KY.-w��Ռu�.�,�[6�^풙r-�����U/�����~��{�$<�Ys���(x����b�`$���X��6�����][���ZV7@�����U�T*Fc'����J�Q"�`S�3@Ti2�QB��)a��C92zd����ۢ�A�m].w:g�I8E���:}V�y>�3��N&A@�0�̥��D�*�:W�J�gRZh��.�r�z���l�z`	)���{]݄F��ň�/��m}���4�Ѩ�@����&�̀�__�h-��i�
?�
�#t�5g%�B>U���/L'⁲�C��(�@*i��l��5��VwD�j��h�����7����\��)һ���M�b`
�(:#����Q睫�6
*���L�
+��V�T�pgɐ��K2@Bv��GC��4�+Z�l���)��i�|Y>��}���!��#LF}*��H1���kj���kk͵��~ӭ/��o��̃'N<D>���;��g���=�~q�5?Y�hxe��.�?~��ih�����)`5��5��Ԙ�Y�H�
+Q#j[��M�H��XOnԏ�o���A%���,��-%��O�^�J:����b�2�H��Ӵ��M�T�
��5r�c�Qۏ"g%����R�_D�"�Ho��2j��G�Xfj|�k�t�
+y~k}6�d���(�|���2d@�"�2$H��l�(j�S�z*Փ�wnt��������ÿ��ϟ�����znl��}������9q^Wς��
+�÷�LW��zZJ���/4�����.��爑)�{���	bʀ��{@�m��Q���0 ����1�
��F==ΐb�E
�
h� C�0d��[�d���~�����)?%���^D�*������^��T>YzGT!�.WbFN�1A���r/��{�KTM���/.^DC���٦\��^S�T����
&��v�p<l���L#�nB.��j(���������I����d�s�n%�Z xN�7��,�$
+
+:>y�gb��i��	�y�A� ��)��6��<�`D?�a�T��
�:���f
Z��P�C?�a��!��������0YzS�#a4|G�E�(�9������8�M�~�o뫆N4(�ۇ����ڧF7�M���!��|�ʐY.xG���/��i}�@~����?�J<��D�ZtY2z�cmwݠ�R8�����2�Z粁
���Z]V�S���M��VKS�7_T}Ex�e����n���NJ�V�	g4�?OH���	K�3���3(���U�v��n�L���]��!��9������M�l����l�tIT!A��H�����&Ko�Gr�U5�J�v�)
AA!7�ͦd�
l�8��R:��[��B��W/���g�{:�R��޶��;�v��T����oZ�x\n��%`n�M�<�!`A�@�EE�5��J���I���\I�TTKF�}�͒�Q�ݘ~�Gk9��)!C'i�E���tD˺�!��cx�lf�=�Vt7s�x���tT��FɈ��
+�=H�jt��A�(0�W��E����*$�j6�P@I	<R0��FȌ�JfvX���zg�C�@a�|�K���VccvNS��M[���
+�5��ˇnl��`���6������s�Nm}~~�򎦆�Cm�jnl_�.l8��6>��#���@��y*Ma�B�[�A!��p�D�/�t�YFT�&FD�ňh��.󲎱g!���5b^��ʫ�����*��l�e��]�{��v���'$��BT����^DbD���S�lӉ8�g\b�R7U��J��P����\7\�u�����L��'��7cel�^�CVv�\�J�"$����L�B,���(�!Ё 8#j�u~�z�f�'�f�d����ӣ�lEy5�
��#(>�-�a�"|���f��p��	��X��H0X��jt�
�A�S7T�f���c��g0���@6�W�lk�h�����UO��.��W5?w�ٍ=��X���j�����7�[Wߢ��\�',���=g��y`��(�{����'�]�C�z��y��4%�"y�f]D����vi���fTA�6�;i*���
�\�YP�A�p�����*�hBѡa�M������K��ޣ4���atbŜ:$�p�N#�9�А3g�Xt�:=�^]ɭ�[�wz
+s�N����/{4���`ģ������aE��e)��B2%Lgˡ{C�il�|�c(��U���^~�ޟ���͟JQ,�V�	V)���c��i�ݛv����~�*�q^{m/c�5$��t�����R�:��I�����f���ۻ��:Ӽ���l��݂V�ҾV3��xt�E^��;Z���Y����x��P-� �K��ߗ�,�p�	���T
l��B[�����,�MRf4k���
�3��x
+>J;<��s�9$����ߓ�reC�x�?T\	�#�t؃�r}ŏ���2�r�g!)>�c)�x����s�ǒ�9��~�wk�.3�h#$E�O����j#!Z菿��P��m�;�?��f���/��R���������=���{90��G�;i�"$CI���e�!ē,���L&C��+N*���n�H��fD���B�(��̏���9d��r!�+"�ܙ�I+�Y�H�߯�f"#�a$7B.�y
�I�OS����U.��2RUL'�u)IOo�N�R��O<��;AnZȧp4ۅ�N؅��|`V��Է�V�[~A��5q&|�Χ����	 �6�
b��"ﮩmɵ��7^?�%�h��=F7�b<6T�
+�/�v�%q��������F�v�Q�z�@~*]�|j����wى'6G����
+��o�+�*zS6R7�ұ����뵣R~c���?��)���!�{����|b�m�
+B�\
�4�]ĉ�w'��,�ߓ���⦵��[�1	4�1C�>�Pv~vI�P�В���T�g;^�d�ש�^��^Ĥl�jH������iH3v��%��r��C�Q���ņC�G}�G�b�:��=�<Y�(���4cg�<Y�-��fd6h��VCS�����<Y��<����R��|4?)7W��}�Áۏ-M�-͉�{@��beEш��I�c|}�ωn�C�i����P݊q9Nz|���>������֮��XK����dF�nq����P�����f�_���?������i�/�}�c�N�Hwb��
+)�.K�][��}5J��u�U�����2�\�i��Q�?^��w��e�9��q��aN�
+����
+��G<sF[������`=�ҫi�X��:LV��y\��Ξy��*�����)��fP�Ms0�!� �'�M�^�����)��þ��̛|��y�mJ:c�tƉ6��TF�Vj���= J���9��)�e�L��`]��
�b����:*��Y>Gw���櫼����,�*���HȤ�,���g*~!��H+a2]_�����W�/�v^?���s����
+-/�*^A������@�T�ūw�kf:��7]���:�?,^z*^<��l����5��o,�N�%��JBV���L
ʹ,HJ���3rQEgj�o�3$ʵ*:cC��D�0zB�!eȺ����|WWr�5�ܬ5�p��k}��}p�dLv�fL7f����*�Ew����7l�z�I��F�"�5Q�����(i*�i��SJO�t�������
Ύ���Uߠ�,},֠�j�Zo���cȷU(���ΥG�ci*v��B��GQM�ނ1ϣ7���%����B#���(�P�P�U���|,��P5P����V����쇦��/|���_ٹ��3/��E�S��7��Σ�ᵏ����7~����-�}����6v�����6ݵ$�Pacn$\§��ij�ʸe
R��h/��r�<nu[�5��8T�E�O�!���S�q�	e7�ᩨ|��zpj�d�/8�A1�<�-�zsp�8�N�S�;�:)N�W������3��b�(�F��'K��C|�d�<=�g;����$SH~PI+��*�ϒ�ŲD<����~�k�!��_@#I=�oYW����z�O��Vygg��B�b��B|m�;|��ԉ�GQ���
+=��[Gk��tbd�oq.|��=�`�
+�oPy�4����)�G���8קpX"��[��ΜH�%��m������<b�Ɣ
+�dJL��n��P�;��$<T��g1�ź����pw��`2Y���1ɞ�X=ű�d=%�5��F����e�+q��N�F��
�
���j��h_�Z��ĝI!�5IQ��Ɠ�2	r�,�t��Qh�5R�3���)��&��M��7�MU@�eM2av�T-�
����O�^����UT��5	�:K63�)�T!OM��0S�O%�p�|��j��L_.u�Zˌ:��zL�勿Z��h��e���NKkӐn�.z�EnX^�Yz��:�����j�	����*��2�>Ъc�u2��[]
+��a��-��qx��⛹&?�m������q�^E��S۹��~N
+B��߉F�&
+��'jH�i�hi&cD�㣙<
�4\P�wO��4^�q�q��r0H�������1;�#�;P��Cbw8������O]3��L3��Lϐ��.Cp��89Ty�{�@-�q:/��N��rw{=�j��~7[X��m�z=+#�n��V�E�B)�k������+�-�"+T+4�0u-)��c��^x#�W���|$�S���J��Y�Ԩ�f�F�Q����l��c<kz��$ς����9lE(.�
+�=�Dk��F1eE���&�s�(ƏQ
+I)��(�u�HV��=�;���NS��H�@���{�RZ*�!����ӛ~8-�2�44�?k�A
�0��4h2�Gj0��uLUjf����)�d�}�Xjf����.N��;w]f�q-e7L�:*&�y�|@���$`&sv���E�|&}�6���Z�C��o��a�׻�<z�=3�r�UN��;
+7�N�>�֛��lY����q�kS�Vmaɖ�>�I0�7�WKоK��u*j}���1
	�N��w�=�DJ�s&=�U�����E��1�G�U��'������P�	��_�+�k5�!L=a"T��a!���G���$�m'�8#ż
��KC|�5�7���SB��Uj^�#fWd+��,�
+�٢���[[������♦V+��Ǘ���޲�����v�|o���6��91{L�L�ɽ���U��G+�������\��ΣRǷ����'G1�<9>����,.{��Ս+jJ��T:.zTKo��*�-
+"���;ćE2'��w�%�DX-�<,�'Jg
+!�>�|9߈�羒OB����}Ԙ��=���?�咊m��Ԕ�&W�a�b��Ͳ�U�U�Q�6��Œ2��l���^��l�䚢�Ҁ�%�!��W:�f�9)�+�p����G�W�_�݆R�+�%��nV��.����iT(\��)��=� .����.R�js��y�+M�xsb����t�\Šb��jӵs䮳���"�Hޏ��(�L8i`B�-6�#3�E�)�v���C�/���^�q��b��"{��;^M����lxgˆ�;9!כ͑u7k��0�#���a����Y�&('KE4�m�<sӗ�^�y')T}N��3��x��gK�qq
�Y�������/}���s~�s+;ŝ6�H���z/�+�;\����C
�
r�z�6��ڕD�������25���s��N�AG�X���K��R��h�7�lޖƆonZzkSCG�W�T�ߗ�<��=b�)1�N�Q���8��8΍�ӧU��I5�^�j�I�#� ��Ã�P�>.&�p�{ƅ���2b��G�C�=����d>Nr'K/UI�K���
+��F�����.wń�X��8F��LN'�D	�
�V���.u��(fjh���,�0��
��w
�k=q=��zl�ux���V=��� ����[�&}�m�ěl�^e&[�Z���u����z�1/�^�/��*H��Q!7*�y
+AŠD��ApAZ��6$$��C��TZ+*�i-�5&���m^����a/fC�D"��+m��z#��
+�S�#Q@��>R���U	�D(���r$
�&1��_,d��?\�S��q��0�:�ث����T���-�t��Z���8�pv��!�&ݬ�y�K�J:���ݾ퓿R��7|�i"�i���߭]i�
+D�'�B����'�����S*!U'�Z��1�n���O�2͂��e�
aJ��2B�X�0��k2����32Y�$i�` ]7W�<p�u��(��b�\�J�	!ћ e���� �v�RYь�t�
+96t�
+��hjU�&�|L˫��[��wUU�J�]wGE��WB-��rF(� � _�z�f�K���݂T������"�<VÐn��ZNt?��/p{`צ��s�d����z��dd�v�K�ɗ|�F���^����ƻ�0#~h]N-:�+�I֐��}Ή��Ht8%h����ꠡ:�T��A���p��-i�Sf���\�!U�Q��nM�np@y@5F�
�E�v7���Ͼ�H?ƿǓ<.��|�N��f���fԣjR�VNJ���"��c|��q��'ő�<Y��s�čj	�.�5������=�P��J�yL�
���Ṁ���g��e�?�LiZ������ϡ@y+����Sm��82U��r������2)�gq� @\G_���أ씧�m���w�I��6����e.�i:�����^�޶�ѿ��֧Z�pyhc$ڼ|qd���k^nC�u��M��6�o�5����Ԁx�^�7�I>f}b��?&���i^��ݖW����x�\�B���W�h�rJz.$2}�(C2欅|j��_x��Ta:��&�e�"�ֶ����w�k�F��-��t�2�j�+=����r]�O�G�u���4E��\ٺ%�7�!��z�.���OJ��J�WAn:G�F~�j͆�F�OE5J�@����\NtzK�)����D�qN�϶��4�Q���R�P���ϣR%(�}�E��3/�B.Ǽ�9?�x�IգZ��NEI$J��g��Uf_��9eJ]g�	��vu��WQ.Q/3������+�S]���s��{�ϭ��i���8=j��J�z��c��;���J_�� =�8�q��jF�(46MTq��b4*֫ eF��V�2$��$�D!Ȉw����*5^�\+Vb8��1m�
+)����|~��*�q��K����K���a�T�%�;� ?4��]J�]
� 1C�!�6�-��aJj!W�J"��ɡ�>@�^�[�f8'�ev�xiB~+Ӱ���א�S--'��8��6C��y>{Z���uA���ӛ����R���vޤw*�r��
+�SxEc�+ֈd+�������|�ڻ�����H����p�{H�n�	�âs(�0b,Rtʓ"�Gje�=�0�,3���lK�Ҵ עAQ‚+@�g�:�FU���q� F��|R��I3�J��:��O-_��
[������z���8�#����K>����Zv|���Tl����~�N+S����k�!
+p
Z�������U��ި�.o/�2�@rNr�e�#Ѭbɘ��,��xF�-^9w�4n�,���ZK4���Tk�Mf�Z�1���q�B[)j���s��م��z5���{Ae��c	"��tJvJ;#`-��*`팀�X��g�3��i�����b���	����l�Z̳�i�ې�;�m{�ݻ
+�]y����<s��փ��w�������W�6o?����,�O�?\�li�J�I�)�gh-���>T|�w�i�	?��]�}K_þi��w�L�
+"�%Q]��Ba��{�����	7�;�<�	]�v*u�9}6Ǽ�c�C̱!�3�W�^��{�TIR�9ݶ��Q�Qjl^R!Qj"�b��he�k�RAڼR]�>E=$x}
+���E{��LU
+U���#g�,���Z^��yᰶ֊RA���>k�����y)sl��i�1fX4rj�8�?q+%�{���R�X�Zx��<�rK�S���o�imJ��-F��l(�56�C��s(�JB,c�E��I�Y�$hS�<�`�� 
++#Kk����M����uf�o����������Q 7��(&���9��e�SRw��u�뤮�n��=}�⌸VAg<�����;ډ�>�魁B�p��4�j���mU�W��R´�sS���xII����]�h�H))�D]Z���_��MF3-p��Y����6��$?:p�Ҫ��-kF��k���)�M-���5[o�����򝽃_qˡ����ۉ�Qә��k�:��
�F ��ihڰnGc����?���,X��}��[>��ޝ�Xi�i���}����?0��.i���B�Kڟ��s*,S�Wk��ќh���Z�3$����*8�O],�V���iTե�]^�i��|A����Pͨ��1:X*��3���]��Sk�Vn���)���N̆��\��]��M!yԂG�/I����|6��7���p<6����blĹ����D�S^O�Q^1�D�١?�'�8���u�턚��W��?eN�:�i�d\�C�>)V-)V-)V-)V-鷅�j1�ժ�L�:!Α����P���ot(c�Rą�5<�y�����B�1��ǿ�l��9�m������G��֞��E�U� ��дn�ִ�Q���fs���?������"ѱ6��g
POK��x]T ����C=�v�Z#m�#��׉h���^�dtz�	]/�'p'���J��XڤP$9�'�I����i�.O�W$Yg�h�Q��u���±?��Hd&�Dp��TcLd&�Dp��<��� ��
+�d%����j��Wi޲x�1I5����E��z��th�gY�����]F�d��3�W��v/��^�Ͻ�v;̜c���jz�+(��/֜�+\�����śfdF���M1���u�����'��O=�@q(�>H��V��5,�~��y�'
+� ��_<X�wM,�	�&P�N{�F�1�I�t-�X�W�\l2Y��ݤQ�ds5�jn��ҬK%�?�#Y;��vy-Q��ee
[�����u��,6���u���s�ۖ�+���l��m▹W������Z����d�M��q�/�P1�����
@"��Ц��ek�b�Ii\��Ҫ��B���+��{���LF��a<`$�����pCc����@��j�A�|
+Q+y�e	3��L!xj��V3Ia
+����̲�Z�W,ڬ�g�1sv�tC˧J�3�E7�H�6�����l�|c��»���t�����>�.�_iܗ�q-;��:�bG:������f]��&NzU���ޫ�ϻ���S�q�F�l�2(�;�6�4Dw7a�F����a�uSh!��֘��ٛv_�rl1���h��E�U�
WU�)F�Z)?��~��y��ֳ�����_��Cd	�aT���Ŧ����8�BJ�ʩ��ӷ�׼�•���I>�u�R��$�E�N \.V�I�M�J%km��[���$R�����fYY��������gZf_Ѳ�~C�"�m
+�Bh�e,�lᚔ�n\-Q�2���̃�P��d�TO�[�H���~�h��}���,uJ]εR�#D.f
+��
+=y�<���le��"�[��f��֩�jD�ͱ�Qi[�K�et6[<ѱ���L������o�-j�9_��C����.p{W������ni�(M� �yE�
�J��쒈-�l��k�]�
+
+mmr�¬���?�(3)M6t<�_�������t&��|8���HQhrͭhDji�0A	��jO��MDZ�CthWK�[������*�Riɓj8�s��r:	z��hDw�OB�$���TZ�x�wJ��=n�����lV�p��4��#�jw��aI-x¹�U2�x���&81�{'�]�Q{t���3��C�_�7�@�fyGMX���ȥR�j�R(�/�����$��G28^z>.
�v�BP���P����
+��-1�F��x}�ep��ↆ�����^4D�T�^'ߕ��qѠR:�F:@�z��IAt<F�P��z
~��d�g\c���J
��]��Hq8Ug0�͗+��b�lG�R��R\��J�8t2��ү�U3��Aʮ��b�/p�/@�+�f��+9�ʒժ�/K�]���Z�j�4�Ӯ��yv��MC=���������.IZ�������6v,���:6�gX�fi�y��
铫��8A��0/
�2(Հx[Ԏ�;�$I��V���<xW7���+�����
+
RYRq&L�Ռt�0�~�b?k��V�6rZ�����*D-�>�h��!�S�GK^�6�xgʬ^��pj��m�B/jd�e��<�D���$����Y�"�`�|�`ιܯ��Jy�tٛ�`��ٕ�\]�.�GKA��G0}��yŢ�e��.��я���a�ݱ��ܠ#������f.�֔�4o��h��Ђ�X�U�?R1�ׁ>�m8Z@�L�-��
+�zD%��'K�O(��B"��-Aa㠅��K�K[����~j�?ؕkڝ38=���rjamm�4�p-�M�-]M]-M�������UC����㻤���XL8]��D6`j�
'�
���֧� 'ۄ6�<�RiN��ZT��)��/��b��f<�C�d_���(�C\J���w�2cL}�1��,�.H��3�1��_|�?F����i��3�c+;�bY��H��s�=�6��5_�}���L�����-��5� �%����M�y���䚺�^�P�A{�Y���ŷe�M�5���67v^?���r��b�EP��R�Wq�����H�M�l\���/LXs��Z[��,Zl��,��r�Z�hps�����Z�����[�Z�����@���G�i�`��2��
}���N.tk'.�t�%-�F��3��5��:I��V�����̌N��U�~T=��J���r�$��U�:U�?����-wuNc$�
��m�)
w5M#�;�u♖�OMђ��|�>w�:d����Ӫ���s&�,�`����h���l�y����nM��l�\f�ͩK4f�g���Hm��d�����:�?+�jj6f�MAv�I�V���h�
��7�B����;`Sk}r6	4-H��
+B�tj�&��ZIM(|�ک��9�D%9�@BT�6��40QM����L�73�Z�k
?�����`2�@섟v
+�^�
+�:�`p�`�u�r���W�V�'�q"ȝ��Jm�o<m@��_1��ÙQ���/�p91����#�#��N�38E5�S�dp
+�϶���f8�2Q�Q�R8>?��]���yH~&M��O�]���(�����:������=�C͞���N-�v�=���3xz[O����%ޭ��b]�z-<]\��I@��JW|r���U����Z�~�b
Y��|�AEQ&}�G��
H~$9z��☜���ioy^H	�\��T��u��DR���m7R��)�P1(PY��z`/5G��J�P�B�������uTv��b�\�������SAͩ�d����ifJLxJL��&4%Q�Ή����N-5�Ƌ�k$C!�'UWO9�]�������._Tn�ea%���C�Y��D�)����h�?ze���ܒ3CJ�f�g�1J�~Tܵ�����6/��U��eP�(a�|���������:�Z�H5Z���^�����:�x�����G����\./S�Ifvw�y���+�Eh�TJ�� �Rrξ�㫾����VF���u��K�H�	,@�E��+�ٽ��˄��h �
�h�o6�&*vW��S�LA*���%��>�Ć��,Ù��;dh�i��##08��]&��D�4�!��ˣ�84�MFD�^�*?����m3�]�y�h��c_�dx���#��q��Ǎ
�l&�ؖk�K&�
�gvm���5�??�D�գw�]
M}�k�l���3+޼����l_9��Mw�5<�v]�ϝ�%�yPS:(ZUL�`@��� �"�.����-��^ʾ�9E��/s��L䲗�X�����B��}֓\ф~9��~���븰c�Y�8���!�Ϧ�ZV���\�6��+W
+:�G�4�� }��ԍ�K�m�z%�����6b��T<A^�@�!�?��,"V4�h,�ciEh6�+�*���w𞔣������Q}��QjԱ�3vA&������1D��vh��m�L⑑�X��b~��������ޯ�b��}�Y7�Y�j�rW����� 55
��(Y�?�#��Xc�f�dBH@OBH����ā�TOߪU��bvl���7�����`��gx�Gi^����`+�|ԋ�E��t��5yɲ���-/�����O��sf��o��$�]!��`߱L�2�g�C�؉�y�_����ŷ۞Z����U�ҥS
wl?�|s��x��� ��
�@���9:�I��RB�̑�$s�\6��ޱ.�k�F�tl��\;�1��P3��vg�%l9A��4�N+S'��6��j\��0���K��,��Q��b��ʼnkF�YVW���&ayTh*FA2�<���6���`:��vJyJ7s�ێ��u3k?t�vtյ�gg��u9V��N�(
+�*h���ZA-c�����S=�n�f&�nCeX^~���?����paԺ��[|��6��rM���tι�����Qv��@����xP4oXq�C�ݑ#���%�ƻ	�B+�K�4C3D�;�];г'R��2����=�HX�]��~�i�9�r]��Ai�WA��1sk�V��8�6Ϊ�.��@F����;"gCgkLg�g
5��vO��„���X^�
r9fz�Ab{i<�
+�hb��;E�i�;ssnw$7�bq���9�j#M����IԦP8�Z�踺�%�H=
+��!��7��8:KnhDp0RmS�L�^.��H��i�ԣ��2�3p��\f�,��„��Ѣ
�Qɑ�yr�+�f>[w�����w�$3a�VO���c�����c���
+i�
>��Yf_���k��<ϭj{j
���Y
tZ�80
���R���Nzm��ؖ��
C����-����;��-�=���#�;���� 9�0&��5�(Q���l�\���2
���;Ȉ���^R)�毠v��� ��W���j?o�]� jt�Ӟ��2�!Q%U�l��r��S7�M�bg*ōqR���DWJ�Mt�д��t���w��DC�
1G�ˀ��^�
+��g�9�
�
+�d�e�
+g�i�F=՗y�T�O
+�!
+�������?�B��:/��[}>���ާ��������و^�2��b�nٳ�n>�c=�u,�!�f��΍�{�>�P/�g�'��nN0ju��k�%��>�Օ���Lt�8n�9�nA�Vr4A��X��J����	ׄ�7)UF�RU�	�nԨ��=[9T�i����#��:�=*�����:���J�\y!sZ|��l�ss���,�3}�n�Z��~j<x_�,�Ztㅏ�6�
�R�4Q�R��&HP�qKd����W��*�JP�嗜�W%��Rp��5�[��?�`ֳ� ��+POЧV�ɴ��-F��.:�Ŕ�Ú��
+?��upv��3��=p��
�ϝˍ>�է_��j�!��tS/_:w�A�	5�Rf���#įr}6���������-���r�b%��2�WGE���Z�
+ճ�Һ
+	��)D�
+������?�+����*݃krN���N��FJh��?���R�����_�4��>3^
@���9'�"���Ǐ���;ڤ
�
�ƛ�J�T�lH44&n�<F���2!��0�R
�@���`��nP.��.��=%St�cbSv�:;:�H<���;:I�,�oi*}�:�%���mmTx<B�M��9��<_�^An 5�xI��ox�s�K`�ȮP�+��s�gnh��Ϥw�n>r�,��ڀ�ř�z�cf�#�h��ޛ�Z�D~�R�J�g�5�>�s�:\pZ�
mj�=��O�����c����t:��e髡v4ۈ�ww7a5K�{�g����pZ�r���HOP�TP
z�F�>.�e2) �j�j��z&�DB�4"z�&FWR2R}0
+VA�'2�G (0,P ��	z���̝�4��JE��\��C3Q�5�!�6�Og��l�}��s���R2?t���`�8�����.��A���S���?�������b�����Kww�� ��'(	�p�yJo���h�V�%!����8A0��)�&�R����*�AD�FB$VR�f���ID�fb
���L\O�B��#�B�
/��I�ܔ�	*�°U��}��vy=.3$��bO�
+�A�L%��2�����,_�$4s�o��x����i�,�GP��Op�B3����&�^u�C�c�f���S|�.��N�����ֆ���"n�3Ppm��V��jl�6��G
+3,��}~y��H-i�[<--�G�_H��u�N�̦S���Q�2#�ɼL��~��.��3^��žy6j�!C`:Xf�$?)�t��k�b��=��+
��>'q��Cb��d"���c�ݜ}����W?y�������-��D$�z��&��W���s�c�?Ҝ�k���_7�[x���K�we�H��8�W)|
��,i!���ڟ��zg�r+X���\�^e��`|9eN���W�-�ɕ
+(��9��IޤlΗ�S�y�����܀�f��Q���A�W��{-��}��c}j��&���q�ǜ?pc �R*ᛤ
+i�}�jZCkNED�S$5��M^1�Z�٬ج�Nz��:�54CkT�_!�Y�d�C��tRH�ɱ$eA]�R�
�Ӛ��R���`��=��j�I�����ϧ�<nf���;����|�;A�����H�竬4�φ���3�~u^�;�6'�3��۶�x����W�
+�mkw��H	˞�p��~�F����Z��*��o�����νI:��t���h�In�E�iH�1!1A; @0&�tTt�:::���Q`|����0��:�(��;��:;�t��vd�s�t���N����͗�����9u�Nխ:�]��\+$��*���x:��T{�m�C��aZ`�e31ɖdN3/��R.��[a�~�==�ȶ[�V�$i��<A�W�7�jX��r�4[�H��*�9��qs�gϭ�[�..@�gv�K�1�k�Z�6͖/�&83��F=�8�����C���%��"
O�<���ز�<��=��a]Ye�`��F�i�+63]��ݰ���ٸ������׷mʰ�vh��~ޔe)�..��q��S��C^(�N�t�)	\�}/�Y�N��´���e�M�A�کݠU�<
�7館�k�_���y]�K{�dIJ�7$�2f%%�P2��fe��կֆ�N=b��&$6���㝷���kr�ZfEm26|tå~w�Ŭ�b���J���t�����*(.&��o@���"�wB�&����:�@0���`5[�D�������"
+bKX����{�L�2,���'�'�6���o�����9
����@��H�C�R��o�	ј��LH�Ԟ�*D%��׎3�
o�9��}6�N� ��b��H\}.���G�UT�+`�&͂�����%�Z��-Z+�TM��HVB�*�D�%�I���d�դ@m!>uiU��ru#�Wctl�RT����_T_$E#��MQMZ��:T��6�E��/+�05B�)A�5!f�Bx��L"����q4>-AH�aJG�? ��*�\�_ɹ�+e_�,��~4e^�S8�s��`9z����/�)�쭃���_�$�"�(�ej�EDr¤�=AL8�*q3�"��!����r�7=��Ԓ�%Ӑh��h�Җ �孭���e��ɏ'� o�A>
+���s����)��F�ظ�!�%���9��RoJ�6��HwEoT͔���(@S�S���0嘈dN0g����+�����0�RȠ�rɯ뺠�y�UIqL���̓����z"�b��� 	/�L�q��������Ƥ{�,kM��c7n	^(P<i���bc�#�|����Ӣg��պE��o5a~ގ�e�'���CTO����Tӊ`]p͘�m��km�4�l&�1���Ǥ|B�;�ʀ&H�%Th��֊��Q"*�dzK�%!���7B��2��%��l�VK�dI�t�Ji����`������������u�_�>�
d�O�lbZ��6J�f��Taآ������[ͦ���ӧX������$Ng���h,8��}4ϧ8���F��T�p�#p�>[�ڏh��RR}�R*[��߄S��ܚ���{��Ԧeh��������#��RV���T���b���u��*���=�oX��^#�y�}�����=2ugvwl��T�7�3�
+=M����������([�2����ޤ�e��0դ�E+l�P1+��ј�h�Ĩ�cZ���6���)�Xn����;��_�~,;��$u��lU`�@8�`�r+C����H���%zIuI�T��l�1o�g�]j&��3�j�v�N2��ϔh��ԖT�Z�k�-��s�@[�Z@�Q넺=~��6�ͤ�ڴ6W[k��6��vVw�t���O�>��S������^�ρ��ق+���r(���66��i�{�F�����Ix^^|��y�I�e=��\S��r�$-�|U�4k�z�uN�h�k�Ǵj-VQcԼ���ht4�ؒh�hTͅ�����Rl91�&�\��a�cRͅj��V���ij�e9:Fޙ8'��,)�1��."�hK~����b�u�A��ّ�F�آ��1�Q�$�M[iʼn1�Q$z��("ۢ�8"�]����~C��|p�bT��*�&"��A9ҭ	��G
+|9�=7�ֱ���)Hn+��/���@��^��L��_p�Π��y���{�x�%��1ˌ�01���怨E�j f�%kpG�2�)�j6B��1�ki+��u��Hw���c��A���Hb�8O�T�HQ%����bx�V\+�*6@����Bzm�1�5�`�t=�7P���&/�^�	�3�MvܰK�L�'���ԣE�b��[,��� �3yN~ͤ�d7����H�]��d�BE����R���`��E��>yɏa�|
��i���R��2r~�����[Q ��-�3X-}=yZA�$c��VD}0yI���(
+��p���/�O~�xNb
9�uQ�S���qPxU�k��$��#IH��d�;CO����g*5�m>�����l�Y�kV�O���m��˚c�8�M۶D�]���hK���u�u�k��{S�!�S��>��r}�v �'�2��̭Y���=����q�t���]��EnQ��|)��z�>^�U�y�������������t3��{!����:�k�� <6C�� �@dz�f)5(6��dG�p��.�	�*ԇa
1x#����0LP�wð�FF���aCDx��ðIb�\LH�a�����c�&�7���D3:�+�1�o�0��(��0;�!Np��>��y�ö�pϻ��N��e�D��pO���y���4#�+������z܆N��+�A���Ћ��w#�D/V�뱊<Gv���g�0#��)\(B

+P
+�Ѝ�X�>���5[?�w"�n���paz��k�
�.��q��
+��]<�t��ǯ� z8E3�*��"�w \�%���݅<��)��(D
+Q�<a���5��9�]�FkyN_�ic�
��w!�9_�.^��q+\[(��cW��j�Q��!j\hG/�-T~��'Fu�m�a
+�"8X�>��S��S磆��ƅ�0��B��}'��r	���m��������j��q�ɟ�y��8�St��y�x�C;`^8����~^G�0��c�.��e��ڿPn虥]��7z[}pa���l�t����c`�Š.~g-���f�=��x��
+�K�U�W`����e��^�	�I��+�[��y�܊f�UQ��/��ۦ�"�╝!{'缏��s���ujeW�LV�
��;�߅;87]�xX����7�d�3:t����y���g�<�m�����c?�b�i.�GHR&/�����R�k�,��tF|/6H�R��(]'UI�R�L��5�w�ez�c}��v�$�&�uՓ��H���X�"��\܀�P�X�Z��
�H�^4��q-?�	���"���(F6� o!���K�G,�p=J��|A����(�~��\$"	%��6ġ
+)������
���Kh�?	�s�s���9y���]��Ч������'���OZ>��)�T���S�N��2Г���-'�_����O$�N$�:��xG�O�Ϧ�=�D�|6}~k6�᳂�ݵ}�v�ھs��} l�w����H�~&�>��A���۞vѣ_�Ӂ�/�旾��H�ܚM�ؚMߚ��lug���6Ŕ.?�h�q<yP�4��ژ�|�lx�c������I)��oݿe?�>A�G��xK��/��VtL�Ư��_�����DK?���⛯Pz�}��"�mٲE�p���x�����z��;_��B[!���@?��L|gO�A�9��x�p���$z�p�{8�r�G������C:}h�H���^^I�����7WPS������#��_F���Ȗ�ad(���R�4������}C
t�P:�w(��Ǐ��3K1$,�:?Dr��o�g��������c\ij*�q&�X�$K�-�o�Dr�ÞCm	94>.�Jd���j��b�QL�h�eT�$+D���"5����b jx�L�VZ��L��}���F��1RRa�(7��ba4��m�G�F4�8���x����E��QS���q����Q��F���h���y����E��#�w�
ax��G���0�d�7etU㍋G�S��1`K��6g tE�]�q�>+���;GϺkwz��J<���v�wuݣ�GFI���߹kw����,[�;�
z�^�w�2���w ��cE��9����s�������8�R
���:�
+endstream
+endobj
+
+96 0 obj
+[/ICCBased 97 0 R]
+endobj
+
+97 0 obj
+<<
+  /Length 258
+  /N 1
+  /Range [0 1]
+  /Filter /FlateDecode
+>>
+stream
+x�u��J�PFO��U�v�D�D@�`]\�
+F��Mk�I���R�B|��&إ�n�ऋ��(�H�r�Tų��9�`��ըa�&�ʺ���:�l
+�3Ŭ�*ުn��h)�&
C|>b��������纝�黓��A��v��Cƫ+	�y')�
+̵�8+�/<����a�Rkp��ش�A�mW
�Q�Ih��
�%��l��E5VW~��d���X��)�ϴ�:׺
��b���k#�&�_¤3�0��
�I�
+endstream
+endobj
+
+98 0 obj
+[105 0 R /XYZ 85.03937 300.80072 0]
+endobj
+
+99 0 obj
+[109 0 R /XYZ 85.03937 772.5197 0]
+endobj
+
+100 0 obj
+[109 0 R /XYZ 85.03937 207.14459 0]
+endobj
+
+101 0 obj
+[111 0 R /XYZ 85.03937 400.93646 0]
+endobj
+
+102 0 obj
+[113 0 R /XYZ 85.03937 485.2444 0]
+endobj
+
+103 0 obj
+[115 0 R /XYZ 85.03937 601.2955 0]
+endobj
+
+104 0 obj
+[115 0 R /XYZ 85.03937 147.02142 0]
+endobj
+
+105 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 96 0 R
+    >>
+    /Font <<
+      /f0 78 0 R
+      /f1 84 0 R
+      /f2 90 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 0
+  /Parent 1 0 R
+  /Contents 106 0 R
+>>
+endobj
+
+106 0 obj
+<<
+  /Length 3549
+  /Filter /FlateDecode
+>>
+stream
+x��]s#��]�b������rݫ���i��i�����vOi:9_�(���wV�\-@\i}m^��� � �~���C�������^��u�W����ׯV�\�JW��4��F�Mmm��V�&����j}���W����a��f����j�ѕ����0���������Z믵6ç>��Y�~��g������
��L�g����p��i5�k�~5p�H�@��Z����w�T7\��Y�u��ݫ�3���(ݺ���Mp����X��Me����ݔ�	r��
Ӕ��
�'�<9�A����[�`ξ�����流6Z�&6u�V�	�I�|�~�U�}���޿��J+_�kUW�V���+�2&v��_��dh�����÷���@�S�4!V�i|E�1F���@��4�u��y�~��-���J<w	����^�J��H]oU��PWAG�l�P��$�$[���~��Ͱ\8Bp8�;���d��H�@�6����LM�_OM�'z�]������ h�-h.`���Euim��vM�����N��I�G�����#�N��q&[��@�p�����y	6y�4�8cW˭GT�y<��YA���9C��~��*/��:�_��L�#�Z��yyѼ�Vgۓ����=A��9VrƴF�h��{��
+�*8�)?pTҍ�e�^`l(n�֪����ٔ���<G33�>����2n�M.
�R[	�`�0�~�L��
�>睲�[�\K�ϗ�KON�}�.U�b�̧�'i�
r��n2�7�Gi1����&�r���pN5��b�6$�U�aYϤM���ϯ��7z�"���4��u�����@�����Yq�N
Hy1Fv�LԨ��''$�}��E7��a؟6}`о���2��L F��iXXK��@�-��`F�hF�Q1���qCH��Zy��
C�����
H�ԨY@b�Ya�X{] },is�0�_�9�x@���=�&�e������I�w<��&>g)�ܼd�S极N
y�ԁ	_���R�cm��̗�Q[�M\��aQ��M��R�R��Q�T��)�
��0Ȏ��OW�9�P�[�L��l�Z32��)�ģ���A0$|��
�=x�3����*L2TB!K�ff��T���du������L��W�&�-��1J��;�@"@.�; �*�q�<
� ӳ�H	�}�b���;;�����k��ų&wh�vh�5�v�x��mJ��v�;�Tt�QYѨ��Q��'Q��K�ke[��E��?�d��
���qe����הSHG�FD�x0�K���S���k��$�D�fʞ1*Hq��Z!�$�8nR �?K����	7�f'��y:�<}�j�\<�<�S�������0L�S�*�!�������Y>{b�����€��+�\6wbD��rVG�� �J�#��zV]!?��n��w�#��֓��5���D�uvĈu��E�9�Gߞ���c����Q$H<���@��~�L�T�c/(�ᕏ�SL4{Nӷ�ڈ���4�1�.Q���2�7�	�]�C�9L�T�7�CM+I�4*�1FP�%
�4��S^(o�g �LO�cl2<�=K,�9��o�!��Έ9B���~�@Oz���1�M^�%�?�턴_�X]K��l[�Mb}�!L)�M�bIdQ�%�< Ǐ���43fLu�Pc�p���Ŧ���l氝��d�U<_�9h�㶙b��z>�p|?\}A&%�k$�sѫ64���_]�0(gZP���%
?�l�X�)���&Q?.0�e8^.�"X��"R�*hĭ�6�l�rx���`g$�Ҥz6�Ċ"��TRJ�V(�j�����Xn��^�J,Go����Jk��ʣ�2ʥ�3qF!�a����q��c���l0M���)���C:�kF!�&/�p�����^�[O���#���m�Flq]RҺ�$1��gg�U��\�V\��d ү y{.���bp:
E(�F��J�C1�vC
+
���ɱ�i+PF�(P��3]Q��*�����>SZ��Ĺ���K:I��@������jlwC��8�Ț!�+>؟	Ӳ.B�6�Q�P��4�S�]��O�����6t%���]�p`��ɜ-����_P��3�Z��g���>�sr�vQ�1XK���>�IJ���J��9�?�ȑ	�J|o8�#`��k��25m�y&Y,4L��(������r�SّH���G/����zG/���լ�[Wl��2X�.ue����b
+}�A	�Ԯ�.Ȍ�'���Q��#8��I
�hH��{��í�"�Ŏ\w�>��
a�~KՑlT���I4]i����6Bv�ar�
+f���4����2�s���L�[�H��uO-d6<�o�(
�f�q���xv=�3��b/���2��٭Nc��Ƨ����:��k���n��������r��E��f_ ˈ��!vdm�{�
��a����%��0�I�^��(����"be���zq��������U�߾y�V����
����ǞD�V��m�U֙���g3�릱�!˂�F`g��3yWef�
��G���k���n�G��$%�]���C&�,l��bj��
�skq�I˳$ր��E��/�i�Rr�r ��pIJ�5A9߶K�R�4v�8m��Ze����2���`yO�I��	�ȕO���1_��,!��͈�5�0��[�NTƀ�x�9�����d@Xz@74j�k��0�8����ށ�`�0c>ԋ�:����Vw�ӣ;���f�՗�UQs��#��r`I7�
yq(C�҂��C���e-�9I<5�W��$̩���R���),.KZ2��^v>���V�G!ѥBxL(�z�|�:���
+&4K�)c�0�q6��XԤt�,�_
�ws�Ɓiˢ
+B!`������X�5���]R�}�GBp����+_��-��=���P:N7��Ǚ�3V�Uݍ�8���w���n�F�!��oNe�BM<��$gQ*|>��F	uj`�'B�P>��;�
q��9kNe����F�u���;d��:��?�Y�Q�b���u�B��Op�I��S7��N|����~�Μ)#��´�¨�Z7J��`��}�?~����jֿ������?��<ٜ�|L���p�c��R��Lc�x�f�kx6&Ѳ�Ts`��CZ�۱�
+�0�h����M���u�MD����%
�\��g�U���r�A�5V=�r���5�㳅hp�R�E�
:�Y���Ӌ�T��Y{��Y������TB˗��M3IY;����%� �V�	�D->�>�U��6�g��h�R^aL�`�D���Gd�^Zk�%��y����&�1с��!a�%�$,\��dC��q�t3��ٝ���O ����඿9�����X^1Z+��g���%K\����!���n����o�W뗟v�O��}����??|��o?}�}�v_������o�{��}��v�r�ն�[nM���u��ucZ6���}<3,�j�݅
+endstream
+endobj
+
+107 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 96 0 R
+    >>
+    /Font <<
+      /f0 90 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 1
+  /Parent 1 0 R
+  /Contents 108 0 R
+>>
+endobj
+
+108 0 obj
+<<
+  /Length 4073
+  /Filter /FlateDecode
+>>
+stream
+x��ko#��~�*.�w	L��5.�	�@��$.�M�ܥ������]�<�ծ��_N�i8�g��"�������/7M�����6vs}ݼ~�f�klc�K�t��Ї��M�w.4�?mv����ec�_n?m^�llss����q���?�����i���וּ���N�����㧍/���Mm�9A�����3Ÿ����3���Gy��!�˧�S�:!�(�ӯ�
+��θ0E�	����O���zWX�S8��ux[��C��-��'����U���
���-���������/�w_���^:U/Co�oS:[/i8�d�eP��N�uٸ�l�h@00�[uܢ�2u?0,b�i6ey��le$�#
�@�zV�ԁ1�]#�X�Q�4Y�Ae�B�� dϨDi� �K���+y�9�Y�R�@�s�~zU?}0.�>/87���,A��ؿ��CH�Vo�U2	����p_�Bl���"�g1�P;l���7��������OOe��(8GAi�N*���0W��.%��}�a��&S�/(mP��f����G�&���h"��kx
+����y��N��i�'Ox�����d��b��d�nu�̉�.�t�Cԯ�[ �[e�q�+hډ#{��Ɏ
Y�ɖ�9WI��9dSEf]���Q�sOݨ)p�;���W�C�q���x�c��Yw<N��F���8�;�fy�m<�(�Y*8�Ƹ�>���W"�E��~�v��;Z�iA�v�XU[?@����܉����	�SS�ФjhN&E�W8b�
��`�w�����hAA��J�����E�"q@����)(�J����29��bwWt�gȍf9�#z �2���AH2NQ
���Y��!�V�LXA[USoڶk��w/�hbȩGa��%G6d�U��T�@�Q]�O
��������X�J"Y9L�uj����Na�S�V�YU�L�m�Z�рDq8/��q_L�5��imn�'���y���V70�t�5�1(��YF�4��
+�	JJh�|�A@�ģ���,��.�[�Ӆ�1I&a�R�D�N�V�+�S�g��c��-�j8�ids�'��LG���R�������d��bJ�G؅	�]i�{u�m6Ά��͗e����>�9�Q(ʣ3�
4XOT�����8`d����׹@�G0�$�@�}��YM�@Q�D�}������tى�i1Y.�
��q�!��b�W�4t�O��b�ڥ�zg|�K�T�7<֗��>=��}��+��j}o�Ұv��dK�dQl�.�C�^	Rr�z�q{m��*yu�3��q�(#�=�-H��0�a��P�N�d^-Y�g]�M�`���Z�M*)'Z����#��S-�Y���i�	}h�����\[^:a�4�r'�>�$��ȣ�}
d���i Z�s�a����	�?�E�' �i5a��#E|U�����漒{�Y�v��ԛ�������'z�q]�\���h�K~���*̄��^��d���W���v+��$)8���yFb��<�{u�OЖ�v�Ӊƽˁot��'���Jf��n8�\���)eêE�:)�Ԏ��i]^�U�{F����mk�dۚr.E�n>�N�bk��RO��F��07U�y�
wϫ6�J��'#t@d�/H�*�F�c�@?jԜf��k,�LNI�Β�����<<>@T����^ǟk����
�n�Z���ag���5�G�&.Q)�C�/j[Pc<ׯ'��x"�|�*n�'U��o5Rggf�T0W�?����G��3]�KJ��%]��
�d#K��M�c�S�x�ZA��w�a�F@s�����*�DY5_���[EP�����"?'mK�.��W�۵�I (W]��V]�o�7%���֯��9�A�6�,��b�K�0`�0G�����4�j�J�(��&󀾝�`VP����‰
�_YA)��%�:���bϩ]1'�:��=
+��I��I1v��v�R��;��돢��g���ߑO��5���Ⴛ��i�Ŏ2��[$�p��ʅT~��
+r�M�L�;m��Eh�%g�H��̰��"��LVQ��&�d;�ة�7E�������������w��[���������|������
+���%��zl�.(:V_˹t���:�W���J��J�2��rtOp�٫�I��LNij�k�j��%�UNTQ����~��\�����k��T3P�������LK����j3@�����9S-Ǚ;���?�q�B�j&��~a���f�_s�]1;Ch+j%��(%��u>��l�.OV.+0��פ
+�� �.Om�e�\�A�vr�oa��U��	D�Mj۸�Q`����#b0]�c����񌊛@�0�����d�g@���ʃ@5��X�;�
+\���2�N�u̹�G�]z\��H#rƪf!�[�M�] :g�>�����\4]�]�|����E�$���8:������R����K��� Z�c��S/��$��>e+�)�p/�xڬg�[�K�@����\(K3��%UU�B�L�â��
+��lJR9Z|kr
+�WV$4���_ɨa��`�\އ��"d�k�h��Ӣ ��G�,��+��%_M>'J&B�+R��ⰥY����mU;
+B�M�\^�k�I�Bk3��E��q��.�~$q���Ę�$M�ޣM���xF%B�l�-¾ZQI�Y�B<I0"�Y��R�5�[ ����.�y�Jޙ�}HPO�d�p��z��3�ζ��Nm��h�����N�,��#ci����CFES�T+
+v"|���ԗna�C]ʼ��U��V������D�
+����t]s~��W�B쌷]Zt��1[�v�<������ۚ�:o�y:���l�֥���hЍ��6a69}��Y�3��B���7�8C��B��Wl�`&)%mS�'����~A��5md4�<�L�s�����J!���:�;������S�3ɥ�d��:an�k�q�u���{��?��a7	�%㹕����p�p�JN��V7L��%p�+O��ŋ�����#�C/p��i2�<U�׺[T����[�ZE�}2m\d%
+�=�N��M�1w̅.�����MكV���2,NENɒ�	‡��t���Dr����Ȍa~W�����K�\��z°����Z��79{����F�.������2_��(���@P^d��T����!TRb�*�S�N��Э��=)�+_]�΍U��i��Ahg�B��R�o��]���_*�ۜ��ָԅRE�Z-�#�',�������ً��]��w_ȔK���9��,�q�ABΨu���@-5[v'P���7i�Z���3���3����+�4���}�څAϒI�v�ɴSW5�KGC���@���M�`HS�T�9��:%��C
�k\��ZG��N����7.�Eݖ���^4.��v]x�
+(D�ޭ:�Y|��1��1��щ`Q����yPB���!��0���Z,v3�WU�n�]k|���嬴Ѥ.7�@C��(E�-���c
+R%�5?���HɄ��i��0?���_.h�4uNA��N�@��h�dTF�@�
q�A���uP��j��[k��K��'|���>�{�	S����3�t��wW�5=�t�g9q�p`N2�9�?��ú��'� ��>���Β�ei�E��_�h��c
�i[,�����`�w�~��y�����\C�_N�� �������]��ŖA��P�oϻ�4WUrA�.ucjI��
+G�����.o���
+n��+R����Ky��/y�3�1�R�����*��
��!Tb�
��*�R�X�ǚ�r&�L�;4�+@�۟<��Ќ�~T"��|�ZbTw*��X�S�ZS/B��zK�>��L��L��?��^��2�h*��O��>L���[������tR�h��:PW�`�y�R�w�%�������ۻ��������?���v�����矾���o~���??lv�?��x���������o?~�p���Ol�Q��v��&�{�#��1)���?X�%����ߪ:���/���
+endstream
+endobj
+
+109 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 96 0 R
+    >>
+    /Font <<
+      /f0 78 0 R
+      /f1 90 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 2
+  /Parent 1 0 R
+  /Contents 110 0 R
+>>
+endobj
+
+110 0 obj
+<<
+  /Length 4259
+  /Filter /FlateDecode
+>>
+stream
+x��]o��]�b�jR;�(~.w��h��Eh��yp\_���[E
�_��v��I�X�n89�C��7�z���⋋�����������݋�^^���u��ݕ��dl�B�r�����ݏ��l�������ŋ����]\�l�|w�;�~�q��ŷO�Xk�X�?��gw��?�����
ǿ���x�����p�˿p��Y,D��\zB��p<̵��{��fHճ��?]�����ū�_^\c;��!�a
+��c�9����Z�uW~�4�!uGo邎��|�}��
݁=�ct��e/��G"DD���+'fP��|q� 	��`r�����!�#<��p{9�3�����H��o�d�f<��p =�i!�sB�[x��N�,@tE޼*on2��n��-�[84ށ�C�I�����>-��q��H�S>�?R.P�
+��RPX�d�����pT0wU���~�^��}�0�p�f6��Y
+y��YH�*�6�)ٴAH�O&O.��=��fa!��EY��
e��P��VԦ�jH�@���y�{�]RY�Q��R���EgM���<�M�OW��c��B��A,���۹��
+�="�*8��٣�C�C�T��IiԤtGc�8m1��$gO[�5C69�a:,���?󞤒�C�G��c�jX0(h\e�8'g3�؜r
+����@��DU��̄'!Q�&�?E�\�޾.�5���H�}�Y"@Qa��z�)�V_���JY�����b֟4&T޻B�t�
�^gF�}&�Gf5�2^��9�G���t$��1T����HZ0�v�9�)TC���,}�A`��V}�B��QAMnFn�M��M���w�?�������w_�~����_~�|�\���>���cO�LjVE<
&��7���x
�I&�dc�5�,�7�Q�'XL�=�3�A���|X��<���0b(�5�E�M=I%���a�^�
��k~����s�fP�����.<�qb�O��NQE�GU��5q�~�_��GXkѲ�9�NVh�RD�j	
1f��ݦl
���H�s�-��Y#z����s�'��>�y��8���h]_7qz����A���S�B�VDmRE�G�&�@�2�ږ�(��%k��C`T�(���}�%�f��
$�r���(�q �l���T��`��Vd
ᩐ�~�M�+�_�ݩ���t��bŔ��x\��ʞS˳�M�P��e��3U%�Vz-Λ8LS�
+z`?�D#I&@Ǝ,�j�uC��w�q�s���i��ӌ�	1j�"¤-�d�l�ygh�<p�M��ߋ�p&XM�@A*�?H�kҨ�������4EFFA;�ۼ�ҬCS���<�-�r�)q��'3�)�	u��/�jF}���ZΨ���B$���D3�"Ny�MX�f�u��B��B�f��
�R\�Y:m�G�� ޕj�&�j�3���4�-�N�4��<y������{��G�H�K���O޻��dN�'��K!�ȥ�ӄ&�������`�cr��Y��wP�l5M���Р��Z�",�؟`�R
�.�j�3Ek�������L��� ���"�޼%��=�F&�xl�\�f45����"�{G���U����F�n���x9I4h&͎՛��j�J�3[i9
g�R��&Mt���x{���D�n4C��)���;������2�Q,�)r�p(
+��K�iYn-+[�D:�Qc�Fd:`}�]��8��@0цG
��/ȕz\��Y����oMBUB�hRt����c�X����N��l�P�!A&�f wd��m*��ZX�/�DX�e$�; ԠKjbv[/�p(pXD�`tjj8W4�Mw����)VP�B��6�8y3�t~a�c�G���s��bOф��PT��Sr%�	Q�R���|8��2
+�������+�Ƌ��)�֎–�&v�=�F��$fǠ6!��p�ƶ5������YD�։���qC����p�r'
#�:���z<�*ݝ2U�zH�0!BɎ
��|�t(!�w����2Rp�L(M�(�;D\�3�B>�N�J�h��C���4K�څk&?�-m�3�Rλ�ɸ���*����o_K�&��NӄB"�W殕�Dv��)I�&P�h�ej�rc�� ��������mIgq7dv˰م8Vv60V�v�Tm̈1���:�+��}�T��)�
q)�O>ku�`�*��Jz�����-J�`Ҕ[|jۯi)a'[]Y����ķ~�p�
+�~Ib
��;pҟ瓐`�nMб�ҖOK�hh"M���%s�`�8)y~�R_�7AN��"�VM|զ��pc�vKSI[�X���RIm%+��5+B��-�JH����G�8C�W�T"�X8�#B{

+��
�g�V�]��Q�j��*��:��lA�����a$��|��˚��m#a�&N�(��"
+�r~2S�ơ�������X��BMZd�s����"��Yjf�tE ɣp��p��#<��)
+U��h[�U0�B�۰�_�B�a��L�Y>�֐�3�q�����c'���d#fr��
�
AIw4CF�
+��5ܝ�譻76`k�T��*�
�\o�\'1J��se����o�2_��H�mQ���z�F��Y��:�5�TD�`M�ٟ�7��*�<�ɧ	=�
�&v-�-���2CY���m�jE�j�
+�
+A��aQg��PJ(�R�2L�C�W�]bs��y�]^S��!�\ťn������nHՆ��s:ߎ~�P)�69D?�ǎP΄+ۡ�l�P/�K:O�>�Ow2v�a�vBx5s>Ѓ����C$��A´��q�Ы���X �c�J�<o�Re�k�Q��b�e����HՄO��~2��
O%>Rl���-u�֯��7\�x!ITx�"�pL�#����g�B�7iN�RDH�#�i��@�R��E�J'ChE>�l����К��n!^���BC���u@���_�G"�cp���^�fQ�~�&8{��q7z8�%��N�
�2��vX�g�5�r�'b7
+�J\RH��{��=�;�E��T�Y
+V?�L1�P�o�a����!^R���}�F��
x�J�=��[�k�A�m�<��W�]��]�c�ʛ��{׹w��(0[0�a�n���}�����zj�̷�PV
+�{e8w݇L
=|���@�>R���Fr��*ٴ��=�2�������O��Y����5i���{�ŭ���8�¥U�����c�_���o��!Ҵš�T�B��ʕRX|��/XWNM�<;�Ѩc��&0b�.���T�~4���6~OH�bnC�@Mp�.���m)1��ϓqcJ^QY�R}e����)hG0�X�8�^Y#)6מ��/L3&�Js,$�X������K���zu�yM!�ժ��>�MV�1��L�꣏-��`yk�}����ь>���X��Q6�,߼�O�	8�y�����8�1X7|��/{���=�nV_{��
�p�(���S��a7�₝'!�E�\z;��G�a�5ԏ�Z2uٚ��M<*E>o��U'Q�&��3%I!z�9Dunz�A�ogb�kL�\腠0!�^��hd�
����Q���Usʜ3&l����(|���iY2k�0���Q
G��r�����>��	��zP�N.E��96?��\m'��6���4|#i f�m�@r>JeҦ��DT���O�ag<�*T�i��+J�&���>W�5;��P(5`q-)7g�&WS-��0��S<�B�
+�t���>�i[~�#>"�$�oq�H��;xRh����H�� ~��y�J�Y�&�/��(W�cɫC�`��R����R�W<0"'?;ݣVL��&۰�/�3>�~Lⷀ��I¥��;���;W�Z��@��o��!�Ĕ���#���n�Ԟu�QV���עk�<���
+4N��K~����Ր"�D�s���S��f�o��.�5xB�«`��O��]���-/�G���7���_�����*#��a���a�NO���*��ױe</�����ǥx��m����M�{�u�VK��5�`�-_kY+pwG� s�f*�{�{��UgE@��%���F��0p��Jĩ��D��	����o瞥����b
+䩶��4��1R�;S��������w�s���������k�������������������~�������?����ۿ�����?���d|��K���W�"mR����@d{s[��:�?]�i�
+endstream
+endobj
+
+111 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 96 0 R
+    >>
+    /Font <<
+      /f0 90 0 R
+      /f1 84 0 R
+      /f2 78 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 3
+  /Parent 1 0 R
+  /Contents 112 0 R
+>>
+endobj
+
+112 0 obj
+<<
+  /Length 4098
+  /Filter /FlateDecode
+>>
+stream
+x��]s���h7�=���;���qlO�ifډ��<(����ȩ�̴��C�w��xG�ŲH`���]���w�^�5/^��f��W_�i��ի����V�Y��6�y�>r�.E��ޅ�����67��l��������6W��ͭm�k�n��w?�~Y}���������3�z?w-��~������i;���h���ֻ�o{0ka�0*Dnnq8���;�h]��q���S�X5�����̀o8L��y��.ҩl��o�M�8���<Y0�YY�~B;�{�Qs�����۫��Wo��j����T�
�Xߦ4[}א��i���vʺ�ϣ5�O��E��C|c�}ĭ�g���o
��.VX8Z���4@�F���<O�Hs��إ2�����M��0k�Dg
�W�Ȅ�6��|��
+��/���L^�4֫�q)�n���E������|L������ĉ0Ot
��0{@l`��M�mAe��&�gs�S.@�<&��k��|���20�Ud��k���:�9Xd����6��ũy�T�P�4
+��
sdJ!���Hn�A�A��#�5�".;�X܁�R`����YD���QS�oMl]��,��������������@+aṡh��;�.lA��v&v@�B��W�d-v�Gy��̰�xG��ƨ�u?�1/�=^Y���ՎbA)����5)[�@)�����q��RLH��B�� ��!�A8ԝ_+�#WTE��J�̎$P���W	Ů�zzn��fMS��I���S��l��l)����le[���q��`V��i�5"/q�'�3��	��`��VV�u1.�ԍ����K���t�z��h��>�p��L6�Ŧv_�Š��
+8ӸOr��0u�u�`B\L���
+�0t�9$
��{�T�`?QW��^�@��wza
+t*�i��o��==Iw��3�
uA	b��?�Gr�8�8����i��I��4z�PE�*3�MX�p�b"����$���W�[�
+k��ډ�{�ȗ'�ZVU�z�R�g�4l�=fC�����6���r��q�l����\���1�S�w_�]��|/RSh�Nx��,@�67YFa���7�=HP�<wm�+C��D>JU=�*�����/q*��L׵!J~�&�p��r����',vjfS��|��NCA����79%9��\uu�S�\�ynh��?(����F%K���!�Q���F�RI;բ{ꬉ�.��i{��������j9I�
��RI�wc-�������v�
Ǟ@	���@Tm!���OR,�D���̆Ȋ�Ї�Q�Bj�f��l�o�K����S�&�얗�QE!�'z��6�t�ٚl	���GD��dM��h O8�B]{�h�<�ZtR�:17��`�{p[A`a��V�Tr+(����S�\�_'p0���ql���yI5�~�zӶ]�ߐ�w���HDb��br��:o�w��C�t�K쇡)�Ӌ�+�5ѭ핼N!�F�"�~[�
�Z�0�4�.Q�1��.
:�M�����gkE�  ��@a��}��NQO�֊d{���z:�������_}��y����w�~������]<k�������ZqNޛ.�<�
+�ư1�|ʧU'@��7�u��|b�~+�&s@�!�
��HPP�Z\����Rx
+�¾ܶRҫv�߿^���<�p��ܐ�Im��Jr]Ir��v<	ޏg]��rC6���Y�m���>��(o�3L�Av�K\Q�Iɶ&�>s/ڱ`��r ��t�b���1��=��}�)�c��D��h��jҰ#�g]'L�Ʃ�)���I�[���V�	j�0
nM�Z��h_�9��=��'E0Z� ud���1�����HW�2U9�:mW�'c�#N-	�+��Q�RIZ�W�D1&4�_��a5�XU���{5�;k|p��0���I�$\���]̎�ɔƜ��V�풸�Zz�eJⴄ��A�w-Zލ����#��:*w�
+�Ȩ�=xrbA��>a�
+I��{�ՌtLф��%��Pz���x���#�G��E"��-a����K��_��'���H�����el��B�	p�q�P^l
p۾v�h�@�E
+�aO9�p�P��5��mr�*�RMG���vI:�
+�ԧ���e4R6�k��
�	������TZ��
+���'���"0�D�A����Y	����K�]�lV�Cà��0AZ������Ay}�3��O���7�k���-N|,�[/��^MI�ޛ�k��o�����.?0k��aI@�#p�f˚^��,9��_�8.����
���w쎍��PoƄ6�l�oٙJ=��S�;o��`�����3i1HM+h�v�K����U��22L�'<I��sP	��^{Á_��a�-6�[N�o�>P����cW�W��
+{��Z
+�7>��1��WCp7���I���m�>�0�k|�C@�?��8�8��.��#�`�%z(�e�����O�KՖ0]s"�ؘT�O,���(T�A�;�,ғk*��'��z�.�7�������
]�~.�,ڼF�2)8ko���٥��
��wu��"	�z
T�3�R��dj�$T�o�w9�3	{u������X%���R]��wj�.��Ĝ�%�5�m9���i���l���� ��{ϸ�O^*��Q�e�6���R
+�8A�)^�v
�~�n��|��t��A�!^<��E>�c�g$ߡ�Ű��(��Ŕ^S/IkM룝$������Թ�h]��`�V��	&u�B�т��a¡��f#�-�׸#]�����8E)=�m��QӑR.J���W�~z׼��ju���t�Ͽ|C\t�u�0��<w�8BtU��ز������<����Z&�6���-|��&��T���y�*�I��8���2*�rX�GL�M|�=���Gb�xwj�=��h��*t6/��U�d�U�t|�2Tu�i5�����⣏��)/x���b��SK��]'ř�S�u]�\����v��C]E�`R�),7�!(U�+|�v�1����X����
+m0�F��a�5	m���i�<��R!L�Y�T�O�:�7�_�wK�uئ[��M�g��@�M���Ŝ!��z�ON���!�z�p��5,�f��I�6,�dDm�=7������$S�
+yn���9x[KʩVؽkMpz�]�Ω��fS|�����_�Nf�E�k7�3��H�G=
+$��AAW��Р��`��%j2����0��\�~.0_ŝo����Վ�dI�l@[�EY%V4��&��.i�[2>\�Lo��;Sil-�)wB�u�>q.Я_U�s��l�aS/{
���?�ϱ��1g~��j-^��Q�
��bNI3�.�G��`�ty`�HkN�ci����`��t��]�:x�7���)Q03?rɥ[Ʈͦ��U�)�2C$�~��G7�IkT�8D"WvԸ6�J
+pi��v�d/��C�~8�/���6LX����me�$����Ɏ#���]��hp�]�{��=��H*�T���	����p����v����`���L�ܗ>��kju��7�{�"���"В���h��c-��<�Fy�'9�䪞����n��8s=F^�cx<�t�����#�����_��=� �r*�r�?Ӧ2Pr�$�����
�=�)� ����y&
*:�B"������ƫ��K�3�-K�N�O���h�#-�rg��b�r�`��z�?/*T�.�k
+���G5K�\kܮ��=��%R$�3.��`|L9U�-�P���(d���ޒ��m��XYŹ)������M���	���H�O����O��@���&D�xY���o-�f���h��J&BM�;k�o�[�?���
z�D�,r�q4L.��^C	�K�2^����Š���t��4���R��y~���!���]�e�j�t�!�(W�Y�1-�ն�m��|Q[���t��_�o����kl��7�����������_~�}���?n���f�������GW������﮷?�c)L�%ׅ&��"�M�ٺ�bw�H�����?!#�+
+endstream
+endobj
+
+113 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 96 0 R
+    >>
+    /Font <<
+      /f0 90 0 R
+      /f1 84 0 R
+      /f2 78 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 4
+  /Parent 1 0 R
+  /Contents 114 0 R
+>>
+endobj
+
+114 0 obj
+<<
+  /Length 4002
+  /Filter /FlateDecode
+>>
+stream
+x���n$���_Q�$�a�\�,A��
�c �u�}�(#�A�qdH�>襪뭬� i�E>��o_8�o�{lnn6M�����6v��U���Ϳ6���m^��$cCr�S4�+.4�?o�����uc�_�7�o7��}�ll�\s�p���u����ﭵ�?�Z�����\>~k�cl��y[c��17�`�
	�������~�ox��!���q�?~�cx/��2���קş��O��������wa�.nk=�ê�
+��B�nE���~@;�K�Oб��E?8��]��э6����Ȁk= �=���?m��n��y�͛�s�S9*t��6��u�E���A`��{M~�iu{B��Ĕ��q��P�ݏ�ݏ?�۠��4�(�@Y�y�=��I���G�)�@i�������8aoJ��I?GL�� i0�H�a�8f_k�=�{&��ߞ���S9Ы�q)vy16��2�'���밁`�������
�X�Ӈ[�������D��Pʏ�k�$"�+2F2��T��b�b W�4HH$ǃ����fՓr3/y��-�F"��{�ʇA�C��/!.7-�8o��&��N�$��O�Vs3�[-2����)D���Z9P��r< �Ģ�7��0�_i���j��s�˱TaF|��0�� Mq��}���*�5�k;g�se5㝅�Y�OD6�[�Ec�����"0[�:�#����u���\�߁->�O�����1HXt�u k�ɪ����w�?6��dP�7]�B;G��#����/�����:���S��;�Q<<���'G�O5�d�u�����P{cmO�j�[v%��SJ����S�<��,>f9h�����7��*[�@���-'pSQeL�&�Ҟ�:��G>�5m��6�4`�3���9B�N'��G�F!��k��F���c�{�It,�"�<)�猽"c�p�����r]�n�RO���j��9�~e��l�ک��)1��q��֬KѴ��Z>z��?��އo!)�a�'�'���H������
+�{h(j�/Z��$`}
O*k��=�COz�z{ ϊ�8�,�6��
[����մU��r�+��ɠ������P�3�h��/���1?dc^,��e�i�1BO��/��-$��$�09�#[�@4�YG����|n��ň�.`���V���ه��Ö��T3`����Э�M�6фևdsc�`�U�%K)n�It�w�'0����-J!�<�("�������|��
O J�R�:�?)t�����zQ\p�kR�?�1t�����R�	�e�qK���(��x��9�6Jg��@�X�66�	:F	�O����9�_Q3'�
og�}0>�.��4�ę�������2 �D��@�s63����'��%�r�.�)-�?��ѣ(��fde+(!7�5# f�=�-����R�F��B��̛��P����Lh��$>)*��+gQ ����@��\��6箔�
��UJ��'M�<���C�nܠ�͖� N���+'�?�D��塍� oj[%�V%�M
+�[�e�;����I.t����`D*��
+Vs�h�yk���TLȎ(�PG�E���nxD�<J�`��uso�{�]���I�p��E�ʫ�Ɂ0������ʭ����iE-��^�>���B�Xi?�
�a˷3A�&�7�áf]�����'{6BٷNʸ0��6��t�fv?2P9�F�Q�_�^e;.G�l1�Ģ���W�Ͽ����㿟����޿�����z���������-*{U�C�{��ԭ
+I�E�}�ϙ�t��{����a5���@ �

�f�ʶ'�ّ�̬%㏱8������j[�N��[Dڸfғ�	G 3��5TP����ɡd[���׿�r!�M,�͇9��-W����c�;[�i�)�8��
=�`]އ;��=o��)<>���\�f&#j`��*��F�����U��)|�!l��K
��V�!OF�Rn�Jc�pR
}�C�r��U�bN��h�u�`B÷l��M��� ����0��=qU���yZ�K褋A���
ǙT�ŶN�p���ƃ�I��+Ju��=1v8z�
+������]�"���5���hK��>|/n�`LD�O[�}8����T32�9�}:[9v�M��Sx\�=pA��l��Q�f��%��A��$H�U7��}UY�K9� h�j$��K�*���!O4����Mh(������F�#���J�:
+hl�bN��<����ɚ�%SRX��~1������&��1|3&�޻��j4~��V���4[B�mE��%�.O�B{��{��S���X�&H��Ӱ�������]'� '����r2���f�B�LW�]��A�u�䘻�L���$�%��:�n�V̭��d��L�m>��	�H��'>��-����L�!l�bI��bI�b$���15W�ˬ�U�����ڌ�h��qy~��I*��ʹbP��e��UBsR!>�����VO�wJ^�oY��9
+���_װ�p)}s.��!F'��O��٤��h�Ï$�S19�H���Ү�u%��+Ҕ�R[N)�7��±�2�8�la�ۘ�P�VO�ȼ�C'��I�ZaGi��6իP�"���@
�7�v��T�S{�D�#��Mh�_���d�'5��!�2!3�GV�A [��O~�2�m���z���<ۜȌdҗ���,��
+�~eyL�����=���*��U��.*{�3�,hp��]jW���j�6�yG�%���Q��i���t]4)�Py��.
+� G���,��nR�U�څ'�r�
cV����X�I���]n��Jz4�.�(�S_��Q-4�qE.�����3���p��2�����&6mi�v#�59��\��O�0H���{A�k˜Kr��cE!��_� ^�o@i�&���"|�Cx��wϸ�?j�շ�t��xDSO���g��D��j�;l�5xJ��H��Q�c.xP7+ؒƒ��W'�N�*ed�x3n�����]���2@������|�Rs�>cS��c�Ӱ0���w8%{�$�&0I4ı($yu4�ͫ�*��p��>V���Q�!�����{I=_9��uΥ0�LW���ba���A�/jF�o\ɫ�����bSǫl�9Z�r��B`��y�9��/�������@�����^ڥ��a�*���H.��V,r�Q��H*�������r޵&��/��[�����-m��H���5'5m�v���7���6f��oaG:k�Dͨ�tώǭ�ǽ�gc��޷��NL(
+�,�{��3"�X3�Nh�<2�R�ި����U���Zcl����)1�(gi��˔s
+���+�8���$K�Xu��d�G$✈�q�ĵ26��1��B0dN��5���q����
�8w�֬���M����E
�D�r�kZL���P��zaw�\±�1dS}�^�Һ6��ܪ� ���5]	._� ���-��1w!J��^���L��)Ԃ�(�>hh5� �o^�kV 
+��0%Ȋ�5��9��F#�N��L���Q>����1j���e]̦䮬���(���S9��g�>�uD�+�ؙ���Ͼ�C����S��+�6�b1t����
Oyp��6
98.�@ȽHe(j���Q-Cp�k˚*�3ee��޵cǮ�0C�`5���r���!t�l�M2��
P�f�hcP��x
�j;������mI�^<������KԨ�Z��\kܮ�pMZ��d\c���%D��=�'�m�f_19��
�� ���Iޣ�:5ϐ
+�9��L����������]��������|�}������?��������c�}����ǧ�W������ǻ�>=��"����I����L��S��um ҉� ��r��
^�u-
+endstream
+endobj
+
+115 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 96 0 R
+    >>
+    /Font <<
+      /f0 90 0 R
+      /f1 78 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 5
+  /Parent 1 0 R
+  /Contents 116 0 R
+>>
+endobj
+
+116 0 obj
+<<
+  /Length 3853
+  /Filter /FlateDecode
+>>
+stream
+x��]ݏ��_���i]��w�ຈ��
�"�����Z��*
+����N��Ώ�!wW�K�ș�pf8_�v?���c��ɦiv�?��Fo�>m��x����4���c�tAi׻�i�W]���~��nus��F7��~�<����f����Ƙ�������͇͏���~��;}��=��~}w��=}��[s�����3��`�6�}��סo�x���|{F�H�H�(����έ0b�����ŏD�8#���o�,j/�v�U��l�i��݅��k�0�J9>n�[�0����
+s؍Q�����,�+~3�L���O��_6/o6߼���f��gD�s��6��B�N����*��~!�I�'j����M����@
+~[Yb9��B��X��
+6l ��F�~^Y*����iDNtv��������9[ܦ�����2Zt�\�;YK����=���?7�~;4O^�|�J��^< ��~�67���0O�
�u�߷�
J�%��&�*���K5)`�Bܥh�Y��?A�*���T
�9:��j0>mzDm%JG�Y��pQ�����L������!8�baMh�J�`6�ؘ�܋
+�[e;��[K'�d���he�n�>D�O'�E�ó��ew�
��s�lx�l%���}:��=�������1Y
u�	{�`e��)�8*�����D�Y�Z��\���;T'��A������p?�Ur�NY��F��\��O�xJ��O$�z�l@ÇH<��p	ZA�4�C޲˅O��8�KR"1����Y
3��x�qEUn�1��k�镦I��)��1&
���<p>k+2��>�U���6��6�Z���ی�9�o�uֶyS�#.�5%��cv5O�V"�l��4$c�!�'Hpn5��b���x
p��
������%�R9!��x0s���ϱ<d&H'�)
�0CrQ��I&���
+މ��:��V�ɉ^(
��B[7_�2S�Y�u��4�g�7/�H��f���P֙�@&a��dHSz��?�v��WĨ���
�ة��vǥt����ZU�EU�V�m���a���ў�p�rS�7d�9Dӽ���?��칿�E'��߱����_C��͚>uBq==#֧CgU��^�1�[.�M6���<���[*n�#�]�<c�24����D,
+�h�V���5���_��|���qn�1FYcC��������0׹5�#TR¸Sj�Jp,xB��A8zݼv��]�|��
[���Hul(0����䊯���x����@tR�`ݖ
�";���߇qRI����.����h#�j�%�<����5ա� ������@�;�����q�U��$!��k3!š��n.��螎�F�iB�a���H��.��h̤�2".�1�1`�3���Vi�X�Ϊ�F��Of���g��:�R�Yb5��K�2;�`�T8?���``��'���>5��3�ş��e1a&���pYB,����|J61��"�hI~pX;:)�~y��脤J��搴�0�rp��<��B��l o,�,�,��/��r|*�}�vN/���S����ؾ
��(	gdc�4�[.9�3��D2��QP�"TaIi�G@�r���s%�2��K��E4���k���\E/A�B�.��4HoaZ�E�Du}���K̾��D��<Q篦���Z%�o�^�#��A�1��$�]���X��l�
+_�J��t��$'=��K�`'�J�0����Y�$`.���
+�W����-����m]飲�u��L�/��6�"a�P&k������Uɼ$��9��R�|�ެ21�sl���+a��E[��Ru����wʛ�-��8'/h��lt��c�E��I�c�&2��"��X�-ù�R?+��G�
��MzRfe����-�t����B���/�2CK�@�����W�X��Ϊ�c���"Z���������d��ӑB��e8��ɮ��l�LǙJ�|;�>s�eh>>�ë�N`�����Xd���h��\I`�bv��h������(�]�'��*����������k��S�
+L��.�z��W%еj��,P@�
+�&f��"%�ň̖
���E���,H�K�����)\���g�05�5��嗿�P��b���Zuڭ�.��/5��J�7���mUe�ZlJ�Y+�ǎ.i;�˯����S�X�ɤB��)LQي5+v���W}g��7eĖ%�U���sHW��Aa}��l��:���:1�أ
+�QԾK��Ye�]Ο@�m	%cD�iE�a�����R��EI�A��5�Qp���$>�\wIU���2F�U흗{Մ0�핎�,�A�A����7��"^	D�o�9*���
+���32��B�
&��|�P�?9w�i�I���S_���^8V�;_��b/�󝲮[q�����l�������*���?e��y�hd`ӕ�N!t�q�B0o�Qk���k�U�O��'��lh�H�6���"Ê�|�w"���E��ŲcQǂ􈓨c_�'+�8g���]�St�����
LT�u���n�C���%�aļs^�	�>�mf2�R �$#�E[�%h4��G<x�����=vİ�f�r>����=%"�8��C\�3zͨ��n����*c��P|�#�m�����F�3�~�=�$Rp�u#2�Bj�-';�$Yت{.<IaU�脪,�
���C>0O?��<uR�%���V��Um��6�c/n�[u�&јVB�������E(�\~�����#�����a��q5y�l��Z)�ٝ[s���g�E�%�����
vU���j�蔳-y�E�a�� B�u���?w59�͹��~�[Z"6�,M}ɚ�M%V]���2���t̉�w;�uzEg��ƥ*�!�������E��X�rby��N�л~�T\Ux�2�7V(��JG�ek兄�6Q�$�<��T��C���z��Uh�KV�w1f��ع�R��̃@,uI+6U��Ȓ򆐴�-�3ۀ�X9��*ӵˣ�
4�)���-�6�_�H�P�i�����	�j׍�-�h���2�)�Y����!1�b��D�8����H�	�)��X�&*g���[���ki�˼`��
+�Nuw�s�ro��
�	\I6
+4��L���I�BtB4'�t�لY�,�U��	�mE6I�'�^��%K�e���eƯ��bm�j���qy��e�T�F�����-6, Z`%.�6;��:�P
2h��b�p�BzF�!�e�[�O��A�I�*R)��5JdH]��!tz
��X`3�9~����bYYTϯ��@bY�t^���U��x����w�[e�Ï;�����=�p�S]���7}���&�J]�eŋ*�5�=�@,K�ڦ��y��O�L�\�
+b���N����
C�l�/��Iې�O�
+Ю�$���)�~�<������i=p�-kܚ���b��أXD���Joَ��2��R5��=7�%m�G#�@�8�[��'o��T�he�������d��Ǣ����Z���;`����B����|վXI���4��a���Jߋ2
+Y�C�W��3�x��kA�bh��<�
�=3�%z^P�dA���.FJ?R'��z��uݪ7��0?�Pð�q�U�����/�M3��c�b�W�N��e���fS���J�A��38���|�,���^���\w`���)BIt�d]����mN���߽�=������_�����{��p����_����?�kv�>}:���ts������o�?}d#͠���o�S�7-u*U�6}t�:��H���J���?�d�
+endstream
+endobj
+
+117 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 96 0 R
+    >>
+    /Font <<
+      /f0 90 0 R
+      /f1 84 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 6
+  /Parent 1 0 R
+  /Contents 118 0 R
+>>
+endobj
+
+118 0 obj
+<<
+  /Length 4263
+  /Filter /FlateDecode
+>>
+stream
+x���r��}�b6J�T���0
+�tX��J%%�Y~��qʢ�5]I�>�ܝ��	��Z΋(r�F�/t7���?\�6Ϟ��f���_5vuqѼx�r���klc���钱���)���\h�?�6׶��qe��oW/.W��ܮ67�q���y��q�q��וּ�����=��?C���n
+�>�Ϛ��51���{L
D��ǹ(�w[8�:�>����棂�e}u�n�3D2�kw �1�Mҹx�:n�8�~)#�k|5�G��tzn�++�	�7m]�G/Je�0�ξi.����r��՗o^�6XM���v&t���t�����OW�_~��]�����_[���ŀ����Kxr�\��B�T�Z���۔N`Q�"0
��w�x���B������X�䀗jAְ�|�q;��	4s�7�:��Xn�GX� �{J6*�N1�XL�n���y@�����o�L�a�0m
�� '�(
Sf٪����u���3$d��t,�{<�Y�=AUh�K�ϳz�J�D�s(.��b��e;^�H��#PU-��/��lD��\@����M7��)�~
+� 0򏙤�?���4)3Mv�
�G�Q��G
���A�x���F	w4~T�rTu�f�?�*w�kS���-�@������
A�y�ύ"B�ňgXfq"�e��q�F�"Gm�^�j5Q�%�i�U�iͺ�m�*縐��iE=A��G�/(t��흉�u�8�Kn<k�3.�>8(-N����$.��q-b��D��ө��'*��0��
�eM=355\���A�L��
+�`!F���t�,��M��Ҿ܌,|t��D(_A��
+l�&k�E��?=mU=�ɤh�|=U6�3m�w��	�q�N�ϥ���nC�3i3��	��p-��`�A]�d��E*g��7HI�ێZy��1ڄ��pͧG�N[6s*�8O0�l�����yR�ݬ�n�M�v���y�7ܔЩ��b�KRrB�6
�rklH�&8��G�<	)2�3����;�оg�2�$�$T�$�r��R�h��AO�����l�J2V�~q�-��.b'����:~4@V��>#�����zo��b{�h�V��L3?�36[��֯l�FW�����\]��xZ�k����O�ɉ����jl��ŒV�rm|\-$�M�����V�H�?�X�Լ��%,�ApW���Z�nm6Ά~���v�{�I|�U$oDx�5I��si2����Jv4,8DfG��gy��CR�͟4&L��~�r���a�m�E�l�l��r~�8�U�%h2�f�dl�q];L�E��J�4�O�3>��
.P���q�=������)�zO ج(��9�=F����A9�hy�[9�C��s�����x,�/t�p��˴�����|�n��~7A:�=٣�y����N�I9����u恔�|AFr|u&�5�3��-�L��4�@1�>V\K��T��8��g�jX0L^�)���C�b�]6��X� 4�;��,��p@-S����Xt�����'�_��j�GJ��]�$�F��[cm�-�q�fZ�A#�8�]ӧ����P
+�������!��"W�
4Q�O��o,0=��+��j�
��uY(���)�G�a+�G}i��ؘ��q�%��u�L��K!H
�%=>�A�cY��g7��0~t�nz�
�Kx�7���O��A1{��Eɼ�ٍW��LU%��
+NmSHޛܧ~~�3.	�g߶31��,�2
+��Ay���FM��i9��G�RX{Sl���o�!mM(�I;��V�����*/�6!$ۚ��0�[hv�g�&�`�LF��r1TX�2xz1ݎ�i�Aל�{Ĵb���]���,lfȽ!�EW
+H�`5����+v�qhLi�����R�
���3O�M��\T��]�Q��_�qjs@읱�Ϸ	��1Tw�q�o[r!~R#XpD��D*��������F�8�����5m;L���cKe�=+	j���$���m��UV\�S���rHsJ[��%�v!Ĝ����V�%'�}u7e�N��*���NN�k]�,�V\�%}6B��a�
€{�4�a?Q�Xǫ��n��@
�,�CW͢�%��V4����xպiL�	�o��P�yoR����'{�+%����=�`�|�%��)���=��t��16w�W��R>������\ ���;�p��p4��ㄘr�`�o���(�#][�Ò9R+�1C��1���VO�+������N|1m�@�5he�EFo�1ʛ
+���
+����>R�%���[%��362��	���0T��jY6�lR��EM(*ӗl<TY,�eL�.�3DEՅġl�.��X-n�nЌF��S�$#��.c�:���
)� �UF�
+����E�n�6���e���1v�E�7o��ξ�0��Y�k��'%��!����L���t*��פ�誝J)>����P�lCah�4����m
�"��f�|AYf���E8/�T�c\IO�n��rXpD����“5��|���0&�2�<�p��EC^�C lՊm�G#����?��:�?h��M6,FE���|)����F#�]�.�%���Kj�uCN�Z�$ �;��o[xE�QH�hv`%>g��ʄ#�*��ϥQ�������k��9�\o�TO"�]�E_~��X_	��芰�%�	�����e5S/���좇R�.��U�m����}H=ۧ����L��A�?�WE;���ά���$��S� e�w�B
�o1n�y�
�/�L#��}..���j�C���ԥ�˫�̑��ۜ�#!kBIF!Њ��^L$"���K����\E��!-p���$v�tc}��!�U�z��^�'H����3��P{�FG����#�M����\C�sn�@ ��&�*��q*7$1��25���ΐ�Iy��{R����!v�r����B���avi��ڥ�|�G�s&���	x����xk뮺cX�L#��߸�yr�G#S�E$��J 1��إ�k�LL~�!���2�15l2��C�2���cq4�Be��?�[K��)�Q%�Tn|�L��9�c�!����B����p�2"����)��1�aKUA��9����8���ؾ�S� eO��ב#V�tL�DL&c�y�3���ge�}�L�r�%��^��|<M\�	@��>t��(��&��.C\��9Jڪ�J�ԙ>�v���~�B�I.8_�8���`�W�,���62���bi���/����@'wu�4+�9�$�/�r�g�r�@��8KK�tFZ~;~�j7���j�������EI<6Մ���;e@)���خ
���\�GJ�( 1H��B2�WУ7
+
+����V̧ ���FF{*���e$�(K��Q�Pշ�q�� f�
+�T;C���ۼ�]S�6��楃��x�Y�k�.�@ .�����}|P�,0��se�(k�D>�=�
+��(%��4�,B[Q��d�_I��%P�|�������v�K���ΙR�$W���=�K]0��u7
+
+6�]:I�z#Nx"��W�4B`�JΔm��� ����D�q��,w� 
�|���f��TQ��p}2��K�?
+���&�i�.��j��Ѵ��%�zڲT�4Ē��@=b�����a?�ߖ"���O<�g]/��?C��bSuɠ0��!
+9�@@��������ABCob�5T$��ە5������/���&6�^���j7��*�d\�v�������f���~uu�m�Ň�3UM\0���zbvߢ�u>"�qMoR���턉��C;�z�.���wҫ�wx����VFz���ഊ�,��]��wl:�cV����LXk  �A�
+M�@Ŷ�?E蠒pYv�R�U�2���V�$��EBY��O��2�s�=��AЛ��rg�k9��_ߛ��w'�v5�E6(�4�e��������B4�wh7�,�h�!�wN���Z#���v�&X�܂�x�7��C���\���?��b=QLGZEk���po�cP^n�J[��+Q-;w�ƀԇt�
+���6���y9�JQ�۵��6����¬_�M<8�[���+'��jH$��D��i�Ì����_�W�!��c9Vf
�
G:O�k�|��Q����i�����|�{���ŧ��O�����O����Çf��ӧ��ݟ.���շ��^�}��}W'��M�E�.G�)3)��ힰ!L�y�8�@��2V�
+endstream
+endobj
+
+119 0 obj
+<<
+  /Title (A Two-Week Sprint for Knowledge)
+  /Author (Jessica Talisman, MLS)
+  /Creator (Typst 0.14.2)
+  /ModDate (D:20260425000117+01'00)
+  /CreationDate (D:20260425000117+01'00)
+>>
+endobj
+
+120 0 obj
+<<
+  /Length 1187
+  /Type /Metadata
+  /Subtype /XML
+>>
+stream
+<?xpacket begin="" id="W5M0MpCehiHzreSzNTczkc9d"?><x:xmpmeta xmlns:x="adobe:ns:meta/" x:xmptk="xmp-writer"><rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"><rdf:Description rdf:about="" xmlns:dc="http://purl.org/dc/elements/1.1/"  xmlns:xmp="http://ns.adobe.com/xap/1.0/"  xmlns:xmpMM="http://ns.adobe.com/xap/1.0/mm/"  xmlns:xmpTPg="http://ns.adobe.com/xap/1.0/t/pg/"  xmlns:pdf="http://ns.adobe.com/pdf/1.3/" ><dc:title><rdf:Alt><rdf:li xml:lang="x-default">A Two-Week Sprint for Knowledge</rdf:li></rdf:Alt></dc:title><dc:creator><rdf:Seq><rdf:li>Jessica Talisman, MLS</rdf:li></rdf:Seq></dc:creator><xmp:CreatorTool>Typst 0.14.2</xmp:CreatorTool><dc:language><rdf:Bag><rdf:li>en</rdf:li></rdf:Bag></dc:language><xmp:ModifyDate>2026-04-25T00:01:17+01:00</xmp:ModifyDate><xmp:CreateDate>2026-04-25T00:01:17+01:00</xmp:CreateDate><xmpTPg:NPages>7</xmpTPg:NPages><dc:format>application/pdf</dc:format><xmpMM:InstanceID>W5dErWG9DV9Lkhue9k6cew==</xmpMM:InstanceID><xmpMM:DocumentID>a9/o7WmJ3Q+3z6a7yg4t8g==</xmpMM:DocumentID><xmpMM:RenditionClass>proof</xmpMM:RenditionClass><pdf:PDFVersion>1.7</pdf:PDFVersion></rdf:Description></rdf:RDF></x:xmpmeta><?xpacket end="r"?>
+endstream
+endobj
+
+121 0 obj
+<<
+  /Type /Catalog
+  /Pages 1 0 R
+  /Metadata 120 0 R
+  /Lang (en)
+  /StructTreeRoot 10 0 R
+  /MarkInfo <<
+    /Marked true
+    /Suspects false
+  >>
+  /ViewerPreferences <<
+    /Direction /L2R
+  >>
+  /Outlines 2 0 R
+>>
+endobj
+
+xref
+0 122
+0000000000 65535 f
+0000000016 00000 n
+0000000130 00000 n
+0000000210 00000 n
+0000000322 00000 n
+0000000447 00000 n
+0000000577 00000 n
+0000000707 00000 n
+0000000824 00000 n
+0000000953 00000 n
+0000001057 00000 n
+0000001340 00000 n
+0000001625 00000 n
+0000001931 00000 n
+0000002244 00000 n
+0000002592 00000 n
+0000002919 00000 n
+0000003225 00000 n
+0000003559 00000 n
+0000003986 00000 n
+0000004096 00000 n
+0000004184 00000 n
+0000004272 00000 n
+0000004359 00000 n
+0000004461 00000 n
+0000004569 00000 n
+0000004680 00000 n
+0000004783 00000 n
+0000005281 00000 n
+0000005401 00000 n
+0000005506 00000 n
+0000005617 00000 n
+0000005722 00000 n
+0000005818 00000 n
+0000005949 00000 n
+0000006039 00000 n
+0000006137 00000 n
+0000006233 00000 n
+0000006341 00000 n
+0000006449 00000 n
+0000006539 00000 n
+0000006658 00000 n
+0000006757 00000 n
+0000006878 00000 n
+0000006966 00000 n
+0000007053 00000 n
+0000007423 00000 n
+0000007511 00000 n
+0000007610 00000 n
+0000007715 00000 n
+0000007847 00000 n
+0000007949 00000 n
+0000008070 00000 n
+0000008158 00000 n
+0000008250 00000 n
+0000008508 00000 n
+0000008640 00000 n
+0000008742 00000 n
+0000008847 00000 n
+0000008958 00000 n
+0000009066 00000 n
+0000009193 00000 n
+0000009286 00000 n
+0000009400 00000 n
+0000009517 00000 n
+0000009624 00000 n
+0000009726 00000 n
+0000009838 00000 n
+0000009926 00000 n
+0000010031 00000 n
+0000010160 00000 n
+0000010265 00000 n
+0000010381 00000 n
+0000010468 00000 n
+0000010554 00000 n
+0000010631 00000 n
+0000010722 00000 n
+0000010808 00000 n
+0000010894 00000 n
+0000011054 00000 n
+0000011914 00000 n
+0000012005 00000 n
+0000012289 00000 n
+0000013536 00000 n
+0000025435 00000 n
+0000025597 00000 n
+0000026475 00000 n
+0000026566 00000 n
+0000026854 00000 n
+0000028115 00000 n
+0000041562 00000 n
+0000041717 00000 n
+0000043125 00000 n
+0000043215 00000 n
+0000043492 00000 n
+0000045243 00000 n
+0000067192 00000 n
+0000067228 00000 n
+0000067586 00000 n
+0000067639 00000 n
+0000067691 00000 n
+0000067745 00000 n
+0000067799 00000 n
+0000067852 00000 n
+0000067905 00000 n
+0000067959 00000 n
+0000068265 00000 n
+0000071894 00000 n
+0000072166 00000 n
+0000076319 00000 n
+0000076608 00000 n
+0000080947 00000 n
+0000081253 00000 n
+0000085431 00000 n
+0000085737 00000 n
+0000089819 00000 n
+0000090108 00000 n
+0000094041 00000 n
+0000094330 00000 n
+0000098673 00000 n
+0000098877 00000 n
+0000100155 00000 n
+trailer
+<<
+  /Size 122
+  /Root 121 0 R
+  /Info 119 0 R
+  /ID [(a9/o7WmJ3Q+3z6a7yg4t8g==) (W5dErWG9DV9Lkhue9k6cew==)]
+>>
+startxref
+100393
+%%EOF
\ No newline at end of file
diff --git a/assets/presentation/docs/talisman-article/en.typ b/assets/presentation/docs/talisman-article/en.typ
new file mode 100644
index 0000000..9dccd1b
--- /dev/null
+++ b/assets/presentation/docs/talisman-article/en.typ
@@ -0,0 +1,160 @@
+#set document(
+  title: "A Two-Week Sprint for Knowledge",
+  author: "Jessica Talisman, MLS",
+)
+
+#set page(
+  paper: "a4",
+  margin: (top: 2.8cm, bottom: 3cm, left: 3cm, right: 3cm),
+  footer: context [
+    #set text(size: 9pt, fill: luma(140))
+    #h(1fr)
+    #counter(page).display("1")
+  ],
+)
+
+#set text(
+  font: ("Georgia", "Linux Libertine"),
+  size: 11pt,
+  lang: "en",
+  hyphenate: true,
+)
+
+#set par(
+  justify: true,
+  leading: 0.75em,
+  spacing: 1.2em,
+)
+
+#set heading(numbering: none)
+
+#show heading.where(level: 1): it => [
+  #v(0.6em)
+  #block(
+    text(size: 13.5pt, weight: "semibold", it.body)
+  )
+  #v(0.3em)
+]
+
+#show heading.where(level: 2): it => [
+  #v(1.2em)
+  #block(
+    text(size: 12pt, weight: "semibold", it.body)
+  )
+  #v(0.2em)
+]
+
+// ── Title block ──────────────────────────────────────────────────────────────
+
+#align(center)[
+  #v(1.6em)
+  #text(size: 22pt, weight: "bold")[A Two-Week Sprint for Knowledge]
+  #v(0.5em)
+  #text(size: 14pt, style: "italic", fill: luma(80))[From Agile to Design Thinking]
+  #v(1em)
+  #line(length: 60%, stroke: 0.5pt + luma(180))
+  #v(0.8em)
+  #text(size: 11pt)[*Jessica Talisman, MLS*]
+  #v(0.2em)
+  #text(size: 10pt, fill: luma(120))[Apr 21, 2026 · Intentional Arrangement]
+  #v(2em)
+]
+
+// ── Body ─────────────────────────────────────────────────────────────────────
+
+In 1986, Hirotaka Takeuchi and Ikujiro Nonaka published "The New New Product Development Game" in the _Harvard Business Review_, describing how Honda, Canon, and Fuji-Xerox organized teams to build copiers and automobiles. Jeff Sutherland read that article, combined it with lean manufacturing principles and object oriented programming, and in 1993 adapted it for software development at Easel Corporation. Ken Schwaber formalized the method at the OOPSLA conference in 1995. By 2001, seventeen software developers gathered at a ski lodge in Snowbird, Utah, and signed the Agile Manifesto, declaring that they were "uncovering better ways of developing software." Not knowledge or meaning. A manufacturing framework for software.
+
+The Agile Manifesto's own language is explicit about its scope, yet the industry has applied Agile methodology to virtually everything in a technological environment --- including knowledge work, semantic systems and now artificial intelligence. And within the same breath, organizations and experts regale the industry with talk of innovation and the emergence of a new paradigm in ways of working. Yet those ways of working are locked inside of Agile methodologies and Scrum, designed for the factory floor and the codebase --- not for the messy, recursive, deeply human work of making meaning out of complexity.
+
+== Scrum Is a Manufacturing Schedule
+
+Let's return to the Takeuchi and Nonaka paper, where the title --- "The New New Product Development Game" --- describes how product teams at Japanese firms developed physical goods such as the FX-3500 copier and the Honda City car. They draw on the rugby metaphor because the team "tries to go the distance as a unit, passing the ball back and forth." The unit of delivery is a working prototype of a material artifact. The feedback loop is a machine that works or fails.
+
+Sutherland's 2014 book, _Scrum: The Art of Doing Twice the Work in Half the Time_, reinforces this with military and manufacturing case studies such as the F-86 fighter production and Toyota. Sutherland's book rebuilds the argument with the same assumption --- that the thing being built has discrete, inspectable, testable parts that, when finished, are finished.
+
+The 2020 Scrum Guide defines the Increment as "a concrete stepping stone toward the Product Goal," one that is "additive to all prior Increments and thoroughly verified, ensuring that all Increments work together." These words are sensible if you are building a mobile application, with a login screen and a settings page. But for knowledge systems, these words are destructive, because the components of a semantic architecture are not additive in the way Scrum dictates. A controlled vocabulary is not a subcomponent of a taxonomy the way a login screen is a subcomponent of an app. A taxonomy without disciplined, vocabulary control as its underpinnings does not function well, if at all. You cannot ship half a taxonomy and call it an Increment.
+
+You see, Scrum assumes tame problems that can be time-boxed, estimated with planning poker, decomposed into user stories, tracked on a velocity chart and declared done against a checklist. If only knowledge were this tame. Knowledge organization is anything but tame. More than fifty years old, in 1973 Horst Rittel and Melvin Webber warned that applying the engineering paradigm uniformly to all problems was "bound to fail." And yet here we are, applying a manufacturing schedule and methodologies to knowledge, expecting that knowledge will fit into the confines of old ways of working --- an artifact of the Third Industrial Revolution imposed on the demands of the Fourth.
+
+In 2016, Klaus Schwab wrote about this historical pattern. The First Industrial Revolution, beginning around 1760, gave us steam and mechanical production. The Second, in the late nineteenth century, gave us electricity and the assembly line --- the world that produced Toyota, lean manufacturing, and the production logic that Scrum still carries as its core ethos. The Third, beginning in the 1960s, gave us semiconductors, mainframe computing, and eventually the internet --- the world in which Agile was born, by software developers, for software development. The Fourth Industrial Revolution, Schwab argued, is fundamentally different, in kind, as it is defined by the fusion of artificial intelligence, cyber-physical systems and knowledge technologies that blur the boundaries between the physical, digital and biological worlds.
+
+The problems of this era are not necessarily engineering problems with stable requirements. We are wrestling with problems such as how to ground machine intelligence in meaning, how to organize evolving knowledge and how to build systems that can explain themselves. These are wicked problems that demand new epistemologies and frameworks for problem solving. Agile may have been the right methodology for the Third Industrial Revolution but we are no longer in the Third Industrial Revolution. The frameworks have not caught up, and the cost of that lag is exacerbated every time an organization forces Third Industrial Revolution principles applied to Fourth Industrial Revolution work.
+
+In other words, this is not a product problem or a platform problem --- it is learning to embrace new ways of working. This is an infrastructure problem that is governed by epistemology --- a fundamental philosophy pertaining to knowledge.
+
+== Why Knowledge Cannot Be Sprinted
+
+An ontology is, per Thomas Gruber's foundational definition, "a specification of a conceptualization." A conceptualization is a whole, not parts of a whole. Specifying half of it produces logical incoherence, not a minimally viable product. For example, ANSI/NISO Z39.19, the standard for constructing controlled vocabularies, requires resolving synonymy, homonymy and scope before terms enter production, because any inconsistency propagates downstream into every subsequent use. SKOS, the W3C standard for encoding knowledge organization systems, requires that hierarchical relationships form coherent structures with transitive closure and OWL 2 requires that class axioms be satisfiable. These are preconditions, necessary for logical systems to reason and express meaning.
+
+In practice, I have watched organizations attempt Agile delivery of knowledge systems in two patterns and neither ends well. In the first, the team holds stand-ups, ceremonies and retrospectives. The taxonomist brings tickets and adds fifteen terms to the product vocabulary. Tickets close and a velocity chart goes up. Nothing is integrated, because integration is the hard part and integration does not fit into two weeks so that goes to the backlog. Six months later, the vocabulary is a collection of inconsistently scoped terms, the ontology has orphan classes and the knowledge graph --- if it exists --- cannot reliably answer queries. But the burndown chart is immaculate and leadership is blindly satisfied because boxes were checked.
+
+In the second pattern, the team borrows from other methodologies while still calling it Agile. The taxonomist conducts multi-week domain analysis and stakeholder interviews --- that is ethnography. She does card sorting and concept modeling --- that is design research. She iterates with subject matter experts --- that is participatory design. She ships, eventually, a coherent model. The sprint cadence was irrelevant to the work. Martin Fowler named this phenomenon in his 2018 keynote: "faux-agile" --- the Agile Industrial Complex.
+
+Ron Jeffries, another original signatory of the Agile Manifesto, told developers the same year, 2018, to abandon the word entirely. Dave Thomas declared in 2014 that "Agile" had become "a magnet for anyone with points to espouse, hours to bill or products to sell." When three of the seventeen authors of a framework say it has been hollowed out, one has to wonder if Agile is appropriate for all technology work and if it is time to revisit ways of working.
+
+== Innovation Does Not Fit in a Backlog
+
+The deeper issue is that Agile and Scrum were never designed to inspire or foster innovation. They were designed to deliver known requirements with increasing efficiency. The Scrum Guide's entire apparatus --- product backlog, sprint planning, Definition of Done --- presupposes that the problem has been formulated by a product owner who can write it in a user story. But knowledge work, like design, is wicked. Rittel and Webber's first property of a wicked problem is that "There is no definitive formulation." The second property states that there is no stopping rule. And the fifth property illuminates that every solution is a "one-shot operation" because there is no opportunity to learn by trial-and-error without consequence.
+
+Given the nature of wicked problems, it would seem that Agile is ill fitted for wicked problems. And for that matter, ill fitted for the dynamic nature of knowledge, where knowledge infrastructures are never one-shot operations, necessitate stopping rules and without a doubt defy definitive formulations.
+
+Therefore, knowledge organization in an enterprise domain is wicked in exactly this sense. The problem of how to represent "customer" in a telecommunications company's semantic model has no pre-existing correct answer. Its formulation depends on who is doing what with the representation, and that changes as you build. The scope of the vocabulary shifts as you interview stakeholders. The taxonomy reorganizes as you encounter edge cases. The ontology's axioms tighten or loosen as the reasoner surfaces contradictions. Don't be mistaken --- this does not bend to the Agile principle of welcoming changing requirements, as the requirements are constituted _through_ the work. The problem is under construction, often evolving, during the solving.
+
+Organizations that manage all work through Scrum are, in effect, optimizing for predictability in domains that demand exploration. The results resist innovation and take on the appearance of productivity. Velocity charts rise while the underlying intellectual architecture is flattened into bite-sized deployments, uncoupled from the continuum that is knowledge.
+
+== Design Thinking Belongs in the Swamp
+
+Donald Schön drew the distinction that "In the varied topography of professional practice, there is a high, hard ground where practitioners can make effective use of research-based theory and technique, and there is a swampy lowland where situations are confusing 'messes' incapable of technical solution." Agile lives on the high ground while knowledge work lives in the swamp. Schön's term for how competent practitioners navigate the swamp was "reflection-in-action" --- the iterative, tacit recalibration that happens while the work is happening.
+
+Design thinking, in the lineage of Herbert Simon, Richard Buchanan and Schön, is the methodology of the swamp. It assumes the problem is under construction. Its five phases --- empathize, define, ideate, prototype, test --- are not a linear pipeline but a recursive, iterative process that expects the definition of the problem to change as understanding deepens.
+
+Research in the _Handbook of Human-Centered Artificial Intelligence_ confirms this alignment --- integrating design thinking into AI development makes systems more user-focused, iterative and impactful, emphasizing data governance, transparency, explainability and bias mitigation. Carnegie Mellon University's Tepper School of Business now teaches a dedicated program on design thinking with AI, arguing that the combination addresses "complex, systemic problems" that linear frameworks cannot manage successfully. IBM's enterprise design thinking practice, developed over two decades, has been applied across client engagements precisely because it keeps human experience at the center while technology evolves underneath --- design thinking is "timeless," while technology is "timestamped."
+
+A 2026 study in _Scientific Reports_ examined the relationship between higher order thinking, generative AI chatbot use and engineering creativity, finding that higher order thinking had significantly greater importance for creative outcomes than AI tool use alone --- reinforcing that cognitive depth, not tool velocity, drives innovation. Research in the _Proceedings of the Design Society_ further argues that generative AI is shifting the designer's role from executor to paradigm-setter, making design thinking more essential as AI handles production tasks while humans must guide framing, ethics and meaning.
+
+To be clear, these frameworks are not different flavors of the same methodology. Scrum and design thinking represent totally different epistemologies. Scrum assumes the problem was handed to you while design thinking assumes the problem must be found and wrestled with.
+
+== Is AI a Knowledge Tool?
+
+This brings us to the question the industry avoids and I, for one, have been itching to ask. Is AI a knowledge tool? How do you classify it?
+
+A large language model generates plausible sequences of text. Returning to the Bender, Gebru, McMillan-Major, and Mitchell paper mentioned earlier, the authors defined a language model as a system for stitching together linguistic forms according to probabilistic patterns, "without any reference to meaning." Without reference to meaning, as LLMs out-of-the-box are absent of external grounding. They optimize for plausibility, not truth and therefore as a standalone, can be best categorized as an information-processing tool.
+
+It becomes a knowledge tool only when grounded in a semantic layer --- controlled vocabularies, taxonomies, ontologies, knowledge graphs --- that provides reference and provenance. The industry's own corrective measures and associated marketplace confirms this as fact. Retrieval-augmented generation, knowledge graphs for LLMs and what recent discourse calls "context engineering" are, underneath the packaging, the semantic infrastructure that should have been built first. But it was bolted on after the fact to prevent hallucination that was pretty much guaranteed the moment the model was deployed without one.
+
+The RAG paper from 2020 called out this problem, identifying that parametric memory is opaque and cannot be updated and that retrieval from structured sources is how provenance is captured and encoded. But it seems the industry has a memory problem or it was too early for most to hear the message.
+
+Ironically, provenance is a core element of knowledge management and ontology work, as provenance is a principled tenant of the discipline and problem space. Without provenance, knowledge cannot exist. Organizations are only now discovering they need it, because their generative systems cannot explain themselves. Call it lineage, call it traces. But if you are building for knowledge, call it provenance, as provenance includes traces, lineage, citations, the temporal and the spatial.
+
+So if generative AI is an information processing tool, I posture that for AI to become a knowledge tool, there must be provenance and descriptive logic. And to be frank, nobody sprints their way into provenance --- it is architected and sustained.
+
+== Knowledge Is a Spiral, Not a Sprint
+
+The irony does not escape the moment. The same Nonaka who co-authored the 1986 paper that inspired Scrum went on to develop the SECI model of knowledge creation in 1995, describing knowledge as a spiral between tacit and explicit forms, moving through socialization, externalization, combination and internalization.
+
+The spiral does not have an end, as is true with spirals, emblematic of life. Each rotation creates new knowledge that becomes the substrate for the next rotation. There is no Definition of Done because the definition of what is being done is itself being constructed. Scrum took the team structure from Nonaka's first paper and discarded his next decade of work because it did not fit neatly with Scrum. The mechanisms of knowledge work, as Nonaka spent a decade arguing, are metaphor, tacit skill, apprenticeship, shared space and the patient conversion of hunches into models and models back into practice.
+
+The Ontology Pipeline™ mirrors this spiral. You start with a controlled vocabulary --- a disciplined process involving agreements as to what terms mean. You mature it into a taxonomy by introducing hierarchy, structuring terms into parent-child relations. Thereafter, the taxonomy is matured into a thesaurus by adding associative relationships and synonymy, encoded in SKOS. The thesaurus is baseline work from which an ontology is developed, adding formal semantics using OWL and RDF. You architect a knowledge graph, assembling components to include the knowledge assets developed through the Ontology Pipeline™ iterative framework.
+
+Each stage prepares the ground for the next and stages cannot be skipped or bypassed. You cannot ship the knowledge graph without an ontology, and the ontology must be logically consistent and coherent to make sense of data. A knowledge graph without vocabulary control risks becoming a confident liar or failing to work with natural language. Because knowledge assets are dimensionally interconnected, knowledge is indeed a spiral, unable to be contained by manufacturing logic.
+
+== The Pattern and the Cure
+
+Organizations budget for platforms --- Confluence, SharePoint, LLM subscriptions, vector databases, a RAG vendor. But they do not budget for descriptive logic and knowledge, perhaps because the problem is not solved by a single tool and is seemingly too amorphous in an Agile, data-centric world. The controlled vocabulary is cut unless it delivers value for the database. Often, the taxonomy is cut because it does not demo well and its value sits in the plumbing, delivering its punch over time. The result is an ever-growing pile of documents with a chatbot stapled to the front, confidently answering questions in the voice of a system that has no reference for any of the terms sprinkled throughout the documents.
+
+Then the system hallucinates. It contradicts itself across sessions. It surfaces outdated policy as current and surfaces private Outlook messages from deactivated employees. I have seen this in real life more times than I can count. And the optics are worse when a vendor tool is responsible for revealing private chats and email transactions. Usually the executive sponsor asks why the investment has not worked, ignoring the evidence of faux pas such as the surfacing of private communications. All in a day's work, when burndown charts and velocity set the precedent for ways of working and reward systems.
+
+Executives ignore the fact that the system was never grounded. It was never grounded because grounding requires the slow, expensive, disciplined work of knowledge organization --- work that does not fit into a sprint, does not demo well and fails to produce a velocity chart. The organization chooses to bypass investment in the descriptive knowledge apparatus because the lines are too blurry to clearly define according to Agile speak. Because knowledge is infrastructure, not a product or a platform. Knowledge is the gas in the tank, the fuel for meaning, the material that justifies logical reasoning.
+
+Design thinking gives that knowledge work a home. It is comfortable with ambiguity. It treats problem discovery as an equal to problem solving. It certainly does not demand a shippable increment every fourteen days. Instead, design thinking demands that you understand what you are building before you build it --- and that you keep revising that understanding as the work proceeds. Library and information science, with over a century of standards for knowledge organization, provides the methods while design thinking provides the epistemology. Together, they offer what Agile never could --- a framework suited to the nature of knowledge itself.
+
+A two-week sprint to deliver what? If you are building knowledge systems, two weeks will get you nothing coherent. Knowledge is not built fast and broken into shape. It is built carefully, with provenance, governance and patience. Knowledge is not a product to be manufactured and no --- you cannot buy off-the-shelf knowledge that perfectly matches a domain or organization. Knowledge is nuanced, specific to each organization, domain and person.
+
+In the spirit of VC-speak, knowledge is the moat, and that moat is not Agile.
+
+#v(2em)
+#line(length: 100%, stroke: 0.4pt + luma(200))
+#v(0.6em)
+#text(size: 9.5pt, fill: luma(120))[
+  _Jessica Talisman, MLS_ is a Semantic Engineer, Information Architect, and knowledge infrastructure strategist with more than 25 years of experience. She is the founder of the Ontology Pipeline™ and publishes _Intentional Arrangement_ on Substack.
+]
diff --git a/assets/presentation/docs/talisman-article/es.typ b/assets/presentation/docs/talisman-article/es.typ
new file mode 100644
index 0000000..c1a1025
--- /dev/null
+++ b/assets/presentation/docs/talisman-article/es.typ
@@ -0,0 +1,160 @@
+#set document(
+  title: "Un Sprint de Dos Semanas para el Conocimiento",
+  author: "Jessica Talisman, MLS",
+)
+
+#set page(
+  paper: "a4",
+  margin: (top: 2.8cm, bottom: 3cm, left: 3cm, right: 3cm),
+  footer: context [
+    #set text(size: 9pt, fill: luma(140))
+    #h(1fr)
+    #counter(page).display("1")
+  ],
+)
+
+#set text(
+  font: ("Georgia", "Linux Libertine"),
+  size: 11pt,
+  lang: "es",
+  hyphenate: true,
+)
+
+#set par(
+  justify: true,
+  leading: 0.75em,
+  spacing: 1.2em,
+)
+
+#set heading(numbering: none)
+
+#show heading.where(level: 1): it => [
+  #v(0.6em)
+  #block(
+    text(size: 13.5pt, weight: "semibold", it.body)
+  )
+  #v(0.3em)
+]
+
+#show heading.where(level: 2): it => [
+  #v(1.2em)
+  #block(
+    text(size: 12pt, weight: "semibold", it.body)
+  )
+  #v(0.2em)
+]
+
+// ── Título ───────────────────────────────────────────────────────────────────
+
+#align(center)[
+  #v(1.6em)
+  #text(size: 22pt, weight: "bold", hyphenate: false)[Un Sprint de Dos Semanas para el Conocimiento]
+  #v(0.5em)
+  #text(size: 14pt, style: "italic", fill: luma(80))[De Agile al Pensamiento de Diseño]
+  #v(1em)
+  #line(length: 60%, stroke: 0.5pt + luma(180))
+  #v(0.8em)
+  #text(size: 11pt)[*Jessica Talisman, MLS*]
+  #v(0.2em)
+  #text(size: 10pt, fill: luma(120))[21 de abril de 2026 · Intentional Arrangement]
+  #v(2em)
+]
+
+// ── Cuerpo ───────────────────────────────────────────────────────────────────
+
+En 1986, Hirotaka Takeuchi e Ikujiro Nonaka publicaron "The New New Product Development Game" en _Harvard Business Review_, describiendo cómo Honda, Canon y Fuji-Xerox organizaban sus equipos para fabricar copiadoras y automóviles. Jeff Sutherland leyó ese artículo, lo combinó con los principios de manufactura lean y la programación orientada a objetos, y en 1993 lo adaptó al desarrollo de software en Easel Corporation. Ken Schwaber formalizó el método en la conferencia OOPSLA de 1995. En 2001, diecisiete desarrolladores de software se reunieron en un albergue de esquí en Snowbird, Utah, y firmaron el Manifiesto Ágil, declarando que estaban "descubriendo mejores maneras de desarrollar software." No conocimiento ni significado. Un marco de manufactura aplicado al software.
+
+El propio lenguaje del Manifiesto Ágil es explícito sobre su alcance, y sin embargo la industria ha aplicado la metodología ágil a prácticamente todo lo que existe en un entorno tecnológico --- incluyendo el trabajo del conocimiento, los sistemas semánticos y ahora la inteligencia artificial. Y en el mismo aliento, organizaciones y expertos colman la industria con discursos sobre innovación y el surgimiento de un nuevo paradigma en las formas de trabajo. Sin embargo, esas formas de trabajo permanecen encerradas dentro de las metodologías ágiles y Scrum, diseñadas para la línea de producción y la base de código --- no para el trabajo profundamente humano, recursivo y desordenado de construir significado a partir de la complejidad.
+
+== Scrum es un calendario de manufactura
+
+Volvamos al artículo de Takeuchi y Nonaka, cuyo título --- "The New New Product Development Game" --- describe cómo los equipos de producto en empresas japonesas desarrollaban bienes físicos como la copiadora FX-3500 y el Honda City. Los autores recurren a la metáfora del rugby porque el equipo "intenta recorrer la distancia como una unidad, pasándose el balón de un lado al otro." La unidad de entrega es un prototipo funcional de un artefacto material. El ciclo de retroalimentación es una máquina que funciona o falla.
+
+El libro de Sutherland de 2014, _Scrum: The Art of Doing Twice the Work in Half the Time_, refuerza esto con casos de estudio militares y de manufactura como la producción del caza F-86 y Toyota. El libro reconstruye el argumento sobre la misma premisa: que aquello que se construye tiene partes discretas, inspeccionables y verificables que, una vez terminadas, están terminadas.
+
+La Guía Scrum de 2020 define el Incremento como "un peldaño concreto hacia el Objetivo del Producto", uno que es "aditivo respecto a todos los Incrementos anteriores y rigurosamente verificado, asegurando que todos los Incrementos funcionen en conjunto." Estas palabras tienen sentido si se está construyendo una aplicación móvil, con una pantalla de inicio de sesión y una página de configuración. Pero para los sistemas de conocimiento, estas palabras son destructivas, porque los componentes de una arquitectura semántica no son aditivos de la manera en que Scrum prescribe. Un vocabulario controlado no es un subcomponente de una taxonomía del mismo modo en que una pantalla de inicio de sesión lo es de una aplicación. Una taxonomía sin un control de vocabulario disciplinado como fundamento no funciona bien, o directamente no funciona. No se puede entregar media taxonomía y llamarla un Incremento.
+
+Scrum asume problemas dóciles que pueden enmarcarse en el tiempo, estimarse con planning poker, descomponerse en historias de usuario, rastrearse en un gráfico de velocidad y declararse resueltos frente a una lista de verificación. Si tan solo el conocimiento fuera así de dócil. La organización del conocimiento está muy lejos de serlo. Con más de cincuenta años de antigüedad, en 1973 Horst Rittel y Melvin Webber advirtieron que aplicar el paradigma de la ingeniería de manera uniforme a todos los problemas estaba "condenado al fracaso." Y sin embargo aquí estamos, aplicando un calendario de manufactura y sus metodologías al conocimiento, esperando que el conocimiento quepa en los confines de viejas formas de trabajar --- un artefacto de la Tercera Revolución Industrial impuesto sobre las exigencias de la Cuarta.
+
+En 2016, Klaus Schwab escribió sobre este patrón histórico. La Primera Revolución Industrial, que comenzó alrededor de 1760, nos dio el vapor y la producción mecánica. La Segunda, a finales del siglo XIX, nos dio la electricidad y la cadena de montaje --- el mundo que produjo Toyota, la manufactura lean y la lógica de producción que Scrum aún lleva como ethos central. La Tercera, que comenzó en la década de 1960, nos dio los semiconductores, la computación mainframe y finalmente internet --- el mundo en que nació Agile, por desarrolladores de software, para el desarrollo de software. La Cuarta Revolución Industrial, argumentó Schwab, es fundamentalmente diferente en su naturaleza, pues está definida por la fusión de la inteligencia artificial, los sistemas ciberfísicos y las tecnologías del conocimiento que difuminan las fronteras entre los mundos físico, digital y biológico.
+
+Los problemas de esta era no son necesariamente problemas de ingeniería con requisitos estables. Estamos lidiando con preguntas como de qué manera anclar la inteligencia artificial en el significado, cómo organizar el conocimiento en evolución y cómo construir sistemas capaces de explicarse a sí mismos. Estos son problemas retorcidos que exigen nuevas epistemologías y marcos para la resolución de problemas. Agile puede haber sido la metodología correcta para la Tercera Revolución Industrial, pero ya no estamos en ella. Los marcos no se han puesto al día, y el coste de ese rezago se agrava cada vez que una organización impone los principios de la Tercera Revolución Industrial al trabajo de la Cuarta.
+
+En otras palabras, esto no es un problema de producto ni de plataforma --- es el desafío de aprender a abrazar nuevas formas de trabajar. Es un problema de infraestructura gobernado por la epistemología: una filosofía fundamental sobre el conocimiento.
+
+== Por qué el conocimiento no puede ejecutarse en sprints
+
+Una ontología es, según la definición fundacional de Thomas Gruber, "una especificación de una conceptualización." Una conceptualización es un todo, no partes de un todo. Especificar la mitad produce incoherencia lógica, no un producto mínimamente viable. Por ejemplo, la norma ANSI/NISO Z39.19 para la construcción de vocabularios controlados exige resolver la sinonimia, la homonimia y el alcance antes de que los términos entren en producción, porque cualquier inconsistencia se propaga hacia abajo en cada uso posterior. SKOS, el estándar del W3C para codificar sistemas de organización del conocimiento, requiere que las relaciones jerárquicas formen estructuras coherentes con cierre transitivo, y OWL 2 exige que los axiomas de clase sean satisfacibles. Estas son precondiciones necesarias para que los sistemas lógicos razonen y expresen significado.
+
+En la práctica, he observado organizaciones intentando entregar sistemas de conocimiento de forma ágil en dos patrones, y ninguno termina bien. En el primero, el equipo celebra stand-ups, ceremonias y retrospectivas. El taxonomista trae tickets y añade quince términos al vocabulario del producto. Los tickets se cierran y el gráfico de velocidad sube. Nada se integra, porque la integración es la parte difícil y no cabe en dos semanas, de modo que va al backlog. Seis meses después, el vocabulario es una colección de términos con alcances inconsistentes, la ontología tiene clases huérfanas y el grafo de conocimiento --- si existe --- no puede responder consultas de manera confiable. Pero el burndown chart es impecable y el liderazgo está ciegamente satisfecho porque las casillas fueron marcadas.
+
+En el segundo patrón, el equipo toma prestado de otras metodologías sin dejar de llamarlo ágil. La taxonomista conduce análisis de dominio y entrevistas con partes interesadas durante varias semanas --- eso es etnografía. Realiza card sorting y modelado conceptual --- eso es investigación en diseño. Itera con expertos en la materia --- eso es diseño participativo. Finalmente entrega un modelo coherente. El ritmo del sprint fue irrelevante para el trabajo. Martin Fowler denominó este fenómeno en su conferencia magistral de 2018: "faux-agile" --- el Complejo Industrial Ágil.
+
+Ron Jeffries, otro signatario original del Manifiesto Ágil, les dijo a los desarrolladores ese mismo año, 2018, que abandonaran la palabra por completo. Dave Thomas declaró en 2014 que "Ágil" se había convertido en "un imán para cualquiera con puntos que exponer, horas que facturar o productos que vender." Cuando tres de los diecisiete autores de un marco dicen que ha sido vaciado de contenido, uno tiene que preguntarse si Agile es apropiado para todo el trabajo tecnológico y si ha llegado el momento de revisar las formas de trabajar.
+
+== La innovación no cabe en un backlog
+
+El problema más profundo es que Agile y Scrum nunca fueron diseñados para inspirar o fomentar la innovación. Fueron diseñados para entregar requisitos conocidos con eficiencia creciente. Todo el aparato de la Guía Scrum --- product backlog, sprint planning, Definición de Hecho --- presupone que el problema ha sido formulado por un product owner capaz de escribirlo en una historia de usuario. Pero el trabajo del conocimiento, como el diseño, es retorcido. La primera propiedad de un problema retorcido según Rittel y Webber es que "No existe una formulación definitiva." La segunda propiedad establece que no hay regla de parada. Y la quinta propiedad ilumina que cada solución es una "operación de un solo disparo" porque no hay oportunidad de aprender por ensayo y error sin consecuencias.
+
+Dada la naturaleza de los problemas retorcidos, parecería que Agile está mal equipado para enfrentarlos. Y por la misma razón, mal equipado para la naturaleza dinámica del conocimiento, donde las infraestructuras de conocimiento nunca son operaciones de un solo disparo, requieren reglas de parada y desafían sin lugar a dudas las formulaciones definitivas.
+
+Por lo tanto, la organización del conocimiento en un dominio empresarial es retorcida exactamente en este sentido. El problema de cómo representar "cliente" en el modelo semántico de una empresa de telecomunicaciones no tiene una respuesta correcta preexistente. Su formulación depende de quién hace qué con esa representación, y eso cambia a medida que se construye. El alcance del vocabulario se desplaza conforme se entrevista a las partes interesadas. La taxonomía se reorganiza al encontrar casos límite. Los axiomas de la ontología se tensan o aflojan cuando el razonador descubre contradicciones. No hay que equivocarse --- esto no se pliega al principio ágil de bienvenida a los requisitos cambiantes, porque los requisitos se constituyen _a través_ del trabajo. El problema está en construcción, evolucionando frecuentemente, durante su resolución.
+
+Las organizaciones que gestionan todo el trabajo a través de Scrum están, en efecto, optimizando para la previsibilidad en dominios que exigen exploración. Los resultados inhiben la innovación y adoptan la apariencia de productividad. Los gráficos de velocidad suben mientras la arquitectura intelectual subyacente se aplana en entregas fragmentadas, desconectadas del continuo que es el conocimiento.
+
+== El pensamiento de diseño pertenece al pantano
+
+Donald Schön trazó la distinción de que "En la variada topografía de la práctica profesional, hay una meseta alta y firme donde los profesionales pueden hacer uso efectivo de la teoría y la técnica basadas en la investigación, y hay unas tierras bajas pantanosas donde las situaciones son 'enredos' confusos incapaces de solución técnica." Agile vive en la meseta mientras que el trabajo del conocimiento vive en el pantano. El término de Schön para describir cómo los profesionales competentes navegan el pantano fue "reflexión-en-acción" --- la recalibración iterativa y tácita que ocurre mientras el trabajo transcurre.
+
+El pensamiento de diseño, en el linaje de Herbert Simon, Richard Buchanan y Schön, es la metodología del pantano. Asume que el problema está en construcción. Sus cinco fases --- empatizar, definir, idear, prototipar, probar --- no son un flujo lineal sino un proceso recursivo e iterativo que espera que la definición del problema cambie a medida que la comprensión se profundiza.
+
+La investigación en el _Handbook of Human-Centered Artificial Intelligence_ confirma esta alineación --- integrar el pensamiento de diseño en el desarrollo de IA hace que los sistemas sean más centrados en el usuario, iterativos e impactantes, enfatizando la gobernanza de datos, la transparencia, la explicabilidad y la mitigación del sesgo. La Tepper School of Business de Carnegie Mellon University imparte ahora un programa dedicado a pensamiento de diseño con IA, argumentando que la combinación aborda "problemas complejos y sistémicos" que los marcos lineales no pueden gestionar con éxito. La práctica de diseño empresarial de IBM, desarrollada a lo largo de dos décadas, se ha aplicado en compromisos con clientes precisamente porque mantiene la experiencia humana en el centro mientras la tecnología evoluciona por debajo --- el pensamiento de diseño es "atemporal", mientras que la tecnología está "fechada".
+
+Un estudio de 2026 en _Scientific Reports_ examinó la relación entre el pensamiento de orden superior, el uso de chatbots de IA generativa y la creatividad en ingeniería, encontrando que el pensamiento de orden superior tenía una importancia significativamente mayor para los resultados creativos que el uso de herramientas de IA por sí solo --- reforzando que la profundidad cognitiva, no la velocidad de las herramientas, impulsa la innovación. La investigación en las _Proceedings of the Design Society_ argumenta además que la IA generativa está desplazando el rol del diseñador de ejecutor a configurador de paradigmas, haciendo que el pensamiento de diseño sea más esencial a medida que la IA gestiona las tareas de producción mientras los humanos deben guiar el encuadre, la ética y el significado.
+
+Para ser claros, estos marcos no son distintos sabores de la misma metodología. Scrum y el pensamiento de diseño representan epistemologías completamente diferentes. Scrum asume que el problema te fue entregado, mientras que el pensamiento de diseño asume que el problema debe encontrarse y lidiarse con él.
+
+== ¿Es la IA una herramienta de conocimiento?
+
+Esto nos lleva a la pregunta que la industria evita y que yo, por mi parte, llevo tiempo deseando formular. ¿Es la IA una herramienta de conocimiento? ¿Cómo se clasifica?
+
+Un modelo de lenguaje de gran escala genera secuencias plausibles de texto. Volviendo al artículo de Bender, Gebru, McMillan-Major y Mitchell mencionado anteriormente, los autores definieron un modelo de lenguaje como un sistema para ensamblar formas lingüísticas según patrones probabilísticos, "sin ninguna referencia al significado." Sin referencia al significado, dado que los LLMs sin configuración adicional carecen de anclaje externo. Optimizan para la plausibilidad, no para la verdad, y por lo tanto, como sistemas autónomos, pueden clasificarse mejor como herramientas de procesamiento de información.
+
+Se convierte en una herramienta de conocimiento únicamente cuando se ancla en una capa semántica --- vocabularios controlados, taxonomías, ontologías, grafos de conocimiento --- que proporciona referencia y procedencia. Las propias medidas correctivas de la industria y el mercado asociado confirman esto como un hecho. La generación aumentada por recuperación, los grafos de conocimiento para LLMs y lo que el discurso reciente denomina "ingeniería de contexto" son, bajo el empaquetado, la infraestructura semántica que debería haberse construido primero. Pero fue añadida a posteriori para prevenir alucinaciones que estaban prácticamente garantizadas desde el momento en que el modelo fue desplegado sin ella.
+
+El artículo sobre RAG de 2020 señaló este problema, identificando que la memoria paramétrica es opaca y no puede actualizarse, y que la recuperación desde fuentes estructuradas es el modo en que la procedencia se captura y codifica. Pero parece que la industria tiene un problema de memoria, o el mensaje llegó demasiado pronto para que la mayoría pudiera escucharlo.
+
+Irónicamente, la procedencia es un elemento central de la gestión del conocimiento y el trabajo ontológico, pues constituye un principio fundacional de la disciplina y su espacio de problemas. Sin procedencia, el conocimiento no puede existir. Las organizaciones están descubriendo ahora que la necesitan, porque sus sistemas generativos no pueden explicarse a sí mismos. Llámalo linaje, llámalo trazabilidad. Pero si construyes para el conocimiento, llámalo procedencia, porque la procedencia abarca trazas, linaje, citaciones, lo temporal y lo espacial.
+
+De modo que si la IA generativa es una herramienta de procesamiento de información, sostengo que para que la IA se convierta en una herramienta de conocimiento debe haber procedencia y lógica descriptiva. Y para ser francos, nadie llega a la procedencia haciendo sprints --- esta se arquitecta y se sostiene.
+
+== El conocimiento es una espiral, no un sprint
+
+La ironía no pasa desapercibida en este momento. El mismo Nonaka que coescribió el artículo de 1986 que inspiró Scrum desarrolló el modelo SECI de creación de conocimiento en 1995, describiendo el conocimiento como una espiral entre formas tácitas y explícitas, que avanza a través de la socialización, la externalización, la combinación y la internalización.
+
+La espiral no tiene fin, como es propio de las espirales, emblema de la vida. Cada rotación crea nuevo conocimiento que se convierte en el sustrato de la siguiente. No existe una Definición de Hecho porque la definición de lo que se está haciendo está siendo construida en sí misma. Scrum tomó la estructura de equipo del primer artículo de Nonaka y descartó su siguiente década de trabajo porque no encajaba limpiamente en Scrum. Los mecanismos del trabajo del conocimiento, como Nonaka argumentó durante una década, son la metáfora, la habilidad tácita, el aprendizaje por proximidad, el espacio compartido y la paciente conversión de intuiciones en modelos y de modelos de vuelta en práctica.
+
+El Ontology Pipeline™ refleja esta espiral. Se comienza con un vocabulario controlado --- un proceso disciplinado que implica acuerdos sobre lo que significan los términos. Se madura hacia una taxonomía introduciendo jerarquía y estructurando los términos en relaciones padre-hijo. Posteriormente, la taxonomía se madura hacia un tesauro añadiendo relaciones asociativas y sinonimia, codificadas en SKOS. El tesauro es el trabajo base a partir del cual se desarrolla una ontología, incorporando semántica formal mediante OWL y RDF. Se arquitecta un grafo de conocimiento, ensamblando componentes que incluyen los activos de conocimiento desarrollados a través del marco iterativo del Ontology Pipeline™.
+
+Cada etapa prepara el terreno para la siguiente y las etapas no pueden saltarse ni eludirse. No se puede entregar el grafo de conocimiento sin una ontología, y la ontología debe ser lógicamente consistente y coherente para dotar de sentido a los datos. Un grafo de conocimiento sin control de vocabulario arriesga convertirse en un mentiroso convincente o en fallar al trabajar con lenguaje natural. Porque los activos de conocimiento son dimensionalmente interdependientes, el conocimiento es efectivamente una espiral, incapaz de ser contenida por la lógica de la manufactura.
+
+== El patrón y el remedio
+
+Las organizaciones presupuestan para plataformas --- Confluence, SharePoint, suscripciones a LLMs, bases de datos vectoriales, un proveedor de RAG. Pero no presupuestan para la lógica descriptiva y el conocimiento, quizás porque el problema no se resuelve con una sola herramienta y resulta demasiado amorfo en un mundo centrado en los datos y en Agile. El vocabulario controlado se recorta a menos que genere valor inmediato para la base de datos. Con frecuencia, la taxonomía se recorta porque no hace una buena demostración y su valor reside en la fontanería, entregando su impacto con el tiempo. El resultado es una pila siempre creciente de documentos con un chatbot pegado en la entrada, respondiendo preguntas con confianza en la voz de un sistema que no tiene referencia para ninguno de los términos dispersos a lo largo de los documentos.
+
+Entonces el sistema alucina. Se contradice a sí mismo entre sesiones. Presenta como vigente una política desactualizada y muestra mensajes privados de Outlook de empleados cuyas cuentas fueron desactivadas. He visto esto en la vida real más veces de las que puedo contar. Y la percepción es peor cuando una herramienta de proveedor es responsable de revelar chats y transacciones de correo privadas. Habitualmente, el patrocinador ejecutivo pregunta por qué la inversión no ha funcionado, ignorando la evidencia de los errores, como la exposición de comunicaciones privadas. Trabajo de un día cualquiera, cuando los burndown charts y la velocidad marcan el precedente de las formas de trabajar y los sistemas de recompensa.
+
+Los ejecutivos ignoran el hecho de que el sistema nunca fue anclado en el conocimiento. Nunca lo fue porque anclar requiere el trabajo lento, costoso y disciplinado de la organización del conocimiento --- trabajo que no cabe en un sprint, que no hace una buena demostración y que no genera un gráfico de velocidad. La organización elige eludir la inversión en el aparato de conocimiento descriptivo porque las líneas son demasiado difusas para definirse claramente según el vocabulario de Agile. Porque el conocimiento es infraestructura, no un producto ni una plataforma. El conocimiento es la gasolina del depósito, el combustible del significado, el material que justifica el razonamiento lógico.
+
+El pensamiento de diseño le da al trabajo del conocimiento un hogar. Es cómodo con la ambigüedad. Trata el descubrimiento del problema como un igual a la resolución del problema. Desde luego no exige un incremento entregable cada catorce días. En cambio, el pensamiento de diseño exige que entiendas lo que estás construyendo antes de construirlo --- y que sigas revisando esa comprensión a medida que el trabajo avanza. La biblioteconomía y la ciencia de la información, con más de un siglo de estándares para la organización del conocimiento, proporcionan los métodos mientras el pensamiento de diseño proporciona la epistemología. Juntos ofrecen lo que Agile nunca pudo --- un marco adecuado a la naturaleza del conocimiento mismo.
+
+¿Un sprint de dos semanas para entregar qué? Si estás construyendo sistemas de conocimiento, dos semanas no producirán nada coherente. El conocimiento no se construye rápido y se moldea a la fuerza. Se construye con cuidado, con procedencia, gobernanza y paciencia. El conocimiento no es un producto que pueda manufacturarse y no --- no se puede comprar conocimiento estándar que encaje perfectamente con un dominio u organización. El conocimiento es matizado, específico a cada organización, dominio y persona.
+
+En el espíritu del argot del capital de riesgo: el conocimiento es el foso, y ese foso no es Ágil.
+
+#v(2em)
+#line(length: 100%, stroke: 0.4pt + luma(200))
+#v(0.6em)
+#text(size: 9.5pt, fill: luma(120))[
+  _Jessica Talisman, MLS_ es Ingeniera Semántica, Arquitecta de Información y estratega de infraestructuras de conocimiento con más de 25 años de experiencia. Es fundadora del Ontology Pipeline™ y publica _Intentional Arrangement_ en Substack.
+]
diff --git a/assets/presentation/docs/talisman-article/un-sprint-de-dos-semanas-para-el-conocimiento.md b/assets/presentation/docs/talisman-article/un-sprint-de-dos-semanas-para-el-conocimiento.md
new file mode 100644
index 0000000..ec637a8
--- /dev/null
+++ b/assets/presentation/docs/talisman-article/un-sprint-de-dos-semanas-para-el-conocimiento.md
@@ -0,0 +1,101 @@
+# Un Sprint de Dos Semanas para el Conocimiento
+## De Agile al Pensamiento de Diseño
+
+**Jessica Talisman, MLS**
+*21 de abril de 2026*
+
+---
+
+En 1986, Hirotaka Takeuchi e Ikujiro Nonaka publicaron "The New New Product Development Game" en *Harvard Business Review*, describiendo cómo Honda, Canon y Fuji-Xerox organizaban sus equipos para fabricar copiadoras y automóviles. Jeff Sutherland leyó ese artículo, lo combinó con los principios de manufactura lean y la programación orientada a objetos, y en 1993 lo adaptó al desarrollo de software en Easel Corporation. Ken Schwaber formalizó el método en la conferencia OOPSLA de 1995. En 2001, diecisiete desarrolladores de software se reunieron en un albergue de esquí en Snowbird, Utah, y firmaron el Manifiesto Ágil, declarando que estaban "descubriendo mejores maneras de desarrollar software." No conocimiento ni significado. Un marco de manufactura aplicado al software.
+
+El propio lenguaje del Manifiesto Ágil es explícito sobre su alcance, y sin embargo la industria ha aplicado la metodología ágil a prácticamente todo lo que existe en un entorno tecnológico — incluyendo el trabajo del conocimiento, los sistemas semánticos y ahora la inteligencia artificial. Y en el mismo aliento, organizaciones y expertos colman la industria con discursos sobre innovación y el surgimiento de un nuevo paradigma en las formas de trabajo. Sin embargo, esas formas de trabajo permanecen encerradas dentro de las metodologías ágiles y Scrum, diseñadas para la línea de producción y la base de código — no para el trabajo profundamente humano, recursivo y desordenado de construir significado a partir de la complejidad.
+
+## Scrum es un calendario de manufactura
+
+Volvamos al artículo de Takeuchi y Nonaka, cuyo título — "The New New Product Development Game" — describe cómo los equipos de producto en empresas japonesas desarrollaban bienes físicos como la copiadora FX-3500 y el Honda City. Los autores recurren a la metáfora del rugby porque el equipo "intenta recorrer la distancia como una unidad, pasándose el balón de un lado al otro." La unidad de entrega es un prototipo funcional de un artefacto material. El ciclo de retroalimentación es una máquina que funciona o falla.
+
+El libro de Sutherland de 2014, *Scrum: The Art of Doing Twice the Work in Half the Time*, refuerza esto con casos de estudio militares y de manufactura como la producción del caza F-86 y Toyota. El libro reconstruye el argumento sobre la misma premisa: que aquello que se construye tiene partes discretas, inspeccionables y verificables que, una vez terminadas, están terminadas.
+
+La Guía Scrum de 2020 define el Incremento como "un peldaño concreto hacia el Objetivo del Producto", uno que es "aditivo respecto a todos los Incrementos anteriores y rigurosamente verificado, asegurando que todos los Incrementos funcionen en conjunto." Estas palabras tienen sentido si se está construyendo una aplicación móvil, con una pantalla de inicio de sesión y una página de configuración. Pero para los sistemas de conocimiento, estas palabras son destructivas, porque los componentes de una arquitectura semántica no son aditivos de la manera en que Scrum prescribe. Un vocabulario controlado no es un subcomponente de una taxonomía del mismo modo en que una pantalla de inicio de sesión lo es de una aplicación. Una taxonomía sin un control de vocabulario disciplinado como fundamento no funciona bien, o directamente no funciona. No se puede entregar media taxonomía y llamarla un Incremento.
+
+Scrum asume problemas dóciles que pueden enmarcarse en el tiempo, estimarse con planning poker, descomponerse en historias de usuario, rastrearse en un gráfico de velocidad y declararse resueltos frente a una lista de verificación. Si tan solo el conocimiento fuera así de dócil. La organización del conocimiento está muy lejos de serlo. Con más de cincuenta años de antigüedad, en 1973 Horst Rittel y Melvin Webber advirtieron que aplicar el paradigma de la ingeniería de manera uniforme a todos los problemas estaba "condenado al fracaso." Y sin embargo aquí estamos, aplicando un calendario de manufactura y sus metodologías al conocimiento, esperando que el conocimiento quepa en los confines de viejas formas de trabajar — un artefacto de la Tercera Revolución Industrial impuesto sobre las exigencias de la Cuarta.
+
+En 2016, Klaus Schwab escribió sobre este patrón histórico. La Primera Revolución Industrial, que comenzó alrededor de 1760, nos dio el vapor y la producción mecánica. La Segunda, a finales del siglo XIX, nos dio la electricidad y la cadena de montaje — el mundo que produjo Toyota, la manufactura lean y la lógica de producción que Scrum aún lleva como ethos central. La Tercera, que comenzó en la década de 1960, nos dio los semiconductores, la computación mainframe y finalmente internet — el mundo en que nació Agile, por desarrolladores de software, para el desarrollo de software. La Cuarta Revolución Industrial, argumentó Schwab, es fundamentalmente diferente en su naturaleza, pues está definida por la fusión de la inteligencia artificial, los sistemas ciberfísicos y las tecnologías del conocimiento que difuminan las fronteras entre los mundos físico, digital y biológico.
+
+Los problemas de esta era no son necesariamente problemas de ingeniería con requisitos estables. Estamos lidiando con preguntas como de qué manera anclar la inteligencia artificial en el significado, cómo organizar el conocimiento en evolución y cómo construir sistemas capaces de explicarse a sí mismos. Estos son problemas retorcidos que exigen nuevas epistemologías y marcos para la resolución de problemas. Agile puede haber sido la metodología correcta para la Tercera Revolución Industrial, pero ya no estamos en ella. Los marcos no se han puesto al día, y el coste de ese rezago se agrava cada vez que una organización impone los principios de la Tercera Revolución Industrial al trabajo de la Cuarta.
+
+En otras palabras, esto no es un problema de producto ni de plataforma — es el desafío de aprender a abrazar nuevas formas de trabajar. Es un problema de infraestructura gobernado por la epistemología: una filosofía fundamental sobre el conocimiento.
+
+## Por qué el conocimiento no puede ejecutarse en sprints
+
+Una ontología es, según la definición fundacional de Thomas Gruber, "una especificación de una conceptualización." Una conceptualización es un todo, no partes de un todo. Especificar la mitad produce incoherencia lógica, no un producto mínimamente viable. Por ejemplo, la norma ANSI/NISO Z39.19 para la construcción de vocabularios controlados exige resolver la sinonimia, la homonimia y el alcance antes de que los términos entren en producción, porque cualquier inconsistencia se propaga hacia abajo en cada uso posterior. SKOS, el estándar del W3C para codificar sistemas de organización del conocimiento, requiere que las relaciones jerárquicas formen estructuras coherentes con cierre transitivo, y OWL 2 exige que los axiomas de clase sean satisfacibles. Estas son precondiciones necesarias para que los sistemas lógicos razonen y expresen significado.
+
+En la práctica, he observado organizaciones intentando entregar sistemas de conocimiento de forma ágil en dos patrones, y ninguno termina bien. En el primero, el equipo celebra stand-ups, ceremonias y retrospectivas. El taxonomista trae tickets y añade quince términos al vocabulario del producto. Los tickets se cierran y el gráfico de velocidad sube. Nada se integra, porque la integración es la parte difícil y no cabe en dos semanas, de modo que va al backlog. Seis meses después, el vocabulario es una colección de términos con alcances inconsistentes, la ontología tiene clases huérfanas y el grafo de conocimiento — si existe — no puede responder consultas de manera confiable. Pero el burndown chart es impecable y el liderazgo está ciegamente satisfecho porque las casillas fueron marcadas.
+
+En el segundo patrón, el equipo toma prestado de otras metodologías sin dejar de llamarlo ágil. La taxonomista conduce análisis de dominio y entrevistas con partes interesadas durante varias semanas — eso es etnografía. Realiza card sorting y modelado conceptual — eso es investigación en diseño. Itera con expertos en la materia — eso es diseño participativo. Finalmente entrega un modelo coherente. El ritmo del sprint fue irrelevante para el trabajo. Martin Fowler denominó este fenómeno en su conferencia magistral de 2018: "faux-agile" — el Complejo Industrial Ágil.
+
+Ron Jeffries, otro signatario original del Manifiesto Ágil, les dijo a los desarrolladores ese mismo año, 2018, que abandonaran la palabra por completo. Dave Thomas declaró en 2014 que "Ágil" se había convertido en "un imán para cualquiera con puntos que exponer, horas que facturar o productos que vender." Cuando tres de los diecisiete autores de un marco dicen que ha sido vaciado de contenido, uno tiene que preguntarse si Agile es apropiado para todo el trabajo tecnológico y si ha llegado el momento de revisar las formas de trabajar.
+
+## La innovación no cabe en un backlog
+
+El problema más profundo es que Agile y Scrum nunca fueron diseñados para inspirar o fomentar la innovación. Fueron diseñados para entregar requisitos conocidos con eficiencia creciente. Todo el aparato de la Guía Scrum — product backlog, sprint planning, Definición de Hecho — presupone que el problema ha sido formulado por un product owner capaz de escribirlo en una historia de usuario. Pero el trabajo del conocimiento, como el diseño, es retorcido. La primera propiedad de un problema retorcido según Rittel y Webber es que "No existe una formulación definitiva." La segunda propiedad establece que no hay regla de parada. Y la quinta propiedad ilumina que cada solución es una "operación de un solo disparo" porque no hay oportunidad de aprender por ensayo y error sin consecuencias.
+
+Dada la naturaleza de los problemas retorcidos, parecería que Agile está mal equipado para enfrentarlos. Y por la misma razón, mal equipado para la naturaleza dinámica del conocimiento, donde las infraestructuras de conocimiento nunca son operaciones de un solo disparo, requieren reglas de parada y desafían sin lugar a dudas las formulaciones definitivas.
+
+Por lo tanto, la organización del conocimiento en un dominio empresarial es retorcida exactamente en este sentido. El problema de cómo representar "cliente" en el modelo semántico de una empresa de telecomunicaciones no tiene una respuesta correcta preexistente. Su formulación depende de quién hace qué con esa representación, y eso cambia a medida que se construye. El alcance del vocabulario se desplaza conforme se entrevista a las partes interesadas. La taxonomía se reorganiza al encontrar casos límite. Los axiomas de la ontología se tensan o aflojan cuando el razonador descubre contradicciones. No hay que equivocarse — esto no se pliega al principio ágil de bienvenida a los requisitos cambiantes, porque los requisitos se constituyen *a través* del trabajo. El problema está en construcción, evolucionando frecuentemente, durante su resolución.
+
+Las organizaciones que gestionan todo el trabajo a través de Scrum están, en efecto, optimizando para la previsibilidad en dominios que exigen exploración. Los resultados inhiben la innovación y adoptan la apariencia de productividad. Los gráficos de velocidad suben mientras la arquitectura intelectual subyacente se aplana en entregas fragmentadas, desconectadas del continuo que es el conocimiento.
+
+## El pensamiento de diseño pertenece al pantano
+
+Donald Schön trazó la distinción de que "En la variada topografía de la práctica profesional, hay una meseta alta y firme donde los profesionales pueden hacer uso efectivo de la teoría y la técnica basadas en la investigación, y hay unas tierras bajas pantanosas donde las situaciones son 'enredos' confusos incapaces de solución técnica." Agile vive en la meseta mientras que el trabajo del conocimiento vive en el pantano. El término de Schön para describir cómo los profesionales competentes navegan el pantano fue "reflexión-en-acción" — la recalibración iterativa y tácita que ocurre mientras el trabajo transcurre.
+
+El pensamiento de diseño, en el linaje de Herbert Simon, Richard Buchanan y Schön, es la metodología del pantano. Asume que el problema está en construcción. Sus cinco fases — empatizar, definir, idear, prototipar, probar — no son un flujo lineal sino un proceso recursivo e iterativo que espera que la definición del problema cambie a medida que la comprensión se profundiza.
+
+La investigación en el *Handbook of Human-Centered Artificial Intelligence* confirma esta alineación — integrar el pensamiento de diseño en el desarrollo de IA hace que los sistemas sean más centrados en el usuario, iterativos e impactantes, enfatizando la gobernanza de datos, la transparencia, la explicabilidad y la mitigación del sesgo. La Tepper School of Business de Carnegie Mellon University imparte ahora un programa dedicado a pensamiento de diseño con IA, argumentando que la combinación aborda "problemas complejos y sistémicos" que los marcos lineales no pueden gestionar con éxito. La práctica de diseño empresarial de IBM, desarrollada a lo largo de dos décadas, se ha aplicado en compromisos con clientes precisamente porque mantiene la experiencia humana en el centro mientras la tecnología evoluciona por debajo — el pensamiento de diseño es "atemporal", mientras que la tecnología está "fechada".
+
+Un estudio de 2026 en *Scientific Reports* examinó la relación entre el pensamiento de orden superior, el uso de chatbots de IA generativa y la creatividad en ingeniería, encontrando que el pensamiento de orden superior tenía una importancia significativamente mayor para los resultados creativos que el uso de herramientas de IA por sí solo — reforzando que la profundidad cognitiva, no la velocidad de las herramientas, impulsa la innovación. La investigación en las *Proceedings of the Design Society* argumenta además que la IA generativa está desplazando el rol del diseñador de ejecutor a configurador de paradigmas, haciendo que el pensamiento de diseño sea más esencial a medida que la IA gestiona las tareas de producción mientras los humanos deben guiar el encuadre, la ética y el significado.
+
+Para ser claros, estos marcos no son distintos sabores de la misma metodología. Scrum y el pensamiento de diseño representan epistemologías completamente diferentes. Scrum asume que el problema te fue entregado, mientras que el pensamiento de diseño asume que el problema debe encontrarse y lidiarse con él.
+
+## ¿Es la IA una herramienta de conocimiento?
+
+Esto nos lleva a la pregunta que la industria evita y que yo, por mi parte, llevo tiempo deseando formular. ¿Es la IA una herramienta de conocimiento? ¿Cómo se clasifica?
+
+Un modelo de lenguaje de gran escala genera secuencias plausibles de texto. Volviendo al artículo de Bender, Gebru, McMillan-Major y Mitchell mencionado anteriormente, los autores definieron un modelo de lenguaje como un sistema para ensamblar formas lingüísticas según patrones probabilísticos, "sin ninguna referencia al significado." Sin referencia al significado, dado que los LLMs sin configuración adicional carecen de anclaje externo. Optimizan para la plausibilidad, no para la verdad, y por lo tanto, como sistemas autónomos, pueden clasificarse mejor como herramientas de procesamiento de información.
+
+Se convierte en una herramienta de conocimiento únicamente cuando se ancla en una capa semántica — vocabularios controlados, taxonomías, ontologías, grafos de conocimiento — que proporciona referencia y procedencia. Las propias medidas correctivas de la industria y el mercado asociado confirman esto como un hecho. La generación aumentada por recuperación, los grafos de conocimiento para LLMs y lo que el discurso reciente denomina "ingeniería de contexto" son, bajo el empaquetado, la infraestructura semántica que debería haberse construido primero. Pero fue añadida a posteriori para prevenir alucinaciones que estaban prácticamente garantizadas desde el momento en que el modelo fue desplegado sin ella.
+
+El artículo sobre RAG de 2020 señaló este problema, identificando que la memoria paramétrica es opaca y no puede actualizarse, y que la recuperación desde fuentes estructuradas es el modo en que la procedencia se captura y codifica. Pero parece que la industria tiene un problema de memoria, o el mensaje llegó demasiado pronto para que la mayoría pudiera escucharlo.
+
+Irónicamente, la procedencia es un elemento central de la gestión del conocimiento y el trabajo ontológico, pues constituye un principio fundacional de la disciplina y su espacio de problemas. Sin procedencia, el conocimiento no puede existir. Las organizaciones están descubriendo ahora que la necesitan, porque sus sistemas generativos no pueden explicarse a sí mismos. Llámalo linaje, llámalo trazabilidad. Pero si construyes para el conocimiento, llámalo procedencia, porque la procedencia abarca trazas, linaje, citaciones, lo temporal y lo espacial.
+
+De modo que si la IA generativa es una herramienta de procesamiento de información, sostengo que para que la IA se convierta en una herramienta de conocimiento debe haber procedencia y lógica descriptiva. Y para ser francos, nadie llega a la procedencia haciendo sprints — esta se arquitecta y se sostiene.
+
+## El conocimiento es una espiral, no un sprint
+
+La ironía no pasa desapercibida en este momento. El mismo Nonaka que coescribió el artículo de 1986 que inspiró Scrum desarrolló el modelo SECI de creación de conocimiento en 1995, describiendo el conocimiento como una espiral entre formas tácitas y explícitas, que avanza a través de la socialización, la externalización, la combinación y la internalización.
+
+La espiral no tiene fin, como es propio de las espirales, emblema de la vida. Cada rotación crea nuevo conocimiento que se convierte en el sustrato de la siguiente. No existe una Definición de Hecho porque la definición de lo que se está haciendo está siendo construida en sí misma. Scrum tomó la estructura de equipo del primer artículo de Nonaka y descartó su siguiente década de trabajo porque no encajaba limpiamente en Scrum. Los mecanismos del trabajo del conocimiento, como Nonaka argumentó durante una década, son la metáfora, la habilidad tácita, el aprendizaje por proximidad, el espacio compartido y la paciente conversión de intuiciones en modelos y de modelos de vuelta en práctica.
+
+El Ontology Pipeline™ refleja esta espiral. Se comienza con un vocabulario controlado — un proceso disciplinado que implica acuerdos sobre lo que significan los términos. Se madura hacia una taxonomía introduciendo jerarquía y estructurando los términos en relaciones padre-hijo. Posteriormente, la taxonomía se madura hacia un tesauro añadiendo relaciones asociativas y sinonimia, codificadas en SKOS. El tesauro es el trabajo base a partir del cual se desarrolla una ontología, incorporando semántica formal mediante OWL y RDF. Se arquitecta un grafo de conocimiento, ensamblando componentes que incluyen los activos de conocimiento desarrollados a través del marco iterativo del Ontology Pipeline™.
+
+Cada etapa prepara el terreno para la siguiente y las etapas no pueden saltarse ni eludirse. No se puede entregar el grafo de conocimiento sin una ontología, y la ontología debe ser lógicamente consistente y coherente para dotar de sentido a los datos. Un grafo de conocimiento sin control de vocabulario arriesga convertirse en un mentiroso convincente o en fallar al trabajar con lenguaje natural. Porque los activos de conocimiento son dimensionalmente interdependientes, el conocimiento es efectivamente una espiral, incapaz de ser contenida por la lógica de la manufactura.
+
+## El patrón y el remedio
+
+Las organizaciones presupuestan para plataformas — Confluence, SharePoint, suscripciones a LLMs, bases de datos vectoriales, un proveedor de RAG. Pero no presupuestan para la lógica descriptiva y el conocimiento, quizás porque el problema no se resuelve con una sola herramienta y resulta demasiado amorfo en un mundo centrado en los datos y en Agile. El vocabulario controlado se recorta a menos que genere valor inmediato para la base de datos. Con frecuencia, la taxonomía se recorta porque no hace una buena demostración y su valor reside en la fontanería, entregando su impacto con el tiempo. El resultado es una pila siempre creciente de documentos con un chatbot pegado en la entrada, respondiendo preguntas con confianza en la voz de un sistema que no tiene referencia para ninguno de los términos dispersos a lo largo de los documentos.
+
+Entonces el sistema alucina. Se contradice a sí mismo entre sesiones. Presenta como vigente una política desactualizada y muestra mensajes privados de Outlook de empleados cuyas cuentas fueron desactivadas. He visto esto en la vida real más veces de las que puedo contar. Y la percepción es peor cuando una herramienta de proveedor es responsable de revelar chats y transacciones de correo privadas. Habitualmente, el patrocinador ejecutivo pregunta por qué la inversión no ha funcionado, ignorando la evidencia de los errores, como la exposición de comunicaciones privadas. Trabajo de un día cualquiera, cuando los burndown charts y la velocidad marcan el precedente de las formas de trabajar y los sistemas de recompensa.
+
+Los ejecutivos ignoran el hecho de que el sistema nunca fue anclado en el conocimiento. Nunca lo fue porque anclar requiere el trabajo lento, costoso y disciplinado de la organización del conocimiento — trabajo que no cabe en un sprint, que no hace una buena demostración y que no genera un gráfico de velocidad. La organización elige eludir la inversión en el aparato de conocimiento descriptivo porque las líneas son demasiado difusas para definirse claramente según el vocabulario de Agile. Porque el conocimiento es infraestructura, no un producto ni una plataforma. El conocimiento es la gasolina del depósito, el combustible del significado, el material que justifica el razonamiento lógico.
+
+El pensamiento de diseño le da al trabajo del conocimiento un hogar. Es cómodo con la ambigüedad. Trata el descubrimiento del problema como un igual a la resolución del problema. Desde luego no exige un incremento entregable cada catorce días. En cambio, el pensamiento de diseño exige que entiendas lo que estás construyendo antes de construirlo — y que sigas revisando esa comprensión a medida que el trabajo avanza. La biblioteconomía y la ciencia de la información, con más de un siglo de estándares para la organización del conocimiento, proporcionan los métodos mientras el pensamiento de diseño proporciona la epistemología. Juntos ofrecen lo que Agile nunca pudo — un marco adecuado a la naturaleza del conocimiento mismo.
+
+¿Un sprint de dos semanas para entregar qué? Si estás construyendo sistemas de conocimiento, dos semanas no producirán nada coherente. El conocimiento no se construye rápido y se moldea a la fuerza. Se construye con cuidado, con procedencia, gobernanza y paciencia. El conocimiento no es un producto que pueda manufacturarse y no — no se puede comprar conocimiento estándar que encaje perfectamente con un dominio u organización. El conocimiento es matizado, específico a cada organización, dominio y persona.
+
+En el espíritu del argot del capital de riesgo: el conocimiento es el foso, y ese foso no es Ágil.
+
+---
+
+*Jessica Talisman, MLS — Intentional Arrangement*
diff --git a/assets/presentation/docs/talisman-article/un-sprint-de-dos-semanas-para-el-conocimiento.pdf b/assets/presentation/docs/talisman-article/un-sprint-de-dos-semanas-para-el-conocimiento.pdf
new file mode 100644
index 0000000..b5bb569
--- /dev/null
+++ b/assets/presentation/docs/talisman-article/un-sprint-de-dos-semanas-para-el-conocimiento.pdf
@@ -0,0 +1,2119 @@
+%PDF-1.7
+%����
+
+1 0 obj
+<<
+  /Type /Pages
+  /Count 8
+  /Kids [106 0 R 108 0 R 110 0 R 112 0 R 114 0 R 116 0 R 118 0 R 120 0 R]
+>>
+endobj
+
+2 0 obj
+<<
+  /Type /Outlines
+  /First 3 0 R
+  /Last 9 0 R
+  /Count 7
+>>
+endobj
+
+3 0 obj
+<<
+  /Parent 2 0 R
+  /Next 4 0 R
+  /Title (Scrum es un calendario de manufactura)
+  /Dest 99 0 R
+>>
+endobj
+
+4 0 obj
+<<
+  /Parent 2 0 R
+  /Next 5 0 R
+  /Prev 3 0 R
+  /Title <FEFF0050006F007200200071007500E900200065006C00200063006F006E006F00630069006D00690065006E0074006F0020006E006F00200070007500650064006500200065006A0065006300750074006100720073006500200065006E00200073007000720069006E00740073>
+  /Dest 100 0 R
+>>
+endobj
+
+5 0 obj
+<<
+  /Parent 2 0 R
+  /Next 6 0 R
+  /Prev 4 0 R
+  /Title <FEFF004C006100200069006E006E006F007600610063006900F3006E0020006E006F0020006300610062006500200065006E00200075006E0020006200610063006B006C006F0067>
+  /Dest 101 0 R
+>>
+endobj
+
+6 0 obj
+<<
+  /Parent 2 0 R
+  /Next 7 0 R
+  /Prev 5 0 R
+  /Title <FEFF0045006C002000700065006E00730061006D00690065006E0074006F0020006400650020006400690073006500F1006F002000700065007200740065006E00650063006500200061006C002000700061006E00740061006E006F>
+  /Dest 102 0 R
+>>
+endobj
+
+7 0 obj
+<<
+  /Parent 2 0 R
+  /Next 8 0 R
+  /Prev 6 0 R
+  /Title <FEFF00BF004500730020006C006100200049004100200075006E0061002000680065007200720061006D00690065006E0074006100200064006500200063006F006E006F00630069006D00690065006E0074006F003F>
+  /Dest 103 0 R
+>>
+endobj
+
+8 0 obj
+<<
+  /Parent 2 0 R
+  /Next 9 0 R
+  /Prev 7 0 R
+  /Title (El conocimiento es una espiral, no un sprint)
+  /Dest 104 0 R
+>>
+endobj
+
+9 0 obj
+<<
+  /Parent 2 0 R
+  /Prev 8 0 R
+  /Title <FEFF0045006C0020007000610074007200F3006E0020007900200065006C002000720065006D006500640069006F>
+  /Dest 105 0 R
+>>
+endobj
+
+10 0 obj
+<<
+  /Type /StructTreeRoot
+  /RoleMap <<
+    /Datetime /Span
+    /Terms /Part
+    /Title /P
+    /Strong /Span
+    /Em /Span
+  >>
+  /K [19 0 R]
+  /ParentTree <<
+    /Nums [0 11 0 R 1 12 0 R 2 13 0 R 3 14 0 R 4 15 0 R 5 16 0 R 6 17 0 R 7 18 0 R]
+  >>
+  /ParentTreeNextKey 8
+>>
+endobj
+
+11 0 obj
+[78 0 R 78 0 R 77 0 R 76 0 R 74 0 R 72 0 R 72 0 R 72 0 R 73 0 R 72 0 R 72 0 R 72 0 R 72 0 R 72 0 R 72 0 R 72 0 R 72 0 R 72 0 R 72 0 R 71 0 R 71 0 R 71 0 R 71 0 R 71 0 R 71 0 R 71 0 R 71 0 R 71 0 R 70 0 R 69 0 R 69 0 R 69 0 R 69 0 R 69 0 R 69 0 R 69 0 R]
+endobj
+
+12 0 obj
+[67 0 R 68 0 R 68 0 R 67 0 R 67 0 R 67 0 R 67 0 R 66 0 R 66 0 R 66 0 R 66 0 R 66 0 R 66 0 R 66 0 R 66 0 R 66 0 R 66 0 R 66 0 R 66 0 R 65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 65 0 R 64 0 R 64 0 R 64 0 R 64 0 R 64 0 R 64 0 R 64 0 R 64 0 R 64 0 R 64 0 R 64 0 R 64 0 R 63 0 R 63 0 R 63 0 R 63 0 R]
+endobj
+
+13 0 obj
+[63 0 R 63 0 R 63 0 R 63 0 R 63 0 R 62 0 R 62 0 R 62 0 R 61 0 R 60 0 R 60 0 R 60 0 R 60 0 R 60 0 R 60 0 R 60 0 R 60 0 R 60 0 R 60 0 R 60 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 59 0 R 58 0 R 58 0 R 58 0 R 58 0 R 58 0 R 58 0 R 58 0 R 57 0 R 57 0 R 57 0 R 57 0 R 57 0 R]
+endobj
+
+14 0 obj
+[57 0 R 57 0 R 56 0 R 55 0 R 55 0 R 55 0 R 55 0 R 55 0 R 55 0 R 55 0 R 55 0 R 55 0 R 55 0 R 55 0 R 54 0 R 54 0 R 54 0 R 54 0 R 54 0 R 52 0 R 52 0 R 52 0 R 52 0 R 52 0 R 52 0 R 52 0 R 52 0 R 52 0 R 52 0 R 53 0 R 52 0 R 52 0 R 51 0 R 51 0 R 51 0 R 51 0 R 51 0 R 51 0 R 50 0 R 49 0 R 49 0 R 49 0 R 49 0 R 49 0 R]
+endobj
+
+15 0 obj
+[49 0 R 49 0 R 49 0 R 49 0 R 48 0 R 48 0 R 48 0 R 48 0 R 48 0 R 46 0 R 47 0 R 46 0 R 46 0 R 46 0 R 46 0 R 46 0 R 46 0 R 46 0 R 46 0 R 46 0 R 46 0 R 46 0 R 43 0 R 45 0 R 43 0 R 43 0 R 43 0 R 43 0 R 43 0 R 43 0 R 43 0 R 44 0 R 43 0 R 43 0 R 43 0 R 43 0 R 43 0 R 42 0 R 42 0 R 42 0 R 42 0 R 41 0 R 40 0 R 40 0 R 39 0 R 39 0 R 39 0 R]
+endobj
+
+16 0 obj
+[39 0 R 39 0 R 39 0 R 39 0 R 39 0 R 38 0 R 38 0 R 38 0 R 38 0 R 38 0 R 38 0 R 38 0 R 38 0 R 38 0 R 38 0 R 37 0 R 37 0 R 37 0 R 37 0 R 37 0 R 36 0 R 36 0 R 36 0 R 36 0 R 36 0 R 36 0 R 36 0 R 36 0 R 35 0 R 35 0 R 35 0 R 35 0 R 34 0 R 33 0 R 33 0 R 33 0 R 33 0 R 33 0 R 33 0 R 32 0 R 32 0 R 32 0 R 32 0 R]
+endobj
+
+17 0 obj
+[32 0 R 32 0 R 32 0 R 32 0 R 32 0 R 31 0 R 31 0 R 31 0 R 31 0 R 31 0 R 31 0 R 31 0 R 31 0 R 31 0 R 31 0 R 31 0 R 30 0 R 30 0 R 30 0 R 30 0 R 30 0 R 30 0 R 30 0 R 30 0 R 30 0 R 29 0 R 28 0 R 28 0 R 28 0 R 28 0 R 28 0 R 28 0 R 28 0 R 28 0 R 28 0 R 28 0 R 28 0 R 27 0 R 27 0 R 27 0 R 27 0 R 27 0 R 27 0 R 27 0 R 27 0 R 27 0 R]
+endobj
+
+18 0 obj
+[27 0 R 27 0 R 26 0 R 26 0 R 26 0 R 26 0 R 26 0 R 26 0 R 26 0 R 26 0 R 26 0 R 25 0 R 25 0 R 25 0 R 25 0 R 25 0 R 25 0 R 25 0 R 25 0 R 25 0 R 24 0 R 24 0 R 24 0 R 24 0 R 24 0 R 24 0 R 24 0 R 24 0 R 23 0 R 23 0 R 22 0 R 20 0 R 20 0 R 20 0 R 20 0 R 21 0 R 20 0 R]
+endobj
+
+19 0 obj
+<<
+  /Type /StructElem
+  /S /Document
+  /P 10 0 R
+  /K [78 0 R 77 0 R 75 0 R 74 0 R 72 0 R 71 0 R 70 0 R 69 0 R 67 0 R 66 0 R 65 0 R 64 0 R 63 0 R 62 0 R 61 0 R 60 0 R 59 0 R 58 0 R 57 0 R 56 0 R 55 0 R 54 0 R 52 0 R 51 0 R 50 0 R 49 0 R 48 0 R 46 0 R 43 0 R 42 0 R 41 0 R 40 0 R 39 0 R 38 0 R 37 0 R 36 0 R 35 0 R 34 0 R 33 0 R 32 0 R 31 0 R 30 0 R 29 0 R 28 0 R 27 0 R 26 0 R 25 0 R 24 0 R 23 0 R 20 0 R]
+>>
+endobj
+
+20 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [22 0 R 31 32 33 34 21 0 R 36]
+  /Pg 120 0 R
+>>
+endobj
+
+21 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 20 0 R
+  /K [35]
+  /Pg 120 0 R
+>>
+endobj
+
+22 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 20 0 R
+  /K [30]
+  /Pg 120 0 R
+>>
+endobj
+
+23 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [28 29]
+  /Pg 120 0 R
+>>
+endobj
+
+24 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [20 21 22 23 24 25 26 27]
+  /Pg 120 0 R
+>>
+endobj
+
+25 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [11 12 13 14 15 16 17 18 19]
+  /Pg 120 0 R
+>>
+endobj
+
+26 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [2 3 4 5 6 7 8 9 10]
+  /Pg 120 0 R
+>>
+endobj
+
+27 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [37 38 39 40 41 42 43 44 45 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 120 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 120 0 R
+  >>]
+  /Pg 118 0 R
+>>
+endobj
+
+28 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [26 27 28 29 30 31 32 33 34 35 36]
+  /Pg 118 0 R
+>>
+endobj
+
+29 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 19 0 R
+  /T <FEFF0045006C0020007000610074007200F3006E0020007900200065006C002000720065006D006500640069006F>
+  /K [25]
+  /Pg 118 0 R
+>>
+endobj
+
+30 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [16 17 18 19 20 21 22 23 24]
+  /Pg 118 0 R
+>>
+endobj
+
+31 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [5 6 7 8 9 10 11 12 13 14 15]
+  /Pg 118 0 R
+>>
+endobj
+
+32 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [39 40 41 42 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 118 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 118 0 R
+  >> <<
+    /Type /MCR
+    /MCID 2
+    /Pg 118 0 R
+  >> <<
+    /Type /MCR
+    /MCID 3
+    /Pg 118 0 R
+  >> <<
+    /Type /MCR
+    /MCID 4
+    /Pg 118 0 R
+  >>]
+  /Pg 116 0 R
+>>
+endobj
+
+33 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [33 34 35 36 37 38]
+  /Pg 116 0 R
+>>
+endobj
+
+34 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 19 0 R
+  /T (El conocimiento es una espiral, no un sprint)
+  /K [32]
+  /Pg 116 0 R
+>>
+endobj
+
+35 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [28 29 30 31]
+  /Pg 116 0 R
+>>
+endobj
+
+36 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [20 21 22 23 24 25 26 27]
+  /Pg 116 0 R
+>>
+endobj
+
+37 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [15 16 17 18 19]
+  /Pg 116 0 R
+>>
+endobj
+
+38 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [5 6 7 8 9 10 11 12 13 14]
+  /Pg 116 0 R
+>>
+endobj
+
+39 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [44 45 46 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 116 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 116 0 R
+  >> <<
+    /Type /MCR
+    /MCID 2
+    /Pg 116 0 R
+  >> <<
+    /Type /MCR
+    /MCID 3
+    /Pg 116 0 R
+  >> <<
+    /Type /MCR
+    /MCID 4
+    /Pg 116 0 R
+  >>]
+  /Pg 114 0 R
+>>
+endobj
+
+40 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [42 43]
+  /Pg 114 0 R
+>>
+endobj
+
+41 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 19 0 R
+  /T <FEFF00BF004500730020006C006100200049004100200075006E0061002000680065007200720061006D00690065006E0074006100200064006500200063006F006E006F00630069006D00690065006E0074006F003F>
+  /K [41]
+  /Pg 114 0 R
+>>
+endobj
+
+42 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [37 38 39 40]
+  /Pg 114 0 R
+>>
+endobj
+
+43 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [22 45 0 R 24 25 26 27 28 29 30 44 0 R 32 33 34 35 36]
+  /Pg 114 0 R
+>>
+endobj
+
+44 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 43 0 R
+  /K [31]
+  /Pg 114 0 R
+>>
+endobj
+
+45 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 43 0 R
+  /K [23]
+  /Pg 114 0 R
+>>
+endobj
+
+46 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [9 47 0 R 11 12 13 14 15 16 17 18 19 20 21]
+  /Pg 114 0 R
+>>
+endobj
+
+47 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 46 0 R
+  /K [10]
+  /Pg 114 0 R
+>>
+endobj
+
+48 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [4 5 6 7 8]
+  /Pg 114 0 R
+>>
+endobj
+
+49 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [39 40 41 42 43 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 114 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 114 0 R
+  >> <<
+    /Type /MCR
+    /MCID 2
+    /Pg 114 0 R
+  >> <<
+    /Type /MCR
+    /MCID 3
+    /Pg 114 0 R
+  >>]
+  /Pg 112 0 R
+>>
+endobj
+
+50 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 19 0 R
+  /T <FEFF0045006C002000700065006E00730061006D00690065006E0074006F0020006400650020006400690073006500F1006F002000700065007200740065006E00650063006500200061006C002000700061006E00740061006E006F>
+  /K [38]
+  /Pg 112 0 R
+>>
+endobj
+
+51 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [32 33 34 35 36 37]
+  /Pg 112 0 R
+>>
+endobj
+
+52 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [19 20 21 22 23 24 25 26 27 28 53 0 R 30 31]
+  /Pg 112 0 R
+>>
+endobj
+
+53 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 52 0 R
+  /K [29]
+  /Pg 112 0 R
+>>
+endobj
+
+54 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [14 15 16 17 18]
+  /Pg 112 0 R
+>>
+endobj
+
+55 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [3 4 5 6 7 8 9 10 11 12 13]
+  /Pg 112 0 R
+>>
+endobj
+
+56 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 19 0 R
+  /T <FEFF004C006100200069006E006E006F007600610063006900F3006E0020006E006F0020006300610062006500200065006E00200075006E0020006200610063006B006C006F0067>
+  /K [2]
+  /Pg 112 0 R
+>>
+endobj
+
+57 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [38 39 40 41 42 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 112 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 112 0 R
+  >>]
+  /Pg 110 0 R
+>>
+endobj
+
+58 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [31 32 33 34 35 36 37]
+  /Pg 110 0 R
+>>
+endobj
+
+59 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [20 21 22 23 24 25 26 27 28 29 30]
+  /Pg 110 0 R
+>>
+endobj
+
+60 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [9 10 11 12 13 14 15 16 17 18 19]
+  /Pg 110 0 R
+>>
+endobj
+
+61 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 19 0 R
+  /T <FEFF0050006F007200200071007500E900200065006C00200063006F006E006F00630069006D00690065006E0074006F0020006E006F00200070007500650064006500200065006A0065006300750074006100720073006500200065006E00200073007000720069006E00740073>
+  /K [8]
+  /Pg 110 0 R
+>>
+endobj
+
+62 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [5 6 7]
+  /Pg 110 0 R
+>>
+endobj
+
+63 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [43 44 45 46 <<
+    /Type /MCR
+    /MCID 0
+    /Pg 110 0 R
+  >> <<
+    /Type /MCR
+    /MCID 1
+    /Pg 110 0 R
+  >> <<
+    /Type /MCR
+    /MCID 2
+    /Pg 110 0 R
+  >> <<
+    /Type /MCR
+    /MCID 3
+    /Pg 110 0 R
+  >> <<
+    /Type /MCR
+    /MCID 4
+    /Pg 110 0 R
+  >>]
+  /Pg 108 0 R
+>>
+endobj
+
+64 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [31 32 33 34 35 36 37 38 39 40 41 42]
+  /Pg 108 0 R
+>>
+endobj
+
+65 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [19 20 21 22 23 24 25 26 27 28 29 30]
+  /Pg 108 0 R
+>>
+endobj
+
+66 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [7 8 9 10 11 12 13 14 15 16 17 18]
+  /Pg 108 0 R
+>>
+endobj
+
+67 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [0 68 0 R 3 4 5 6]
+  /Pg 108 0 R
+>>
+endobj
+
+68 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 67 0 R
+  /K [1 2]
+  /Pg 108 0 R
+>>
+endobj
+
+69 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [29 30 31 32 33 34 35]
+  /Pg 106 0 R
+>>
+endobj
+
+70 0 obj
+<<
+  /Type /StructElem
+  /S /H2
+  /P 19 0 R
+  /T (Scrum es un calendario de manufactura)
+  /K [28]
+  /Pg 106 0 R
+>>
+endobj
+
+71 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [19 20 21 22 23 24 25 26 27]
+  /Pg 106 0 R
+>>
+endobj
+
+72 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [5 6 7 73 0 R 9 10 11 12 13 14 15 16 17 18]
+  /Pg 106 0 R
+>>
+endobj
+
+73 0 obj
+<<
+  /Type /StructElem
+  /S /Em
+  /P 72 0 R
+  /K [8]
+  /Pg 106 0 R
+>>
+endobj
+
+74 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [4]
+  /Pg 106 0 R
+>>
+endobj
+
+75 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [76 0 R]
+>>
+endobj
+
+76 0 obj
+<<
+  /Type /StructElem
+  /S /Strong
+  /P 75 0 R
+  /K [3]
+  /Pg 106 0 R
+>>
+endobj
+
+77 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [2]
+  /Pg 106 0 R
+>>
+endobj
+
+78 0 obj
+<<
+  /Type /StructElem
+  /S /P
+  /P 19 0 R
+  /K [0 1]
+  /Pg 106 0 R
+>>
+endobj
+
+79 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /MJFNAB+Georgia-Bold
+  /Encoding /Identity-H
+  /DescendantFonts [80 0 R]
+  /ToUnicode 83 0 R
+>>
+endobj
+
+80 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType2
+  /BaseFont /MJFNAB+Georgia-Bold
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 82 0 R
+  /DW 0
+  /CIDToGIDMap /Identity
+  /W [0 0 1000 1 1 833.4961 2 2 689.9414 3 3 253.90625 4 4 648.9258 5 5 657.71484 6 6 520.01953 7 7 353.51562 8 8 397.46094 9 9 663.08594 10 10 571.77734 11 11 833.9844 12 12 635.7422 13 13 512.6953 14 14 1015.625 15 15 595.7031 16 16 344.23828 17 17 715.33203 18 18 531.25 19 19 595.21484 20 20 684.08203 21 21 328.125 22 22 1023.4375 23 23 685.5469 24 24 676.7578 25 25 393.0664 26 26 701.1719 27 27 648.4375 28 28 571.77734 29 29 346.1914 30 30 566.89453 31 31 635.7422 32 32 645.5078 33 33 631.83594 34 34 576.66016 35 35 721.1914 36 36 689.9414 37 37 548.33984 38 38 445.80078 39 39 758.3008 40 40 679.6875 41 41 548.33984 42 42 562.0117]
+>>
+endobj
+
+81 0 obj
+<<
+  /Length 14
+  /Filter /FlateDecode
+>>
+stream
+x�������?��
+endstream
+endobj
+
+82 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /MJFNAB+Georgia-Bold
+  /Flags 131076
+  /FontBBox [-116.69922 -216.79688 998.53516 756.34766]
+  /ItalicAngle 0
+  /Ascent 756.34766
+  /Descent -216.79688
+  /CapHeight 692.8711
+  /StemV 168.6
+  /CIDSet 81 0 R
+  /FontFile2 84 0 R
+>>
+endobj
+
+83 0 obj
+<<
+  /Length 1194
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+42 beginbfchar
+<0001> <0055>
+<0002> <006E>
+<0003> <0020>
+<0004> <0053>
+<0005> <0070>
+<0006> <0072>
+<0007> <0069>
+<0008> <0074>
+<0009> <0064>
+<000A> <0065>
+<000B> <0044>
+<000C> <006F>
+<000D> <0073>
+<000E> <006D>
+<000F> <0061>
+<0010> <006C>
+<0011> <0043>
+<0012> <0063>
+<0013> <004A>
+<0014> <0054>
+<0015> <002C>
+<0016> <004D>
+<0017> <004C>
+<0018> <0075>
+<0019> <0066>
+<001A> <0050>
+<001B> <0071>
+<001C> <00E9>
+<001D> <006A>
+<001E> <0076>
+<001F> <00F3>
+<0020> <0062>
+<0021> <006B>
+<0022> <0067>
+<0023> <0045>
+<0024> <00F1>
+<0025> <00BF>
+<0026> <0049>
+<0027> <0041>
+<0028> <0068>
+<0029> <003F>
+<002A> <0079>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+84 0 obj
+<<
+  /Length 11716
+  /Filter /FlateDecode
+>>
+stream
+x��{	x׵�3�}��b�H#����bK�1X`����X��
6[;�@HҘ�:!&dOp�4!@�n������)�k_�4�K�޴I�M��M�4�/ ���H6fI��������s�=۽
+�$�7�f�n��Wx�a�������8
 �����mz�/�
 �৷������b�@|˖�^�m�y���޲c�Ɵ_����>��w��Py7���{� y�4
+P�`w�����@�@Q=804,:(iH&
���]���V��e���v�@t�����⇪�0�6���	Ϸg��F����	��H�������>tmD�(
7�(��� JB��:`'���za-����z�
��M�k�n4������M�ڈ��8
�� �����!#2B34�Y��-E��m�zn��HB��>��W�ex5�at�x	���@?CV$�[���r�_��"����c�h��xE�y�a�vD������yb��ڮ�D��2����P0����y=n��dG��f���MF�^��5j�R!�I%bE�([��}�"	؜Ng:T|�^��%=��,�.kd�������+����Y0d���F<�)h�]�Yd���fAW���ƥ�f-
}==�&����l��|a�S
+y��/�\��5(BA�6q��P�$DS��
RU(����'���e��z���kt:��`V�fb���s� ��@���↬D��ݚM�f�{*xn��	6��}\_��,ٛO
�Im��|L῞-l��p��Ö%=lj;�av���dIט�Y
+fe
�w:�ٲ���;SYm ��5f�o���c���,~���]�=�։��t�$d�R\jKOc(�ڶ(D%|(X��Ȁ��mx�m���6v�P����B��.����j56���R}�}x�,�i�&��/�Z��������t���`Mw�jz�x�CAԺ��
#��6�
+�>[�S,i]ޝ��d1-�dO���fay7�%<���_
c��q�Q(��y�WV�9v��E=�ԟ./�-��=��
�M\S��X�6����NL����47v��ul0��f��;�z'���e��Ng�-�6d�4-﮷9���ΙW�J�b,­�1�������f����6daewږMvuc��;m+|cA��>W�N��l�<ꯞeOCt:�t�H†PЙݿ�����۳���,уk���W��35��{8g(x��4f����ڤOm��"�?��/�g�
ݤ���o�&l$��
�f��@(�-���:��YQ�9[]���Y���[��.[�ͦ�f��PR��A�Kq�[Ɗ[	�)����e�����5�gh�`W��"z�ӡ	Xl�Yg�)��IO���	h,=2 �[��UA�Mmm̢���N�*�34
�A����\�c�Z���&vKo_���˻���4�faE��‚f�i�,؟N׆& �ǡ�q��l��8�6a��4�M��`+�%���˺��m�dc��t������F�S d�,�,���֒"������La��٤-鱱������q��	8wE
�+�ł	F$=�	��S���9���8'�Lۜ��������~@`A���	�&&(��+o�"<"a�֩�8�Nb�x�.�o�/*�}������!��	s �ḑt�/��t�oJ&��%�2eB"�;y&�D��t��L�:U���L�&3m�L�&�~�~�\����l�eґ0R#��p�P����U��*�q����/��)���2�o�/��=�浶�`��ֳ�N���f%6��F����sￗ�D��?�M��
+����U�y�����#v��jA��,���&�J[�"}��7tȦ|���,@e�I�2A<%]p��l����גd2)W%�B[�SA������L=���d.:�	d�&s�
��5�S����"�U-@��ƹ�5�h
&����K5��1&�Ģ&c��hC�i}\mԵ�������`����QKLǖ�V]‰��P�N������M�w�b���C�7m.1�t�����ɖB"VTx�&�N�����~K���u�Ѽ�TIԤ ��S�N�|x.��ɸE��b7y�fw�Xg�F�πg�Cz<n�hܤ� x?C��w�{ăb�X,R�z���=�I��	O<Z��יh��"���^��'�c"��د�h@��|憩?EtÔ�������55|B��3�S:A��S�e����x_twT�I���̝t�\$��[&�Ș�KU����2^`�9�sB��v �
�1~�;����U�k��?����o�:���C�S�^ݷ��y�WKW9��"�Wm�k�K>Sw��U-�ox��&�K6t�����g�(�9`ţ�k/t�����㵩�֬��D���y`ݗkꅀ�E���H^�r�k�u���򴻳|��f�Mq�c�y��%�?�k��&�v����������'4'�1�(s�>i>SW���d��p
+�sQ�u5EID�\�2�*�8!���G'��U;�C�b�w?���L��m��
+K��5��a�~��.z�����tra{�إ�$���߿��̞�=W�T�%K�XQ�"��Q�JK�XUDÒ�/��zW<��s�7��m�tR�r��
!�K�I��Gt�ב�I���ae�1@s��7<��4n
�P%x�y��'
'�T����`Gv>SبXEM
���8U�������=b���2�_n�X�D9�/m�S��LV�FSn��#]%���lW���6�d1W�%ꢹ_p*�]څ�mqv��ijb%X�
�2���R�f�~CK[�q�~�j��Q�$:Q��+ ERi�W���3����$�(��1~*��y���"gH�EMij��֨���%����a7����\�{��e�X�(C�����&����q�ND���ok۾�L�K�D
+0CG2L%e��BA�$��5f��aҤ�@'������(oAR=*�JΔ�E��(���>�c���jjb|$��U�UŴN���,���J>���#�'P
+������֣������#>���s�ܽA*���5����J�N��(�H��A�͕"�x��h�v��(���D��t�(������{����
�U�GS���F'y��>���<���2�*TUV1W1ݣ�\�ĢqQs^��D�Y�����������	��b7.��	��dh���7���M��4���U(��7Q�ǂ�=u&=�b�?~�[�Q��V��\m1��yu�mUUu<�/��oS;�����D���������Y��~qٸ���A�`�҇�&�i�
ˋAM�v{���z@�B��(���qA�`1�x�U�\�@��YX*4�O����������)՚j|m�ݧ��T��=����T�_Uj��BK�Q��w{˺���yæZ���j>�r�+?�܁Ru����ǾPН
�~�!�C��a��͒��u�"�%�-d�A�q9�)��$ȉ��ʞ��J�?RrTqT�\�K���)০�5i�
�DLqWn~g�"��K�qvn?z��=��b{h�����4ۇm�P]m�����&�������+��r:���+W�#�Ds�v%�Z%Ѿ��W;ߺ���/_C�/�F��kf
_#s`|I�s�H��e&�
��H8�鉐e4/�2�W�~*F3F�Ź�
+5��G*��}钎_�镎��K��쎻t�d4JV.l^i6�j�yբ�U�Թ���4�v655|�_hiZ�4'�>���te���5���q���A���ȟ���'o��c�RK��7�M�U���f~SxwhT�P�x�S��b��c���F	�3�Wp6h߸=HG)`�K��~�;�`�v��O{�w��;��8�b�'���E"�������f��ikj�)~���W��5S؃�f���/y��h�Y�]բi
+��Rk��\Y����2�:���Jc��}q���ۆ��t~�U�6�ݡB��K�+������l;|�c���Җ#ɚ�FG�7�8=k%9�l�'��Љ�t'��vz����7Rt�*��ݴ��v:f��7 B���F�d�SҰ���֧��G�mȆ�mO����$*�j
+�6��:Q�E����>�/)S'\���Ї2��2��'=�
�@Or�Eww�*��y��_�|�W,��l
+�&eKc��^�*�f�g�}GWU3e:��ظ]����.�pjף����+ۗ:r�TJ�LII���y��`�kr�w�T���L_��*f+���/��Ö���ݸ��b���=��zBʂOM@��;�)2���I�i�T�C���Ԭ	o�l�ɸ/rs�"u$h���Y�Y�[n��-�*�W�(��щ`���!M��L<.⼟��y�I�2���;���h^��O�/,�?�B�X.��
+N����w@X�LQ}n�/h	������G�jx��(�btW��ԩ��
7Ůz7��}�������\����u��<i�i�5�V���'�z�-\�"�ٵ-���z'��7W&�mF=���
+kMƻnX��c�:g��Dz��-*�r�6�1��-�N�:�f77z�
�DB�*)���#6��y�M!��C
Ⳃ_�H:�
+�C5�L1x/`��S6k��®�dћ�g����akDϚ�a��Cc-5�L�P]ޅZU	���w=����<�s��SPے�F�"�:SD��.��*I۶����d�%{�j]RM'�鴤�
%�r1���ul�C}L�i��t�?�|J�5����c1~���t����Zs�y��J]"ÈY#�0kH�˿߿q�ЖM�q|Ϸ��u��q��o7�k��kVv.�����g=2>��l��q'rO�*�\�G_�KK[��g*�E݅m�|2!�L2�����ƕ�`*+s?���#[�5؅�etB��;��x�y@�B&Lp�5�;,8�9���f�u�瘶K���(�j\���?n˴�����-a�3�b�̪����eM��ڵ���Շ�,�^\��������<��k+������@;�s��!��ڤ�G����b((;^N���ҕ�D�*}9&�Ui�G�#�$+�	�80@0!ӭ�M�|f*S���Y�M31�>���L����+P*1�ӬwB�S^��̫ۖ����w��[�9]#�cn���w�����2E����Oϯ��s��6�&wT��/�$��N����F �.:,Kƀu����)�<�"5��ZtJG,�1�}ڣm������� ORA'��e��):7�a!.�Ec|&	�B�@W���M�N����]��zo��{]�+ݪ/���l�]�R�u��׾.���q'h��jo'R:�6�)��k�#����76�z`�Z�m��
+J��oR��Մ���rS<�H�cK�
+:�R\d����j�Uo���2�����ֺr���u˒��sp�ߊA<�쩖7�Zԛe{e"��W�]&7�dr_��_~�Fa�h
+y��/Ҡ����'����:. �ojn/{��M��r  C2�&�����天��xI���:sM��y�F�+ER�J2���I��'s�:s͝
��oߩ�(�Hͱ��:\�3�$CPbN~N����8�ڨ���[eͦ��N��6k�q�
�r;˖w��h=*����mb�,�2�ZK;��j*�)�������.>y�m��lo'�V����7�+-��l����O����P�Uծ�zwD'�I@��":^hI��q-�� ��q�{L��M2:
P�Q>PN�ZG��Q�seE���)A}୅ȫh��Z�8������r�5n��Ϭ��
+v5[f�Y���G���Y��|M;���e�{n^��
+�(W%��k���
+=;�T�W��߈V��mJ�CP����NP�2ϸ��2�0n|�D�8�=,7���x	�y�(����0I��	>ro��t�*5�W�	�MNe��d&6�
+����H8 D����e�Sz�i��s�ka(�v��'CҊ�Ut��o�Z-��������R�
f^������[e�FE�X5�͛7,�6W\��₋��V)�1NB�'�
+�Л�o�#b$�fX,���r>�֘��F�T�����AqI\jPC�IZ�G]��DN>3����'c1��3�����)�9��N1R�T� \�w�b��џh����cM�<��?�KKڙ\F��/��`a���n9(�h�3Z�;�k��c��\GY@.�\r	�n����=��v��� �\�k�E�?�ɠ��H��R�=Ns�.�5@�l�����-��0G���3�f,�8`�dN�xq"��SgD��D����pu���k~߰�kX�\�.�їYJ��3�AzD�o����_��o��xu4������fUH�
+��*�d�W~������ms�Q�J��հ>Yס�N����j�j
�Yh�6y�qd���R�E��+a���a�c(���+�S>����hJ��ؤ��B:�~&Q�)�%��(�3)l�g4�F��5�h�01V�6Q�X��4���������9Lm�\^b1�V��l�:������%�L*�t�AW�f�~�rW���_n���S���H������,�_��-\U�7B�y�����!�	 ��h@Y�	Ta�Mk�D@��!i�"	��8�9��o8�A5Rc�(��j�����[/됍�H��'eI#2CAM�	
�ȂٍM�tZ1Q��f3����˼Z���`Y���ܾ��H7�W��}e�~�>�P3��=��h]�.�a[M�޽�G._�t*3���ޙ!�~w�?���������hXٝlۛ���bg�2�n�S���,v�����rP��V#W�e���H��!)Y)ta]�nPG�tZ���!�3�������O�s�s���mɌX$V ����aij�*,-��e��*�ΐ2ќ�Y_�G�|����B��9Xd
+h���4v�Ą���Ѭ��jҙR=�͝��7ݹ��N_kLW[\�b���ѵ
2�+�Th���#�x����:���|\�҂�&q��-���w����Hg���D�%�|M�K%�@
�&)'�	����2X<�����+�o����d�G��m[m7�n
+���7� ��އKi�<<N�R���X�T���VR����t����P�j��sQ���p*S�R�<��@e�!�k����]�h���Lv�+���DƪoaП5��,}���f��{��NV�Ԗ�#�������Ԣ�{���[�t?�y�B�R�c�ܞ�E��LC��x|ayY�a��j��`�H�|�������]��%�Z~/�D�#�ۈ@��h��R�F�aQ�
G�HR�>��-
�jDj��I��O��AMTg��8^�J(��ѸD��ej�Zj;j�
+�
�L�P�`E��'�'�gj�!���
+҇�D���i�e
+V-��v�E�4�������Q��,��a�?1����/I",�Y4?�[}_���ЅKr[�z�\j׷1诚%g����zt�o���膧�2�*������
+���9��:�uJ/…,:Y��(]��9e��l��|p'�RXI���ϐa,�up8�N^n)/w����k���7ۥ������L��l�x���"����2�l�8k�'�t��U}���+X�n���"Q%E�׬����'����h��b�P����8m��(|&6��u�rE�L�$�f��sO9֍�n�L�Z��h܉fb��U_������lI��=Ǖ�M�֞-��寨ki�x�x�?�~�u�}���v�Ҽ�8�D���,-��i�mM��~޻���}owp�V�v�n�J��ee��� ��9��mwDb���wl)�͍�Z��&UD���`�,��h�A�Z�q������(��k-u�Rg_�GPt���d<\�t�XԾ���2B&�%�;����#
��Cz���q�F[�?�Bn�щ@�^S�Q?XO�l#uGCG+�[P���O�3S|.Z�����q�&���	1�N|L{y�Y,�5WG
"����&��ܫ$�x̼��M19��{@o��n�6x�-�����ǯ�{^��o]_B��r.{Z#!s��z�Á\�?R��7'[
z%!���G�M�RJG�B��4��N�%�~Mx1��p������U���O��2��
+���dD�
��N=�Qr]�@�H	U2i4�5��*��jL`�ͬ�4���d���g�B$�{f���u%.Q�T	�2���'���s_��:>P̸ഡ7�39!�"�y��XԤ7P��]U�G1�l�e&cN�j_:4��}�\��3���$����pO���u���������v�𵡣�;����U�9��v=�\iZ�#��zr��0�,i�i��$CG:H��	8j��eWi���8	V�HG��T�d��]@�̇�(��0-�&���"�%�_��m5xlVo�z�w�aʬ6����NH���������/Ezw��i	 �� n=
>�ד]�2��ɵIC�
�l�Io�X.��o���Q�4�`Y����+�Z��k�/f۸�L7����>��VdB������j�{|�H*T	.+U�	0 ��(�����!�>DP��Ƃ���0��#��	(82
�R�9��dΝ������Ι�﬽HX�>f-*=3R q���Hɾ�I���%�7��:W��j����޷ծ*�ؗ8r�W_n��ԠE^��@uTΆ���VS�F�u���'M�}���g�r�X	`kr���<������=ԕ7zN��ѣ��FO{�iC\e�.m.���Y4s�b:l�6�����z���79tAv�F��M�=��z�!��]�Si'�ӕ���ΎWz�J�n��ߒ�}J�7&��K��e
Z��C{J%��i%�$I�H�C
+�wm�()S$D��ל&�7�������"��~"�c���z����%�^^�0W*J�8� >��
+�
+qǢ�\uB�#��߉�����%r�H|�H��j�J���uѨ�v����@�Ҟ����η����60hz2�H� ߆y�/��M!ds�B�P� )&C�;��!�"��fP�q��6�YP�Ҧq�4b��H]0_�����;��'9΅�K��/楯?f�[I�C��sc�:sW`��?x�lWP�?��r�m��.���C�I�鄆�
�����+q.�
r3�ΉP�@��"T�*`Lx��b�2v9Zi_�����Xx}���BY�)��5��e��>WZ,u}jł�>�e�������q4P�޶H*�S�����o�U��F��5��3����R:� XH�&�!�50P��KG���B�k:4�(_֘aҘ4��5b^\/�Sb�6�3�k����7���A�B^�p�%�;��2Q\7��e���|�bue�kK�-��y��yF�w�#��A�m�t��E��&3�Y���ZJ��JЀ���?L��tB����/�*!���
H
+~| #\�)�v.aB�}���;��=\��6�Tt:�(ew/��>���&�~����i����oI���?��/k&;.|6� }X�C@ �3dU������I�p��l�'(�ɻ/A�h��@���
���yp�X#C3� 4��
"4SVh����
�7
Dg��JB3����q��)�6А���(\��δ��k�]<��?�'�H�s`}Rj#l'�z�v�{�A ���A�LOR�E�����yѯ�m�o�$y��n������.�G�ci����\6&{Q>(?��V�P�T�@�KU�ڤ���az�6��vX���=��۰�pވ�
�/���mD@� ~.�7��k�:���H @-D1��
j
ƽ�,�X!.�b @�E��`	 ��E��`)����0v`���"�@�J�0
R�-�$,GE�������f�C	z�K�c�j����^����MV�D���K��1D,�8D�`���0.q�+1��%E���r
V
�x[���$��0�g�0�����p�c~���
�#.�c���������(pK�l��%�E��d�[��,B�{؆�W�l��*yQ�K1>��l�b|$�`Fh�A��/
+���������%�!�^��S`�.ưt��s�,���t��9�s�+��_9��v�Ͱ�az��av�f�
+��^Q;��
+v�{W��T�	����Bi�0��U��r� �u���7ȯ���lsY���(�!q`a)l���``
B�A���
+�*������a+l�-0C�[?A?�=�W̿`�5�)�,���b�
+*o�T
+-0��@"�D��G��]�+��m�뇭0[�Y��(n3�B)��"�V���]p=��h�>׮��?Z
�~�aa%�p+̿�(ex-�;��"s(��v
+mZW@��S��EvB?,�,��+�_j�Fb��
�)(�W��&aLQ��>(��!`=����M8u��}wͩx��a�0b�{��6–O�����n���[X��j
}B� l/RP�Za���0�0�/<�.��n\�]��a+���
fg�V;���ϣK�c�w��Lby�xo���k�~I�.�k�`J
+�`v��"�@k�^��
A��Mi�Ͻ�� eWK&��0
+O<~?��������
+a��ǩ��a��B�����B�x�}�X[�� l�}��R1��.��O5}��
���=0 �쾪��Z����j9S���v�>�=�캆�����H̍�O�R�5��/�����a�`��nsymg���aCQ�]��Zm�r�j|/��j
5�j��T5���S�T͕=�٦��𥚂E���mŔ�켪�RM����kP1��M�g��j���)�$�mp�5���V
ˀ�8�t�jh��P�atB#t�� 	�@x��A�RH�>HA;,�r��A��
+.X?���2�\W��w�@��#����A���&�_��.e^y]ʼ�����>���d��k8����t\�(�(�轏>��3�i��>&e�B9��~�����>N=3.e������1��K���~��b����,���C�>��C��P�!�����H���@<��L�y
��W�׳�?��g��x򬜎��>z���	�Apf������O<������s��:�#�_п :��:��>��Ss�)�9�ǼyT�<y���գ[���Q�6���?@���й�������'��?�3�Oџ~bg~���KZ��KZ�/i����e4���D�鯜&~}�~I˼4�e�w���x���P9sh���k��)g�J�;G���F���
+��\��s�80�1?�2O���s�H����7Y��G�̭#,���!�o��EC@]��x��n��ũY����m�RU�Ū2��*�ޠ������Ϙ�~F��3�ch����Z�R�V��
+�X"U��H	�P2�A0bT��
������ ��(I����
�l󤌦Vʐ5R��Lgeu��ڵ(�G�кbQ6h����l4К�u�-�Z6d������	����5k�'�W�
+?\?���?z���U���@i��uEwv�4��b��4����������9���ޛ}�k:�����g�5��a���/BžÁ���nC���p�u( �0<S�g
+
�Á�@pkܧг�������m
+endstream
+endobj
+
+85 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /GVVWBF+Georgia-Italic
+  /Encoding /Identity-H
+  /DescendantFonts [86 0 R]
+  /ToUnicode 89 0 R
+>>
+endobj
+
+86 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType2
+  /BaseFont /GVVWBF+Georgia-Italic
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 88 0 R
+  /DW 0
+  /CIDToGIDMap /Identity
+  /W [0 0 1000 1 1 749.02344 2 2 471.6797 3 3 241.21094 4 4 670.89844 5 5 572.7539 6 6 297.36328 7 7 285.15625 8 8 572.7539 9 9 609.8633 10 10 589.84375 11 11 431.15234 12 12 879.39453 13 13 347.16797 14 14 537.1094 15 15 575.1953 16 16 589.84375 17 17 814.9414 18 18 461.42578 19 19 538.08594 20 20 653.8086 21 21 575.1953 22 22 701.66016 23 23 822.2656 24 24 561.03516 25 25 453.61328 26 26 383.78906 27 27 618.65234 28 28 562.5 29 29 328.61328 30 30 975.58594 31 31 527.83203 32 32 471.6797 33 33 553.71094 34 34 374.02344 35 35 642.08984 36 36 389.64844 37 37 578.125 38 38 559.5703 39 39 517.5781 40 40 269.53125 41 41 927.2461 42 42 603.5156]
+>>
+endobj
+
+87 0 obj
+<<
+  /Length 14
+  /Filter /FlateDecode
+>>
+stream
+x�������?��
+endstream
+endobj
+
+88 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /GVVWBF+Georgia-Italic
+  /Flags 131140
+  /FontBBox [-176.26953 -216.79688 1089.8438 756.34766]
+  /ItalicAngle -13
+  /Ascent 756.34766
+  /Descent -216.79688
+  /CapHeight 692.8711
+  /StemV 95.4
+  /CIDSet 87 0 R
+  /FontFile2 90 0 R
+>>
+endobj
+
+89 0 obj
+<<
+  /Length 1194
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+42 beginbfchar
+<0001> <0044>
+<0002> <0065>
+<0003> <0020>
+<0004> <0041>
+<0005> <0067>
+<0006> <0069>
+<0007> <006C>
+<0008> <0061>
+<0009> <0050>
+<000A> <006E>
+<000B> <0073>
+<000C> <006D>
+<000D> <0074>
+<000E> <006F>
+<000F> <0064>
+<0010> <00F1>
+<0011> <0048>
+<0012> <0072>
+<0013> <0076>
+<0014> <0042>
+<0015> <0075>
+<0016> <0052>
+<0017> <0077>
+<0018> <0053>
+<0019> <0063>
+<001A> <003A>
+<001B> <0054>
+<001C> <0068>
+<001D> <0066>
+<001E> <0057>
+<001F> <006B>
+<0020> <00E9>
+<0021> <0062>
+<0022> <002D>
+<0023> <0043>
+<0024> <0049>
+<0025> <0070>
+<0026> <0079>
+<0027> <004A>
+<0028> <002C>
+<0029> <004D>
+<002A> <004C>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+90 0 obj
+<<
+  /Length 13362
+  /Filter /FlateDecode
+>>
+stream
+x���	\׹7�̢��f�o�i�����XB ƶ��6`�b����Ɖ���)n��!�K��I��f��,�io�m{��&mS'�m���&m��o��M��%
��	�7���0����9�y��?����v��bl��؟�پ�ñ����i���{;��P=Pubh�ෛ2�{�vhh�ߘ4T�o�О���9E	@�	xj��~���
4T��=��!��<�p;p{������h�
��= [.��� �8�p��5�m���
7�
pP�!+M�;������o��ks�R����珲�����1��?n����/�vx��G�Yh�_���=h�
��fx�`%�]���Sp
+Fa
�z��Ch�|�Fh���r|��+|�������1~�q0���	ہ����C�%��p�!M����­0	�[�W��PO���B��60��!�;�+ǘ�w�u���(��;�^xǝ�f�
�n�'�)�?�G�A��@9@��+���5Ց�ʰP
+��J�>��w�8�YR�۬��h��Z����*�B.#	� ��6u��)���*��W�s����+��*9>֨�c钏�����90�Z��f��yhy3�f�
zfX�}�Qz`'�ޑ�5
���Z�f��r-�
+��H}�/R7�M��PΫ�����Pr-��y�e&xK:~�&��9ܛF;s�3}9��7�\�P0g�R25?���E��(C��r�Bz/�#�����|pz��Sl�P�@�����
+�›�B|L���!.Gz�f���^.=č�顾��{B�k懂9US�i״#�o�>�����|sn��7�xں�C����\���.���鱆��x�O�5��革P�
+�`~N��D��ُƙ�ɍ����yiR�������j�����@�zM�6�]�t�"vp�\�����U��۝#�����ʡ ־��	
��ov�}1���Ӿ�;�Pȡ��}9n��5�|�F�c0
+�ۢ��z�P���J�������_ ���3o_��_ȑ{��"[������k�?���~�|{��H���AGw럚�tƑk�|O���� �$�eMw����,$;��S4��H��� .�Ho[ᣟ�AW��k����G.�Ս��G�	���t�����6ģ��"{�
+�˅���T����܉��|����� )zrx*�^(1�C%'J���P�i@.ӔS�i�lH�s��S���9CS7������q�(u�g���%
+�J�Ͻ��@N�=����]�h���흽�\z|Q
+�9��"9H�i�h��JH�1.��q�dS+/@=
+r(�%��S�T2��װP�<��<vK��$v����wKW�S8�7��zzBS�.�>����vk�6?6�%AĖ͡)X������
�)X_�
+MAw�k��Ís�m�\7�?�#���n~p�G�r��{G~�s��"9��MAꇔ���Z���vJ=����)�
�s9����ٝ;���%�{.��Mwt禛.i"G�q�;��1o
+N����d���%9�ϧxW����c��YHO���20�xF��1R��7=��萊N�.$�i�Żz����l��]PT��H)@@����qП�_�HzT�]:���ҹ0��'d�!��2ɛ����%;,T�`2�kQ����%;�;�G�G�
+��RK�Z�w����#�x$F����H\�L���z�F��T�S�03���	c̛
�F ˼?�=��Ȏ2߫IJ
̈+�$��j�e�hX��T���_tU�-NLQ���r��LFs��V���޺�������}���U�M�h�}���{z_���wP�m�#ȁ��\��za�=���Jm�'f���UD�TdY�_z����Z;U_y˾��pHϿA�K�(�I�Zo6�<���(�eS��(F�O�O'�"F�N
+�������{��0G����1kP%��5��1:��`F@�1�������٪�."dg"B3!����cBv�2d1�\����[����T#nE��
+�Y�W,��eq����?��7���wO�>�r��o��̔Y�/5}�GK.���p��˵�><�������c�ے�0�Q�������F�� �����`����u�eض�y�8�<Tb�Z�M�c�):��x�c�=>����f��_�Y��z��;<���=:��y<�
+lċ!Üt�(��*��
�i�Ҹ��8n�	K� I�93�XBbH��W�l6"�c������������r��b�
+�A̋�K�)$��(\�\Qb���,N̢ v��=�iv��e��9�k�Q���A��'m_[�7z��[��bc��Ɠx���N��"��lQ>}��s.�q��R)����0=~�GF��ր�$����f��x*a9��t��8�k	K�Ҩ1j�.�vD?b��Gb7ę0�/#�"�TR��S���+bDC,3�̭��y�y�9k�i>j��S�ֶָ�әDjŨ�D���n�2m}m�;/�:��^vJ�|+Fct�a��R���b^op�ر z�C��1�H"%v��f��+!�^��ޥ�h-�B�����E�N̂Dwf(H���*N���W�J+6��gcf�+ZO����^��#kw8��Z�[��8�S<V]��k:���f���HRg2�Ŗ�e�q&nv����ؽ�k�޾��	��Tc
5l�V}U4�]�Z\g�g;W�bO�6���3!�Q��[�^�[��I���15����(u�"(uuMu<\3R�W�C���}�#�d) �j5�F�Jɗ�e�T����7�/�&�\�N�fH̾��}S��Y"��3ـ����ӳӕ�+��BĪ<gM�%F�`��E�̸�޴��ѡT�T�ޘ��@��ѳf�R�KMGb��+�Ȋ�X랤�dm��ʇ=�wG�ZY�w�N闱s���X�Z&���׬���-�[N
ۓ~ ƌf�+Vd.3G��ͤ�f\n<l|�H]nW|č�NL���Q�t���p��a�2�ˊ	m�	��X��>�0}�h�?��#x����b"1c�H��7t�U)���Yͦ�K���f?\�[��Hz��L�֪���*9�`
Z�� ����upWr�J)�dZ��ј���:o[�'�3~T{H7�?ZA|P�ԍ���S����������^��
&9K70&�dj2
}
D�)���`����$ð��'�ף
+3$f�ܗ��d�#YI�‚fe���Ѫ�8�w4iJan(�o���M��^Y~��5�ԅc��۝h�:�5�����hh�7���^�ټ/=}��={|V�����1�J)E��u�]�v���s�/�6�}�L>V�=�����R���*�聗Kw	�ERQQ*O�F��<�@�
��){4�۟�V��Q
+�!r�眗��@��J��L~!v�\�B� �F�>�;��g�d�L�ձ��
+`D��V������Ï_�㌕��g�m�L���`�\���������6�s����w�r����H5ͭPf�ZAa��VjkxR�2�MnE�/����W�qq�ٛW!�2��G�� _(*łS�/%�*FTQc�?��}%^�7���fo�X�-]��,��d`Xd`���R=]/��\]̒2���Ź�9i��S��^ЙD3�kSS���7{��%��u
.G'��� H����:I��@`_$/yY�UUB +I ���E�={MSD�7K��'�y��� ���H�L4L6=ziBf����z�g�f\�����vQ]*]Ṵ�@��Z�`��`-���+g��-֍�Q�������{��Bs���Zbe�S��o�6�=�����"m��T�#QE5b����r��E�B
+�hхR�"Z�b780�?�����#����d��1�����x��C���h1��s�'��=�����	AK},&���l���-1���g��de�I����\�$M�U�g�ny^�MF9������K+��EI�T���pz��{��[+�Đ7�0F��mܸ��P�c]i��fo��Jvoj�tS��\w���z&QU��9˜J��֖��c��o/Ȩ�'�{�
+��T�k��������J+&ע*F���cf��k�y�����s�y����|?��Ֆ����LX޴�,S�����-v;/	5/	us��2��p����)����x0�N�qwA��؅US�;�ō�Y^.���[V2!2p�I���|�b��[�)ẘ	d�-tfzvzC��?���+��R
�O���"M<�t��K��T�����G/M(i�hNK��ф�M��V���ێt�	���$o}E'�g�W͕T��6�|Gq�� ��О�ؾ�g�RQKk�{�-�$�
+%0�\~���E�HiݘZ!/�;ceE�E=E;��"��u���}vܮ�+�v���0AEE�Fq����-�|x��X��b3�*ُ���^�"�K�ʃa–n��+O��my�����:���V�f���FӪVK���F�;��l�}շ�&F4�6�_�՚�C���@<�I�J!gn��H��D�I�qN���S�M�7�ֈ�S��=�_g�_g��AJ���H-z�|�J�:V�z�����{E��N>�o/nY{��̷m�g<����ݳi�ߟ{sÆ]���z�u�~�{�ّ�]�B!�xNNAZ`*ys"�)+���%�ZIR�֮4Pm ]�nOoL�J����i��Jh�y����.ͭ�d}�#�x���◷.����z���V=��O�S�Ox9Ϸ
+��־V�����%;/��
+q�a�/�'*^oZ��c���3W�좮��C
+E��kE#K#2��C6Â��:0�����w���PϪ��GӷW�e%vk�[�M���Z�Q��
+�뾀ۍ���U�����
-q�d�s����������{�ǎu9�z�O%v􇝮u�X�H�����%�����X�Ꞇͥ�@j
��|��~��@
��!H���F�:]�F:dA1��8��_��eD@�op!�)E�f�]�"���nO7��R%%'��Q�i!���������v��ޙ���;b������(>K7	MxSSi�j��ع8���dٹ��Fw��:�V��4
+}�G��)d����V)!c��+1�c�:��kEB�-4u-ɔ�Ew��n�q\o�3���ڶ�����fGq�`�*�י���4����:�+��3)�Qh�܇S&N�ң�I��Tyq�+R�FOђLR�ԩ��1�BӹF��NivI��� .�AKV�0��0f4�xS��l�2]g5�U?���mN�)�L)��Nc���ޝ����F��c����.yc!Y����T����ʂTe�A����TpN�����@im���D�[�k���$B�h�����7�M���&����Ů�i����j�k��mذ}
c0��A��oHZ��N^���*-*�*Z��K��s�Q�&\���`05�Hҡ�E��
kH�4�9�aD��R1fF�G(Qe�l���c�'�
y�MĊ���A~Vd"3�&FF���7��� s�H��7ut�v��Դ?fj�ӫ��t�R���j��s6�Q[��2]+�뚈���^�G,q5���G}zY���h1��sg|-��F�N�>�~q���j�b�[��ީ��� 4@<��K�)F��jF�@�nA���}���g䙐"�<��z���N��`�GU�%l�eD���АX�\�u�ؔ�sMg��L�Z�|'�x<n����\��N�9K
#�t^������ʎ]�L�=`Rp�$�$Ŕ������B�HUm>�*h��G\���X����iG���Ͷ3�>�U����T�X�����p�P�s�:�����G��fo�)�k�G;�N��:��6���R��������)\��V�\mP��"��S�9�x��f���f�lv���ͻ�؜Ek-��X+�0��'�o�2x*�5��Dy��)��|m`G�RV1"�c��ъ�
+��w����k�z=}��%bt�J$�h��љD:�TI��"ٗ�2I̒D+m�$�®�D5�C!��q��D/�LԿ^z�wJNaA���?��UC: ��m�U��SA���m>�]yEWYxu�o��^��x�Ⱗ�]��Qk����'��&�L��?����{�����^k�k6�Q)L����sky��iuS3�E��X�V'�M�{3�ӿ	 땝���R2e�R�1Әm�n<^�Q�G���?VSY#O,S(N��F����Ԥjz*�jvT�\_y�F?T�����T���b�����)n{��ғ8�ɓ�����ɟ��᜾�󺡏-�YKL�J1��'�U�|��������(�ê}|����Һ���܉:�v���}��m2z|y���a��c�o
�4��'��[�5�u�:�>8��/����N�~�����*�2�[-�\�
+(�Im�TJO������_c��s'c:���\�v�wR<�󿑯#ރ.x)Y��߀�JNE�5S�](���T��둙�@f�_Ĉ��1��V�k��ظ�����xD$ELm���IƟE����������
��p۬�bw�%��(�$�b��(���N�^����i���#��3�7
+2{�W��WdF�N2�폜��p�|��N5
+.bE�{����\�w������˓��sx*+;+�9��*�:��*�ɪ������'�/(h�a�ڛo��k��R�{��M�m��9�-�<�qVZa�J�n�6��J��Q��}�
Eb���ӫ�߽��`���-��쟠
޾��ג�h�6��<`�C�:�F^��VW��l4v����*�쫺Qc�H�H��MF�d��p4��l�FUn�q��{&jH�eIB�'��r�Ls�1.�&��ї���p��6b�m��h��^���M�l"�9� =	!�)����5�(��p��츌�叧�n!�XJY��A�@3zK,!��}���}~o�\�X�6#5ⱨ���
c����2oy���&�^���ͽ{׵4gK�}b�`'as�m=tݲ��ѓ�Vu�[ICi��ɆD���V}��V�_�{�.�\o�+[�`m]ӹ��9w��vr�Ωe���i�
��{���Qb6/FZ�aD���kI���4F��Iø8�r9��Oh�ҸMc���P�QR���	��/���;�;N�{���d�f���E��t�W�5/�}�mH|��h1�[O��*�~��m����AM�ʲ[�|w���Xz���o�]OϾ�7��j]ܰ�a��><��|�}h_������k(4&�h����h�$�p�k4$Ռ�r
+������*JԇmXV�"�L�����ƖEK�%:��ΎX��!D�?b]�����d!ٸ�dOJ1�]��9-#�{�1k����b�v+Ni4'Y�Ȳ\(ƥ�PE(�?��2���b��
��D  8^��lE(�� a���J��lUD��1�8E�,d�S�.>%�B-`	��c�>���E��A�dO�1:��̖&����6�3��q�*1u��Wpu˛
-=�3s��݁���$?<�4��,�� _A��jy!#���~/v�J��l�S ��z�����ehs�#ɕe�W����d����J���Ԛ+-IK�e�rȢ��p���U.��FDŽX"�������
Ǿ���;1E`3B�9sR�}��Ϛ�2���Yn�v]v��c�M�2p�f	μ/����_ڷ)���fl)�@\en�
+eNN4=z	+�)�fC
+���'�?�]���k��kV̺e
+�x>DϮ\�q����皉���dr(�q*��1��j��Vz�փV��L��H��0U�)keUe��5��M
+�E��n�[WX7YI�ϰ縇�x��vT���>���P����ų�XL"'�Y"g{# ����6i�I�aQ�l�H�#y
]�%��Z��"
+읜0w���Օ����U{J��k���:�j|he���j�Bo���;�Mġ����so6E�b�.n��Y�oc�K��6��0ۂ"���
�g�������0H��z,>������
l�����SE��?�U�{~���	=4���1����%���?��
����O(�
+�'�ż'Q+��N�*��ޢ����M�.�Q�+ڥ,�Aq��V������g��������2U�P5R��iz4�f�v�v�����jA���t#:R��*tl�Ҫ��;�0�[�Y	@����}��`����T���D�%�I�$����DŽ@D���1I�O3�Y龂�ӌr&�o�i�te�.�7r x�����f��1�
X�P�/D<�r����e���s�]�90{���)qc���N�u��N',:�k�ᣗڂ���c�Y꫿=;4�
���K��=�sA�N��R
+�ݸ�,=���������F�ܸ���e�=�����b����2%�.�Ƭ��veZ���v���#�Q�ъCa��Kň�%��1�J�RiJJ��򛕘�Pee0T�cYN��8W�VrI���q�L�jD���{4�kz�R���X�����be"2�bT���I�^���ϲw�_g�e�lR5�-Dpl��嗃
�WI�M$�� B�*eI�U�R���e�.��l�d�-��r%"���b٥�K	��]�F��h���i��te�[:�`v!�)���*^ˠ�5��W�'L�2Ş
��.��Y��X^�d����]9�V�O�8���,#���zvc�ߪoO�ʔϾ<������c2Œ3o��w�\Ȧ*c�V�|�^��}�Kh��
,P�@q����,���cp �)cM�I<Rq�Vf��*�$�84^MDӬ��l��(]���uB]��:��iRT�.7sT�:M��a�e+��r2Փyi��hЭ���E,
1�Z~`��Sf[d�2�,��t����
+a�
Ñ3(pn���\��?6�-u�2W�h#����̙*��sw�������mq����
?y�f������ư�b�fֿ��G���n ���b?bA@����
+
�+\��%�?��>��wONJ�40�����ǠnNV��^�L�C{U
ӷ(wY��[z5�nX\���:�~�"}>�,tS���h���)�����9?�7<��cz��y ��t�@��ʟ+/ĜٙB������|�^F�����Z�BX~0��&�^�>�$~xϿ���=1�:���h��5v��h�
/�Z�駹�[�zl�m+.�[X������k&���S{v���k?���_��<'a�_��
/�'�e���ó#�h��X�>�CW|~_��'����~�n��h?���	�k�UȜJ��b4���J,�he? �L��r��]7���l��I�x��/U ���Ț��Z��?+�
+G)�]#�K�ҫ�N�2!��*��5���{�8�TWȪ�զ*s5�:�<�:آ1b�d�J=V\\R]��O�7[�V�nZ�Zs��/YĈ�屒T1:򹋫��/�N1Y��O���
+��d�'�f��a���!U�K��K�-��wE���T��.g>q�R�&� ٻy��v����~�m�M;�����X��J-G{�G-�*e�WVM>:��"��d^��b��bϪ�8>z�����ٹlA��S(�x"���J1�[U��dI�wy"��'�O�����]�?]5�o�m��n���eě��J2�RcK�_�_�����W,/��L>CH��(F�IR�(���Ҳ�P�4
�W2F�(UK�P멃��5F�T�]��ŀt�S���L��p7�F��1FU�(q�J����;a�	�����7�ň�:��.5YU���xd�ߔ�@��Uc
d�_���G(�߽���H���;�F���"�
+h�J�ܨ�6&��4|�q�1ߡ����
���+�Ռީ�Ǝ��6ޱ�;�;FCM�ck+�J�J�ck��?z���3-V���H����h]��3�����(;��Ť��Y��p��&��ڵ�
�Y?���BiT(�U1eJQ����Q��PD�fA�/*bDK��$��LU�����*	��q���Sk(D��r������1�q�
YܷX,y�<��"]����T�L�mL?6��G�u�W��8C��l����70Bo�X%��Qk�==�4Q�X������+x���7E�����eF��M1:ɸ`@�݋���I��m�OY��Mh	��
ŷ�ii�*%�kŠg�����ٙʰ7?l^�g�ݳ�!~c���̕������ܖں�l]���dKp���
�������4���3�s�h�
+-#OP�hXA1"$�D�s��_�;<������b�i���ds#����'�K������d��*&N�8��,�,9�ĜN�Xl��,��TUU���Z�!����ޒ�U���-�xY�P^�L��&��Ngkn8� ���j�\�'���jSW�sΤJ+
+N,��0o�<ފ����ilAd��4f- ��ӳ
�����
,���?£
�2LT��.�E����ޤ�hgfoc<��|@񗢏�_�)��Q�V�?�bf���KK�%ߐ����Kc��
+^�V�W��M��akQ���C�>��o~4Y3���䖥R2��q���ū�t�O}�3�hd7�p&N>/;e��G¤aD?†LH���:��F�P�NU$���T��*���
��\���H/��S��@y\�L��p�H�w�pt$=�r�f��`~aħ�b/��!����
+7��!�ٸ�A�h��ߣ��R�s(��+������WTFE��T
+�՚8�Q_��-�з�
��B�'n��@�#��h�:�������$l�z!|��_�8�ٶ*�7�\!٥�J�g��{S�� `�J�1�U�s`a]��H�iq�����i?T�I+gY��̚s�L�i3]�-O�qy0@��0f
#�E
(2g��g�aK,?r�C+��d����d��:�o�VǸ�:\�\/�t��5�t�M6������ȵ%l\��qL��t~��6���v���/�����'T�������;{�('
�*�)��%�O~��9P��U��5�ds�%���
d����	(���
o�~rZ���ONA��B���\	M䗡����O����du�%���{�ߔ�	�����_�&B����� �_��[�cx�����K�,>�s������#�'=�?�K�R��|e��$`�`+�qs� V��b?ĿJ����Z�E�!�S�T�W���V���Ϩ�ԓE;)��ߚ���z��~���.��00�a��������&�V��������5[R�u~

+8
+2�
�F�@��B*���Q$
8h�i�D���D�VZh��h9�@Aw�&�$Z!�(�(�����)�(�i�V
��T�1�b�
+Kh�c�4	�؃ZNl�@ˡ�u�V��Z	v����b"?/h�I��@k�K�/�^��>
F��@�����R�#Ϗ�8�f��J��M�U~�D3������1�6H�?,�(�m\�O�~���-�$8�D�Q��]�Q~X�m�E~�6ԏ�W�R�C���D�,yoɒ��R?h����ɏ�#��H��g�W�V.�r	��KƯ\2~jI}jI}j	���7�A���A?,�A���v@?p+�%����5�^]�kI�k�̗�J�%e�����c��"q����*{8��0��8X;`�a�a�HsBmG�g?�0{�8h�ݰ8X;`;�
�R�0
+��� @�����vKo�����
�~F��*��~BP-�@c��JA%�!UP���~z�W�}�XwH��p��^9)?JuFန�x��
�{0{���.���m�]z��hEЈ�3?��^il����f�Rh���@(�`x���0��҈vH�+�IS~6�
+#���J�3�����f��'��z��$���yw���Q�����F�0�Rۅ9�a���B��KJF$
�A�&����a�]�`��7�Fu��n8(�@~��
}�*�݅乖׎B�_�B�.q@Ҋ���R�
e��쁭�o�֨�~���>���ƿ�d�#��E���ܯH���[�
4��\��,�"�??W�����K�}����O�R�I�D\=#��?���.��j�B�N�~� ���G�ֳ
+���\/�p�4���AX��(WY�aim~�u^�sž샃0�)�+_ތu�n8��z�y˰��Oz���W|� \�)��.��z8|
Ovu)�����1\UJ��&2I6�Q���5�*k�TOzU)^�ٮ�׼����~I��^c�W�������o_Z���`�
���m8�������
+X	h�t�z�
���g ���
5���% B;�AV���
a-l��
��N��@��R����_0����\�M���k��1���[�tU�5����%���/3�$�S����^2�j/a�~��~��v�+_
�_=do�����{��w���J���k��>|���{��N�ec'w�ͱɻ���}.XQ{��l�}�^�E�[{���"~�Y;�rJ�z}��o�����[�~�`.b}G.�����w/ʒSZ�69���.U�Q����h��)��gTT�kO����C,��C6����X���oǿv��=q?��1�
�2[;���d~���s�{����;,�����o'��g���[�_�±9Ʊ�?)�~����4�|�^\����ǝ�;��`g��;����٧�.���\8o�3r�y
S{ᘃ�s�����3���q�?&���)��c���`���ٛ���gO�XU�T�^j,��l�Ύ+go:��#'N��O[�?�f�J\��1�[���a�ı���c����é
�@p 1��2�k�Ѐ2����2���z����r.��j�%�ͮ1�-���Q��Y���5��Y���%�2���Y��Q�FK��E�\��RF
�S�Sp�C�;��4-�	:C�	���3&@20��N��q%KĔ,D�lG��ۡ�+�3`�о6��ڧ��&Whϩ:6濰#	��[�0�ʑ�L�Е�7�n��l�xL����6��������(�
���΍���q{I��P��$��?�E�/�����o>��_I_ē�5ߜ�[zG.y&G��r�ϫ�0֤���j�F�<�uq=���%=_���������V�ڊ�������b����J�G�/���
+���\�
+endstream
+endobj
+
+91 0 obj
+<<
+  /Type /Font
+  /Subtype /Type0
+  /BaseFont /OWZDVF+Georgia
+  /Encoding /Identity-H
+  /DescendantFonts [92 0 R]
+  /ToUnicode 95 0 R
+>>
+endobj
+
+92 0 obj
+<<
+  /Type /Font
+  /Subtype /CIDFontType2
+  /BaseFont /OWZDVF+Georgia
+  /CIDSystemInfo <<
+    /Registry (Adobe)
+    /Ordering (Identity)
+    /Supplement 0
+  >>
+  /FontDescriptor 94 0 R
+  /DW 0
+  /CIDToGIDMap /Identity
+  /W [0 0 1000 1 1 558.59375 2 2 429.6875 3 3 241.21094 4 4 574.21875 5 5 483.39844 6 6 503.90625 7 7 560.0586 8 8 409.66797 9 9 292.96875 10 10 286.1328 11 11 613.76953 12 12 565.91797 13 13 279.29688 14 14 389.64844 15 15 590.8203 16 16 345.21484 17 17 539.0625 18 18 670.89844 19 19 509.27734 20 20 880.8594 21 21 653.3203 22 22 565.91797 23 23 596.1914 24 24 269.53125 25 25 814.9414 26 26 535.64453 27 27 618.65234 28 28 575.1953 29 29 454.10156 30 30 582.03125 31 31 291.9922 32 32 767.08984 33 33 571.28906 34 34 581.0547 35 35 737.3047 36 36 609.8633 37 37 749.02344 38 38 496.58203 39 39 374.02344 40 40 725.09766 41 41 581.0547 42 42 432.1289 43 43 539.0625 44 44 642.08984 45 45 492.1875 46 46 599.1211 47 47 710.4492 48 48 504.8828 49 49 443.84766 50 50 559.5703 51 51 325.1953 52 52 269.53125 53 53 517.5781 54 54 561.03516 55 55 292.96875 56 56 551.7578 57 57 694.33594 58 58 483.39844 59 59 744.1406 60 60 603.5156 61 61 528.3203 62 62 756.34766 63 63 927.2461 64 64 670.89844 65 65 503.90625 66 66 856.9336 67 67 615.2344 68 68 590.8203 69 69 666.5039 70 70 564.9414 71 71 312.5 72 72 575.1953 73 73 502.4414 74 74 701.66016 75 75 975.58594 76 76 575.1953 77 77 468.75 78 78 601.5625 79 79 539.0625 80 81 410.15625 82 82 653.8086 83 84 478.51562 85 85 942.3828]
+>>
+endobj
+
+93 0 obj
+<<
+  /Length 13
+  /Filter /FlateDecode
+>>
+stream
+x���M��
+endstream
+endobj
+
+94 0 obj
+<<
+  /Type /FontDescriptor
+  /FontName /OWZDVF+Georgia
+  /Flags 131076
+  /FontBBox [-73.73047 -216.79688 990.2344 912.1094]
+  /ItalicAngle 0
+  /Ascent 756.34766
+  /Descent -216.79688
+  /CapHeight 692.8711
+  /StemV 95.4
+  /CIDSet 93 0 R
+  /FontFile2 96 0 R
+>>
+endobj
+
+95 0 obj
+<<
+  /Length 1796
+  /Type /CMap
+  /WMode 0
+>>
+stream
+%!PS-Adobe-3.0 Resource-CMap
+%%DocumentNeededResources: procset CIDInit
+%%IncludeResource: procset CIDInit
+%%BeginResource: CMap Custom
+%%Title: (Custom Adobe Identity 0)
+%%Version: 1
+%%EndComments
+/CIDInit /ProcSet findresource begin
+12 dict begin
+begincmap
+/CIDSystemInfo 3 dict dup begin
+    /Registry (Adobe) def
+    /Ordering (Identity) def
+    /Supplement 0 def
+end def
+/CMapName /Custom def
+/CMapVersion 1 def
+/CMapType 0 def
+/WMode 0 def
+1 begincodespacerange
+<0000> <FFFF>
+endcodespacerange
+85 beginbfchar
+<0001> <0032>
+<0002> <0031>
+<0003> <0020>
+<0004> <0064>
+<0005> <0065>
+<0006> <0061>
+<0007> <0062>
+<0008> <0072>
+<0009> <0069>
+<000A> <006C>
+<000B> <0030>
+<000C> <0036>
+<000D> <00B7>
+<000E> <0049>
+<000F> <006E>
+<0010> <0074>
+<0011> <006F>
+<0012> <0041>
+<0013> <0067>
+<0014> <006D>
+<0015> <0045>
+<0016> <0039>
+<0017> <0038>
+<0018> <002C>
+<0019> <0048>
+<001A> <006B>
+<001B> <0054>
+<001C> <0075>
+<001D> <0063>
+<001E> <0068>
+<001F> <006A>
+<0020> <004E>
+<0021> <0070>
+<0022> <00AB>
+<0023> <0077>
+<0024> <0050>
+<0025> <0044>
+<0026> <0076>
+<0027> <00AD>
+<0028> <0047>
+<0029> <00BB>
+<002A> <0073>
+<002B> <00F3>
+<002C> <0043>
+<002D> <0079>
+<002E> <0046>
+<002F> <0058>
+<0030> <0078>
+<0031> <007A>
+<0032> <0071>
+<0033> <0066>
+<0034> <002E>
+<0035> <004A>
+<0036> <0053>
+<0037> <00ED>
+<0038> <0033>
+<0039> <004B>
+<003A> <00E9>
+<003B> <004F>
+<003C> <004C>
+<003D> <0035>
+<003E> <0055>
+<003F> <004D>
+<0040> <00C1>
+<0041> <00E1>
+<0042> <2014>
+<0043> <0059>
+<0044> <00F1>
+<0045> <0056>
+<0046> <0034>
+<0047> <003A>
+<0048> <00FC>
+<0049> <0037>
+<004A> <0052>
+<004B> <0057>
+<004C> <00FA>
+<004D> <002F>
+<004E> <005A>
+<004F> <00F6>
+<0050> <201C>
+<0051> <201D>
+<0052> <0042>
+<0053> <00BF>
+<0054> <003F>
+<0055> <2122>
+endbfchar
+endcmap
+CMapName currentdict /CMap defineresource pop
+end
+end
+%%EndResource
+%%EOF
+endstream
+endobj
+
+96 0 obj
+<<
+  /Length 22792
+  /Filter /FlateDecode
+>>
+stream
+x���	xչ?|��h�kF���K#ydI�$˱�{'��$��D���$!!@h$�-�
+%$�PڂC8moo���R(��-_�ͥ\�i�C)�H��s$9�{���3���j�{�����
@��$���}�GԶ��
��Z?���1~���ذ���}�,��5liٸn�����s9�L�6n\�J��������W]���ݖNv��[�n_�JuU�zv�
L^�j�9)��8��m�U�������Ho۾󚃦�M��5൱�׍M���
+��$�=�
+}U�a���x@��j@c�W���ʯ����
J����(M���?�����\�u�{���<X~
V��������۠��<f��`;;�0\��`x�v�
���OA/�
��A@�C,����6��
��^�4�@���7�<���;���`�
���0T��������,� V�%x�~��w#���_&��`�D��� Z@1�8�I75�����P��p](�s>���r:�6��l2�:-�h�*�B.���$ ��	K�����{���Xu�v��`��N��?q����O컦�L�D7�щ|t�9
��0
�_���@W��k�f�kӄ�c���D7�ɱ���
+կ��}F���:�)bQpF���:��(�������V�DwW��Ա脎� ]�w�xxt�p�^�7��_>3Y>��S`B��F��N�R�w=�&�U��L����'Y�z�W��֮Z14A��E�2еq1�c��虠\'��'Ȁ�k����ѵqt�p�ñ�g�E'�C�������]Z~�������<�e��A���L_84�m���-���P׵q�3����B��VީJ������ܼ
+}Ϯ͞C����z;��Ү�\צU��U�u��֮Z������x"��kB\�9\=T�`���όv�Y�Ea��P�bܪN{eڧ��V��u�Nz�7�3!�Nx�x&��7A2h�.��`���X����]��y�L�Qn�+������
vsݣ�us��C��VM����<,w�Lo�Q�����&��<l��}x���cQ��ݫ����v���c�F,�;��0A�T?Vy&��!��c,�O����x�а��i�|>3<<\%�Ѻ�4y:�C�q��I��E��U�=`��i 
+��1�Μ��1.Ag���L�>�ycѳYE�,8��aM������?�^W9?��"�bx}�a'�H�s��e��Ǣu�!���M���d輽e��j'���"�w�!Oסi.���)�.��[��PU��O��f^X�%g�~���.<#�[-:��u���$:Fۇ�c�`I�w����@x��$,���s@ȑ��I0�x�6uN���$X��olE=ݕy�=�<��=���l\�v�
+�ρ!nݡa�3

m���8l��n�M�a�
+?�а�۳�����	���bl,��z&�`��¡����	����k�|����N�����o��t��d�~��I�<���S
M��	0|�Pe��N�?t�~���O��8�'���?�tM�����~΋8���rh&�;c�`e�wQM.
��G
+hH ��h��Wx�w/�MCܫ�j^�|�_>F�`���4�mPfj!��b�D
+�&��
+���u��ۘ���a�J��ƕ�4�A�M� �&����|��m���T*"&��D_�d	H�BA�x]6;���,�����Ph��|!���1�|�����SM�V�L��Z��蟗2_Zjm���[™�lj��)5�[M-��/���7�q����$��=?f�Qp��J�<C�`�����X����O�i-�C4�2L-6p��T<�T�j�'�翡T����U�JQ�NV

k8���28.����g�z"'�,�-Zd:����d���;�(g�|A��
+<�v$�S�� hS�|R�j�׈Ѩ����tR�Z���f2\\�����Lo�j`I�����B-U�b�w�׷=�U|�K��:0�|��O�@̂�s _>/z�7̋J�'���l��|���Y2�i�������C��@�3�s�$����i�2��'5�D�1�r�����d������G�R�$r^z��AɲD��A��N�KgY�F��D�ZM������h�l�Y�����j�]�YI͢T�*DN�
+=C=Y.=���V+�
_��*L�/��������T~J���y����x
�|@ji ����4���)�j��h�a�"v�@#>�L4Q�K�3���?Ԝ��6�̊�q��1��G�M?�U���ow.H��xO�����ΤYA4�����ޮ-�gw�sv�4ҥ�k�V3��Ϲ���[�6[���mm�|}���r��n����7ɳ��"�b>*"E�DFEDZ�
ϖ�Ѣj�d��Wm�l�GU��c��n��XD��\$B��1�Mn\�%r��>��y��M�L�=�E7};�D�h_���?/G��F$��+g�%r<Ar<5�o�U����J
+S	:��@����&�:��F�4���<����k��ܜ���~e�������@]h���[��7����o��}���緌��۳�3�96ܾ��-�������|}�Sb:���D�K~ZA/Ԝ���E�^�̤��"s��]�A'&˯��\�V�t��)�`�&D4=z�IPl�� �� %K��@��Y::�	l�I���&ws]�,�UlD�
+��t�L&o:N5z�Cu���A�ݍ,b��o�w����]��K">��y�ZPj�Sb�y��z�����b")Le+Zg*�S�$VAՃX	�ڤp��21�&�qz*�O�d3V�U��X9
&<�
+=�5�v��wz<�
��J�aNG}��gB[*����d#mw0m���%�Ӊ�=[:�qN��y�n�.��v�m�z�7�vk
�\sO�~��2�k���͡zVN���q�� .Z�% �F��Z��@��P%Zu5K��x� "6T$i"_@�z��aJwZ����
+�@�*4�Dr��X�4v�4u�A��Λ
�0Y~CT�;�= �A3	���4
}
��1��$���
��J�=%�nW"E��VN��PQ�J
+M�;=Q)C�B���r���Y�%�ʚ���dWg~��E�
+��#��
����������hRu���ֺ`U�5��:�𡮹��m���f���hϳ�����m�߽)Bs.׼����w_+������}i����Q�	��w���7ޝ�@�yvZ(~�":�;�cd�|����PRa}د�<$$E%C�HR~�*d�Y����7�6�7d��l݄�A��ӆ'C��S�G����"���E4v��2D\�7��c�|�,"�c��6*�s��
��b�"��~
+�D���|ŏ�H�MK"X�l�5>M[c�K�29))�ڝ�>g�Cs��aw��)���������ɾf�����.~��*
+����hy�O�r��TJ�ͤ9�p� ��XA|���ccB<݉^>�9���SDǝN-s��z"NJ��X�0+��g_ֲV�4�<d��G��W�2DD�҉��e">�����;�{:&�D��X�����;$��BT�}�2A/"��<�"�SDܩ�8b����'m1Y#t2a"B%�F��.�Bcw�"�}��doo2;g�X�F�J��u�\�R�$$�쒅���|�w��V
+xZ43X/�xk�4cN�/kkJ
�zD%-@4��y>���v�����%������"#e��#8M�?TI���,�2=Q����a��춪��
+;
+�f��*b�?S�0�����p[u�N�Scs�x{;!��c~j�?�S�L~�S)���oP7Q�p`����f{�w���8$_j��4��ʭ��HD	�ݕ$K$�?/zK��c캭��,�1X52z�),F���BCb�!��3@��G
�dBG�b~Ag���7�4�ޖ�z�m�;O�~ǩ���2���w���=s��.��u��x�e+J�w!]���?�����nIɓ�
+��1.��G]g��F��ZԴ��E��n5��ܨ�ߠ�����A���,��
��#
P)���K��fd!�3���<C1T�A��1pz�c	�U�!E%��t�H��4���B$鋐�!VU��
+�L��Qa=U��x�O]���b����LC�R���ty�y�!=����93��l^�_��t�~3�
>^8=��{�������^{|a����y��
��o7,c�K����Z��x>ӮZj}���]_��ï�P
+R@
�
T�
5T�
+U��A��%��<*t�I�t'�NT�i�6,��j6��,ѥfӺ��u{
^������Qqo�w�K��`�dU��x�*O�T�KQ��KX�cASWd����������h�����~U��k��55��y�F�z��e
+��;�q��'hsr�o!��cu�	��H��@�B�����M22=��9�6���gE}����K� g��ӑ�y6fWݛb�������&�R�tNv��ь
�v�Q�e3X� zdL�a
z��:��3��U�\Ƈ���,�)�12��-
+b^������S�G}��_���jj
*��a��V�>�l�&�oas�;���\^ðOr��)s9"y��xx���p��b<�@�"궒Ӝ�9�kz�|ސ��06�J�+��e�'�^��6��¹�F
~��j����⋪�0��lʹZ:�ϙ��2ǫ2�
+u���W$r`GD����Ҕ(g�#�~Q��l�{I�`iw"�8����
+٘t���j��I'�oa:K�[�D���y����N��ƞ��ʨ�d����L
+ _L��l6) �8�,�&�^��B!�_i�f���=���^��\ˊM�U��{5��(L�^-��u���K�
�.����-�S�����',���6�&�>����^��7��0l��^�t��}_�W�ߺ���:�%C�nV�R�P�T��+��8L��P����4��W�L����̀]�>���+�ho�կ��/�\�����yln�oü���l�Iw���Ύ���������ͭ���g=8g��(�y�S�'�#��AD�qQ�z���
E��Oڻ��9����#urٍ^�7Qhnr8�^�\�v�7��5i�[�1����;�n��2FE�&
@T�B#l+"�Qq"D$bt��$(�IcE��9��C�� k�l���������|�k(~#���lZ��:�OB�ᛡJ��,�@Q�K��Ө���o�X��-�X��}�����y�G�2�m
+�ڽ1�7i�TO�g��T@��L(n��!,�⫀�R'�g��zA
�U��v�y���'�w-��ȁ9�b�1!=1$E	$-�X�*B<�aӢ͙N���
8�����z��YDC�cB�3ͻ�D���'�����\H⌈VW:bB���B$��-�&q�#�r�-�㈜
+_�Rup����c�N�����M�H,u��T2)���#Y܁��B#W#�.�
z��/�8�T��*��k��2���l�Q`�����&ﭫo�w|�7^;�9�h�yz�n����p=z+�p�u�D�����/�52(t
�U��S��'wn]�
+p������hO��z!u���%�]g�F�=L�H���:�(�oL�tV�BR�uDp���[h�~�ᢐ����58�n��@z3�{��`�YܴFdX"�G�`������ٹ�E���Ë�7+;Nw����,lTݻ�
޻�MZ�U��6�nd �1v�bmn�K���2����Q��х��G}�G�c:O5>�2Y�(*��`��e��[��Z�nB�܂t,�j|}�����l�,�*NC�2S��xO����^�{�k`�Ef�z�%рݍ'������>z�1�[R�X���Y�M�?��U}[�a��[�ZX�}
�FT�/�����S�*���n�Mn%��
�������A6�2ӿ/�͂�a�i�;C
+��p��f~�X���0F�%o�Zg\p�Fj���6�=��c�
���X�昣f��w�즤��#.��˧�6:ᶛҚ��|�W��u��N� {~��i�x6��]6ϩ'J��#p8�D)9fABT7W����d���
+6��(_(jUj:@ ^�/��cBr6�Ď����s�8(�CK�����I-)�< �?��f���'� w�/*4i��}��~��4ÄЌWD�� /�#��}�xD>���U���ȇ���d�]Q��X�L�gjh8�J���bS�ȣ�N�X ��21H~r��Zk+j�K�
+���G**�����(��w�e<Vg_n�����ӯ��c7�%���w����l�y�����4�Di%�a0�yɛd����ļ��(��N�s_�e|�%d)
�s�z%Eu���,.�Ay�
`u�'b��I�
I;)<�$ϙ��>���gZ"_fZ�� �ʴɷSκМ\�`�Q��h���Ҡ�ж
j�"
�
+0�)�)/%�R�˹�, �O��S��󐠅Z�s����k�W�r�)��	|
K���s�
�����t�.]�ȳ�11�Z�<��SM�gۆu�ߝ�-�������O�
�_x�_AAܽ{ܔ!	�z���5WѲ�턶�t����"t��W��6��p����<e1i��n,�F$H@�h�W2�:���xQŤe��I�c�Ȥq�h��I[�)9%��
��O��F�11�ٝv&6�G�۝��5���^�W2&�F=�3_���Mr��v�1�}ޯY�����<�5R����q5N�!R2T�A�j9|��a��\�台G���jz]�P�:tP%W��?�����T>5�KQ��,
i��,<��c�H8
+��Cj�(��Gv�q5EY3��@�Š���7U(����J<���>h��w�����}��;�:�ro{�)�����e_���	����^{�Ư��<���o_zh�3�=_#����U��Y�rqc~$\ħ�p���uU�����0
^�ӿ0���B#U?
�+ZVKhK�B��O	@(��J^5T�i�
~�-����Y\:��ۙ�f﯏�5v�Z\�@�;��	 X
��%�I��-.K��U�?�C�=/��듩&?�=�#; ��'�)x<*�7�	�5Y�+F�@!�6K���,����l�%t�]c]�;�+��|��i�h�Z�Vӊ6�1`�!�m��
+f�3�� ��[L�_MS���6Ӝ��uC������+s8
x�d@��߰,׺�uf�v���g���;���������=|�T
!�ϒ����:�S=B
O#��dd����4>y�K�r�s��xD�*�A��m3�Rؕ!��[����H�%��m'�#x�x�aY�
I?"������][��qgRCI(x�)���8���E���Qܭ���$n��Iv�ǚ(�O4Qi&|:z�g��/#�=6���b��*I�&ڄ�5�4iwܑ�[�h<�$� ��
c���iv7��fJ}�#9�|R��������ҁN2b$�XK
�
���' mR(�K�p
Bf�SE�9[�Y�y��b�:�ÃlQv�!��B��'$BR�t���Z�iv��t�(�~����
��u��z��vx�f�3�2��:���}�����hw�H\?�m�P�#TP��mZV�P���|�[��m����2,����f>称Na��(9��Ȭ�Vrۨm�>jG��d�uѠT�C!��h�3�4�E�Mㄙ�a���KA��<��p
�Y�~�]v��"Zۑ��mGB���
Q�>] `�����#�8O]v�ߟb�ߟ��4�\܊o2�=�A۪[��RN���%���{e]�����W;v���f�����媗wS�P
+��h}���>�ۛf�$�*��!L�!%TO,��oDt���6�o�t�sS��S8�@���ˠd�J�֝�a
��/��*��ߧ0�-���Bx�S�:ĤQ�D��D���6�c�(���	�(�u�V�o�ܩ;S��JQ��Hj<uW�/�r�R�I����K=��-�����w5ܯ�j�=5�\w�aC�Xư�z:>Q3�ʋJ�Q=
����T����˨ΙUk/�?��T+/8<	��������JgP�J��Ж�Pl���o��tM.8��H:�a�ê����êK�t&�^k�lw|�$���
�yMqі�>�Q0�7�SIбK�ŵJ)*�����!�B�a;���0�y���B$�yG�=�XIl'�	��/2��i��f�o9&g�~�}�@��MX `��RU�R�PP�P.�P��LX��}ᑰ�u3�g���F1p�
�!	��������r�I\߂2�ح�!��G̬��B3 1Df���kk�=���g��\������KÁ�GoYusFl��
:���=K�I�o	�FLnc8�xb�7��red���m��V�V8OTWm�<������5M�bir|6Ѝ n\��1Ր�+�m�b��5X�-qe֖3� Y��]|X$��x�X)A�5��a�/"=
z˷�`V��}#�}e���;}��1�~�_|�;�A�U��ᩩ �W���b��͐�����Q����x%��d���~���0g06��|~��aӖB��
��Pަhp���;�(<��&�3�����S�V�D~�7�xN?��qE�M�N��;�
x�s V>/�q�ف��x�2�E��\ؾ���Q՘�T�HK������<��nD`ϸ�Q�H�Ͷ���9��9���Z+��s�p�`�T�/U*(��
+I�Q��+�*�1!6#�„��J4\p����������>2"w�i�l�g$��Sbc����e�zw\�z�ymn+{�ޮaj��m��w����6�Q�h���'�S��V�~af��R~��I�(xA�J|!�ONr>ߍr�A.�͑�"V$"���t�"h Z�h"�&�^���t=�i��lZ����[���iGȊ9ߊ�+�Hh�wZ����G����r���"Q@zN�H�?�C�X�8����W(u�AS�b1����z���>��jՐ�����y�pzD��X�֥�N�=���o�����c�����.��1�%�$���Z���Q��7|�a
+"��i;ߪ_a�
+D��?$N~�I�'zp��[�����Eu
,�݈L���ڭ�lPH	͍"����4bR1i����AV]'��eۏ�M�i��&a�f�'ǼI֜Nr��|�II\'�N�~ě䷻�n���{��2<jQ�b~8�O��=ѣ��G�
�j���K�y��b��:�uM��H!]NӡB�bb&V���p�;3A��� .2����C���bǩ�W�e�e�Y����첸,:��5���Vܿ�.����n��V��Tf#���ܵ����5��^��?��kaGAK	eue�*5�fRm'Z�\�g�u���-V<go6Z�9@�'��ȯ%0��X7�4�W�7�V�6�F��
�T���$P"1�T�W�d+,�b�1�P��ly�
)����ə��j(=$�zUp��T��5x�x�'�\A�.�]������}��?.������8�t�@���
ԛ��
n��4�j��0��yA���Y���W�7��gh|�]����(�r�O�v��Je�
y�/������A����1�a�攽���디D�PJ}�Mi�]��W$U�ƼЧ�Pu���"�㈰Q�Q���^q��z��u�]�=>�B-c�>��p���[�
+��3s�����J�������[�9�A�f�j�:*ߣ�X��㕓R�J�rm���؞ HMP3�:ܾ�B�������O#rL��BR0gg&����{頬(9��AV6uPv����;
+�0<��M�`@� 1���̤3�p:��\���DV��i]�,������;����ΗJ<�K��K<R��l��Z�
���h韧R{�į����3'�&I{74�)��T|��a�󖏷�x��!��D���@��
��b�H�q��~��o�'��H=�1))������nR�=��b���L*Xs�I���H�t����y*1cb�A%�iٌ
d�3�2�rkuoS�'�J�����&X�\8��/�?���������Ö�:���ډ[���K>���]���������f2��ר����V�ԉ��F�$z���=�!��R�#8'V 9#9��<��E��C�L��{�{�=���qG�'M��ߟUi����Ι�*M�DF��NJ����(���gQ������g�*��i�+>E��0�@���K�k�	������&�X�L|���B�)����<�b�"��6�h&Yͦ�噵�Ir�֫��޽�x�Y`\v��=�Fl��ኽw�}ޜ����`ݼm�;��,�oϔ>X�dq�R�N�)�kb
���1\z�{w�ip`%	?�������릩�?(�*���DU�����9��D#.�tDyPֳ�0�Ph�l��������)�e�%��uؼ��rI�V��zP�0��
+��K�%
+}]P�i-��a5�V�I���V��(J��<9�M�G���4ʎv���E���AM=/��>PSo��vC���>-�M�i�5z)Sl�3����/9>h����)$���Uy��k6G5�:�!YC�-������gx�����nh6�Vv}鞱A�l,Z�B� �
+f����
+�
��'��܄@q(�~H�q	2EI��"���%�R������ʯK�]�h��9�`
노 
++"����
�&Ӻ�n�:�ڷX�����݄��"�G�ȟt�P����<����8�k;iT����vyN��믗3i7�^��B��"�WoB|������5Hz���k�"W�ɤ0%`t@��&��Q����Nh0�i
+9A�1�K7!dݜN�+��'j;ҭp�M���-�.+I(}8~륕#c[V�@�}{������

GVoy�uV{[k�7w��A�߻�Cl#ҤZŤ��,����W�@��и~���pg�x8��]:�����e���&Wjs`A��ab�A�{�KU?�C��a�&z��:�:m`eoPR��g�u��E-��j�����t,�a��dƘ�䓕j�d��G����3��z���:�n��SĬW��y���d��=�:���Ru���k	�lm��
ҵASmЈ����tI�m��3e"T�f�����W7��q��;<����ߟ}���n~��O���#�p̗0�`��U��H�J�?F�E"����-8��eB>Gئ��_�ŭ*���]щ=gާ"���n��ǔJ�c*���c?��(�Mm6��S(���)BR(�Q�FE|G
�9�ZZHʑɚG[��:���Y������'�vz]�T tU��F��;�v����_l�.h��t�X�v��w�������V[\���,]�e6t�/����]�T���4�Z��d̠L\�EDh
�4X*�x��!���zB.yԃG��I4���|V��g���x8����a16�X���m"�*�'v)�����Mo׍��t:-i=���0�w7��'��	mT�?���q��Kc�@c�@c�@c�@?'�T�AET�V�,�ɴ�Wj��O�i��0��ˆ7o\���#����$���痬:=���K�ں��Db��w�tѻv�J�<�4�]�5%t��_o�?�x�G����%��Ht���d�J���|d`���n=�
+x����Pt�Tz2�9Q~_�=r:t��xZwZ_���d�-\��r5W�q@>�NM�SS��Ps$���"ڰKؙ�T����8���57�bq����2*è����IT�P8��U����%�H�K4��R̨��*����H
�L�_
��G"�B3dp����ң+�2���j�SMY�eRQ�i-dVy�'�ёV���m�DjC�1yc�鰧^���������\��t�A>��Ej[��\�-מ�����u�K�I�-z&��_���\5��`֔�m���[L"�JrM�wJ�[n(qέ�~POI�$xM�#l���A�t��M���*i��i��vTۇ�)\�!j4t��\���m8� �����=��F`��Ì�+T��i��G��}
�p�D�}��#5�12�;F��y&�ι.+���W�:p��ŽZZ�_q|�O=�k<0˽��ݓs��d�<�3���'m^f��x�k-�W*=�����r�*���1��cր�d��X�
���Y�����U�}>lo����6h1wj�sq������$�)�Mj�R��9[-���&�5�q*��fw&�gg����<i��D���Fk��"6��vy�sg/�.����^k]�J�2��.�^�ٖ��a��ln���R~%������p�i=�j�lO}J����*��M�g��}B9�sKPæu��qa�]/3�3�9*x2Dj�C�����Z��+`gi>azs�-�?_��,D�DVR<�>��M�u��M\Q�٤���L��rxE*q:��fə�Ɓu���mMCo��X|׶Ж�{��o>��o[�X����&|�`�S�ÄֺОJ��G����o�,t�v��Y�z�5�sN�U�*Վ���n]%��Sz����!�6�HS����?x\*�WY�������k��-d4՜�
+�D�
+z�ʪne�N����,�g���&�o=�ԩ��,��8��U�R9g6��������BNhR�X��|W���篼A\kK���cW�e�'�-ʵ�tz!����(�Ҷ�N����fK �<��&��A�G�h˵v�z[��Kl�[w�nh�g��@��B#����V.�0���\���3>�A�B��\�Df"v�N��T�i�3�)y��Z@N�A�8����#P�H3�d�O	W�n��8��|:/]a#�Tj��
+��|j(.e�����E�{�P�h��1~�Ӫ��
+M\!Dɺ�p��A�w��򫛝=���QdC7/o��Kᒛ��Ր��ݾg΅���.��Mj�}���R��hE�����O-�&)�3i�%��
N<K�(�D �چF��a�� HZc'x¯�)M/ѩYN,�l!n �Ѩw*�R�!���`��]���`PU�
=�9��(Sj���1�(0b�r��:8W�
+�\�5ţ�ȣ\]�pPR�܋��JD՟A�h#��h�z&�\�Qh�갣A7�x���`�nig]X�t�Z�ŴV�q1—�CHPD��\�}�N�&���
�v���}�~ܣ�A��ozK笙H&�kɤZf�t�u׸ohSp\�1��u���XS=q���������ub]�T�7T:�UM�ኝ:\|��XNV�}l���\0H8��:؎�"{3��J~ۄ3�&t�0�n2a���o�V{����
�Pv��T6�^�}�Zkf�J���v��ܟ�:7�lh-w��/7�T������l�S77���\���=fi�:��U�$K�mN�\s�^Y��Xظ�awпm�����*T
�:�1�{�:t�s]m�ZuXz�l�^�jCqb�u���y�ŷ�2�P�2H�A�-jF���8I�8�z��j*��0�ON�_����(W#�'���,a�[���u�U{��g̶�V7pk h� ��5�m��u�i��IV�g���>Z�Bo-���N�zY��� �2�F/*��)6�9�����t�g�be�P�</*>�M���cg��P[�FN�4~:��5&S:�n�܂��Su�)��NK������阮�I�L���|��F�-\��N�|4�{hF�+��߁�K@	B��-��E�
+6
���	ZP�ըtn
+-�Q�us�s}/�d�ߝ��>4D�T�_#ߕ��qQ�T�&��z��� z2F��@$�������d��}R��M
+5��]���؍T�qJ��T�����
+��i\B��|:X�d��Ϩش��"#�����E\���2q��t:��<gu镚���$��:9}��d�䦓?��h�cc���8�;�`6��~w�E\-�D�p��ڿ��9�/un�H{D����MtdR�V��Ź8V��H�$_"�@��c4�
,ـP��E�}�
+&�L��&U�i�D05;Z� ≃VO\\�]ܺ�6�Qc������μ��ޗ����a}}�Էr�M����\w�`N�ku;(��:
+��	A�C�'b1�D}k���v�3���W�SY���v�*N(��U\%��U�J(T��9�e�E��"�0s@��|�%�F�ԁ�$�p����f�����_z��#����5����]<��H���h������_�����t��ܲ�[�io�NN\����s�9��ew�s�~�w�y�Ѡ-�hz����ө\K]g)��-�]���.�%���r�ҋ�q�e@)XZ�

��&�#$�t�"��`1��8_N��VRg�9+�f��,
�s�,m�{,����-��b��hps�����������[������A�	65E̫�H\rx��A�)��z���m]8T��Klt��x�l��"i�m`� �r5�
\4:�L.G�0��$�m4���~*[��+��S8�B��u�V�����0�W�5�3]*����!Z���\p�2⛍
+�i僙̬xcK�HKvO�ز=m0��[n^z���[SF�G&5�g564g�����H}���p;�fƓ����(��6��rA�j�ܥbtu���6m������;nUi|2O!��K��
+B�Tr�.���[XI](|��
�@�}�a�>�d�,)KYy���BRH哽����kP\/�/��M|��j���C9���A���<���8B��P(6N�
��?�:�6R!��΋�2�����w��%���5ǍӁ���Z���$Q\����y]�d;/�Α�4�dc�8�$U��dNj��FР6�f��g�f���ul�i��,�u�|�4�R�3�`0N�&N;�(4�~�ꝥ?=`�z٥��7�C:�ޱ�E�‹
>�aٓ=}sGTCH�Z�0q��
ol;��1~�m[`{�<Z�M�2Y%�`gV��=:�����i]!M�8ۮ�6|��#Mm�ZΈ� ��-����1��M��������pDTɽt��5�(�=Fk��!�pk6�5.3�F*vO��K�DN�-�
+�%rr�}/.��֎TTܜ�Al��@��2^�.3���Jr��X�y$��:��&�р�=�S�.Z��5�ە���>8����-۰ud�?o=�Qs&�M�۲���DC*�t��-_������ǿRzg����kVA㗿�5�7n����o�q���KF��z离�Z6�fm�^��%�����+�-J6�ף�!�:�
�V�p:B.H�x�v��g
�u�m���������/��"���*�jO�*F�g������r�B����Wq"��3/�K��Um1/����&|�>VZVM���R�ʌxN'U5{/���1�M��o5c��+%/R
@�!q �W�X��1��t��"/7�3J�����=������~�r�(5j��WC6���#��!DPg�h|3�"��##��	���-9�X?*����_�����hM;ˌ� �̓NŦ��x@�����u���g��'�9s&I6
�� 4��F�hs�\Tj��6�	�Z��cTP��<
+k�K�T�h-�eTO��^����zo��ҫx�D�%o�e~��7!^�����Nw,!��x ;:4�!��ԩ����cے��
ϨYjQ*��d��mGεtۈ��.
y�1���k���0�˜�����Dcd�z�z��z�ę��v�NFLh*+����L��h鷳�$�Rc�$:�ֿ�y=CdD�6�A���F����Ӱ��/W���[�#O�e��Z,t'M瘜�۞#������zQ߯G^y����O}-ۥ��0�����
+I?Y~�|�K>Q]�e;Y��)D���<�+��p��s�?��哒�c��g[���T��wvt�9�+��Xb𙭁�-���ˇ}7w���tB��|���w?�i:�٣�d'4��=#��mJwwg�v��ۏ���@@��9H�
K����Ⱥp](|�BiP(�u��p�F�ʠV�кɡ��T/���
�4"�U-��j6��9���Jn#j�e)��e����`���\X��*k��[�6�d�ע7}=Ӣ�R�TG)s}�d�4� AYN�+F��@fA~:F�V
+J��W�)%��p�P�3�:s��gT�X�r�_z9��D'�TS<��bp;��#QJ���!c���5v��r��E�Ex\��C��c��g��t+�=P��7(��=�˗�ܰ�aD�����V�W�~��F�|�k���-��-@
+-Dk�T�,��#�@���a�z;g
/�Ҧ�Po
��r��F��t��G`D�_��y�t�/4"}���ޅk"JT"%^�9H@�V-л��7��ۋjJ^
@���b��������#����v:Ӝih��VhZ�i�47�ȹ
�6 �lD�6�fӴB�.@7���`��ʨ�E��rB'��D��WBb1u:�:�P4������E��<���60Z�L�kh��
+�m�T�d���B�V���RłD<9�p�ĉ7<�e&�*��+��
+V�L���M�i�P���eC���_�89�Voq�c�䷇�-�4y�^K��(�_�R������G}g�NI:�Ue�����Iq�������̌�;��I�0�$u
Ԍf��b��!,&�e�u����Q�v��!�IȞr��n��e"��� 
�V`Y��ā�

+�}�
+W��}���d\r����`��mPDI�л�B=�"*f��0O����i��E��=���M�́k�=�qF�v;��h9�vިR)r�<"')�QP�4Zo5��
+4P�i�R`���l���TN��K&>(�I	�Es�ʢB�RP�ɯ�9q� *$G�ON�I���|H2n�k�
���XtW�����kow��dd������U�}�N�`%����cGJG�������)j��P��K]�
���+�w��n�BA� ���t�j�z���JKj�ꈔ6H�4 h�T�Hp�	�DBJ�"Z^��RRRu�r����|b�V�cf9Ҽ�yD��+�5àX` -*
(.T�CyGT��<�@	?��j�َ�C�����q�,��%X:���zaR��j!�����Ŀ���Ү��e�,����K�ww��{���	���I�
������q#a|�Q��U�V�:��j��L�Jڙ��Orן���P<rI���/�Y����J���E"��?���P=G@ЕTG|#�7�KG)	����9�o�v���Aڒ�;*9� ��HR�N+���*�������M�!HD�h!�+������ ���]ru��ǐ�e��*1�D>d��/��c��!^KD�/��Ng!(I1��x�t���aw�z�5���>���-��hV �1Ι��������n�Ӷpג䳍�����۸�qM��ƫ���W���D�{��5��8�N�H�`j�{h~��42��V�V�#�ֿu9�W��g�̏�"tT�^	�q�<���ކ�0��
�	�L�0�$�%?.�4�ž�M~�2w��jQ���Aܹ��b��DCl߼M��E{8�[�/|��C�?��J���n%�&��!z��������k������9���㜼g2̜��If2��L2y��!$�C�
H"��4�J�*X�@��+˫IA,�^��^��]Q{���.i���]m����I���aq����{�������{�ه�V�[��x��뇇��]���s�n��I����7�'G%9Ęa�����s��;)`����g���f^rG�3�yQw�[o4pZ�ޓ����k
+�����fy�����y������
�{�w�������'��V�f��
������o����&�1��Js+��E�h1L1,C���Po���Z-L۵�M�-�DI��<^�V��ހ*&���2�,Z/r���VT
+��z�=b�6x��'�����D�G�}4�$�c��c��T��s��2�mA,E��ڎ�[�.�ML
y���uS<Ai�O6-=8��}�{^��e���gλS����p�Vq����6sl=O�$��]����ѣB�՟p����C�ɉ טgl�st����_����l�==>�Sl���Sw���5a+Sl�]:�N�j���h

ǽ�L��!��2Tt�`�L��i�T��p�l�R�����e*�7��I�BE
+Y{}����O��}�����������8��fA��>�-��}"Q?4��	_+��!��HTUz]z��T������t�m�ֿ�p�r�`4���c��h�f"�˜��>9���b�Tel5IU�(I�!�a�Ov� �b�x�	��$*� -b7Y)�$qѪ��g���,Ҏ
ڷ�)s�S�`�a.+D�T%�v��Q�O��
%=5}S�b8���ƶ�eɫ���yY\^��E8�^��1z�z�R�t`VP�]��}m�9猸�{e����Z˹;��x����9�Dø� ��v�m*���TKա�huK���x��Gu�p0����L~����Q��h���q��,��]9���i7�s�G���&�i�}Q�EZ)Ie���X������WD��/ǘ�k�e���*u&s�dn�Z��҆�ώT�ͽ�_�L���5.�sR��ç��ʧ��lY6��@pj�T�O�gLJJз�\������-Z����;������L�w8��s������'W����,�`�Ғ4�#ש�zE��e:��=���dC�%�7���W:݅M�EEfE����"QS\��_8C���s\3*�Յs�,�9��I�>f�U�粢�5�3;��{.R��Ϭ��k�,�n���o(YY�����l���L�'���|y�1��v�N����}i�L%V��T������(�Ta��i�%-�:�q�p��2=.KN�I
B?7g���ŕG��G�G�&���5@�� m��G[����H{q�{�����-��[�R��~�ْ��L|�����,�8`����$�p��2}a[��g�\_i[�)��]s����vGj�U��lk�v�|���JG�΅}�?�d�={�����X�W-	��-��O��˩���*������W���\p���A�]�(0����F͑��l,w�}B6h��	ex4�m�Y|�],�ǾSޚo�Oz�q���t�ҝ�PshE��B��ӳ�Nr� hϱ�V1۩C�!��(>ꭶ\-�g$[��[�mN�2՗�E�夛�q�W�����3����"���?~\Mƿ;�$?P�>4w��ʪ���,r_i�z�=%
�&-:��D#�R��W�����j]u�w������F��w-��a�-��O��x�7�}N�5ϡex,�M��#=o�7j%j�G���9S���j�zG�l�kC����A��׽�^:�+���U�ź!��%�I�P>e��J�BW�u	.ʤ�&X��p6�Ƃ	^%�q%[�'�+â���&��)R�ߜ�ᇯ��RS�dw��S��UJf^5N�1�GI�g�X��^�7�*}]��d�V�f����7e��j9^��5i>Ǭ%�K���׊�_�M�%ܶ�U��E���b���'rt��]���O�o!(���bR|�
+�"R�*X��V������/��D��6�F�/�.VH.���"�i/L]�Bl*�I�yd�E��k8��d�E�J�JkK��VM���h� 156_�2c�����b�G�j�
[F�<"9̨6Q&��A����=ji�<��I�)�7�)7��o��N�o+9�W�颣�H�F�,�Y�,Z$����פ��f�b���~w�d"�9��$����4���C�i�쀵� +sFg�Dc�Xf�k�!�G��r��d�W�_杣߭�����ˈ�d)�pKm
+ݒ�G��.{����6�N��鈁;��Z����l��Sy��h�M|:�4���l�~C�kγ��M�
o�$Bx��Z��lz�")uS�&u<�o(-m��G�C��pcc8�?;s�BR�C��\�P߻�5��8]��@���b*bNz�e?�,�](�xK4'
+B�}s��Mu��Rpx@������-�z]��Ϸ�!�e�j��{�=]~C�F�<�3��R���2�:9F��.C�r���`B&��R7Tm����x����dO�@��$�ZcMq���}�]P^�/�JY�敂�O�K�T���Z�]Y����2�\iE�#h~��J7""Q�]ׅ� F�H��a�Sr���ar�|D�̌bQ,!K���@�U�B3��Rq����ۋN��K{nS*\����ً�}wl�u;K��2��g�!��Q�3��,��D���B�&<��S�Nx�������[F��lo��~�ڏW�S��D�����.���8��Z��}���
0Ш
ɘYе�F��2G[W��RK�� ܍�$�]��q4��(��"�;v���4�A~.x�3��:~*��Әƿ
+�ЂY��c�2ЙQ��A�� Z�װ���C;�U�h�?A�C�B��/��mc]Zl�^�	���擱��o�T3���oc��Q,Ҽ��3���֏ȕ��›hӼ�`��<�*�4�a��xD��$��r���
Qͷ��Ҥ�3Z���J�A�ւ��;bx�Ѣ�����c�Q�4���Q��e����b͓X����5��9�
+���F4�F���by��=l�=�
8.�s
+�
��N�<�9#w��.yXX.\��*�o����~�2�
'�L��OXE^�.�ӊ�Υʨɜ�y%�H���?���ۋ�د�sd;F�HS�9�����yB���G�ox�z뼭��e���[�^X�
+5��`IpV��E?.������Q��Mm	��JB�J�˄���%����W�W�Tb�f)NA�Ѐ�(]��k?a�^�%
+	<��g0-)��`-K��`K�3X���l��I���%aV�N���yR�h��$�MI���\I�#�O�k�
S��$l��'a+Z��0ؘB��⩙�`sJ��š%f^75kLc:�kz��R>���
g���Y��v�2?d�3���{Y�.f�QX���>�}sJ�y���VlB6��������asJ���E6c
:�k�A2Lΐ��3�9M�óPP�B������jlE7zЍ�腂���8�v�]��f�@�l�F(X�.��z����u��؊�Љ5(�lt���Ѕ��}��0��Wm�nO�(a�ݍb���R����e(��&����aډ.�b=�����7���R)�
+z�k��x�wBa�%��:w����#zM`�`63��'�!��X'���Ġ;����F63��X��d8%�QP��`3:1��������(����0��CA{�	��[�Z�u��8���a=��*Vw��,�<�H�ݚ�g<Z�>�f-&x����뿤��=-��Ǟ@�iuC��g�l��}u%[���P'���E�i�F�хB6f6a�DO��j�M-��<��:�3�%�$���#�ִ�E��M��$A�a��X��'h����1ʻ�Ⱦ5�	>wL�ib��<2)W{gW�~'�b�tN�CKnD�<!*ch;5P�b�W�8:�=+a57AE/�3
+�1�X��P'�聚"I��울�-6N����d)����7c��J�&a�0Mh�'�D߸��]��TB��VlM��lO�~cM�eu�<� ��F�'ߧfd��WQ���,�s�?��
A,@�1M��R܎g�a��8�6lG
���X�Z1i(F%z���Q6܆�]8����ȁ�#�A?�1
��^D>���·�~����f���N�$�Y�x��?V?棿WbDQFX���g~�#����A������;��x���|A���|Ax�MI~�|����̴�s>�^��>��������O��ɏ~��)��"ʡ#�x��P���G����}�����c�����|�S(�hQI�ѽz��{}�>�;{}��^OAe�'k��W��B��N���F�R:ŭ<?5pꏧ��b�o�I��2�r��{^&x�_�_��"��Eև�E��r�g��O����ʕ?�y���e������.��={�����?8b��p8��>v䘀U�*��9���G���se�ğ�?�ϝɕz&W>{&W>q�)��wȻV�o>�����C�z���F��'�'s�o��殺
�]ͻ�����s��=��G��`��§�9>00���?W���w�1~��ߝ_�~ne��J?)*A��P,�Y����cU�D�luLqZl�\KVv�%=#Ӣ�d�- gg��,�BY�(�����`4��:��38�,�T/k9U�6k���Cs��o&u~ߌn�Z�,��eR��q�^n	sC�Mhj�����t{�P8�tR��Ce��!C��D��pp��$��!ᡓ<Z��g.Y�~����A�}��Nr��<�����A�К��ۇ���Ce��\��9�$~)nx'��/	4�{��%O����X���=�C�k���"
+�~�6P4�,��zzi��`O��/\޳<��h��7�����]�{z�}/�{{��D�ğ�	��}��4X;
+endstream
+endobj
+
+97 0 obj
+[/ICCBased 98 0 R]
+endobj
+
+98 0 obj
+<<
+  /Length 258
+  /N 1
+  /Range [0 1]
+  /Filter /FlateDecode
+>>
+stream
+x�u��J�PFO��U�v�D�D@�`]\�
+F��Mk�I���R�B|��&إ�n�ऋ��(�H�r�Tų��9�`��ըa�&�ʺ���:�l
+�3Ŭ�*ުn��h)�&
C|>b��������纝�黓��A��v��Cƫ+	�y')�
+̵�8+�/<����a�Rkp��ش�A�mW
�Q�Ih��
�%��l��E5VW~��d���X��)�ϴ�:׺
��b���k#�&�_¤3�0��
�I�
+endstream
+endobj
+
+99 0 obj
+[106 0 R /XYZ 85.03937 221.44281 0]
+endobj
+
+100 0 obj
+[110 0 R /XYZ 85.03937 617.16705 0]
+endobj
+
+101 0 obj
+[112 0 R /XYZ 85.03937 717.34656 0]
+endobj
+
+102 0 obj
+[112 0 R /XYZ 85.03937 183.7146 0]
+endobj
+
+103 0 obj
+[114 0 R /XYZ 85.03937 205.52747 0]
+endobj
+
+104 0 obj
+[116 0 R /XYZ 85.03937 253.14221 0]
+endobj
+
+105 0 obj
+[118 0 R /XYZ 85.03937 405.88647 0]
+endobj
+
+106 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 97 0 R
+    >>
+    /Font <<
+      /f0 79 0 R
+      /f1 85 0 R
+      /f2 91 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 0
+  /Parent 1 0 R
+  /Contents 107 0 R
+>>
+endobj
+
+107 0 obj
+<<
+  /Length 3426
+  /Filter /FlateDecode
+>>
+stream
+x��]�9�ݿb6���b�����݄*(BAe���a�d��.Α�
+���홱��jɞ�(^�x�jI���.��x��^�Zt�廛߿������������Nv+�E'��M���u��������ݗ����v!��ϋ��i��>�>n?.�~����)��O�[e�Q҇n��ç>����5�c��w�͢���a�����԰W3~�194�F
�j9�|���a��v���w7�K�EEpQ�(���w��7���V	�F
+�����5|�<|��H;R�J�;�0��tC�(�shc�S�4
=I$G�Q4'�&B9Ž┎x@�S�ψ���i�
+���p�8)���y�������~	��!�y�K~�~!E�.Fm��7ZH�.l�n��q��J���K\6����n�]���#�
+>v���Wʳ2�+�&#
J�Q;��[��I3�(�	��
N�p��
�҆��ם^��{�
v����|u0�Z�)[\����BKc���)=�Ñ��m1Z� k��x>��#���Y��_g�^F�uM��oT��{�'/�I�x`$7֨��kH�/^3J�g1�¨a/f���{`l�:�W�f�f�Q�:VК�
W|��{@���q�	$�@8�8]
+�t>��HZZ��I�!�y:��KM�<�oNFa��s�m��_��n��LXeWA�؇�}���f������:Fo�5�S^��n5n�5���*�^��G�l��7�<���ё$�q��ӯ��G��j4�	2���k�AȀ7�E�hs�vm[����l�گK��`*-���-����:��\��90:�������n?�k�=_M�R⛖RH�T�[�w7�a��	8��TqFPr	x<Aj�w'/��;Z}���D��AHrgo4�2�Œ[�a�0`ئ�WUP\P^�h@=�.��&'���-�jX#��K�Q�F�ŌvR�g�i�hXly��
+n"�)փ��5X�$ۜ�Q�r�腲޹�A�aի�B�n��ZZ�|���sPn��P$�f#G��
�E���PT��x�7�:0�Җe��\G��|�]�6"���D
+$�V����~Q��@���j@���z�<L���s��C{P�4��A
+l���SsZ����(�I���S.��b���AB
D�.���B�2 �u����V�H�5�d�Yz���b\`��n��Nj<Z�hL�����ws�4%/��w*���E�3��v�jX�uVXiL<[q{�l�J��Ð$���R��Bs��q��d9`�%e��Yi�H��xh�
ӯH�)�����}rj�j�Vs{��5"�ܯ�`h�_�:�`��G�g$ZU�,�H��kM�4��ͭ�U������H��������7TN]eun
+f�� 1�@(��UzJ�S(�����0h.t�)3S*����ڗY�`�v�<��M�z1e�e	�k$��/�sc5�Y�Zx/�>[IA�i����GZ|��%W3���`��R*�q��jp�Us��r� 0,���M�V�P�$�;���`�I�m�)/1=���qɌ=]5l3�����].���t\V��]l1�J/�>�|�)�o���}I���x��g��Ò^���:-��R�hJk��r`r�4�ψ�c�㔦bpP6I�-O�*%���&S�Ejgx��VM�"�X���WA��Ђ�W�'�T�q�<E��g���띝�˽�VRxc�V�H��cq�7�L>��91'�^f� �LlPXs�
+ 	'y�nY"�H�Y�\q�@���х�2�UmE(��1+Q��&u̒f���Y���5vv��&i+d/mt��:�@�t�
�X��6l���5�8�+�0��=0:4�tU��9
+?'V�t��U�O-�:<'X ��A��h�;�g���8��6O5�Ԙ���qzN�g%�\ӱ��ܢ�2�?�_�V������
��?�Y���ӥ�MJ��)�;�;�XO�������LȀA����7����nY�)��0�aW\7$M�YY�]���Q�����K�q�q�pt����� 
A|w�׋$������bܯ���p����a@RE���mo�U�|!���!�ǹKࣉY)ő?N�����Ȍצ�
��.[�5�/�97$� ��]P���F�)ߋ�A��::��]�2���Y��;=;�'ýb	�{NzO0�#�s�tN�T"�P9�h���2��Q��鬙wՔ�rʧ{)�	s��8{�U��I�O`Ylh�`F�A��2Bpw�%Ù��p.��)L㢝_��rFmL)d��k��7�=�e�t�"z7�n�2��
�ވ�yW��O#{��Ky2.������/�
��q�D)�D�����M�0����W
+�k�g���)�k�F+ۈM�T��u�.�������'t3A�b�²
@�U!lP���f*�����)�
3���: y�hT�>vJ
+5���R�V�{��pV�-D��Y����e���ڷ�uM7�N	m�PZ�(�rԹ�;�=<ve?��I��uӛ
�y��pݴ\�[i�s>�ئ�f+�Z�M�����������t�i�l�DI���o|�����3�F��F�M�qU��i॥i��*O؆
��ʇ���9��o�D�}o
+]����
+訐'�z�0�U����%�l�`�������^�jy�&F��řp�>wQQ3�v(E�nF<��ƬW&VN	i�Nέ�L�a��_N|]�@ت����Y�Γo��;%�N:8��Oz��JJ
(��d�lvN����EK��t�aK��i=��M9�2�$�d�Z}t/���r�#8&L�-eWb
���0��F�5����h�Oo\�~i��…h"��Fq�3�w@�M�&N���c�V�4%�����Yy� �2o	\e��pȫfI��
KڨB�}M��Ne��RKu��7�K��d�Z8�}F�GJm�x:�zUL�Ș&�S��,�pqG�k���W]W��%?i�(_�/	�
+`�Y�"����pycm0�5�d��J+���s�Ⲕ]�e�=;��qJK.+�I
+����@�eIt7A�Eeu�ڈh���祦&G%�+�!��묅ɳ�x2ɗ�c�������~>	���d��^�r��wVЉ �^���l��Nh�K�߅�|�_*O�(��.TZ=0��ᢲ���U&<cĸB�\Һ��@J��پ
}a.��&�b� Ԟ�Ϋf#Z���e�秪,�:�
+!I�r�T�J�� ��Ġ��D)qp*�#n)ρ��J���ñn����ۇ�v__^�n?}�v���?�����Cw��O��>��t�������m���!����^��� U� �&����Ɉ�_�إ����
�S�
+endstream
+endobj
+
+108 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 97 0 R
+    >>
+    /Font <<
+      /f0 91 0 R
+      /f1 85 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 1
+  /Parent 1 0 R
+  /Contents 109 0 R
+>>
+endobj
+
+109 0 obj
+<<
+  /Length 4392
+  /Filter /FlateDecode
+>>
+stream
+x��]���}ŚP�^�:}�f�m �@UR���α	��&f�J�}jow�ԟ��셗\̴Z�Vw����������OW���7/��lmWϞ�_|�r��[۵]oݺOƆ!�uN��C����O���v����]��z�zq��������v�������ß�V�=~e�}e}���;�ˆ'�m??xۯ��'�|���4��ì���I�_�	��ӿ;0�Px��2�e&	��=-…ӿ���'RN$������W�ϯW[}�����)w�_��~X?~��f�|��'�|Ӿ�ʾ����ރ�|{�ꕅ��\x���y�+�n��&��{݁]N
t�Q��\}�',yÑđ�!k���@NZs4J�LuE(}M(�mI��]JKd2�d����!���9��5�V3s���I!�O����X%
+4JW'b:�o�5+�$0n�s�j
+쑀M9p��Ę�&��,�����KHpj�8O�i�.[ԉr`�%�~��o�[�wp�Va:�(ģr�3��e������O�U4/�&��R��C;U����z
g�
1�啫z�z��Ϳ���g�~�=<��؛�[$����y�+iL�&
1vȍ���1�nե�G�+@�"�'jxoô�sxPZ�rAr[J�����"&Јxd��j��.p��63G%�Qifs
+���7�G���O2��}�Ƴb%�j%l6�q���d�v%��u���\mqrS���tc��q{^@���X���)J�-O�
X*`h6�s¨��;�x+QR��3J�1������#@���ԡ�7Sg�M������������Lt�����_g���,|�)$뉪1>����,	���29[��g3�S��뭷�ɮ?{�@Nr}�^�A*�n&>5B�pn��
7P:��Q�
 0���I���9Z�{<�A��G�j��~6��
+Kږ�2Ȃ�RL���x��x)�w�����go ;��!�i���.
���$���phK�`�I���%��9Fxm���&v?��4 8��b�;�4X\*t8e��Z!e��`�ijwUӷ�ѷd{��/ѷ2����jI���c�r}Rvi�ro�s}��#��
+CL�� ��o4�>4�L�I�h�eA��,����B�D4�fk-������I�x(�D��M�zp�������`ޛ!�8�<U�
�b���3�h켱��{����Pt�0$�L�TP���`� 
+�b�{E�8<��y��ȑ�4��2�3��7z�E='�lw<C~DjSj�^U=���ax�;������151AC��W�V\��L��
+F�p7��ӼY��lt�@C�e�sp�J�����L����Dy��쌲kl:S]�p�״1hژg|��{g
+I�G�+�y��C+��^�_%�������Zf.	}mG2��LF�Pqhm����E"�2�U���:������٪�N/�)JŠAkʦ��SN&tvI%Y��C4���ef'pGB����±Td���#�6����*���֜�FC�I4��G�π爾˜�45����t8�)�rE6Xp?
�X�����O-Ҧ4��n~������0((=U,څ`rt��ʆ����Yl���,�d��\�A���^��qq
l���N���>�0_�kj«��)
r��������Z�M1���E]��A˝����V��.���Ĭ�A��q����|��azb�s`��\"�%��ک���3��C�v��U��'�M~�,�zTVM�wI�0�D
+�+�DP���i���QB��'��	a�{�FHc��w`	1� �$c�wTlⴳ�X5%Tk��9�縤w����:[���5�����5��D�U���lS�E}Q�\�E�r��De��v��-Q���+w��/7hRgl�����CK�;�}�[��z������QKSRr�%H�ĶI�C-sK�����X28����TZ��W�v_rd#CX��'�9O���L��T����{Ĝ�������!!
/�P�.:whzq6Z��<޳��k��B�yN�/�
���@t����6"8;6��f�\D�I5�ru�ԬH�L���hݞs�l�N��a,l:ǫUMtZ��Mlj��ji<��;tKn��U��`��Dk��݌�\tU�:����i0�		b,$�-eH���ߴ\�-bC�Z��h
+3�;�%�)%O*^Ni ,,�0y%��&��x�p�Y��c&�~A�r�W���m�Hx��50���>n�r7<0rx��w�aq�Z�*K�)�,TmzC���ʐ(�����t��&��e[#7�c���Z�O�s�j�=�lR����>���'7�
y�����R��f��l4�r�jh��3�#�����	��z��/��k�vԚ�Ӗ��
+,�N����b(�(���1x��{
�m<�E9/�'�ʋ(ξ4�Ղ|t�tC\��Z*����C�.�C�[���V;���R�^�G0R|��|12�t���.�[�64��O�~�Bؒ*�E2���J��њ��5�p8�}��~PI�I�L��挵��s;	�ЛD��zD�R� Vr	�x�ᢑk\_:���yɁ�G�哟��$�]��H�$����&	EQ�L�ť����RU۲��Z�mm)(��`����\��7�|0>�>���B_@�h#��B��87q�=���d���W�:�M�a<�#B�T���3ִR����ۄ�����j�5t���.h]m���I&
!%���
+@8!A�4Z����PH�'�����B��향�C��˨�W�q��.ؠ��. ��>��U9vw��c[ꏐ�{'�\���4?���n���O���c�
��6ˏ���柵��\�KZ�`�����
2+�S�0`u�ԋ��g�nX���n�>��:κ�>�|~���c�Q�£��8��'X=���k��r���I��!�L/՝3"dB�e�)]><_���+� ���ܶ"W��RƵ�QjQ;8g�Ka~���L9��M̾��E�r�.4�k
Fh��/t{A_�G۲��z�,-�h�&a�|S'�k�&�4�
+A�ET�);h+��y8�o��by9���%�4Jl�H�D�d��@!9���q1�tr��O��4v��C(�	��ĮL:�\/F� �wZwi
�*
�Uˢe����tòN>��ŝ����Gx��Rb�'�Ĭr�P���T�ů�q�^L؁��Ep�����{�
������K)���g�f��[*ѤWő.bT�n<���/#%ڙ1ۅk�'|���͏Yc/��'�|�k��P�G������n�{)��.ޛ�!��5�����tA
�����
Ur�I�Ungz��1
+�mF$�
>-��S=|��\�r�	G~"(qlE�/��h�)�>����dbn>��q{c����}>Ej��`1��X��\��ѻ��F�́ɬ�{�̂�R�ݳ}�'��ʟ��Io����1�����
a|�q�	bD7�s��.�n�����h.�E
�8$`��nE�5��x����%Y�������@���_�?#>�O��L.�a˷Sy#�ٱdR)�(���xEt\�BL(�4���/�]g|���-����������Yc2ǒ�az*�R����Z� .����#�R\�>��>+sP! |���(t`Ha{U�Է��TSXT��Z]\�����)m�3��ѣǫ��6_�CC�
j�ݓR�2��9���\j4��TSeY�8�����n�l�3�kz’i�
.yĖ,
+Q˴��7�*����Ө64�>����N��O&���c��Lm@�Q��7��΢�o0�_�C3j�Gi!�1��
+��)�5s�@l-9���9z^O㇨�TΚ���k�9�9���G��7]v�;��G��_��4�*�D���_�jM��l�x�㽰(M�7d���7�/q ~���]K]��~ �LQ^GPa�,��脟����{r��rhr���
(�|�ݰ����Ԋ������~�Z�nf7�=��Թ�?�����3<�n<P��A��
��[U�w�(>B�P@?Q�G�qk���"\�.���wW��!㙚⩥_�:�l���(|:p-&㺡�%�$K��\���L�A��8���x���Gn�s�Ⴭ�{�$�
+1n/�I�~=�ZWO���5>��-�K2
+1���p�+�@;PG;�qh���1��G;�M�M�N�H^+�c�O�w�`�	�=0t	'=���
D3v���^���-LKC(��rLR�{�~��ۛ����=��o^������]�x�߿���������~����ݻ�����t}�������f���*O�;U}�N���e��&����;zXe1g~�����С��
+endstream
+endobj
+
+110 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 97 0 R
+    >>
+    /Font <<
+      /f0 91 0 R
+      /f1 79 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 2
+  /Parent 1 0 R
+  /Contents 111 0 R
+>>
+endobj
+
+111 0 obj
+<<
+  /Length 4338
+  /Filter /FlateDecode
+>>
+stream
+x��]s��]�b��n��(~s7����Ӵ���y�T�M�9��e�����ݞ���;��HA@q����z�=yr�u�/�����^<}�=��ſ.\g;�m\�'c�JWR4�л���pq}g���.l������ͅ�n�_\���s��ۇ����.��䕵��u����Ӧ����/y�yu�%\v>�����)���_v`o��e��w��[��a�Y�&�(��p��=�-M��1?�c���AF,i�����G*��)���`ǔ�Bf���7h���1�Ow�@4F)s
+[�R^�0ry�,`Э8�jB�i�7��w���_�\���˗_\\ct�
+��X�SZ����`r�%�NP�N���]�4PO�<��5|UmB�7�!n�G�� �x�f,�"��乌`F4�Z�[ ��nk,�H2��9:�*#���;4�W5,:�0[*�U[���>��Pk�a�������GE�m
���v�"†s��5�h�5��,0��HS��C6*���H<������*ԔS*(0�iS8Ӟ#��&�ưT��z̭�������VSTS`��}�+\gpL��iَ��ʱ�I��Tb>�`75(�����%y�V29n�,���`B��ׯ���'�Ƴ��G��v��t��L����G%�;Sγee��2����\�X?C���e�5J�Ji�&
֯1�#��q�E�N'o�8��x���
+�����G������ �_Z���D'�-Se���v��A�s(�й�m���U�ȸg���T�*:�UN���:����b�C�&�!��(1$t��2j�u�y�@<�A”ѭ|�w�MWU�D�	k��Z���)�cc�~ގ��f��w�VUԯ��z��bW���=���+��=���#���޿���w��]�Wx�����M��L��O�;8[�;��ؤ���
�$N��??��I=��z��&�i�M��vz��yMz������3:e'KD���Q N����`�� ��'��m�7n�tˤ��-W0�
�)B��^�6��}���
+�ս�I��J@oM�n��z*)�e����{r�%�9�O���`��#Hl:L�G�f�m
��L�r�+Wq��Ѭ!���)��]!_:)sF;%�D��L
{�)�����F�y�:�m8h�o\-��S�&�aM����r2e��!����N-�]l�B7+px@p)S���`9�Y�$��6�a2��j (6�	�BuEr�j�jN��؛4�~U}����hs0i���2�A�#�mGy�p�VLsNp���	�TS��(��͸�x���bev
/�К����*��{��y�Y�3w�v�S�)xS|Vo!��}�7U��MH��_}���-�;��#�4�L��JV����E��1&e�?_��<���ǯe\���XA!�X��R-�Y��tn��Ԏ^���y��`�&�a@.�VO&6�/ۻ�vq��d{�دҮ{Bu���C�^͆%�M���""GR��&�&f��[�)����8��xV<�I=��|��a��N;��M.���P@�nj�
2'�!���v�p�i2Ù��{��P��q����p<`��5���5C�~D�BJ�^��F��*�,?J9��pS��Y�f����̥�ى�ᬸ��SR�78�pb����5 �C<�P_3�3�A\S�XM�Ԝm��>e�&gq�%�F��+'ܼ�0��aDx�y1�\N�P�C�C@�>C�Ɩ�H�b��8yȟ�9��”��G�f0ֺ�+���$�V'8c�jʧflc�ð�чK�A
+��bm�|Ӗ�%I���
͹$H#
����D�Q���A`�jV�D�z�sV`��T���~Q�L�xk�g�4��rL��܇U�t
+�f$|38c�+E��!��	p��8�F�ϯOg��<PTN��xy��4�� �N%�RR��Q$
��:D*���rY
�
+�Qt�!)d���!�D�JT�v�X1Մ��!���P�#�Yn;	���?�K�`�xn�
+)Z��kL_0S�e��[��
��Xfs�4LА�=���Q��N͢G�M.ɕ3׳'obBR�v"�	��N@Cn
+ϊ~������K��"�u>��'��/�T�Gn���ڂXhͷ"�w2JshR�8�U��xNK1�Q���W��Z�۰��˿�dR��Z�a��$�aϔ���h%�đ; ���Kpnd�`N�$�B���lK�ˌ?ys��<*ӷl
	7q���7��w�;�4@o0��(�;c0o4Z�T�>i
}4CX�4��v6țR��n���޼����ds9'����zj�^ʃ�3ևz.��HЯ�XY�i|�

�<o��7�x�Z(uO�-{BLT��tS��Țc�2���h��OQ�!��;��#V.ڔ#/M\`�d�fZ�UX����~������M��ѐc��k^�J�.e�Jo���f@�\�"�!�H芪�("���2j��Ĕ�L
�G&Є�l4��8
�]���J`�m_
�������J��uk�^M������\<2g�]I����51��<���D
[
+t�+�<4Wz3�Іztir�m��"�"����/F@��։�J��qUF?}B��J\#ibu	��߯Ϋ���	����4)��6%�(� �y����x���ࢌ��S����2�`�㘾�����PL��x}^�G"D]��b�B�Cݔ�+�u֋+�㜭җ�8-�մp��Ę���8��D�bM\��Sk�fޡ>���:yr���/p�ɉ����l���S.cb9"���\)ĸQ��
f��p����ΨR,�L��YQ�4q��A�����Q��?f6q�e^&�42�c��H��fF�^��&E��ul�#W}�O7�l�����織��O�揵�,FS�&�[���O��Oe娄�GDR�������`��+꫸V���B�C���ms�����#w��j�Q�d@qҺ�}䜊!����=ME'\�[�8���cQ֍W8X�Q�̇m�����/�}�M�����
+]2i�6x�0l�E8�*%���^Y@�� ����:�A8��x%�
$�:�L΁��ݗ^��}�׶j�����P���)k�5���$����
+���%�T����'��B��N��Tj&�����c��`4��IP��Qe7�$��,ʤYA���}�Y�PjZ��>�P���(h�C�lu�|vӿޤ�<	�m�~���
+^<rIL�by�,�Y�Px`1!R�!J�F�(@��er$
+��9![�>^�F�G� _G� j:��-����Z��u�:fW�b��[�EpuHl%�Cs����	Yzj@��=�{!�c�H�0fe�zf�˱�����j%c�*�Cq`�<�����jᎷɄ!�a}��dW�G�c.C�?`^��W?�-���B�1�ľ�2�G����@m_ف󼕔�"b�|����v���!�$�f#ˡC��O��e5�L���uX��� �r�<�=@Kh}��ٍ� K��$��~S� �U#mhFko��J�B��}Z坒�A#PL 1W�<�a�D�C�e9*VE�hRS5��ŕ���t/ .�$|¾dǾ(/3��������/�)�}'_�1ޣ�k��&ή�w>�-Y�lŁ���v��;�-ɤ�T0��'(��S!g�����o���j��K��~X�}�x��`�h��%\�&9�x�,a�"���c�ǰ�㮕+I�\�q��H�i��������ڒD��l
}}_Y��sG��S�[���Cu(Qۆ���)j��������M������6�L��.l:D=�����:�Q�OJ"=U����K>����i*$z�����g����S
+�+����O�l$ζq��i�����2睱}.k:��u֛�sv�2|%�4�~�[�Ky$g3����s�t(���]\Q1$g�(SX/}_���<�h�E'pdi���Vi'ﾁ��Z"�5�TkM�Mƻ��=#��
+}Y�o;ٸ�7ŕ$��K�@U!(�׿:�V#j
��7��Y-:L���H���Y�^D�e5��7kDP��*�C5s���E{3C���~0!��&���"q8}��pq��%s���@n�9h��t�l�@p��F�(J+D��xZ�A8s��:@{��N�D�?�X�ml6����xU���#�~�~����w�18�۽�����v�^?�۽���_���v���t�_�{�{�~��������o�o_��e�$㋍]
+��D�XȤ8X7�@,� v�AG��9�2&
+endstream
+endobj
+
+112 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 97 0 R
+    >>
+    /Font <<
+      /f0 91 0 R
+      /f1 79 0 R
+      /f2 85 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 3
+  /Parent 1 0 R
+  /Contents 113 0 R
+>>
+endobj
+
+113 0 obj
+<<
+  /Length 3891
+  /Filter /FlateDecode
+>>
+stream
+x��]ݏ�
߿b�$n����=3��v�i�&Ƚ%y�v�Y�����}�w;{C�GJ3si_|�[��(�◴�?���}�l��^��/_l���Ϸ/�x����m��n�ܶOƆ!t�.E���׿m�_��������yq��ۛ���v�������ӏ��6?~����'���g��y���p�7��?�ӭs&�������v
+@>�|t���<����q
+���q
+�0>"�l8 �[��:�eŇ*V>G]}��������ѱ���l�e�<ɕ����
 .�*�HT����&�3,��6�������=�vP�ӟ�7_o���|���ۗ�k��N��0�sJ�u�0%뢑����c��(�#�,�)�b�=^cOc�w
+w��y��uTH.K���[���:�Tϰ,��)�-�w�C,�������-
�k�*�{k��+$��B��R0����f�&s�B����n�v�xd�,X9kqrO�O�XC��G^������pR3r�sf<��`L=϶�^3!kci�o�|��'U�{œ(h8N��q'f��X����hr�5�7��LJ��ٖ��XZ��Ba#wt��Y.�lj�s�	�la�>��1�ua+i���W=-mo�Hj�&="JYY�2ϕ���Ɉ5~�MeM�кd{�دѺ[Z��t{�Oe�|�M�u!��y�q&H�ө�h�Ol����iDU���S�UF
+&W��	:�92{f\�-��>���f_�a7+R���*_��?ʆ�r�r�pG��q
+�A���m;a8������T�
�6uy��~�Z�x&�Y�g�
}TC�(8�� �$t��"n�7���G�}�	���bݥ�P�[t<�U �qA֎«�����ԱR��%�0���u�RQ�^U6��_�%oʑ<�ۘ�S��JX��2J��쑡U�],Ʒ:k��p��qJ��kYsG��\?N��"���N�p,O׸��:��>[���v۬��������C�\%��X����G�=+A*1��4�N���R5B�"�g�n1�L�?,݌Z�OH�p��:{���9�|� Z"\���Mum_�mI��|׊j1����-V�I�1Y3�Νm�_�6���[:�P�A
+�x�/�n�ϊ��<�~�R�^]��/����ŽE���Si
+`|(�z9�.#z��Ko
+�����f�S+~)&�!.��k��֤�Kv���F0W
+(NC�<H1	����/%;e,/2H9R1)��b����2�1�W���ƃ�a6jW�/��(�����D�����j�9�`R����ǥXCn��:1)����;��`�����S	�BT�;S�A�K�$;Îe�xl���ڥ��{�-��$��d_5���c�Bgr��t�v������
e�"Ꝑ����]S�S���eӧحN0kF<{�{�vLr���2��6��U�=NaB˲"\2*8��cb���.�No?�y<�S���64���X���ՙ�d긁�F����6&U�5C��
-
�t��CC�C�l�*��^����
+��а���:�G��#�r�Y"�$L���|�t	|k���=��-V���(��=`���+�_"��8�ִ'����۫`����񦶲�2�~
�Q6Y·��
+�0���5��{�g����F�ȉ�X�D3�?�.�L.Cš���T�
+ð����1����ң���@`�j�q�v�z��g�v
�[ɠf��0��v
+H`x�v�h��.0+Ol)����B˖�m+Z�leZ���5%U{&b
+&�>�9)חa���7)�n�J>}���c�cv�"S䂛�
a&"��LT��5���y��b&������:d�K�✕�,�MT�a���P�
+F�;dE"�=]��brt��.��y*Z�-:L��U�-��Do��!�QH)�;��Z�G1=�>EA�62�� t��T��D6�hIe�	19�����
+
Qvp����W���Z�۰���$����!_��L��J�R{K����*ʄ�������0�Fq�&&-AP�ܘr�|�Ԇ�	�&�bJ�5�R���f˛�aY���~�&��즧��]Hl�S4��'�tL�5�Ȼ�A��:�gn�1�Km�|��T�LZ�"�_��`�Q�ֆC��vU�?��/��Y�,�!M�x�S��ԹTj���⦠l�1Ej-�x�D�R,�;�x�$���'��>��gё��F��� y.�	��_��%=5�T�B
+��}��ª`�*/�@�!���$�[j�O��Ĺer+���A�U����RFy�g-�l�z��N��:^�|j��X��R7���0A3WTA�c������d�R;NJ��ѫ)�ڵBg��V\�y&���`�0X���ij/M�:ؠʠ`3/�$�Pt�Ǯ�	~9�����X�|�٤�t9�:M���n�b9�I�a��'���}���Uu0S9R�U��ڣk4���	��p�w&Ɣ���!��h�j�sm0)�.�w�n�p�ۗ�j�v�U,
+��^<�/�s�����bP
dM]�c�m?L:8��𥇩Ȑڪ�+݉�Ķ��!"��O�S��|�Zv��l2������.������1�3Ct�I�.Z:��;���+���JL�b~6-���4����o^~�~���S �a��uC�[d��2������,_��;���T،�3��?���]�J�t�CQ�ǧ�qH�:�~8�|.��XUȬ��KBޮ<BSdS*u:����q�O,;�bwx��
+�RZ+��i�I�\�����nQ.d���j�QP�?>�����<
�L�y�`�0L����RAd3�wB"e�k]��;�0��P�)`^�LK9_�;ӓK�H�|��P�o�	|B�$>W�p� (�e�5�J�O�tR�z:����E�S�#ɛ�G�\P��[��/GoH���Q!
+B	5��n'�S_O���=�
+R�aI�Z5Ua��K��O�]��ZƬ��8��:�B���"��@4���d������N0�yfa���O�	<g���2&��}-P��0K���$�T{�I%m�䟵|9ڔ��ح�M��{j4�5�T�9�;�s^�l��mEOo.�)�ᒽjj���qt���tʩ�ë:���W#�dz���4̄�y�\�]Ij�3Ư�P{"��[٘&�i���G�x�q]Ͼ�|j�f��5�l�Z��6�0Ĵ��
�%()��5d*M���m.�n�Z!��0/"��`�,~i�f���z�B-��L�V��[�G��^�?ͻ^�,TS�=�RV�!\�&3��m�Uɚ�Bt��&��X
+� �
+$������s��6���~ŷ@}�&^�%O\,ߙ�G���g��oY9��Ӈ� p�x�+����.�����H	7���8�/��&�eP[��>nE�d�R�R�5�sD<�#|_F5.�K����<V;��hiӱ�"��X��5��]`oU��[B����n���M��l
+���1
+��3Hk�-����-֖�LQ������J(��-r������P��ߐqF};�V�lY{GT+L�%�b
+���ҋ�$Y�L�5��>�B+�'О3����/2Ռ9�C)�wDKh�q<Y��xR/z�E�l�wB��%$4����@zˮ�w��V��Ԓ�-�@k�LU�u���Z�㻰��R���%��Бk��gyAM�o�Q�����ڲ&񽈱r���:�X
����� Ň����� ��f��X�n/P���HN���3*�WW=L��?������Y���^���߷?^�xw<�����_�������l��z������O7����_����;@�>�\�)���ԟ9Q_��`ݐC�ǃ��E|���*F�
+endstream
+endobj
+
+114 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 97 0 R
+    >>
+    /Font <<
+      /f0 91 0 R
+      /f1 85 0 R
+      /f2 79 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 4
+  /Parent 1 0 R
+  /Contents 115 0 R
+>>
+endobj
+
+115 0 obj
+<<
+  /Length 4240
+  /Filter /FlateDecode
+>>
+stream
+x��]�#��ݿb���5X�����HU�PI`�8֛;B*��a����c�[�-͌7��s��juK�)�~��îy�r�4��|�/������/V�^�F7�٘&�]ﺦ^�>�<���}���+����[�ݯts�~u�V7�4�o/��?�?�����������W���:��Z�������_w�8�:����p���m�g�!��/ �����N��������F:�0���������O��@������;���g�y�����{a�*w�E���mf�0�4����_2�������~���W�|����jDeu�Ҷ
a��B�c�v7�7�{o�e���@���Ӗ�w�	<7R�C�]
+�f����=����=�V����#��市g̉�hciHRK�h=��.��m�����}D#\����f���tG�IU������ݿ�Ͼ�xs3��
^w��Q�|t�����(cT������b��S·@RşV<!�S&���N����4:��^��Zd"+m��
���ތ�8��K?�%�|��"z�
���C�(O�8Ş<)�ӷ�r�n�����m%��g���-3f�%GZ�1i���u�}����^j��s~������q�����k$T�
F<m�P��'����&'%��a$�jE(gH��3�֊�����3�Eb��)h�f���|O	Q�Ri�6�U�Vt��\̫��Im��*ߚ8_m�+��V��A�of?���=��G��+Oû�r���-Ϧ���ZG>[YՊ��pVH`|&tI�L\�q�PG0��f��Ԑ�5�5�c/�&
+����|$pPֳt�u��*����ɳ
x�l\���w��킞a��'S�Ȳ#ػ�F6^�X=��!�=邟]etRm���6Q^)�Z3?�p�+	
+8
gE���1>(�BX����r���8�L��v���:���)㛥f�:��|�bt]���r��&L(\ki�ܴ'�v�!,���ʓKv}���@Ѕwx61��/��r���g�����`AI���.��;=?�,�I�
+��1AI��?�������]��{�U&*m��Ϫ���ѡ�@�@8Z�{:B\RI<�r3P��Ć��������zC����r]l��jJ���.]��(5��C�j�PA�:���-&���	�MH{���2]>vb��,��)?���f��'�<���X�ej[*
�[��5.\A�8^ҞXz���C׈��V�!<�/Yq&��ŷ*��␴��i�
���Y�����ڲnB�"�@�*,�� ��4P1(*�T�{;��.�4��vT�i>Ƴ�|�Q@8��ro1�4�I�b�l�
+d�,�����Ov\ba%&-��L����#�W2�|��]JxI�f�v}iX�,
+�l�B�@ؖ�5˕f�"P��	�b���D�xW�¦5<�$�#&�B��7zI�x4����b:��q<�\��NRE�6��(�ORL~�%AHa0�,�iKJ>���\�k����줛��v�B�K��Jա@��9Y�Y^כ%����G����Vc��L&nH��q�X
��d6م�0y @�O)E���i��El�„L8 �p�g��š�Q6<�{
���e%�w����dP�i�'������j�.,)��̋'*uwU���\j�o�ʷ�h �S�һ£��k{1�.dx8�
+S�`q)�^�oR��G�|%��󬲦t	���\�P�^�ČT�Vu}�秤>X��H���U]�2�SK���n��8���Ɛ�5���qK҈�ͺ���	��I�f�Kˆ"5dbA=c��̲��vl��e��)�6�iAL(ݪ�z7��� �p}�1w[�lo#k�"�Dqa�,�gvsf��;�������DiʶɃ�n��;��ļLo=��0c��I��#��4}A��Q���QD���4pe��r�E����7J��2��Q�&2�k
��������"S�pAfWLq�*r`t��
+�+	�#���AF�	����&���±3��fX!⤬ZFJ4u�O	��5͊����Dc��i�g�Ti�b�ꭍQ��vL]�(r�cQwp����|J$��2��W[��������PaБ�	]d���Y�ی��-�'
+!��Z�V�ׇ^9ݷK:�����K�2�
�C ��P�.��.V�b�+w���w�U�r������mtA�MU��?���Ǩ�G�`�����Ӿ�cF&��&6�����3��r���"}S�N�J5eVeu��x\**��Pѵ\��+� �=i<�7e�PIT6�x
��QK'�q�?�P�4�v�nZر${�d%�(J Ot�U�eI@�Ԟ�V�������-a^j����t]��0��>1BH�S�GRf�Ȉ30G�� �W�_/����ʀ`1󨦴�g�7���^slúǒRQ�	�$�Y�"�1����
]!�l��̄��TR1�\�q���=�Ӌ��[�y��Vf!�2�u����ҭ(����*z��X<v�!�P��y)g5�e�%h����%M٘
�TՓ�;���:���rB����b���Z��]�M�69�	A0]T�y܋*ɀ�|��Qy�����m�KO���M��zG�S�^�>�*ɑ��x�F�eӥ-�tl�E]N�UQ��t��\����VG1�⺠�6�@��,��	��C�Lۇr^R���׍'�m��<[C��	��b9����X$�$�@�M�D�Z`�Ө��X��N�nd9U����N̳��+���7O�oE�ye,"��Ъ>��g� �����Hc�0�W�'S����1T�Z!�7rcl��!9�풨J"�èvQ���V`?��<s��è�يa�@���Q�loZe��L�
+�Es�*%���%[�J�{ɜw�S9J|e�Ruѷ�X։`8�A]5\����3$|^Z�"۩KI�0�0T�R�U�lR�)��N�_^s�zf�/�ō/�[cM�^�	%m�М픋��a�_�)9���/k������
+
+?#34���Πl�\����r:���P��3-��4t�-�:�X�f�\��h�k5�EbWAq)�4U�e��-g� /��>���m��晿��C��������n_�d=R�w��_R��p-��e��•1�%R�f���꺕f��M$K���q
�TW�&	�H���bC��j��
��1O����eY�rb��v^Ek�~X��[�b�
���&|�������c$d�Q�}x��<�
+�L������@�i��%o]��+�X���a���3Xu�s�S�O,���\�vE��Zvإ���bB͆����%��D�F�.D���s����T��h,�Z�E��
�FL]��汜j�[�sN���K��T!m�öխ!�C_�Q%�S_�;��~Y�!�t�x���o�X��+�N(��0K� ͦ$�j����B#N`�v�gD/������srJ��p��+����$b&��NY�-��,��er/�#f�~0z{��#�6}�!�~�Ѿ��x�vu?@#����);�e������V���z�{�c�~���V���M�����l�16�$ƒ����M!:��sܩ~��ț��} �����/
+S�~���H�?���B0$l�\v㗾�A�&V����}�t�L&.{;�̨���0,a-~����’7�%@m�v��A.�
+��m$�"ȥ(�tk1UYZsϊy_���_��o����
|�)�\���2݌�+0���8e�v�<�O�.w�B%�z(\�t��C�>c���R[�$��p�}Q��l�qVi�U{]�|�Z�X�"HtE��4�1��A]8�N�Bܼ����d���g���C̤�$�Y��"C���介����ž��FSJ���_
���I�${I7�����\�-�Of9�	�r��ģ�+�N>����E�oe�Aمq8
+���m=g8y�sG��!u*��.�m`�G�ݐ����v�
[�����AF̀%o
+�@�i��ؒPq
���/�b.Zn�ث��T�MD���b��h�l̂x)�R�����%!���PFT��PPa�'�2�䣭�����vC�F2�P���3ST��R��ʙY�ŗyV��:QG���W��?�}x܏��~����7o���{�߿����_��m���/o��?�{��~�����>����a�ӻ�m����&8e{���Z~��RV�m��h���$�
+endstream
+endobj
+
+116 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 97 0 R
+    >>
+    /Font <<
+      /f0 91 0 R
+      /f1 79 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 5
+  /Parent 1 0 R
+  /Contents 117 0 R
+>>
+endobj
+
+117 0 obj
+<<
+  /Length 3771
+  /Filter /FlateDecode
+>>
+stream
+x��]���]�b���-G<~sٺ��t�L3�N�-΃ϵ[w&r������]�}�iA	_\�|�ӛ����bn�}��W�^|��������3�A+3�Ai�]R�j̣q��7o�����~y�^��[��������ޟFo�������k��������_m��u����y�f��Z��.���X�h�tX�QE���|J����Q�J�G���O
"������(�?oNBi�|�k�4G���s
{|F��k�<�YL��pX��l?�M���Z^�
作��w�k�})�݇���������w��/��������e�ma�^]T�&g�� s�@����ˣ�Q��0HkB�(���?l@��XM*�
H)�=�F���z("1�%�0&��!�EF���m�ƒ�^����Y�L�9MW��ؐ>��j�3c?=�Ƙ���8�����G�\)����#�X`CO�B�F���� �H��|�,�9y|�X
`;*IC1~�	?�u�VM��s�o3,I�؈h��'�u���.,i7Tى������5��n��:��`�
+3oG���9`Sz8fv���A�u�#�C�KU���Բ'�k��Q\K��mɱ'�bP$Timi�jQ��..0�?�˄�x�-�0'�����KJ�Qޘq��|Gm�U���[S	�T�V!k;y*pw�\�ʇ�X�c�1�����!���QEZ��*��
+Ǹ�,�Je6Su�)&����uτ��VT���n��y�
���(�a�*�1���j���.n�:ik��\���nɥ�S���c��ä�A�p���xk�S�G�-ף&��ڀ
�G���ő��H��v�u˝4n����[K��SA�ʍ~��Sl���p�o��(*��蓞���bՍu*ڬ���򁠄�F�R]I�…t�꬙չ.��8�<��\��WڛN�A
sW0�!=��RU�Q����,*��*�����q�N�`R��*}���'ͦ�t%�;p��`ЪC�l��+z���>�yۉ[&���OZ.(3q'TM��~
�¡����$B
����@���y�VaF�xF���._ O|�	�b��JIi�1���|��ڎ�Ge[:Y� WB�pg�b�`��U@׎3�5���/"!�U��I���ZYbN�e�E�e��?�k�2i�FYggt�z�Ju��Qi?&�-�$y�qY�_H�xD�H-�	�jS
5�K�d�~��E;�2%5�hN(���61�_"����p&�*�`r�ӅA��)��3p���.w��X@
)(����R31�1��b;�C��N�U��>Q�Q���s�r��xx}u���+a�5?���Z�)�"5�+�
+�$a@����P�:<^��o)�X�!+?f3��	"��<'K�Q&�\�>`�I��w��	���4容���J?�a*v�'�*揧=��(눺9��G-'��
�_��
[�Y��PJ�ѓ�� -����;M�ӝWI�|�֋��]P�j�=F� B��%po�[#�l��*��$�,�NJ��������'*���	��#�D�ĸo��v[���-���@���,�;���]A?M���<�Rg4ޟ�;@�|`��6/�.T�Pi����mE�Ė.���2� tT�z7��CtvvS�X���?o�V��H�4�� WCľ�������#��Y6�`�<V-韭��1��v�0WI72R�u`�P�����I���l�vF[�3l�,�Ggt�u�

�m�y;0rd0�m��İ!��/��x͔&C���ɞ�fn�B�ꡁ�u`zgG��y�U��o��V��>eF3���2�֬����&�
����@4Õ#I�&���7	�B#�T۴C a'^
A��|����Q��=�쎵�r����	?=n�)|���g�V-x�ļ�7a�6.�r�&Ui��U>�+mU	���͸܊�p�F��mp��b
+栃73��Q���B�k��•�t��5����.�,�N
+��4odL|���'q1���(3#�}�~tHpտ����b��[�������]�U��Z9��9����
�-xtS
+!��%�RO�mCyޫz�W�	u���eX@�$/`�޾p٦�v/�y
+�:�`!�H�)�o�౮�^�'����L���uT��1]���Dm�:�K���@�{��;_�ɱ����S$�������̴�EY��ƪ�}ܘ�чjJ�cg�T�[J&V�]�jLnF@� �U�A<n�+rHbr��&V֊�h�
+�jr���%04Vʞ`2��V�����Wt�V��4O�h|=c
+�$X��&��%r^K3qՠ���J]�&��"oi�X<w)(�ͬ�d�����*%�B����O=}@����ip�@I���x~��-�'��Θ
�7�y'�P4���a�M9���YfjYA�V5��[�@��Q��b�=7�*\��8=��mW�����^�����;!�k{�����.�W�
��s7�(M�@��%2��an�ղ[�Ը(]ϝ�q]�s�Xmv�)�0+]Ky�'
Ӻ'�J��%�o(��H�7`/	���b��Y���yz7�oF�����Z9$�Y��<e0��D��u���\�Wd�>s1��r�
+�:/G��n/��x���.�V`�[)����*��f_F�nTW�����t��LC��K8/y�B��a� 
+#���#�3��",�u{���n����M��FVZ�]��ub%�f�b��zzq�7X�r��k5���*�UNi�G�S�s^r��Cpρ�7���7
a�T*M�
�&�k�r��CJI�����J��2�,��U, ���h���e�ι��GG�ׇP��׻�4�����1 JU�TVM��3!��5�'���.RJR�����+4���������{HЫ@�E[,�{"��)̓%/�1�l��9���Cy�=�������(֩��ʙ8C��W�������ۗ:�3L��eV���ly����m��"�erP�'<dڥG/�$j�Y�A��P�3�-K�*f�j5�] ���Nؖ��%)�����9y1����ms�
��^Xa#��h���$h�!bN"Gvq��;�-��QC����re�X}s�B�^�s/�r��i�{�Źke������9$�qkS-��-���]�r'^�4�W!9=�K��,\�*WW��� X���a���tV�iO'
+9�5�9��,�#�E4D�S��
"�/-�8��f`G�u��LRA�CqJ��4^,P�8����7��W6)�;ҭ���,v+�����3�x��x�z) m�Me�
D��1��ɽo<���~,���$gt�u�^�Cߐ���Zz*�<M�j�y�K0�|����BI�Y���.����h��Sϋ
+
�W@��?�P��&U
��9^%)��WD��)ۇ��_���z�],��,tܜ�?[�]���LՋ�̓Ԁs��|�Xs2&*�b�qO�|�$���fЏ@?A��,?�r"(� $�+%m��"N����(�u��z=�H�׉�����u��a1�5�V]zW?`Q�$� ��Ԑ���b��RKe�“�Z�f���ZQgۦ/o��r�-�2@���+�|f�(�L�6y�Y4T���ϖ�FOu�VZ���֙,g�c�؛�h��u|���3t��6޿y�ٳp�ټy��w���y�q�����o�����?�n�����ݧ�Ww�����o6>�I��2�f?�l6	fT�Y���!0�����{����o2d
+endstream
+endobj
+
+118 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 97 0 R
+    >>
+    /Font <<
+      /f0 91 0 R
+      /f1 79 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 6
+  /Parent 1 0 R
+  /Contents 119 0 R
+>>
+endobj
+
+119 0 obj
+<<
+  /Length 4158
+  /Filter /FlateDecode
+>>
+stream
+x��]�o��_qB]7>�(~.w#���@Rh�)΃��m��:����}hg��!wWʋ�Cr��=�]}�����ŋ������_��_|�z����\��]��ƭ�dlB^�M?�.�n������_.����ū����xq�ޮ�[]����u�������־�����9������/{�����;<[u&��'�:�����e	f��!~���ќk���Q{\VH`�8Д�M[�ư��
�V:����y�����䧽��|��K���[	�r��#3a��C��)e���N^1�YQ$��y:4�F����V��\|u}�׋��}}q�eԩ2c}����]S�!��N��Y��`�0���3)����f`R:�cvp��}|�����6��3F~ry�8.*<7��UV��(��isZ���u[��%�a%x*O���P
zZ9�/�\�
+�W���R�����yv$2k|+����}b]!�����Y϶������ϩ��(��X;abJS8h�%(IC�,U�z�@r�h�XR&��)�6��C*RT����>��B���b4��C�e���\;P���G(�>$ǬyCeuWj��&Fm;��/�϶�H	�qo��#�w�5K���*p7ch�(����q��Lޮ�Q婀��Rv���F�+R5)�g�s�l)=��X �T��!�`������Ϲ�`�N����0f�y���~���B:s�c�=�$��3���)t��X��I2Ua�S[kG=�G=[��x��v�t��ع>$�s|��<`������؞���$D��{��o2:���Ğ�Zkp40��>d�Pc���
CӁ��I���a��
!	~J�X�z
+���	b��i_8
"���ֲ��t��h�����j�Q��ɇ^�9��È��@;m���C�3#9��)���GM;�%����_/�/j���'A��Z���2#��d��ԁ�q�N���7,T!�I�>#'�h%P����]V�.�����OH"��{3X�34�_s$1Q����:���zN㰩:V�62=�(R񁡂���a�q����c#O�`Ŏ	�e�t�d�U�
+��f�)ηfB&՛!�C��ϓ��EdG~��;gL����_�����w�O6�N���:D�.�x��ݔx������Q�I�fJ��b8�g��(m�	`9�+�E�O52��@^�A<�."X�@U�i��j�a���#��&���-�h��0���5i��'ۛ��~�����\QT������0,�(N�6�76���G�E�s������h���q�؍��"���rc<sBu<i�R'�$��pNP�F�)�t�0se�e�-�^�U�Ւ���/�ʡKDZ|�&�4��a;��z�FC,�"P�*C�6���^P���b�Z��������ˊb`���aZ�F���8�*�J�%ܝ�.�Ha����ѹw����5Ӽ@��,�Z�L9����p���|WT��H�;m��Q�D	��>�f������
+@	;��$k�@�c�����!bB)^7H ��F����Xf]������Cm{��]9�6�N-I�4��n~瀖Lɸ��R�����F�8:1o�
�|+pl[����Q�p˺���8���S�>='��x�NC-��(�f)��g
+F��	�ZHM1���M�K�����
+�S�X�{��4�O�}�O8� �L��qC@m�em �ׁ�A��[X�I�4Ā��/l[�x��f�tӶ��"V*���X>!�)э���Iu�(��ٓ��:�z̦r�溘Sc�vf�1��|b���ʟ	�;�i��')H#��P�3�;�&^ƭ��l�M�ztg2���4-�P�道l��2��7Z_#�A�.*7�B�����M~����{!M2�b�7��}Wk��L\Ś2�f������I���X
+�Y��U��2N�����Lل�+�>n�ÿ(5V1A���CJMQW�V-�-�ڦ��W�m1'�z��Mt���ɽ˰���y(�̖?�	k�8}��Rʼ
���9�(�Q�V؋�‚\���U�*��C��\"ʼn������l^���4�`�nQ�
�%;�c7=>(#u����|Zi�6�֐L ��؅6����;�j����z2�-�Ta]!G&�W��B=RϦ܆������U9T+nK��8�%����_��J3C���v�G{�ޟ{��D���f�p{NS��ԏn兡�y�s-�i]�`w�ٞ��tP�}m�.:�$�[�
+U��ʫ&�jQ.�lR����n!���8�&�S,���mRr��%�� �r��/�.ϼ��?�Z[Z��e�pB�M�^�
+�V����c���oR��ȏGѫ�<�|2�����
��*e�O�c;�G/���~01X����{���I�7�'�[��Aa�<3��]?��y���d����x�����ѧ���Hr�T��Hʖ���j�
+�Kw��9/��Q�|Y�����K��
T�!�����"2�iٙ�
��=���������.P}0Q#BɩF<�fF
+�DD����賉8'B��=N�D^C��\"�5̽��`�>)z�)��Qr�"��Q��N{״���c�&�;Of�c�V$kbH�㙍H�WW�V�a+�j�o�y�����Z�Ġ�ӝ����i��M����#O�6ńH
+Hs4���-q�5��
/4������1u�zI���hs@9�m}�P��mqme�������'h'�� ��_��7�
+�,�1�Є�+
+N���"
+覭Tj
7��Ӳ;�mWfcݏ��XMf�BH�ɤ�C_=v�I]F��<0�D�Fu8h$�0�S<,����3�/��@���j�؎����,N'���Q�"���E޳|{��y<����Q��5�&����L�?Z��
�3?�f能��\�FX�����T�x�Z��A��JF���w�D�����e������g�I�Jx��[凜ζt&��Z`�9�>�aQ�����(x�<)��/Rz���e�h����@Z�0�<b^�I�p�p(x���0�Vr��uR������e]��;�Y�/�f��Uy�N��V�a�����ܼ�j�O�iK��j��w)�<?�C��hAR�Q�#����u�u�W�+}��o�B��)P<|h{��)����1'��g�p����.*�)g����T��W�?���%�y���I)����b���5~�8
j�	G�l2(j�1�����Զ��t9�h�-��9�pD�E�h��2�TV�{�����.V����[~I�AJ�^�;�]��<���)�^k��	"�(��Lh���)��^ ,5	yR~D��6t2ۂM�'�<	�W;��z���d�Ӱ�1+d���jM�|�Fv+��U7�L��0"��>e���ڋ0c[~�d^�Ug�v̿n��|�Kk`�U�>��(@V�L><���bW�e�5)T�s\?�䃝_� 7Q�t3��M�ah{�Sg�N�����j�H=�TzQ���4:�FY�Ρ�����r�oRn�d���'P�����sf�,��|�c~�d8�f=	��I�z[r��5��E����`���C$LG�����w��I�jJSD�/M?E[|ST�&�uFe׮<���s��|i��4�=i�£9!7���<��ѻ��
:!���N�L־Jm)p)���~~6��	�7���磒ƿ�>��Z�
%(��8{/��3|s���fF��\�|*�hNQ
�8-X��=E�Z��;�E-Y����w*��_�WK�.:3�~A�.3e��rO�ԙ4��GV�΁�)�����Dz�}-��(8�Ì}�4���,z�~
�iv"�V�!�jxCtj4�J���a^�a���BQ�MM1�eN�;cC��?�L"�.rt��I�#�����_\���0
+H���d8\\`0�_&	�		|U|%,�HYQ B鏖�0
�wcqN��s8@���FR+@�����ΚЪ�M�q]z�]�N.lIs��&wι)ߠ�!�q���BtI���.H�Zw�DEH�M��Gg_f˭����>�t��1��@�2ޞ|D��v~�����������[�v����|����W�>�v~�a���~�����w��?}��{�q��뻿���?n��~����w2�EV)?���5)���9�;�n���
�?�w�j
+endstream
+endobj
+
+120 0 obj
+<<
+  /Type /Page
+  /Resources <<
+    /ProcSet [/PDF /Text /ImageC /ImageB]
+    /ColorSpace <<
+      /c0 97 0 R
+    >>
+    /Font <<
+      /f0 91 0 R
+      /f1 85 0 R
+    >>
+  >>
+  /MediaBox [0 0 595.2756 841.8898]
+  /StructParents 7
+  /Parent 1 0 R
+  /Contents 121 0 R
+>>
+endobj
+
+121 0 obj
+<<
+  /Length 3238
+  /Filter /FlateDecode
+>>
+stream
+x��ْ��_1[Qd�.���=���Vr��RI�����r#9JE\[b*�w��c���2/Z��h4��s}����y�b�4�o_��M�g/_67o^�~��F7�Y�&�]�R��W���5�f��y�2�͗����n���ϳ��1���p|z�����g��������c�פ������>��M���dZ�tHV}�9�uĖ�_�)|�#`~��j��EuE{a Ow��R��p]l^X���R�-v��2�����AI��z�yO��j�d
=\�1�K�vIIi>�����h�����=�{�}s���7w��Ͼy�zv�m؈6�Z�ma�
�X���u�{���T9AJZ�����J�>1_��h��ڷq�_0.la�j��
�Q�Gw�]�|E��_X�,�>�8.W��g�����	�B@�C������l���s�>�NYN$��BDVV�&�6�K�[(b턹�#@��1�U,؉��r���ܜ���(�#�D�|yU��d��a���o	��\a�T+rZ�$|�"��;���
Tr�(��<'I"=Q�Kù�d��V2z���;D�Z�̙��<�,��tV.�<�R�l}�����@<Hn"�|4yg�#��1HO{{�5�w�VmL�1��@/�I�7+l��&��+
+��I�LF������hMKx�K�k�3ڠ�FIz(#*�y|��4�h�I��j;�n�S�'|��_�f�mVƛ����B2���.K�(� 5�����j��k8�M$
+2&A��%Ϣ~1@\����Q���l�N@�Ԩ����S-2��J&�?1C��~��l�T�M*�p�9j%��~�
&�Q��%"򴇡.�+|�L���*���G�W֣
���O�<��3L��SYW��e��\V�'=���8��9�i�&gl�� �AP`Ɂ��w�}���Q`O�{�<��) &:��-�%�ZKς`��p���L�%�	����D�uy����ON�����[��c+����M�O�~��n%`�w9����`��^��u�
!j�sv�`4���b 2��@�kt@�e~E����+V�CY��eā�$/0cK�/��t.�	VՐ*�2��r9�d�+���db�;ꤌv�َ�W��d|ő^���1
����(�vd�B�h��g�Bq��d�v0fv���%/Ib�L��q�a�}MDe��^�;�!Ge��2�A8�6���
�Tj؎r^�WT���E��M�BaTQB�L�L��V�"Q�x�)�'��Ȩ�,�=X�ET�׽�6|Z��,C%7 P��!i�N��
+�&:��):�(Ѐ`�'��բ&�ȴ�TD��U�c�H��iD	E�-�Ig�dȦG�H�DA䖴9��4�'���U�d*wć�$X3:q~ ��k��Xb֣
+���<�_R�ˏ𹢚
+�^���φT�SW��Ń�������1K95vb
�S�jp�9�R&�.G|
+��tA�1���!<��h�.�S?�C��p�@�oUh��R�
+�\(�-9�l����2���)zP��qūgd�`. �����	��j<0)�P�[�ˮl��G��`�'f��*�a�h,�6Yn�v��@˓�D��i��$�� (YO!ZQh� �<-
x���R�P	�)���\����3[�"�ǡ���@8v����4�(�,�)��`�`��'e����ލq#f%�V�ڞ灅�~���+n���
�2���U�{4J�|/�U��o'sCty���x
+�'3̪ɖ7��H�x7b�da(�I8�1Ĺ���Q�D\���W���Q:�	��ι�b��|�Q'L�?c�K�!�V
�qbz����Ip7���.�7�:`�b��譀��w�9U4i���5>e�9�@頳V�*mc��������t5
+�`i��Y��8Kk�\ใ�Rs��x�-rʒ��(
��)��H�C�ҊnF�0].����o�2��Z�t��'�*Z�/����򧍊�Gq����Cz���ؕO���Q���8��]V>g=��w���e��Ʃ`���]�(��3L�p\|�*��:Yp7�l���
��MQ�8Ј�h���~䒵5j�F\t�CC��jqF�v7��ݻ�bS�[����&&�g��޽�e:Y^���c
� H_�
@��G�*��D+Q�K�!����?YĦ������E�SN�&jJ}��}U=��b7��a(O�!9�V��u&��C����D!��F�G��B�qU���/�z-����2i]�dư]:�>�솾B�.ъ���MI��Sd�U�{ܮ�*'7�|.��N-�!�(t{5�����5"H�kT��|�3��h��^��Ļ�Ԧ�d�{�$,2t�tW���I/�m��i���Q�p�6�h�c%%��TpV'?l�G��g��[�ٍ��8��B`��!����˷�-�4��e������@S�����ps�5*yK��ئv�U��I��<̺�m�I񯀹�UnL�n%�(�G��	�Du��u�a��(��!2�G%!\��Y��X�C��-Nl,;�
y�h%��k�]2YeXl�9���)wމ�ԍ|�J�8��w4�Qשk0o�	c�,���/^O�M
+�Qp��%���2��?Nݱ��/�ud<�@Ⱦ��
�-�s�Q�Q�H�����Q�܄�ވ���s-��K�������a�ܼ�9��(�?���i��w�%�ܾ��L+��o曷��ßf�eژ}����n�W\����_��<{��$&�j�V��
!9[��e�V^^_��K�;���� �n��e�>{��<.�t�~��tf�������NW_J5@L����;I`Z���������ۻ�mF/�6�F��!��4�y�w�
�>p|��H�Q�B��
ɒ�b���#�EcНe�gz�G�`,�1�.�t��;tb7�:�l
+ٞ�t����*���/�P��:�R����ޱ��K�Ab1w��_5s,[�.�Uftu���>RN�&��[Qm�y���I�8�
~������_��5�&���u���X��1�Wc?�r7c6/d���T��'=�?�w�8�?��L
͟�{I@1�h{Y�\U"�6��gw�ﭭ�l�FJ#�Мp��m�����Z�������w�7�����7���w���������?o���~���W�돏+24{����	N��$�	��[m6���l���N���/��b�
+endstream
+endobj
+
+122 0 obj
+<<
+  /Title (Un Sprint de Dos Semanas para el Conocimiento)
+  /Author (Jessica Talisman, MLS)
+  /Creator (Typst 0.14.2)
+  /ModDate (D:20260425000939+01'00)
+  /CreationDate (D:20260425000939+01'00)
+>>
+endobj
+
+123 0 obj
+<<
+  /Length 1201
+  /Type /Metadata
+  /Subtype /XML
+>>
+stream
+<?xpacket begin="" id="W5M0MpCehiHzreSzNTczkc9d"?><x:xmpmeta xmlns:x="adobe:ns:meta/" x:xmptk="xmp-writer"><rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"><rdf:Description rdf:about="" xmlns:dc="http://purl.org/dc/elements/1.1/"  xmlns:xmp="http://ns.adobe.com/xap/1.0/"  xmlns:xmpMM="http://ns.adobe.com/xap/1.0/mm/"  xmlns:xmpTPg="http://ns.adobe.com/xap/1.0/t/pg/"  xmlns:pdf="http://ns.adobe.com/pdf/1.3/" ><dc:title><rdf:Alt><rdf:li xml:lang="x-default">Un Sprint de Dos Semanas para el Conocimiento</rdf:li></rdf:Alt></dc:title><dc:creator><rdf:Seq><rdf:li>Jessica Talisman, MLS</rdf:li></rdf:Seq></dc:creator><xmp:CreatorTool>Typst 0.14.2</xmp:CreatorTool><dc:language><rdf:Bag><rdf:li>es</rdf:li></rdf:Bag></dc:language><xmp:ModifyDate>2026-04-25T00:09:39+01:00</xmp:ModifyDate><xmp:CreateDate>2026-04-25T00:09:39+01:00</xmp:CreateDate><xmpTPg:NPages>8</xmpTPg:NPages><dc:format>application/pdf</dc:format><xmpMM:InstanceID>0nNryLQ6Gr+17gVxeQIG/A==</xmpMM:InstanceID><xmpMM:DocumentID>nOg+A18w00zIujAFAfSTtQ==</xmpMM:DocumentID><xmpMM:RenditionClass>proof</xmpMM:RenditionClass><pdf:PDFVersion>1.7</pdf:PDFVersion></rdf:Description></rdf:RDF></x:xmpmeta><?xpacket end="r"?>
+endstream
+endobj
+
+124 0 obj
+<<
+  /Type /Catalog
+  /Pages 1 0 R
+  /Metadata 123 0 R
+  /Lang (es)
+  /StructTreeRoot 10 0 R
+  /MarkInfo <<
+    /Marked true
+    /Suspects false
+  >>
+  /ViewerPreferences <<
+    /Direction /L2R
+  >>
+  /Outlines 2 0 R
+>>
+endobj
+
+xref
+0 125
+0000000000 65535 f
+0000000016 00000 n
+0000000138 00000 n
+0000000218 00000 n
+0000000334 00000 n
+0000000648 00000 n
+0000000886 00000 n
+0000001164 00000 n
+0000001430 00000 n
+0000001568 00000 n
+0000001740 00000 n
+0000002032 00000 n
+0000002303 00000 n
+0000002651 00000 n
+0000002971 00000 n
+0000003298 00000 n
+0000003646 00000 n
+0000003966 00000 n
+0000004307 00000 n
+0000004585 00000 n
+0000005012 00000 n
+0000005125 00000 n
+0000005213 00000 n
+0000005301 00000 n
+0000005391 00000 n
+0000005499 00000 n
+0000005610 00000 n
+0000005713 00000 n
+0000005926 00000 n
+0000006043 00000 n
+0000006231 00000 n
+0000006342 00000 n
+0000006454 00000 n
+0000006805 00000 n
+0000006907 00000 n
+0000007047 00000 n
+0000007143 00000 n
+0000007251 00000 n
+0000007350 00000 n
+0000007459 00000 n
+0000007807 00000 n
+0000007897 00000 n
+0000008165 00000 n
+0000008261 00000 n
+0000008398 00000 n
+0000008486 00000 n
+0000008574 00000 n
+0000008700 00000 n
+0000008788 00000 n
+0000008882 00000 n
+0000009185 00000 n
+0000009465 00000 n
+0000009567 00000 n
+0000009694 00000 n
+0000009782 00000 n
+0000009881 00000 n
+0000009991 00000 n
+0000010230 00000 n
+0000010431 00000 n
+0000010536 00000 n
+0000010653 00000 n
+0000010769 00000 n
+0000011084 00000 n
+0000011174 00000 n
+0000011525 00000 n
+0000011645 00000 n
+0000011765 00000 n
+0000011882 00000 n
+0000011983 00000 n
+0000012072 00000 n
+0000012177 00000 n
+0000012310 00000 n
+0000012421 00000 n
+0000012547 00000 n
+0000012634 00000 n
+0000012720 00000 n
+0000012797 00000 n
+0000012888 00000 n
+0000012974 00000 n
+0000013062 00000 n
+0000013222 00000 n
+0000014109 00000 n
+0000014200 00000 n
+0000014485 00000 n
+0000015760 00000 n
+0000027556 00000 n
+0000027718 00000 n
+0000028612 00000 n
+0000028703 00000 n
+0000028991 00000 n
+0000030266 00000 n
+0000043708 00000 n
+0000043863 00000 n
+0000045379 00000 n
+0000045469 00000 n
+0000045745 00000 n
+0000047622 00000 n
+0000070494 00000 n
+0000070530 00000 n
+0000070888 00000 n
+0000070941 00000 n
+0000070995 00000 n
+0000071049 00000 n
+0000071102 00000 n
+0000071156 00000 n
+0000071210 00000 n
+0000071264 00000 n
+0000071570 00000 n
+0000075076 00000 n
+0000075365 00000 n
+0000079837 00000 n
+0000080126 00000 n
+0000084544 00000 n
+0000084850 00000 n
+0000088821 00000 n
+0000089127 00000 n
+0000093447 00000 n
+0000093736 00000 n
+0000097587 00000 n
+0000097876 00000 n
+0000102114 00000 n
+0000102403 00000 n
+0000105721 00000 n
+0000105939 00000 n
+0000107231 00000 n
+trailer
+<<
+  /Size 125
+  /Root 124 0 R
+  /Info 122 0 R
+  /ID [(nOg+A18w00zIujAFAfSTtQ==) (0nNryLQ6Gr+17gVxeQIG/A==)]
+>>
+startxref
+107469
+%%EOF
\ No newline at end of file
diff --git a/assets/presentation/docs/talisman-steal-this-deck-extractions.md b/assets/presentation/docs/talisman-steal-this-deck-extractions.md
new file mode 100644
index 0000000..3bd5f65
--- /dev/null
+++ b/assets/presentation/docs/talisman-steal-this-deck-extractions.md
@@ -0,0 +1,118 @@
+# Extractions — Steal This Deck (Talisman, KGC 2026)
+
+Source: Jessica Talisman, "Stop Betting, Start Building", Knowledge Graph Conference 2026.
+Newsletter: Intentional Arrangement (Substack). Received: 9 May 2026.
+Full deck: `Steal_This_Deck.pdf`
+
+---
+
+## Framing / Intro hooks
+
+**Para abrir cualquier presentación de ontoref:**
+
+> *"AI is a knowledge tool. Not a data tool. That breaks every assumption underneath many AI strategies."*
+
+Adaptado a ontoref: los agentes AI que trabajan sobre tu repositorio operan sobre correlación estadística si no hay infraestructura de conocimiento. Ontoref es esa infraestructura, pero para proyectos de software.
+
+> *"Agentic AI is not a model upgrade away. It is a shared language, from which to act."*
+
+Este cierre del deck es el mejor hook de intro para ontoref: un modelo más grande no sabe qué es tu proyecto, qué decisiones tomaste, ni en qué estado está. Eso requiere conocimiento estructurado, no un token window mayor.
+
+---
+
+## Datos citable para el problema
+
+Todos con fuente verificable — útiles en posts y slides sin necesitar justificación propia:
+
+| Dato | Fuente |
+|---|---|
+| 89% de empresas reportan **cero impacto de productividad** de AI en 3 años | NBER, Feb 2026, n=5.937 execs |
+| Developers experimentados fueron **19% más lentos** con herramientas AI | METR RCT, 2025 |
+| Ganancia neta semanal del trabajador AI-promedio: **−14 minutos** | Foxit/Sapio, March 2026 |
+| 1.7× más issues en PRs escritos por AI | CodeRabbit, 2025 |
+| 46% del código en GitHub escrito por Copilot | Octoverse 2025 |
+
+**El argumento que emerge:** *AI is not saving time. It is generating volume. Without a knowledge backbone, its only measurable output is noise.*
+
+---
+
+## El argumento central — para slides y posts
+
+**El stack de conocimiento (Talisman lo llama "knowledge infrastructure"):**
+
+```
+1. Controlled vocabularies  — términos con significado único y autoritativo
+2. Taxonomies               — organización jerárquica
+3. Thesauri                 — equivalencias, relaciones cruzadas
+4. Ontologies               — compromisos formales: clases, propiedades, constraints
+5. Knowledge graphs         — live, queryable, governable
+```
+
+> *"You cannot skip layers. You cannot start at five and reverse-engineer to one."*
+
+**Ontoref implementa este stack para proyectos de software:**
+- Schemas NCL → vocabularios controlados
+- `.ontology/core.ncl` → nodos Practice/Concept con edges tipados
+- DAG-formalized knowledge → el knowledge graph
+- ADRs + migrations + `state.ncl` → governance
+
+---
+
+## El argumento de precisión — para posts técnicos
+
+> *"The accuracy gap is not closed by a bigger model. It is closed by a defined schema, an ontology, and a validated query."*
+
+Con datos:
+- **16% → 72%** en question-answering sobre SQL enterprise con ontology checks (Allemang & Sequeda, data.world AI Lab, 2024)
+- **3.4×** GraphRAG vs vector RAG en 43 queries enterprise (Diffbot KG-LM Benchmark, 2023)
+- Vector RAG colapsa a **0% past 5 entities per query**. KG-grounded retrieval sostiene.
+
+> *"Defining your terms is the cheapest accuracy and cost lever in the LLM stack."*
+
+Ontoref es exactamente esto: `ontoref describe` y el sistema Q&A (ADR-003) son el vocabulario controlado que reduce el espacio de alucinación cuando un agente trabaja sobre el proyecto.
+
+---
+
+## El argumento de MCP — crítico para el posicionamiento técnico
+
+> *"MCP and A2A are transport. Not semantics. They move bytes between endpoints. They do not establish shared meaning."*
+>
+> *"Two agents connected by MCP exchange text bytes. They do not truly share context."*
+
+**Ontoref cierra este gap:** tiene superficie MCP (`ontoref-daemon/src/mcp/`) encima de una capa ontológica. El protocolo mueve bytes; ontoref provee el significado que hace que esos bytes sean conocimiento, no texto.
+
+Esto es diferenciación directa respecto a "simplemente exponer tu repo por MCP".
+
+---
+
+## El argumento cultural — para posts de opinión
+
+> *"The industry is optimized to ship solutions, not to own problems."*
+
+| Celebrado | Huérfano |
+|---|---|
+| Launching new platforms | Maintaining existing platforms |
+| AI-generated content | Taxonomy work |
+| Ontologies built on the fly | Domain expertise |
+
+> *"Knowledge work IS the maintenance. Yet it keeps getting cut."*
+
+Ontoref formaliza exactamente lo que siempre se corta: las decisiones arquitectónicas (ADRs), el estado del proyecto (`state.ncl`), la memoria operacional (`.coder/`). Lo hace queryable y machine-readable para que no dependa de que alguien lo recuerde.
+
+---
+
+## El close — para cualquier formato
+
+> *"The organizations that close the perception/reality gap first won't be the ones with the best models. They'll be the ones who finally did the work."*
+
+Adaptado: los proyectos donde los agentes AI trabajan mejor no son los que tienen el modelo más grande. Son los que tienen conocimiento estructurado sobre sí mismos.
+
+---
+
+## Atribución
+
+Jessica Talisman, MLS — Semantic Engineer, Information Architect, Knowledge Infrastructure Strategist.
+Newsletter: *Intentional Arrangement* (Substack).
+Talk: "Stop Betting, Start Building", Knowledge Graph Conference 2026, Technology Track, May 6, 2026.
+
+Al usar cualquiera de estos puntos en público, citar la fuente — es un argumento de autoridad que refuerza, no debilita, la posición de ontoref.
diff --git a/assets/presentation/intro_slides.md b/assets/presentation/intro_slides.md
new file mode 100644
index 0000000..33614c7
--- /dev/null
+++ b/assets/presentation/intro_slides.md
@@ -0,0 +1,989 @@
+---
+theme: default
+title: Ontología y Reflection — lo que la IA no puede entender por ti
+titleTemplate: '%s - OntoRef'
+layout: cover
+keywords: Ontología,Reflection,IA,Infraestructura,OntoRef,Nushell,Nickel
+download: true
+exportFilename: ontoref-openspace
+monaco: true
+remoteAssets: true
+selectable: true
+record: true
+colorSchema: dark
+lineNumbers: false
+themeConfig:
+  primary: '#f74c00'
+  logoHeader: '/ferris.svg'
+fonts:
+  mono: 'Victor Mono'
+drawings:
+  enabled: true
+  persist: false
+  presenterOnly: false
+  syncAll: true
+scripts:
+  - setup/image-overlay.ts
+class: 'justify-center flex flex-cols'
+---
+
+<h1 class="font-medium text-3xl leading-tight">La IA gestiona tu infra...<br><span class="text-orange-400">¿Sabes lo que está pasando?</span></h1>
+
+<div class="flex justify-center mt-6 text-gray-400 text-lg italic">
+  Ontología y Reflection como base de un proyecto que se entiende a sí mismo.
+</div>
+
+<div class="flex justify-center mt-8">
+  <img class="w-40" src="/jesusperez_w.svg">
+</div>
+
+<style scoped>
+h1, h2, p { z-index: 10; }
+</style>
+
+<Footer />
+
+<!--
+Abrir con pausa. El título es una pregunta directa.
+No es retórica — la mayoría no sabe. Y el punto es exactamente ese.
+Esta charla no es sobre herramientas. Es sobre qué pasa cuando externalizas comprensión.
+-->
+
+---
+layout: default
+---
+
+# El problema no es la herramienta
+
+<div class="grid grid-cols-2 gap-8 mt-4">
+
+<div>
+
+**Lo que la IA hace bien**
+
+<div class="mt-3 space-y-3">
+
+<div class="flex items-start gap-2">
+  <span class="text-green-400 font-bold text-sm mt-1">✓</span>
+  <div class="text-gray-300 text-sm">Genera YAML correcto sintácticamente</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-green-400 font-bold text-sm mt-1">✓</span>
+  <div class="text-gray-300 text-sm">Completa Terraform que nunca rompe el plan</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-green-400 font-bold text-sm mt-1">✓</span>
+  <div class="text-gray-300 text-sm">Rellena gaps de configuración con defaults razonables</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-green-400 font-bold text-sm mt-1">✓</span>
+  <div class="text-gray-300 text-sm">Traduce intención en forma declarativa</div>
+</div>
+
+</div>
+
+</div>
+
+<div>
+
+<v-click>
+
+**Lo que la IA no puede hacer**
+
+<div class="mt-3 space-y-3">
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400 font-bold text-sm mt-1">✗</span>
+  <div class="text-gray-300 text-sm">Saber <em>por qué</em> ese servicio no puede escalar a N réplicas</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400 font-bold text-sm mt-1">✗</span>
+  <div class="text-gray-300 text-sm">Conocer la tensión que existe entre latencia y consistencia en <em>tu</em> sistema</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400 font-bold text-sm mt-1">✗</span>
+  <div class="text-gray-300 text-sm">Recordar la decisión que tomaste en diciembre y por qué descartaste la alternativa</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400 font-bold text-sm mt-1">✗</span>
+  <div class="text-gray-300 text-sm">Entender qué no puede romperse sin coordinación entre equipos</div>
+</div>
+
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-6 border-l-4 border-orange-500 pl-4 text-sm text-gray-300">
+  Declarativo ≠ comprensión.<br>
+  Describes el estado deseado, no el <em>por qué</em>.
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+La IA genera forma correcta para el problema equivocado.
+El YAML válido es el problema mínimo. El problema real es que nadie sabe qué asumió.
+-->
+
+---
+layout: default
+---
+
+# Las 2am — cuando el Terraform no habla
+
+<div class="grid grid-cols-2 gap-8 mt-4">
+
+<div>
+
+**Lo que ves cuando falla**
+
+<div class="mt-3 font-mono text-xs text-gray-300">
+
+```text
+Error: Error creating Instance:
+  googleapi: Error 400: Invalid value
+  for field 'resource.machineType'
+
+# El plan pasó. El apply falló.
+# ¿Qué asumió el agente?
+# ¿Qué región eligió? ¿Por qué ese tipo?
+# ¿Hay una restricción de org que no sabía?
+```
+
+</div>
+
+<v-click>
+
+<div class="mt-4 font-mono text-xs text-gray-400">
+
+```yaml
+# Lo que generó la IA (hace 6 semanas)
+resource "google_compute_instance" "api" {
+  machine_type = "e2-medium"   # "razonable"
+  zone         = "us-central1-a" # "por defecto"
+}
+
+# Lo que tu política de org requiere
+# → solo europe-west1, solo n2-standard-*
+# → la IA no lo sabía porque tú tampoco lo escribiste
+```
+
+</div>
+
+</v-click>
+
+</div>
+
+<div>
+
+<v-click>
+
+**El patrón que se repite**
+
+<div class="mt-3 space-y-3 text-sm">
+
+<div class="border-l-2 border-red-500 pl-3">
+  <div class="text-red-300 font-semibold">Sin modelo explícito</div>
+  <div class="text-gray-400">La IA inventa contexto plausible. Tú asumes que sabe. Falla en prod.</div>
+</div>
+
+<div class="border-l-2 border-orange-500 pl-3">
+  <div class="text-orange-300 font-semibold">Con Stack Overflow</div>
+  <div class="text-gray-400">Copias código sin entender el supuesto. Al menos tú lo buscaste.</div>
+</div>
+
+<div class="border-l-2 border-yellow-500 pl-3">
+  <div class="text-yellow-300 font-semibold">Con la IA</div>
+  <div class="text-gray-400">El agente genera con confianza. La confianza es contagiosa. El error también.</div>
+</div>
+
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-6 text-base font-semibold text-gray-200 border-l-4 border-orange-500 pl-4">
+  La diferencia con Stack Overflow:<br>
+  <span class="text-gray-400 font-normal text-sm">Al menos ahí buscabas activamente.<br>Con la IA, delegas también la búsqueda.</span>
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+Este slide es el que el público siente. Todos tienen una historia de "la IA lo generó y pareció bien".
+La pregunta es: ¿cuándo lo descubriste? ¿Staging? ¿Prod? ¿6 meses después?
+-->
+
+---
+layout: default
+---
+
+# Tu proyecto vive en tu cabeza
+
+<div class="grid grid-cols-2 gap-8 mt-4">
+
+<div>
+
+**Lo que vive solo en tu cabeza**
+
+<div class="mt-3 space-y-2 text-sm text-gray-300">
+
+<div class="flex items-start gap-2">
+  <span class="text-orange-400 text-sm mt-1">→</span>
+  <div>Por qué este servicio no puede tener más de 3 réplicas</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-orange-400 text-sm mt-1">→</span>
+  <div>Qué pasó cuando intentamos escalar en enero y por qué fallamos</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-orange-400 text-sm mt-1">→</span>
+  <div>La tensión entre el equipo de plataforma y el de producto sobre quién controla el networking</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-orange-400 text-sm mt-1">→</span>
+  <div>Qué alternativa se rechazó y por qué</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-orange-400 text-sm mt-1">→</span>
+  <div>Qué invariante no puede violarse sin romper el acuerdo con el equipo de seguridad</div>
+</div>
+
+</div>
+
+<v-click>
+
+<div class="mt-5 border-l-4 border-red-500 pl-4 text-sm text-gray-400">
+  Lo que vive solo en tu cabeza <strong class="text-red-300">no se puede validar</strong>.<br>
+  No puede consultarlo un agente.<br>No puede consultarlo un compañero nuevo.<br>
+  No puedes consultarlo tú mismo en 6 meses.
+</div>
+
+</v-click>
+
+</div>
+
+<div>
+
+<v-click>
+
+**La trampa del README**
+
+<div class="mt-3 font-mono text-xs text-gray-400">
+
+```markdown
+# Mi Proyecto
+
+Este servicio hace X. Está desplegado en GCP.
+Usa Terraform. Tiene 3 entornos: dev, staging, prod.
+
+## Arquitectura
+Ver diagrama (desactualizado desde marzo).
+
+## Decisiones
+- Elegimos PostgreSQL porque MySQL no nos gustaba.
+- Usamos gRPC internamente.
+```
+
+</div>
+
+<div class="mt-4 space-y-2 text-sm text-gray-400">
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400 text-sm">✗</span>
+  <div>No dice qué no puede cambiar</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400 text-sm">✗</span>
+  <div>No captura las alternativas rechazadas ni el por qué</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400 text-sm">✗</span>
+  <div>No es consultable ni verificable — es prosa</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400 text-sm">✗</span>
+  <div>No refleja el estado actual — <br> refleja el estado de cuando se escribió</div>
+</div>
+
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+La pregunta clave que lanzar al público: ¿cuándo fue la última vez que actualizaste tu README?
+Un README es documentación. Una ontología es un modelo vivo del sistema.
+-->
+
+---
+layout: center
+---
+
+# La tesis
+
+<div class="mt-8 text-center space-y-6">
+
+<div class="text-3xl font-semibold text-gray-100">
+  La ontología no es para la IA.
+</div>
+
+<div class="text-3xl font-semibold text-orange-400">
+  Es para ti.
+</div>
+
+<v-click>
+
+<div class="mt-8 text-lg text-gray-400 max-w-2xl mx-auto leading-8">
+  Formalizar invariantes, tensiones y estado actual<br>
+  te fuerza a saber qué es verdad <em>ahora mismo</em>.
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-6 text-lg text-gray-400 max-w-2xl mx-auto leading-8">
+  La IA se beneficia como efecto secundario.<br>
+  Opera sobre lo que <em>tú ya clarificaste</em>.
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-8 text-base text-gray-500 max-w-2xl mx-auto border-t border-gray-700 pt-6">
+  El valor real: el proceso de <strong class="text-gray-300">escribir la ontología</strong><br>
+  es donde te aclaras tú, no la IA.
+</div>
+
+</v-click>
+
+</div>
+
+<Footer />
+
+<!--
+Pausa larga después del primer par de líneas. Dejar que aterrice.
+La gente espera que digas "la IA necesita contexto". La inversión es el punto.
+-->
+
+---
+layout: default
+---
+
+# Ontología ≠ lista de decisiones
+
+<div class="grid grid-cols-2 gap-8 mt-4">
+
+<div>
+
+**Una lista de decisiones**
+
+<div class="mt-3 font-mono text-xs text-gray-400">
+
+```text
+- Elegimos Kubernetes
+- Usamos Helm para despliegues
+- PostgreSQL como base de datos primaria
+- gRPC entre servicios internos
+- OAuth2 para autenticación externa
+```
+
+</div>
+
+<div class="mt-4 text-sm text-gray-500">
+Captura el <em>qué</em>. No captura el <em>por qué</em>,<br>
+las alternativas rechazadas, ni las consecuencias.
+</div>
+
+</div>
+
+<div>
+
+<v-click>
+
+**Una ontología**
+
+<div class="mt-3 space-y-2 text-sm">
+
+<div class="border-l-4 border-blue-400 pl-3 py-1 bg-gray-900 rounded-r">
+  <div class="text-blue-300 font-mono text-xs font-semibold">INVARIANTES</div>
+  <div class="text-gray-400 text-xs">Lo que no puede cambiar sin decisión explícita y coordinación</div>
+</div>
+
+<div class="border-l-4 border-orange-400 pl-3 py-1 bg-gray-900 rounded-r">
+  <div class="text-orange-300 font-mono text-xs font-semibold">TENSIONES ACTIVAS</div>
+  <div class="text-gray-400 text-xs">Trade-offs que aún no se han resuelto — viven en el sistema ahora mismo</div>
+</div>
+
+<div class="border-l-4 border-purple-400 pl-3 py-1 bg-gray-900 rounded-r">
+  <div class="text-purple-300 font-mono text-xs font-semibold">ESTADO ACTUAL → DESEADO</div>
+  <div class="text-gray-400 text-xs">Dimensiones de madurez. Dónde estás, dónde quieres estar, qué te bloquea</div>
+</div>
+
+<div class="border-l-4 border-green-400 pl-3 py-1 bg-gray-900 rounded-r">
+  <div class="text-green-300 font-mono text-xs font-semibold">GATES ACTIVAS</div>
+  <div class="text-gray-400 text-xs">Condiciones que deben cumplirse antes de cruzar una frontera arquitectural</div>
+</div>
+
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-4 text-sm text-gray-300">
+  Es un grafo consultable. No prosa.<br>
+  <span class="text-gray-500 text-xs">La IA puede leerlo. Tu CI puede verificarlo. Tú puedes razonarlo.</span>
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+La diferencia no es el nivel de detalle — es la estructura.
+Un grafo tiene propiedades que la prosa no tiene: es consultable, verificable, traversable.
+-->
+
+---
+layout: default
+---
+
+# Ontoref — tres archivos, tres preguntas
+
+<div class="grid grid-cols-3 gap-5 mt-6">
+
+<div class="border border-gray-700 rounded p-4 bg-gray-900">
+
+<div class="text-blue-300 font-mono font-semibold text-sm mb-3">core.ncl</div>
+<div class="text-orange-400 text-xs mb-3 italic">"¿Qué ES este proyecto?"</div>
+
+<div class="font-mono text-xs text-gray-400">
+
+```text
+invariantes:
+  - backend-agnostic-core
+  - zero-external-runtime
+
+tensiones:
+  - formalization vs
+         adoption-friction
+  - ontology vs reflection
+
+prácticas:
+  - nickel-as-schema-layer
+  - nushell-as-operator
+```
+
+</div>
+
+<div class="mt-3 text-xs text-gray-600">
+Inmutable sin ADR nuevo.
+</div>
+
+</div>
+
+<div class="border border-gray-700 rounded p-4 bg-gray-900">
+
+<div class="text-orange-300 font-mono font-semibold text-sm mb-3">state.ncl</div>
+<div class="text-orange-400 text-xs mb-3 italic">"¿Dónde ESTAMOS?"</div>
+
+<div class="font-mono text-xs text-gray-400">
+
+```text
+dimensiones:
+  backend-maturity:
+    actual: multi-stable
+    deseado: all-production
+    bloqueador: 
+         missing-nats-tests
+
+  nickel-depth:
+    actual: schema-input
+    deseado: bidirectional
+    catalizador: 
+           export pipeline
+```
+
+</div>
+
+<div class="mt-3 text-xs text-gray-600">
+Evoluciona con el proyecto.
+</div>
+
+</div>
+
+<div class="border border-gray-700 rounded p-4 bg-gray-900">
+
+<div class="text-green-300 font-mono font-semibold text-sm mb-3">gate.ncl</div>
+<div class="text-orange-400 text-xs mb-3 italic">"¿Cuándo estamos LISTOS?"</div>
+
+<div class="font-mono text-xs text-gray-400">
+
+```text
+gates:
+  core-dependency-gate:
+    activa: true
+    permeabilidad: Low
+    protege:
+      - backend-agnostic-core
+
+  nickel-primacy-gate:
+    activa: true
+    permeabilidad: Low
+    protege:
+      - nickel-as-schema-layer
+```
+
+</div>
+
+<div class="mt-3 text-xs text-gray-600">
+Condiciones de cruce.
+</div>
+
+</div>
+
+</div>
+
+<v-click>
+
+<div class="mt-5 font-mono text-xs text-center text-gray-500">
+  Nickel tipado · exportable como JSON · consultable desde Nu · verificable en CI
+</div>
+
+</v-click>
+
+<Footer />
+
+<!--
+Tres preguntas, tres responsabilidades. La separación es deliberada.
+core.ncl responde al SER. state.ncl al DEVENIR. gate.ncl al CRUZAR.
+Cada uno puede cambiar a ritmos diferentes sin contaminar los otros.
+-->
+
+---
+layout: default
+---
+
+# Reflection — lo que se CONVIERTE
+
+<div class="grid grid-cols-2 gap-8 mt-2">
+
+<div>
+
+**Ontología = lo que ES**
+
+<div class="mt-3 text-sm text-gray-400 space-y-2">
+<div>→ Invariantes — verdad estructural</div>
+<div>→ Tensiones — trade-offs vivos</div>
+<div>→ Estado — dónde estamos ahora</div>
+</div>
+
+<div class="mt-5 text-sm font-semibold text-gray-200">Reflection = lo que DEVIENE</div>
+
+<div class="mt-3 text-sm text-gray-400 space-y-2">
+<div>→ Modos operacionales — secuencias con DAG de dependencias</div>
+<div>→ ADRs con constraints tipadas — decisiones vivas, no prosa</div>
+<div>→ Backlog estructurado — tensiones → tareas</div>
+<div>→ Historial de sesiones — quién hizo qué y cuándo</div>
+</div>
+
+<v-click>
+
+<div class="mt-5 border-l-4 border-orange-500 pl-4 text-sm text-gray-300">
+  La ontología captura el <em>ser</em>.<br>
+  La reflection captura el <em>operar sobre ese ser</em>.
+</div>
+
+</v-click>
+
+</div>
+
+<div>
+
+<v-click>
+
+
+<div class="-mt-3 font-mono text-xs">
+    <b>Un modo como especificación ejecutable</b>
+```bash
+# reflection/modes/new_service.ncl
+{
+  id      = "new_service",
+  trigger = "When adding a new service boundary",
+  steps   = [
+    {
+      id     = "draft-adr",
+      actor  = "developer",
+      action = "Draft ADR for service boundary decision",
+    },
+    {
+      id         = "verify-invariants",
+      actor      = "agent",
+      action     = "Check no invariant is violated",
+      depends_on = [{ step = "draft-adr" }],
+    },
+    {
+      id         = "seal-config",
+      actor      = "developer",
+      action     = "Apply initial config seal",
+      depends_on = [{ step = "verify-invariants" }],
+    },
+  ],
+}
+```
+
+</div>
+
+<div class="mt-2 text-xs text-gray-500">
+No es un runbook. Es un grafo acíclico tipado.<br>
+La IA lo ejecuta. El sistema lo valida.
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+La dualidad ontología/reflection es la tensión central de ontoref.
+No pueden fusionarse — tienen ritmos de cambio diferentes y audiencias diferentes.
+La reflection sin ontología es procedimiento sin semántica. La ontología sin reflection es diagrama muerto.
+-->
+
+---
+layout: default
+---
+
+# La IA como actor, no como oráculo
+
+<div class="grid grid-cols-2 gap-8 mt-4">
+
+<div>
+
+**Sin ontología formalizada**
+
+<div class="mt-3 font-mono text-xs">
+
+```bash
+# Lo que el agente recibe
+"Ayúdame a escalar este servicio"
+
+# Lo que el agente asume
+- que puede añadir réplicas libremente
+- que el estado actual es el deseado
+- que no hay restricciones de org
+- que la decisión de diciembre es irrelevante
+
+# Lo que genera
+resource "google_compute_instance" {
+  count = 5  # razonable, ¿no?
+}
+```
+
+</div>
+
+</div>
+
+<div>
+
+<v-click>
+
+**Con ontología formalizada**
+
+<div class="mt-3 font-mono text-xs">
+
+```bash
+# Lo que el agente recibe (SessionStart hook)
+INVARIANT: max-replicas-per-zone = 2
+TENSION: scaling vs consistency [unresolved]
+CONSTRAINT (adr-003): no horizontal scaling
+  without explicit capacity review
+STATE: backend-maturity = single-stable
+  → desired: multi-stable
+  → blocker: capacity-review-pending
+
+# Lo que el agente puede hacer
+- Preguntar si procede dado el bloqueador
+- Generar la config dentro de los límites
+- Indicar que el ADR debe revisarse antes
+```
+
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-4 border-l-4 border-green-500 pl-4 text-sm text-gray-300">
+  El agente opera sobre lo que tú ya clarificaste.<br>
+  <span class="text-gray-500 text-xs">Deja de inventarse contexto porque tú dejaste de tenerlo implícito.</span>
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+El SessionStart hook en ontoref inyecta el contexto ontológico al inicio de cada sesión.
+El agente no es más inteligente — tiene más información real y menos que inventar.
+La mejora no es de la IA. Es tuya: formalizaste algo que antes vivía solo en tu cabeza.
+-->
+
+---
+layout: default
+---
+
+# ¿Quién mantiene esto actualizado?
+
+<div class="mt-4 text-lg text-gray-400 text-center">La pregunta que siempre sale.</div>
+
+<div class="grid grid-cols-2 gap-8 mt-6">
+
+<div>
+
+**El problema real con el README**
+
+<div class="mt-3 space-y-2 text-sm text-gray-400">
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400">✗</span>
+  <div>Nadie tiene incentivo para actualizarlo —<br> no está en el path crítico</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400">✗</span>
+  <div>No hay señal de cuando está desactualizado</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400">✗</span>
+  <div>Actualizar requiere fricción de escritura libre</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400">✗</span>
+  <div>No hay verificación — puede divergir silenciosamente</div>
+</div>
+
+</div>
+
+</div>
+
+<div>
+
+<v-click>
+
+**El diseño de ontoref**
+
+<div class="mt-3 space-y-2 text-sm">
+
+<div class="border-l-4 border-blue-400 pl-3 py-1 bg-gray-900 rounded-r">
+  <div class="text-blue-300 font-semibold text-xs">Pre-commit como sincronización forzada</div>
+  <div class="text-gray-400 text-xs"><code>nickel typecheck</code> en cada commit — la capa declarativa nunca queda rota</div>
+</div>
+
+<div class="border-l-4 border-orange-400 pl-3 py-1 bg-gray-900 rounded-r">
+  <div class="text-orange-300 font-semibold text-xs">Modos como procedimientos obligatorios</div>
+  <div class="text-gray-400 text-xs">Añadir un servicio ejecuta el modo — el ADR se escribe como parte del proceso</div>
+</div>
+
+<div class="border-l-4 border-purple-400 pl-3 py-1 bg-gray-900 rounded-r">
+  <div class="text-purple-300 font-semibold text-xs">Drift detection</div>
+  <div class="text-gray-400 text-xs"><code>sha256(nickel export)</code> — sabes cuándo el estado sellado difiere del actual</div>
+</div>
+
+<div class="border-l-4 border-green-400 pl-3 py-1 bg-gray-900 rounded-r">
+  <div class="text-green-300 font-semibold text-xs">ADRs como constraints, no prosa</div>
+  <div class="text-gray-400 text-xs">El constraint es machine-readable — el CI puede verificarlo, no solo leerlo</div>
+</div>
+
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-4 text-xs text-gray-500 border-l-2 border-gray-700 pl-3">
+  La ontología se mantiene porque es el path.<br>
+  No porque alguien sea disciplinado.
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+Esta es la objeción más honesta del público. Abrazarla, no esquivarla.
+La respuesta no es "la gente debería ser más disciplinada".
+La respuesta es: el sistema lo hace obligatorio, no opcional.
+-->
+
+---
+layout: default
+---
+
+# Para el debate
+
+<div class="grid grid-cols-2 gap-8 mt-6">
+
+<div>
+
+**Propuesta 1 — Ontología**
+
+<div class="mt-4 space-y-4">
+
+<div class="border border-gray-700 rounded p-3 bg-gray-900">
+  <div class="text-orange-300 text-sm font-semibold">¿No es suficiente con un buen README?</div>
+  <div class="text-gray-500 text-xs mt-1">¿Cuándo fue la última vez que el tuyo estaba al día?</div>
+</div>
+
+<div class="border border-gray-700 rounded p-3 bg-gray-900">
+  <div class="text-orange-300 text-sm font-semibold">¿Qué diferencia una ontología de una lista de decisiones?</div>
+  <div class="text-gray-500 text-xs mt-1">Estructura, consultabilidad, verificabilidad.</div>
+</div>
+
+<div class="border border-gray-700 rounded p-3 bg-gray-900">
+  <div class="text-orange-300 text-sm font-semibold">¿Quién mantiene esto cuando el proyecto cambia rápido?</div>
+  <div class="text-gray-500 text-xs mt-1">El diseño lo hace obligatorio, no opcional.</div>
+</div>
+
+</div>
+
+</div>
+
+<div>
+
+**Propuesta 2 — Infra + IA**
+
+<div class="mt-4 space-y-4">
+
+<div class="border border-gray-700 rounded p-3 bg-gray-900">
+  <div class="text-orange-300 text-sm font-semibold">¿Es diferente a copiar Stack Overflow?</div>
+  <div class="text-gray-500 text-xs mt-1">En SO buscabas. Con la IA, delegas también la búsqueda.</div>
+</div>
+
+<div class="border border-gray-700 rounded p-3 bg-gray-900">
+  <div class="text-orange-300 text-sm font-semibold">¿Dónde está el mínimo de comprensión necesario?</div>
+  <div class="text-gray-500 text-xs mt-1">Saber qué no puede romperse. Sin eso, no tienes control.</div>
+</div>
+
+<div class="border border-gray-700 rounded p-3 bg-gray-900">
+  <div class="text-orange-300 text-sm font-semibold">¿Las plataformas internas resuelven esto o lo esconden?</div>
+  <div class="text-gray-500 text-xs mt-1">Una plataforma sin ontología solo mueve la ignorancia un nivel arriba.</div>
+</div>
+
+</div>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+Open space: las mejores respuestas vendrán del público.
+Tu rol aquí es provocar, no resolver. Lanza las preguntas y deja el silencio.
+-->
+
+---
+layout: cover
+name: end
+class: 'justify-center flex flex-cols'
+---
+
+<div class="standalone-slide text-shadow-lg text-xl">
+
+<div class="text-3xl font-semibold text-gray-100 mb-2">
+  La IA puede declarar tu infra.
+</div>
+
+<div class="text-3xl font-semibold text-orange-400 mb-8">
+  No puede entenderla por ti.
+</div>
+
+<v-click>
+
+<div class="text-xl text-gray-300 leading-10 max-w-2xl mx-auto">
+  La ontología es el modelo que te fuerza a saber qué es verdad ahora.<br>
+  La reflection es la operación sobre ese modelo.<br>
+  La IA es el actor que opera <em>sobre lo que ya clarificaste</em>.
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-10 text-2xl text-gray-200">
+  Gracias. ¿Preguntas?
+</div>
+
+<small class="mt-4 block text-gray-500">jesusperez.pro · ontoref.dev</small>
+
+</v-click>
+
+</div>
+
+<style>
+.standalone-slide {
+  text-align: center;
+}
+</style>
+
+<Footer />
+
+<!--
+Cierre lento. Las tres líneas son la síntesis completa.
+No terminar con "ontoref es la solución" — terminar con la inversión cognitiva.
+El punto no es adoptar ontoref. El punto es tomar posesión de tu propio modelo.
+-->
diff --git a/assets/presentation/jj_rad.sh b/assets/presentation/jj_rad.sh
new file mode 100755
index 0000000..6878963
--- /dev/null
+++ b/assets/presentation/jj_rad.sh
@@ -0,0 +1,3 @@
+#!/bin/bash
+
+pnpm run dev
diff --git a/assets/presentation/jj_slides.md b/assets/presentation/jj_slides.md
new file mode 100644
index 0000000..dbc1bbb
--- /dev/null
+++ b/assets/presentation/jj_slides.md
@@ -0,0 +1,1188 @@
+---
+theme: default
+title: jj + Radicle — VCS para la era de los agentes
+titleTemplate: '%s'
+layout: cover
+keywords: jj,Jujutsu,Radicle,VCS,Git,AI,Agents,ontoref,vapora
+download: true
+exportFilename: jj-radicle-agents
+monaco: true
+remoteAssets: true
+selectable: true
+record: true
+colorSchema: dark
+lineNumbers: false
+themeConfig:
+  primary: '#f74c00'
+fonts:
+  mono: 'Victor Mono'
+drawings:
+  enabled: true
+  persist: false
+  presenterOnly: false
+  syncAll: true
+scripts:
+  - setup/image-overlay.ts
+class: 'justify-center flex flex-cols'
+---
+
+<h1 class="pl-27 font-medium">jj + Radicle</h1>
+
+<h2 class="flex justify-center mt-3 font-medium text-orange-400">
+  VCS para la era de los agentes
+</h2>
+
+<div class="flex justify-center mt-3 text-gray-400 text-lg italic">
+  Workspaces aislados. Parches soberanos. Sin staging area.
+</div>
+
+<div class="flex justify-center">
+  <img class="mt-8 w-40" src="/jesusperez_w.svg">
+</div>
+
+<style scoped>
+h1, h2, p { z-index: 10; }
+</style>
+
+<Footer />
+
+<!--
+Abrir con calma. La audiencia probablemente usa git todos los días.
+Vamos a cuestionar cosas que asumen como inamovibles.
+-->
+
+---
+layout: two-cols
+---
+
+<h1 class="w-300"> El problema con Git + agentes IA </h1>
+
+<div>
+
+**Fricción acumulada**
+
+<div class="mt-3 space-y-3">
+
+<div class="flex items-start gap-2">
+  <span class="text-red-500 font-bold text-sm mt-1">🔴</span>
+  <div><span class="text-gray-200 font-semibold">Staging area compartida</span><br><span class="text-gray-400 text-sm">Humano y agente compiten por el índice.<br> Un <code>git add</code> pisotea al otro.</span></div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-500 font-bold text-sm mt-1">🔴</span>
+  <div><span class="text-gray-200 font-semibold">Worktrees frágiles</span><br><span class="text-gray-400 text-sm">Git worktrees necesitan branches nombradas.<br> Colisionan. No escalan a N agentes.</span></div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-orange-400 font-bold text-sm mt-1">🟠</span>
+  <div><span class="text-gray-200 font-semibold">Centralización invisible</span><br><span class="text-gray-400 text-sm">GitHub/GitLab son el single point of failure.<br> Tu código, sus reglas.</span></div>
+</div>
+
+
+</div>
+
+</div>
+
+::right::
+
+<div class="mt-24 ml-4">
+
+<div class="flex items-start gap-2">
+  <span class="text-orange-400 font-bold text-sm mt-1">🟠</span>
+  <div><span class="text-gray-200 font-semibold">Push-to-review es un cuello de botella</span><br><span class="text-gray-400 text-sm">El agente necesita push + PR + aprobación.<br> ¿Le das acceso de escritura al remoto?</span></div>
+</div>
+
+
+<div class="flex items-start gap-2 mt-3">
+  <span class="text-yellow-400 font-bold text-sm mt-1">🟡</span>
+  <div><span class="text-gray-200 font-semibold">Merge hell con N actores</span><br><span class="text-gray-400 text-sm">3 agentes + 1 humano = conflictos serializados.<br> Alguien siempre pierde.</span></div>
+</div>
+
+<v-click>
+
+<div class="mt-6 border-l-4 border-orange-500 pl-4 text-sm text-gray-300 leading-6">
+  Git fue diseñado para <em>humanos</em> que trabajan en <em>branches</em>.<br><br>
+  Hoy tenemos <strong>N actores concurrentes</strong> (humanos + agentes) que necesitan <strong>aislamiento real</strong>, no convenciones.
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-1 font-mono text-xs text-gray-500 bg-gray-900 p-3 rounded">
+
+```text
+Agente-1: git stash → trabaja → git stash pop
+Agente-2: git stash → trabaja → CONFLICTO
+Humano:   "¿quién tocó mi working tree?"
+```
+
+</div>
+
+</v-click>
+
+</div>
+
+<Footer />
+
+<!--
+No es que Git sea malo. Es que fue diseñado para un mundo con un actor por working copy.
+Hacer la pregunta: "¿cuántos de vosotros habéis tenido conflictos con el copiloto o un agente?"
+-->
+
+---
+layout: default
+---
+
+# Jujutsu (jj) — el modelo mental
+
+<div class="grid grid-cols-2 gap-8 mt-4">
+
+<div>
+
+**Cambio fundamental: no hay staging area**
+
+<div class="text-sm text-gray-300 space-y-3 mt-3">
+
+<div class="border-l-2 border-blue-500 pl-3">
+  <span class="text-blue-300 font-semibold">Working copy IS a commit</span><br>
+  <span class="text-gray-400">En jj, <code>@</code> ya es un commit. Cada save es una snapshot automática. No hay <code>git add</code>, no hay índice.</span>
+</div>
+
+<div class="border-l-2 border-blue-500 pl-3">
+  <span class="text-blue-300 font-semibold">Changes, no commits</span><br>
+  <span class="text-gray-400">jj trabaja con <em>change IDs</em> estables. Reescribir historia no cambia el ID del cambio — solo su commit ID muta.</span>
+</div>
+
+<div class="border-l-2 border-blue-500 pl-3">
+  <span class="text-blue-300 font-semibold">Conflictos son first-class</span><br>
+  <span class="text-gray-400">Un conflicto no bloquea. Se materializa en el árbol de trabajo. Puedes commitear conflictos, resolverlos después.</span>
+</div>
+
+<div class="border-l-2 border-blue-500 pl-3">
+  <span class="text-blue-300 font-semibold">Undo nativo</span><br>
+  <span class="text-gray-400"><code>jj undo</code> deshace <em>cualquier</em> operación. El operation log es persistente e inmutable.</span>
+</div>
+
+</div>
+
+</div>
+
+<div>
+
+<v-click>
+
+<div class="font-mono text-xs bg-gray-900 px-4 rounded space-y-2">
+
+**Git mental model**
+```text
+working tree → staging → commit → branch → push
+     ↓            ↓         ↓        ↓
+  (mutable)   (mutable)  (fixed)  (named)
+```
+
+**jj mental model**
+```text
+working copy (@) = commit automático
+     ↓
+  describe → bookmark → push
+     ↓
+  (siempre snapshot, siempre revertible)
+```
+
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-1 font-mono text-xs bg-gray-900 p-3 rounded">
+
+```bash
+# Git: 4 pasos para guardar trabajo
+git add -A && git commit -m "wip"
+
+# jj: 0 pasos — ya está guardado
+jj describe -m "feat: add parser"
+# El working copy YA era un commit
+```
+
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+El cambio mental más importante: no hay staging.
+El working copy ya ES un commit. Esto cambia todo el flujo.
+"describe" no es "commit" — es poner nombre a algo que ya existe.
+-->
+
+---
+layout: default
+---
+
+# jj Workspaces — aislamiento real para agentes
+
+<div class="grid grid-cols-2 gap-8 mt-2">
+
+<div>
+
+**Problema resuelto: N actores, N working copies**
+
+<div class="text-sm text-gray-300 space-y-2 mt-3">
+
+- Cada `jj workspace add` crea un **working copy independiente**
+- Sin branches de por medio — cada workspace tiene su propio `@`
+- Los cambios son **visibles entre workspaces** (mismo repo)
+- Merge entre workspaces: `jj new workspace_change main`
+
+</div>
+
+<v-click>
+
+<div class="mt-3 font-mono text-xs bg-gray-900 p-3 rounded">
+
+```bash
+# Crear workspace para un agente
+jj workspace add .agents/architect-01
+
+# Cada workspace tiene su propio @
+# Sin conflictos con el working copy principal
+
+# Cuando termina:
+jj new agent_change main   # merge
+jj bookmark set main -r @  # avanzar bookmark
+```
+
+</div>
+
+</v-click>
+
+</div>
+
+<div>
+
+<v-click>
+
+**Comparación directa**
+
+<div class="mt-3 text-xs">
+
+| | Git worktree | jj workspace |
+|---|---|---|
+| **Requiere branch** | Sí (named) | No |
+| **Staging area** | Compartida | No existe |
+| **Visibilidad** | Aislada | Repo-wide |
+| **Merge** | Checkout + merge | `jj new A B` |
+| **Cleanup** | Manual | `jj workspace forget` |
+| **N agentes** | N branches | N workspaces |
+
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="-mt-1 border-l-4 border-orange-500 pl-4 text-sm text-gray-300">
+  Un agente IA puede tener <strong>su propio working copy</strong> sin afectar al humano.<br>
+  Sin <code>stash</code>. Sin <code>branch switch</code>. Sin locks.
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+Aquí es donde jj brilla para AI agents. Cada agente tiene su propia copia de trabajo,
+literalmente un directorio separado, con su propio @. No hay colisión posible.
+-->
+
+---
+layout: default
+---
+
+# Radicle — code hosting soberano
+
+<div class="grid grid-cols-2 gap-8 mt-4">
+
+<div>
+
+**Git, pero peer-to-peer**
+
+<div class="text-sm text-gray-300 space-y-3 mt-3">
+
+<div class="border-l-2 border-green-500 pl-3">
+  <span class="text-green-300 font-semibold">Protocolo, no plataforma</span><br>
+  <span class="text-gray-400">Sin servidor central. Cada nodo tiene una copia completa. Las colaboraciones son <em>patches</em>, no PRs.</span>
+</div>
+
+<div class="border-l-2 border-green-500 pl-3">
+  <span class="text-green-300 font-semibold">Identidad criptográfica</span><br>
+  <span class="text-gray-400">Cada desarrollador tiene un <code>DID</code> (ed25519). No hay cuentas, no hay contraseñas. Tu clave <em>es</em> tu identidad.</span>
+</div>
+
+<div class="border-l-2 border-green-500 pl-3">
+  <span class="text-green-300 font-semibold">Patches > Pull Requests</span><br>
+  <span class="text-gray-400">Un patch es un objeto nativo del protocolo. Vive en el gossip network. No depende de ningún servicio web.</span>
+</div>
+
+<div class="border-l-2 border-green-500 pl-3">
+  <span class="text-green-300 font-semibold">Seed nodes = espejos voluntarios</span><br>
+  <span class="text-gray-400">Cualquiera puede ser un seed. Web UI opcional (<code>radicle-explorer</code>). Sin vendor lock-in.</span>
+</div>
+
+</div>
+
+</div>
+
+<div>
+
+<v-click>
+
+<div class="font-mono text-xs bg-gray-900 p-4 rounded space-y-2">
+
+```bash
+# Inicializar repo en Radicle
+rad init
+
+# Tu identidad
+rad self
+# did:key:z6Mk...  (ed25519)
+
+# Publicar un patch (≈ PR)
+rad patch open --title "feat: parser"
+
+# Listar patches
+rad patch list
+
+# Clonar desde la red
+rad clone rad:z42...
+```
+
+</div>
+
+</v-click>
+
+<v-click>
+
+
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+Radicle no es "otro GitLab self-hosted". Es un protocolo P2P.
+Si GitHub cae mañana, tu código sigue existiendo en todos los nodos que lo replican.
+Los patches son objetos del protocolo, no features de un servicio web.
+-->
+
+---
+layout: center
+---
+
+# Radicle — code hosting soberano
+
+<div class="mt-11 text-xs">
+
+| | GitHub/GitLab | Radicle |
+|---|---|---|
+| **Hosting** | Centralizado | P2P (gossip) |
+| **Identidad** | Username+password | ed25519 DID |
+| **Review** | Pull Request | Patch (nativo) |
+| **CI** | Actions/Pipelines | ci-broker (adaptable) |
+| **Disponibilidad** | SPOF | Seed mesh |
+| **Coste** | SaaS / self-host | Gratis (tu nodo) |
+
+</div>
+
+
+---
+layout: center
+---
+
+# ¿Por qué juntos? jj + Radicle
+
+<div class="mt-4 space-y-3">
+
+<div class="border-l-4 border-blue-400 pl-4 py-2 bg-gray-900 rounded-r">
+  <div class="text-blue-300 font-mono font-semibold text-sm">COLOCATED MODE &nbsp;·&nbsp; .jj/ + .git/</div>
+  <div class="text-gray-400 text-sm mt-1">jj opera sobre el repo. Radicle ve los git refs. Ambos coexisten sin conflicto.</div>
+  <div class="text-gray-500 text-xs mt-1"><code>jj git init --colocate</code> — crea ambos directorios. Radicle no sabe que jj existe.</div>
+</div>
+
+<div class="border-l-4 border-orange-400 pl-4 py-2 bg-gray-900 rounded-r">
+  <div class="text-orange-300 font-mono font-semibold text-sm">WORKFLOW &nbsp;·&nbsp; workspace → describe → publish → patch</div>
+  <div class="text-gray-400 text-sm mt-1">El agente trabaja en un jj workspace. Cuando termina, publica un Radicle patch. Sin push a ningún server.</div>
+  <div class="text-gray-500 text-xs mt-1">La revisión ocurre en el gossip network. El merge ocurre localmente con <code>jj new</code>.</div>
+</div>
+
+<div class="border-l-4 border-green-400 pl-4 py-2 bg-gray-900 rounded-r">
+  <div class="text-green-300 font-mono font-semibold text-sm">SEGURIDAD &nbsp;·&nbsp; El agente nunca necesita push access</div>
+  <div class="text-gray-400 text-sm mt-1">Patches firmados criptográficamente. El humano decide si mergea. El agente propone, nunca impone.</div>
+  <div class="text-gray-500 text-xs mt-1">Sin tokens de GitHub. Sin claves SSH al servidor. La identidad es local.</div>
+</div>
+
+<div class="border-l-4 border-purple-400 pl-4 py-2 bg-gray-900 rounded-r">
+  <div class="text-purple-300 font-mono font-semibold text-sm">SOBERANÍA &nbsp;·&nbsp; Tu infraestructura, tus reglas</div>
+  <div class="text-gray-400 text-sm mt-1">Sin dependencia de servicios cloud para colaborar. Seed nodes privados o públicos. Tú decides.</div>
+  <div class="text-gray-500 text-xs mt-1">Combinado con NATS para eventos → ecosistema completo sin vendor lock-in.</div>
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+El colocated mode es la pieza clave. jj y Radicle no se conocen entre sí — solo comparten .git/.
+Esto no es un hack: es diseño intencionado de jj para interoperabilidad.
+-->
+
+---
+layout: default
+---
+
+# Implementación real — vapora-worktree
+
+<div class="grid grid-cols-2 gap-8 -mt-4">
+
+<div>
+
+**Abstracción VCS en Rust**
+
+<div class="text-sm text-gray-300 space-y-2 mt-1">
+
+- Trait `WorkspaceBackend` con implementaciones jj y Git
+- Detección automática: ¿hay `.jj/`? → `JjBackend`. ¿Solo `.git/`? → `GitBackend`
+- `VcsCapabilities` declarativo — no hard-coded blocking
+- Runtime dispatch via `Box<dyn WorkspaceBackend>`
+</div>
+
+<div class="font-mono text-xs bg-gray-900 p-3 rounded -mt-5">
+
+```rust
+#[async_trait]
+pub trait WorkspaceBackend: Send + Sync {
+    async fn create_workspace(
+        &self, agent_id: &str, ws_id: &str,
+    ) -> Result<WorkspaceInfo>;
+    async fn check_merge(
+        &self, ws_id: &str, target: &str,
+    ) -> Result<MergeStatus>;
+    async fn merge_workspace(
+        &self, ws_id: &str, target: &str,
+    ) -> Result<()>;
+    fn capabilities(&self) -> VcsCapabilities;
+}
+```
+
+</div>
+
+</div>
+
+<div>
+
+<v-click>
+
+**¿Por qué un trait, no un enum?**
+
+<div class="text-sm text-gray-300 mt-2 space-y-2">
+
+<div class="border-l-2 border-orange-500 pl-3">
+  <span class="text-orange-300">Extensibilidad</span><br>
+  <span class="text-gray-400">Añadir Pijul, Sapling, etc. = implementar el trait. Sin tocar el manager.</span>
+</div>
+
+<div class="border-l-2 border-orange-500 pl-3">
+  <span class="text-orange-300">Capabilities declarativas</span><br>
+  <span class="text-gray-400">Cada backend declara qué puede hacer. El orquestador no asume.</span>
+</div>
+
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="font-mono text-xs bg-gray-900 p-3 rounded mt-4">
+
+```rust
+pub struct VcsCapabilities {
+    pub supports_workspaces: bool,
+    pub supports_patch_open: bool,
+    pub merge_is_atomic: bool,
+    pub has_operation_log: bool,
+}
+
+// JjBackend: all true
+// GitBackend: workspaces partial, no patch,
+//             no atomic merge, no op log
+```
+
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+Esto no es abstracción teórica. Es el crate vapora-worktree, ya implementado y con tests.
+El trait permite que el orchestrator no sepa si está hablando con jj o git.
+Nota: VcsCapabilities es el pattern correcto. No boolean flags en el manager.
+-->
+
+---
+layout: default
+---
+
+# Implementación real — ontoref vcs.nu
+
+<div class="grid grid-cols-2 gap-8 -mt-4">
+
+<div>
+
+**Abstracción VCS en Nushell**
+
+<div class="text-sm text-gray-300 space-y-2 -mt-2">
+
+- Módulo `vcs.nu` — único punto de contacto con el VCS
+- Detección por filesystem: `(.jj/ exists?)` → jj, else → git
+- Todos los módulos de ontoref importan `vcs.nu`
+- Ningún `^git` directo fuera de vcs.nu
+
+</div>
+
+<div class="font-mono text-xs bg-gray-900 p-3 rounded -mt-3">
+
+```nushell
+# vcs.nu — API uniforme
+export def detect [] -> string {
+  if (".jj" | path exists) { "jj" }
+  else if (".git" | path exists) { "git" }
+  else { "none" }
+}
+export def current-ref [] -> string {
+  match (detect) {
+    "jj" => {
+      # @ es el working copy, @- es el parent
+      jj log -r "@-" --no-graph -T 'change_id'
+    }
+    "git" => { git rev-parse --short HEAD }
+  }
+}
+```
+
+</div>
+
+</div>
+
+<div>
+
+<v-click>
+
+**Decisión arquitectónica: ADR-013**
+
+<div class="text-sm text-gray-300 -mt-4 space-y-2">
+
+<div class="border-l-2 border-blue-500 pl-3">
+  <span class="text-blue-300">Todo VCS pasa por vcs.nu</span><br>
+  <span class="text-gray-400">Hard constraint. Sin excepciones.</span>
+</div>
+
+<div class="border-l-2 border-blue-500 pl-3">
+  <span class="text-blue-300">jj y rad nunca son required</span><br>
+  <span class="text-gray-400">Un proyecto consumer usa git puro. Sin configuración adicional.</span>
+</div>
+
+<div class="border-l-2 border-blue-500 pl-3">
+  <span class="text-blue-300">Filesystem detection = la verdad</span><br>
+  <span class="text-gray-400">No env vars. No config files. El filesystem es autoritativo.</span>
+</div>
+
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-4 text-xs text-gray-500 font-mono bg-gray-900 p-3 rounded">
+
+**Alternativas rechazadas (ADR-013):**
+```text
+❌ Env var ONTOREF_VCS → estado mutable, desync
+❌ Detección inline por módulo → duplicación
+❌ Shim de traducción → frágil, no isomórfico
+```
+
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+Dos implementaciones de la misma idea: Rust para vapora (crate), Nushell para ontoref (módulo).
+La detección por filesystem es clave: no hay configuración que pueda quedar desincronizada.
+ADR-013 formaliza esto como una hard constraint.
+-->
+
+---
+layout: default
+---
+
+# jjw — el lifecycle wrapper para agentes
+
+<div class="-mt-4">
+
+<div class="font-mono text-xs bg-gray-900 p-3 rounded">
+
+```text
+jjw agent create   → workspace aislado + ontoref run start + metadata .ontoref-run
+jjw agent step     → report de paso (pass/fail/skip) al protocolo ontoref
+jjw agent publish  → push + rad patch open (o git push fallback)
+jjw agent merge    → merge al main + bookmark advance + cleanup
+jjw agent discard  → abandon workspace + cancel run
+jjw agent status   → lista workspaces activos con estado de run
+```
+
+</div>
+
+</div>
+
+<v-click>
+
+<div class="grid grid-cols-2 gap-8 -mt-7">
+
+<div>
+
+**Flujo completo de un agente**
+
+<div class="font-mono text-xs bg-gray-900 px-3 rounded -mt-4">
+
+```bash
+# 1. Agente recibe tarea
+jjw agent create --agent architect-01 \
+  --mode develop --task "add parser module"
+# 2. Trabaja en .agents/architect-01/
+#    (su propio working copy)
+cd .agents/architect-01/
+# ... edita archivos ...
+# 3. Reporta progreso
+jjw agent step develop parse-impl --status pass
+# 4. Publica resultado
+jjw agent publish
+# → rad patch open (si hay remote rad)
+# → git push (fallback)
+# 5. Humano revisa y mergea
+jjw agent merge
+```
+
+</div>
+
+</div>
+
+<div>
+
+**¿Qué ocurre por debajo?**
+
+<div class="text-xs text-gray-300 space-y-2 mt-2">
+
+- `create` → <small>`jj workspace add` + `ontoref run start` + JSON metadata</small>
+- `step` → <small>`ontoref step report` (DAG de modo)</small>
+- `publish` → <small>detecta remote `rad` via `jj git remote list`;<br> si existe → `rad patch open`; si no → `git push`</small>
+- `merge` → <small>`jj new agent_change main` + `jj bookmark set main -r @`</small>
+- `discard` → <small>`jj abandon` + `ontoref mode cancel`</small>
+
+</div>
+
+<div class="mt-4 border-l-4 border-green-500 pl-3 text-sm text-gray-400">
+  El agente nunca toca el working copy del humano.<br>
+  El humano nunca necesita saber que jj existe.
+</div>
+
+</div>
+
+</div>
+
+</v-click>
+
+<Footer />
+
+<!--
+jjw es el pegamento. Conecta jj workspaces con el protocolo ontoref (runs, steps, modes)
+y con Radicle (patches). Es un script Nushell de ~320 líneas.
+Lo importante: el agente tiene un lifecycle formal — no es "haz lo que quieras en un branch".
+-->
+
+---
+layout: default
+---
+
+# NCL merge tool — resolver conflictos en .ontology/
+
+<div class="grid grid-cols-2 gap-8 mt-4">
+
+<div>
+
+**El problema**
+
+<div class="text-sm text-gray-300 space-y-2 mt-2">
+
+- Dos agentes editan `.ontology/core.ncl` en paralelo
+- jj materializa el conflicto (no bloquea)
+- Pero NCL no es texto plano — es un lenguaje tipado
+- Un merge textual puede romper la semántica
+
+</div>
+
+**La solución: merge en dos capas**
+
+<div class="text-sm text-gray-300 space-y-2 mt-3">
+
+<div class="border-l-2 border-blue-500 pl-3">
+  <span class="text-blue-300">Capa 1: git merge-file</span><br>
+  <span class="text-gray-400">3-way text merge. Resuelve adiciones disjuntas (la mayoría de casos).</span>
+</div>
+
+<div class="border-l-2 border-orange-500 pl-3">
+  <span class="text-orange-300">Capa 2: nickel export</span><br>
+  <span class="text-gray-400">Valida que el resultado sea NCL válido. Si falla → conflicto manual.</span>
+</div>
+
+</div>
+
+</div>
+
+<div>
+
+<v-click>
+
+<div class="font-mono text-xs bg-gray-900 p-3 rounded">
+
+```bash
+# Registrar el merge tool en jj config
+# ~/.config/jj/config.toml
+[merge-tools.ncl]
+program = "nu"
+args = [
+  "jjw-ncl-merge.nu",
+  "$left", "$right", "$base", "$output"
+]
+
+# Usar al resolver
+jj resolve --tool ncl -- .ontology/core.ncl
+```
+
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-4 font-mono text-xs bg-gray-900 p-3 rounded">
+
+```text
+Solo mergea archivos ontológicos:
+  ✓ core.ncl    (nodos, edges, tensiones)
+  ✓ state.ncl   (FSM dimensiones)
+  ✓ gate.ncl    (quality gates)
+  ✓ manifest.ncl (capabilities)
+
+  ✗ Cualquier otro .ncl → merge manual
+```
+
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+Los conflictos en NCL son especiales. Un merge textual exitoso puede producir NCL inválido.
+Por eso la segunda capa: nickel export como validador semántico.
+Esto solo aplica a los archivos de ontología — el resto se resuelve normalmente.
+-->
+
+---
+layout: default
+---
+
+# NATS como pegamento — radicle-nats-adapter
+
+<div class="grid grid-cols-2 gap-8 mt-2">
+
+<div>
+
+**Tres mundos conectados sin acoplamiento**
+
+<div class="text-sm text-gray-300 space-y-2 mt-3">
+
+<div class="border-l-2 border-purple-500 pl-3">
+  <span class="text-purple-300 font-semibold">vapora → ontoref</span><br>
+  <span class="text-gray-400"><code>vapora.tasks.completed</code> filtra paths <code>.ontology/</code><br>→ publica <code>ecosystem.reflection.request</code></span>
+</div>
+
+<div class="border-l-2 border-green-500 pl-3">
+  <span class="text-green-300 font-semibold">radicle → ontoref</span><br>
+  <span class="text-gray-400"><code>radicle.patch.opened</code> / <code>updated</code><br>→ publica <code>ecosystem.reflection.request</code></span>
+</div>
+
+<div class="border-l-2 border-orange-500 pl-3">
+  <span class="text-orange-300 font-semibold">ontoref → vapora</span><br>
+  <span class="text-gray-400"><code>ecosystem.reflection.step.completed</code><br>→ publica <code>vapora.tracking.ontoref</code></span>
+</div>
+
+</div>
+
+<v-click>
+
+<div class="mt-3 border-l-4 border-yellow-500 pl-3 text-sm text-gray-400">
+  <strong>Invariante crítico:</strong> subscribe-before-publish.<br>
+  El adapter se suscribe al resultado <em>antes</em> de publicar la solicitud. Previene race conditions.
+</div>
+
+</v-click>
+
+</div>
+
+<div>
+
+<v-click>
+
+<div class="font-mono text-xs bg-gray-900 p-3 rounded">
+
+```text
+┌─────────┐   NATS     ┌──────────┐
+│ Radicle │ ────────→  │ Adapter  │
+│  node   │  patches   │ (bridge) │
+└─────────┘            └────┬─────┘
+                            │
+              ┌─────────────┼──────────┐
+              ↓             ↓          ↓
+        ┌──────────┐  ┌─────────┐ ┌────────┐
+        │ ontoref  │  │ vapora  │ │  CI    │
+        │ daemon   │  │ tracker │ │(broker)│
+        └──────────┘  └─────────┘ └────────┘
+```
+
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="mt-3 font-mono text-xs bg-gray-900 p-3 rounded">
+
+```bash
+# Dos modos de ejecución
+radicle-nats-adapter run
+# → efímero: 1 evento CI, stdin/stdout
+
+radicle-nats-adapter bridge all
+# → long-running: 3 bridges concurrentes
+# → tokio::select! + explicit .abort()
+```
+
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+El adapter es un binario Rust standalone. No depende de ontoref ni de vapora.
+Es puro glue: traduce eventos entre sistemas usando NATS subjects.
+Las convenciones de subject (radicle.*, vapora.*, ecosystem.*) son el contrato.
+-->
+
+---
+layout: default
+---
+
+<h1 class="-mt-4"> Flujo completo — del agente al patch</h1>
+
+<div class="-mt-8">
+
+<div class="font-mono text-xs bg-gray-900 p-4 rounded leading-6">
+```text
+ Agente IA                    jj                    Radicle              Humano
+ ─────────                   ────                   ───────              ──────
+     │ jjw agent create        │                       │                    │
+     │────────────────────→    │                       │                    │
+     │    workspace aislado    │                       │                    │
+     │    .agents/agent-01/    │                       │                    │
+     │                         │                       │                    │
+     │ ... trabaja ...         │                       │                    │
+     │ jjw agent step ✓        │                       │                    │
+     │                         │                       │                    │
+     │ jjw agent publish       │                       │                    │
+     │────────────────────→    │                       │                    │
+     │    jj git push          │                       │                    │
+     │                         │──rad patch open──→    │                    │
+     │                         │                       │──── notifica ────→ │
+     │                         │                       │   revisa patch     │
+     │                         │                       │ ←── rad patch      │
+     │                         │                       │     merge ─────────│
+     │ jjw agent merge         │                       │                    │
+     │────────────────────→    │                       │                    │
+     │    jj new A main        │                       │                    │
+     │    bookmark set main    │                       │                    │
+     │    workspace forget     │                       │                    │
+     │                         │                       │                    │
+     ▼                         ▼                       ▼                    ▼
+```
+
+</div>
+
+</div>
+
+<v-click>
+
+<div class="mt-3 text-sm text-gray-300 grid grid-cols-3 gap-4">
+
+<div class="border-l-2 border-blue-500 pl-3">
+  <span class="text-blue-300">Aislamiento</span><br>
+  <span class="text-gray-400">El agente nunca toca el working copy del humano</span>
+</div>
+
+<div class="border-l-2 border-green-500 pl-3">
+  <span class="text-green-300">Revisión</span><br>
+  <span class="text-gray-400">El patch es un objeto firmado en la red P2P</span>
+</div>
+
+<div class="border-l-2 border-orange-500 pl-3">
+  <span class="text-orange-300">Control</span><br>
+  <span class="text-gray-400">El humano decide cuándo y qué se mergea</span>
+</div>
+
+</div>
+
+</v-click>
+
+<Footer />
+
+<!--
+Este es el flujo end-to-end. Nótese que el agente nunca necesita acceso de escritura al remoto.
+Propone via patch. El humano mantiene el control. jj provee el aislamiento. Radicle provee la distribución.
+-->
+
+---
+layout: default
+---
+
+# Graceful degradation — jj es opcional
+
+<div class="grid grid-cols-2 gap-8 mt-4">
+
+<div>
+
+**Principio: opt-in, never required**
+
+<div class="text-sm text-gray-300 space-y-3 mt-3">
+
+<div class="border-l-2 border-green-500 pl-3">
+  <span class="text-green-300">Si hay .jj/ → usa jj</span><br>
+  <span class="text-gray-400">Workspaces nativos, operation log, snapshots automáticas.</span>
+</div>
+
+<div class="border-l-2 border-blue-500 pl-3">
+  <span class="text-blue-300">Si solo hay .git/ → usa git</span><br>
+  <span class="text-gray-400">Todo funciona. Sin worktree isolation, pero funciona.</span>
+</div>
+
+<div class="border-l-2 border-gray-500 pl-3">
+  <span class="text-gray-300">Radicle: igual</span><br>
+  <span class="text-gray-400">Si hay remote <code>rad</code> → patches. Si no → <code>git push</code> clásico.</span>
+</div>
+
+</div>
+
+<v-click>
+
+<div class="mt-4 font-mono text-xs bg-gray-900 p-3 rounded">
+
+```nushell
+# init-repo.nu — decisión de inicialización
+if (".jj" | path exists) {
+  # Ya inicializado con jj
+} else if (".git" | path exists) {
+  # Git puro — funciona sin cambios
+} else {
+  # Nuevo proyecto → preferir colocated
+  ^jj git init --colocate
+  # Crea .jj/ + .git/ → compatible con todo
+}
+```
+
+</div>
+
+</v-click>
+
+</div>
+
+<div>
+
+<v-click>
+
+<div class="text-sm text-gray-300">
+
+**Lo que esto significa para adopción**
+
+<div class="mt-3 space-y-4">
+
+<div class="border border-gray-700 rounded p-3 bg-gray-900">
+  <div class="text-green-300 font-semibold text-xs">PROYECTO EXISTENTE (git)</div>
+  <div class="text-gray-400 text-xs mt-1">No cambias nada. Todo funciona. vcs.nu detecta git y usa git.</div>
+</div>
+
+<div class="border border-gray-700 rounded p-3 bg-gray-900">
+  <div class="text-blue-300 font-semibold text-xs">QUIERES PROBAR JJ</div>
+  <div class="text-gray-400 text-xs mt-1"><code>jj git init --colocate</code> en un repo existente. Listo. Git sigue funcionando.</div>
+</div>
+
+<div class="border border-gray-700 rounded p-3 bg-gray-900">
+  <div class="text-orange-300 font-semibold text-xs">QUIERES RADICLE</div>
+  <div class="text-gray-400 text-xs mt-1"><code>rad init</code> sobre el repo. Añade un remote. Los patches funcionan.</div>
+</div>
+
+<div class="border border-gray-700 rounded p-3 bg-gray-900">
+  <div class="text-purple-300 font-semibold text-xs">QUIERES TODO</div>
+  <div class="text-gray-400 text-xs mt-1">jj colocated + rad init. Agentes con jjw. Patches soberanos. 0 vendor lock-in.</div>
+</div>
+
+</div>
+
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<Footer />
+
+<!--
+Esto es crítico para la adopción. Nadie tiene que migrar de golpe.
+Un proyecto git puro sigue siendo un proyecto git puro.
+jj y Radicle se añaden incrementalmente, sin romper nada.
+-->
+
+---
+layout: center
+---
+
+# Resumen — qué aporta cada pieza
+
+<div class="mt-4 space-y-3">
+
+<div class="grid grid-cols-4 gap-4 text-center">
+
+<div class="border border-gray-700 rounded p-4 bg-gray-900">
+  <div class="text-blue-300 font-mono font-semibold text-lg">jj</div>
+  <div class="text-gray-400 text-xs mt-2">Aislamiento real<br>Sin staging area<br>Undo nativo<br>Workspaces nativos</div>
+</div>
+
+<div class="border border-gray-700 rounded p-4 bg-gray-900">
+  <div class="text-green-300 font-mono font-semibold text-lg">Radicle</div>
+  <div class="text-gray-400 text-xs mt-2">Soberanía<br>Patches firmados<br>P2P gossip<br>Sin vendor lock-in</div>
+</div>
+
+<div class="border border-gray-700 rounded p-4 bg-gray-900">
+  <div class="text-orange-300 font-mono font-semibold text-lg">jjw</div>
+  <div class="text-gray-400 text-xs mt-2">Lifecycle formal<br>Agent workspaces<br>Protocol bridge<br>NCL merge tool</div>
+</div>
+
+<div class="border border-gray-700 rounded p-4 bg-gray-900">
+  <div class="text-purple-300 font-mono font-semibold text-lg">NATS</div>
+  <div class="text-gray-400 text-xs mt-2">Event mesh<br>Desacoplamiento<br>Bridge adapter<br>CI integration</div>
+</div>
+
+</div>
+
+</div>
+
+<v-click>
+
+<div class="mt-6 border border-gray-700 rounded p-4 bg-gray-900 text-sm text-gray-300 font-mono leading-7 text-center">
+
+```text
+Git solo          → funciona, pero no escala a N agentes concurrentes
++ jj              → aislamiento real por workspace, undo, sin staging
++ Radicle           → soberanía, patches firmados, sin plataforma central
++ jjw + NATS           → lifecycle formal, eventos entre sistemas, CI adaptable
+```
+
+</div>
+
+</v-click>
+
+<Footer />
+
+<!--
+Cada pieza añade una capa. Ninguna es obligatoria.
+Puedes quedarte en cualquier nivel. El valor es aditivo, no todo-o-nada.
+-->
+
+---
+layout: center
+---
+
+<h1 class="font-medium">Gracias</h1>
+
+<div class="flex justify-center mt-4 text-gray-400 text-lg italic">
+  Tu código, tus reglas, tu infraestructura.
+</div>
+
+<div class="mt-8 text-sm text-gray-500 space-y-1 text-center">
+
+`jj` — https://martinvonz.github.io/jj/
+
+`Radicle` — https://radicle.xyz/
+
+`ontoref` — protocolo de auto-conocimiento para proyectos
+
+`vapora` — orquestador de workspaces para agentes
+
+</div>
+
+<div class="flex justify-center">
+  <img class="mt-8 w-40" src="/jesusperez_w.svg">
+</div>
+
+<Footer />
+
+<!--
+Cerrar con calma. El mensaje es soberanía + aislamiento + formalismo.
+No es anti-GitHub. Es pro-independencia.
+-->
diff --git a/assets/presentation/lian-build.md b/assets/presentation/lian-build.md
new file mode 100644
index 0000000..2ebf934
--- /dev/null
+++ b/assets/presentation/lian-build.md
@@ -0,0 +1,761 @@
+---
+theme: default
+title: "炼 lian-build"
+titleTemplate: '%s — Ephemeral BuildKit Substrate'
+layout: cover
+keywords: Rust,BuildKit,CI,sccache,cargo-chef,lian-build,lamina
+download: true
+exportFilename: lian-build-presentation
+monaco: true
+remoteAssets: true
+selectable: true
+colorSchema: dark
+lineNumbers: true
+themeConfig:
+  primary: '#ce422b'
+  logoHeader: '/ferris.svg'
+fonts:
+  mono: 'Victor Mono'
+background: /jude-infantini-mI-QcAP95Ok-unsplash.jpg
+class: 'justify-center flex flex-cols photo-bg'
+---
+
+<h1 class="absolute top-15 left-3/10 font-bold mt-3 text-5xl">lian-build</h1>
+<h2 class="absolute top-30 left-1/10 font-medium my-11 text-2xl opacity-80">
+  Ephemeral BuildKit — from <code>docker build</code> to a substrate
+</h2>
+
+<div class="absolute top-57 left-2/10 text-sm opacity-60 font-mono">
+  BuildKit · cargo-chef · sccache · NATS · Nickel · lamina
+</div>
+
+<div class="absolute top-65 left-3/10"><img src="/lian-h.svg" width="420"></div>
+
+<img class="absolute bottom-10 right-10 w-32" src="/ferris.svg">
+
+<style scoped>
+h1, h2, div { z-index: 10; }
+code { background: rgba(206,66,43,0.2); padding: 0.1em 0.3em; border-radius: 3px; }
+</style>
+
+---
+
+# The Problem
+
+**Rust CI: 8 minutes cold. Every. Single. Build.**
+
+
+<div>
+
+### What happens today without a substrate
+
+<div class="absolute right-2 top-4 w-110 box-highlight mt-2">
+  Every developer and CI pipeline reinvents the same wheel — and pays full price each time.
+</div>
+
+```
+push → CI triggers
+  └─ docker build .
+       ├─ FROM rust:latest            # 1.8 GB pull
+       ├─ COPY Cargo.toml Cargo.lock  # layer invalidated
+       ├─ RUN cargo build --deps      # 4–6 min compiling serde, tokio…
+       ├─ COPY src/                   # always changes
+       └─ RUN cargo build             # 2–4 min compiling your code
+```
+
+</div>
+
+<div class="grid grid-cols-2 gap-6 mt-4">
+<div>
+
+### The compounding failures
+
+- **Cache bust cascade** — `Cargo.lock`<br> change invalidates every downstream layer
+- **No cross-run reuse** — parallel PRs duplicate identical dep compilation
+</div>
+
+<div>
+
+- **Registry pull cost** — base image re-pulled<br> if not pinned
+- **OOM silent failure** — exit 137, no retry, <br> build marked failed
+
+</div>
+
+</div>
+---
+
+# Why Not `docker build`?
+
+<div class="grid grid-cols-2 gap-6 mt-15">
+
+<div>
+
+### `docker build` limitations
+
+```bash
+# No runner control
+docker build .       # uses daemon defaults
+                     # no VM sizing, no OOM retry
+# No external cache injection
+# --cache-from only reads local/registry layers
+# Can't mount S3-backed sccache bucket
+
+# No SSH forwarding into RUN steps
+# (without BuildKit secret/SSH mounts)
+
+# No structured events
+# build started/finished = exit code only
+```
+
+</div>
+
+<div>
+
+<div class="box-highlight mt-11 text-sm">
+  <code>docker build</code> is a convenience wrapper. When you need <em>control</em>, you need BuildKit directly.
+</div>
+
+</div>
+</div>
+
+---
+
+# Why Not `docker build`?
+
+<div>
+
+### What we actually need
+
+| Need | `docker` | BuildKit |
+|------|---------------|----------|
+| Cache mounts | ✗ | `--mount=type=cache` |
+| SSH into build | partial | `--mount=type=ssh` |
+| Secret injection | ✗ | `--mount=type=secret` |
+| Remote daemon | cumbersome | `buildctl --addr` |
+| Structured output | exit code | `--progress=json` |
+| Parallel stages | limited | LLB graph native |
+
+</div>
+
+---
+
+# Why BuildKit
+
+**BuildKit is not a build tool. It's a graph execution engine.**
+
+```
+Your Dockerfile ──► LLB (Low-Level Build) graph ──► parallel DAG execution
+                         │
+                         ├─ content-addressed cache (every node keyed by its inputs)
+                         ├─ prunable: unchanged nodes cost nothing
+                         ├─ remote execution: daemon can run anywhere buildctld runs
+                         └─ mount primitives: cache / secret / ssh / bind
+```
+
+<div class="grid grid-cols-3 gap-4 mt-4 text-sm">
+
+<div class="border border-rust-orange/30 rounded p-3">
+
+### Cache mounts
+```dockerfile
+RUN --mount=type=cache,\
+    target=/usr/local/cargo/registry \
+    cargo build
+```
+Registry downloads survive across builds *inside* the daemon.
+
+</div>
+
+<div class="border border-rust-orange/30 rounded p-3">
+
+### Secret mounts
+```dockerfile
+RUN --mount=type=secret,\
+    id=sccache_creds \
+    SCCACHE_S3_USE_SSL=true \
+    cargo build
+```
+Credentials never written to layer.
+
+</div>
+
+<div class="border border-rust-orange/30 rounded p-3">
+
+### Remote daemon
+```bash
+buildctl \
+  --addr ssh://runner:1234 \
+  build \
+  --frontend dockerfile.v0 \
+  --local context=. \
+  --output type=image,name=…
+```
+Daemon on ephemeral VM, client local.
+
+</div>
+
+</div>
+
+---
+
+# cargo-chef — Dependency Layer Surgery
+
+**Problem:** any `src/` change busts the dependency compilation layer.
+
+**cargo-chef solution:** separate the dependency graph compilation from your code.
+
+```dockerfile
+# Stage 1 — planner: extract dependency recipe (no actual compilation)
+FROM rust:1.82 AS planner
+RUN cargo install cargo-chef
+COPY . .
+RUN cargo chef prepare --recipe-path recipe.json   # only Cargo.toml/Cargo.lock matter
+
+# Stage 2 — cooker: compile all deps from the recipe (expensive, cached)
+FROM rust:1.82 AS cooker
+RUN cargo install cargo-chef
+COPY --from=planner /app/recipe.json recipe.json
+RUN cargo chef cook --release --recipe-path recipe.json  # deps compiled, layer stable
+
+# Stage 3 — final: only your code compiles (fast, always runs)
+FROM rust:1.82 AS final
+COPY --from=cooker /app/target target
+COPY --from=cooker $CARGO_HOME $CARGO_HOME
+COPY . .
+RUN cargo build --release
+```
+
+---
+
+# cargo-chef — Dependency Layer Surgery
+
+<div class="box-highlight mt-3 text-sm">
+  The <em>planner</em> stage is cheap — it only reads metadata.
+
+  The <em>cooker</em> layer is stable as long as <code>Cargo.lock</code> doesn't change.
+
+  Source changes never touch the cooker.
+</div>
+
+---
+
+# sccache — Compiler-Level Cache
+
+**cargo-chef** caches at the *crate dependency graph* level.<br>
+**sccache** caches at the *individual compilation unit* level.
+
+<div class="grid grid-cols-2 gap-6 mt-4">
+```
+cargo build
+  └─ rustc src/sizing.rs → .rlib
+     └─ sccache wraps rustc:
+      hash(source + flags + toolchain) → S3 lookup
+      HIT  → download cached .rlib (seconds)
+      MISS → compile + upload (minutes)
+```
+
+<div class="-mt12">
+
+<h3 class="ml-15"> Orthogonal cache layers </h3>
+
+| Layer | Tool | Granularity | Backend |
+|-------|------|-------------|---------|
+| Toolchain image | lamina | Docker layer | Registry |
+| Dep compilation | cargo-chef | Cargo crate graph | Docker layer |
+| Artifact cache | sccache | Individual `.rlib` | S3 / GCS / Redis |
+
+</div>
+</div>
+
+<div class="grid grid-cols-2 gap-6 -mt-45">
+<div>
+
+### Secret mount pattern (lamina canonical)
+
+```dockerfile
+RUN --mount=type=secret,id=sccache_creds,\
+    target=/run/secrets/sccache_creds \
+  . /run/secrets/sccache_creds && \
+  RUSTC_WRAPPER=sccache \
+  SCCACHE_BUCKET=$SCCACHE_BUCKET \
+  cargo chef cook --release \
+    --recipe-path recipe.json
+```
+
+
+</div>
+
+</div>
+
+Credentials injected at build time, not baked into layer.
+
+---
+
+# cargo-chef + sccache + BuildKit Together
+
+**Each tool solves a different cache miss problem.**
+
+```
+Cold build (first run)
+──────────────────────────────────────────────────────────────────────────────
+  planner  ──► recipe.json                             (always cheap: metadata only)
+  cooker   ──► compile 200 deps from scratch           (6 min — sccache MISS: upload)
+  final    ──► compile your code                       (2 min — sccache MISS: upload)
+
+Warm build — Cargo.lock unchanged, your code changed
+──────────────────────────────────────────────────────────────────────────────
+  planner  ──► recipe.json                             (cheap)
+  cooker   ──► BuildKit layer HIT (identical recipe)   (0 sec — skip entirely)
+  final    ──► compile your code                       (sccache MISS if changed: 2 min)
+
+Hot build — src/ minor change
+──────────────────────────────────────────────────────────────────────────────
+  planner  ──► recipe.json                             (cheap)
+  cooker   ──► BuildKit layer HIT                      (0 sec)
+  final    ──► rustc on changed files → sccache HIT    (seconds per file)
+```
+
+<div class="box-highlight mt-3 text-sm">
+  Three cache layers, three different scopes. When one misses, the others still win.
+  BuildKit serializes the dependency; sccache and cargo-chef operate independently inside.
+</div>
+
+---
+
+# lian-build — Architecture  
+
+<div class="absolute right-30 top-8"><img src="/lian-h.svg" width="140"></div>
+
+**Single binary. Orchestrates remote BuildKit runs. Emits lifecycle events.**
+
+```
+lian-build CLI
+    │
+    ├─ 1. resolve runner size          sizing::resolve(.build-spec.ncl → P95 → lang default)
+    │
+    ├─ 2. publish build.started        NATS: <prefix>.<workspace>.build.started
+    │
+    ├─ 3. spawn runner                 POST /api/v1/vm-pool  →  lease_id
+    │        └─ hcloud cax11–cax41 (ARM) | proxmox | docker-local
+    │
+    ├─ 4. rsync build context          rsync -e ssh context/ runner:/workspace/
+    │
+    ├─ 5. run buildctl over SSH        buildctl --addr ssh://runner build …
+    │        └─ OOM exit 137? retry once on next size tier (ADR-039)
+    │
+    ├─ 6. record_metrics               POST /api/v1/metrics (cpu_p95, mem_p95)
+    │
+    ├─ 7. destroy runner               DELETE /api/v1/vm-pool/{lease_id}  [always]
+    │
+    └─ 8. publish build.completed/failed
+```
+
+<div class="text-xs opacity-60 mt-2">Compute and registry are plug-in slots. Callers supply <code>BuildDirectives</code> in Nickel — no caller identity in core code.</div>
+
+---
+
+# Three-Tier Sizing Resolution
+
+**First match wins. Explicit beats historical beats defaults.**
+
+<div class="grid grid-cols-3 gap-4 mt-4 text-sm">
+
+<div class="border border-rust-orange/40 rounded p-3">
+
+### Tier 1 — Explicit
+```nix
+# .build-spec.ncl
+# in build context
+{
+  runner_type = "cax31", 
+  # authoritative
+  # or raw resources:
+  cpu = 8,
+  memory_gb = 16,
+  time_budget_min = 90,
+}
+```
+Validated against `schemas/build_spec.ncl`.<br> Repo-level contract.
+
+</div>
+
+<div class="border border-rust-orange/30 rounded p-3">
+
+### Tier 2 — P95 Historical
+```
+GET /api/v1/p95?workspace=…
+→ { 
+cpu_p95: 3.2, mem_p95: 6.1
+}
+
+effective = {
+cpu:   ceil(3.2 × 1.2) = 4,
+mem_gb:ceil(6.1 × 1.2) = 8,
+}
+floor: min(2 cpu, 4 GB)
+```
+Measured from prior runs. <br>Advisory — operator must approve before production use.
+
+</div>
+
+<div class="border border-rust-orange/20 rounded p-3">
+
+### Tier 3 — Lang Default
+```rust
+match language {
+  "rust" => (
+      4 cpu, 8 GB, 60 min),
+  "go"   => (
+      2 cpu, 4 GB, 30 min),
+  "java" => (
+      4 cpu, 8 GB, 45 min),
+  _      => (
+      2 cpu, 4 GB, 30 min),
+}
+```
+Conservative floor. Rust is more expensive than Go — that's structural.
+
+</div>
+
+</div>
+
+---
+
+# OOM Retry — Bounded Escalation
+
+**Exit 137 or stderr "OOM"/"Killed" → walk one tier up. Once.**
+
+<div class="grid grid-cols-2 gap-6 mt-4">
+<div>
+```rust
+// MAX_OOM_RETRIES = 1 — 
+// ADR-039 constraint oom-retry-bounded
+pub const MAX_OOM_RETRIES: u8 = 1;
+
+const SIZE_TIERS: &[(&str, u32, u32)] = &[
+  ("cax11",  2,  4),
+  ("cax21",  4,  8), // ← most Rust builds here
+  ("cax31",  8, 16),  // ← OOM retry target
+  ("cax41", 16, 32),
+];
+```
+</div>
+
+<div>
+
+### Why bounded at 1
+
+- Second OOM means **misconfiguration**, <br>not transient pressure
+- Unbounded retry loops spend money <br>on dead ends
+- Forces developer to set explicit `runner_type`<br> in `.build-spec.ncl`
+- ADR-039 constraint —<br> changing this requires a new ADR
+
+</div>
+
+<div class="-mt35">
+
+### Retry flow
+
+```
+build on cax21 → OOM (exit 137)
+  └─ retries_used(0) < MAX_OOM_RETRIES(1) ✓
+     └─ next_size_tier(cax21) → cax31
+        └─ rebuild on cax31
+           ├─ success → record_metrics, destroy
+           └─ OOM again → FAIL (retries exhausted)
+```
+
+</div>
+
+</div>
+
+---
+
+<h1 class="-mt8"> Cache Namespace Model</h1>
+
+**Isolation between CI and session actors — the core tension resolved.**
+
+```
+Registry
+├── ci/<workspace>/*          canonical — written by CI, read-only to sessions
+│     ├── ci/lian-build/deps:sha256-…   (cargo-chef cooker layer)
+│     └── ci/lian-build/base:sha256-…   (toolchain layer from lamina)
+│
+└── dev/<actor-id>-<workspace>/*  ephemeral — per session actor
+      └── dev/jpl-lian-build/…        (your WIP session cache)
+```
+
+<div class="grid grid-cols-2 gap-6 mt-0">
+
+<div>
+
+<h3 class="-mb4">Resolution rules</h3>
+
+| <small>Actor</small>  | <small>Reads</small> | <small>Writes</small> |
+|------------|-------|--------|
+| <small>`'ci`</small> | <small>`ci/*` + own `dev/*`</small> | <small>`ci/*`</small> |
+| <small>`'human`</small> | <small>`ci/*` + own `dev/*`</small> | <small>own `dev/*`</small> |
+| <small>`'agent`</small> | <small>`ci/*` + own `dev/*`</small> | <small>own `dev/*`</small> |
+| <small>`'ci_aux`</small> | <small>`ci/*` only</small> | <small>`ci/*` (restricted)</small> |
+
+</div>
+
+<div class="mt-1">
+
+### Nickel schema
+
+```nix
+# schemas/cache_policy.ncl
+let SessionCacheDisposition = [|
+  'export,    # write back to registry on success
+  'discard,   # ephemeral, discard after run
+  'rollback,  # revert to last good state on fail
+|]
+```
+
+Sessions declare intent. lian-build enforces it.<br>
+CI never imports from `dev/*`.
+
+</div>
+
+</div>
+
+---
+
+# BuildDirectives — Caller-Supplied Vocabulary
+
+**Callers (provisioning, vapora, CI) supply directives in Nickel. Core has no caller identity.**
+
+```nix
+# schemas/build_directives.ncl  — the contract surface
+let BuildDirectives = {
+  workspace         | String,
+  artifact          | BuildArtifact,
+  compute_provider  | ComputeProviderRef,   # 'hcloud | 'proxmox | 'docker_local
+  registry_provider | RegistryProviderRef,  # 'zot | 'harbor | 'ghcr | 'dockerhub
+  cache_policy      | CachePolicy,
+  runner_override   | RunnerOverride | optional,
+  nats_events       | NatsEventConfig | optional,
+}
+```
+
+<div class="grid grid-cols-2 gap-6 mt-4 text-sm">
+
+<div>
+
+### CI invocation
+```nix
+# ci/directives.ncl
+let D = import "defaults/build_directives.ncl" in
+D.make_ci_build {
+  workspace = "lian-build",
+  artifact  = { 
+      image = "registry/lian-build:${sha}" },
+  cache_policy = D.ci_cache_policy,
+}
+```
+
+</div>
+
+<div>
+
+### Session invocation
+```nix
+# dev/session.ncl
+let D = import "defaults/build_directives.ncl" in
+D.make_session_build {
+  workspace   = "lian-build",
+  actor_id    = "jpl",
+  disposition = 'discard,
+}
+```
+
+</div>
+
+</div>
+
+<div class="text-xs opacity-60 mt-3">Hard constraint (ADR-001): <code>src/</code> must not match <code>provisioning_workspace | vapora_ | woodpecker_</code> — caller logic stays in directives.</div>
+
+---
+layout: cover
+background: ./jude-infantini-mI-QcAP95Ok-unsplash.jpg
+class: 'text-center photo-bg'
+---
+
+# lamina
+## The pre-baked layer library
+
+<br>
+    
+### *The catalog that feeds lian-build*
+
+---
+
+<h1 class="p-b-5"> lamina — What It Is </h1>
+
+**Docker base images (toolchain layers) <br> + pre-cooked cargo dep caches (dependency layers).**
+No binary. 
+<br>
+    
+No `src/`. No `Cargo.toml`. Dockerfiles + Nickel schemas + Nushell scripts.
+
+```
+lamina/
+├── rust/           ─── Rust toolchain layer (rustup + sccache + cargo-chef)
+├── leptos/         ─── Leptos WASM layer (rust + wasm-pack + trunk)
+├── ontoref/        ─── Nickel + ore tools layer
+├── nushell/        ─── Nu shell layer
+├── lian-build/     ─── Build directives per layer, ctx-test.nu script
+│     ├── Dockerfile.rust           # planner/cooker/final for rust layer
+│     ├── build_directives.ncl      # per-layer lian-build config
+│     └── ctx-test.nu               # local test runner (docker-local mode)
+└── schemas/        ─── workflow.ncl, layer catalog contracts
+```
+---
+
+# lamina — What It Is
+
+
+<div class="grid grid-cols-2 gap-6 mt-4 text-sm">
+
+<div>
+
+### Layer types
+
+| Type | What it provides | Cache scope |
+|------|-----------------|-------------|
+| Toolchain | rustup, cargo, sccache binary | Docker registry |
+| Dep layer | compiled `.rlib` for your deps | Docker layer + S3 |
+| Utility | additional tools (nu, nickel) | Docker layer |
+
+</div>
+
+<div>
+
+### Catalogue invariant
+
+Every layer in `catalog/` has a `workflow.ncl` that declares:
+- `tools_provided` — binaries that must exist post-build
+- `build_base` — which layer it depends on (DAG)
+- `artifact_paths` — what gets promoted to registry
+
+`catalog-validate.nu --check-dag` enforces the DAG.
+
+</div>
+
+</div>
+
+---
+
+# lamina + lian-build — End-to-End
+
+**lamina provides the layers. lian-build provides the compute.**
+
+```
+lamina                              lian-build
+──────────────────────────────────  ──────────────────────────────────────────────
+
+rust/Dockerfile                      BuildDirectives (Nickel)
+  planner                    ──────► --cache-from registry/lamina/rust-deps:sha
+  cooker (cargo-chef)                --cache-to   registry/lamina/rust-deps:sha
+  final                              --image       registry/lamina/rust:latest
+
+schema:                              Compute:
+  workflow.ncl                        spawn cax21 runner (hcloud ARM)
+  build_directives.ncl   ──────────►  rsync context/ → runner
+                                      buildctl --addr ssh://runner:1234
+ctx-test.nu                           OOM? → cax31, retry once
+  --layer rust                        destroy runner always
+  --mode docker-local       ◄──────   NATS: lian-build.lamina.build.completed
+  (local dev without VM)
+```
+
+<div class="box-highlight mt-1 text-sm">
+  lamina layers become the <code>--cache-from</code> inputs for downstream project builds.<br>
+  A project's cargo deps compile <em>on top of</em> the lamina rust dep layer <br>— hitting sccache HIT for everything lamina already compiled.
+</div>
+
+---
+
+# The Full Picture
+
+**From `docker build .` to a controlled substrate.**
+
+<div class="grid grid-cols-2 gap-8 mt-4 text-sm">
+
+<div>
+
+### Before
+
+```
+dev push
+  └─ CI: docker build .
+       ├─ pull rust:latest (1.8 GB)
+       ├─ cargo build --deps (6 min, always)
+       └─ cargo build src   (2 min)
+
+8–10 min every build.
+No retry on OOM.
+No observability.
+No cross-build reuse.
+No actor isolation.
+```
+
+</div>
+
+<div>
+
+### After
+
+```
+dev push
+  └─ lian-build dispatch
+       ├─ sizing: .build-spec.ncl → cax21
+       ├─ NATS: build.started
+       ├─ spawn cax21 (hcloud ARM, 30 sec)
+       ├─ rsync context (15 sec)
+       ├─ buildctl:
+       │    ├─ FROM lamina/rust:latest 
+       │    │   (registry HIT, 0 sec)
+       │    ├─ cooker: cargo-chef layer 
+       │    │   (registry HIT, 0 sec)
+       │    └─ final: your code
+       │        (sccache: seconds)
+       ├─ destroy runner
+       └─ NATS: build.completed
+
+~2 min warm. OOM retry automatic.
+Structured events. Multi-actor isolation.
+```
+
+</div>
+
+</div>
+
+---
+layout: cover
+background: ./images/cleo-heck-1-l3ds6xcVI-unsplash.jpg
+class: 'text-center photo-bg'
+---
+
+<h1 class="font-bold text-5xl absolute top-4 left-4.3/10"><img src="/lian-v.svg" width="130"></h1>
+<h2 class="mt-40 text-2xl opacity-80">Alchemical refinement</h2>
+
+<div class="mt-8 text-sm opacity-60 font-mono">
+  lian-build · lamina · BuildKit · cargo-chef · sccache
+</div>
+
+<div class="mt-6 text-base">
+  Each build: <em class="!text-orange-500">ephemeral compute, content-addressed cache, structured events.</em>
+  <br>
+      
+  Callers supply intent. `lian-build` supplies execution.
+  
+</div>
+
+<img class="absolute bottom-8 right-8 w-24 opacity-80" src="/ferris-celebration.svg">
+
+<style scoped>
+h1, h2, div, p { z-index: 10; }
+em { color: #ce422b; font-style: italic; }
+</style>
diff --git a/assets/presentation/slides.md b/assets/presentation/ontology_slides.md
similarity index 100%
rename from assets/presentation/slides.md
rename to assets/presentation/ontology_slides.md
diff --git a/assets/presentation/public/lian-h.svg b/assets/presentation/public/lian-h.svg
new file mode 100644
index 0000000..242d3e0
--- /dev/null
+++ b/assets/presentation/public/lian-h.svg
@@ -0,0 +1,70 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="28 19 110 50" preserveAspectRatio="xMidYMid meet" role="img" aria-label="Lian Build">
+  <title>Lian Build</title>
+  <defs>
+    <linearGradient id="lh-fg" x1="0" y1="1" x2="0" y2="0">
+      <stop offset="0" stop-color="#e86c2f"/>
+      <stop offset="1" stop-color="#f5a623"/>
+    </linearGradient>
+    <linearGradient id="lh-cg" x1="0" y1="1" x2="0" y2="0">
+      <stop offset="0" stop-color="#f5a623"/>
+      <stop offset="1" stop-color="#fcd99e"/>
+    </linearGradient>
+    <style><![CDATA[
+      @import url('https://fonts.googleapis.com/css2?family=Jost:wght@400;500&display=swap');
+      .lh-flame{fill:url(#lh-fg);opacity:0;transform-origin:50% 100%;transform-box:fill-box;animation:lh-flame 23s cubic-bezier(.4,0,.2,1) infinite}
+      .lh-core{fill:url(#lh-cg);opacity:0;transform-origin:50% 100%;transform-box:fill-box;animation:lh-core 23s cubic-bezier(.4,0,.2,1) infinite}
+      .lh-cruc{fill:none;stroke:#a8a8a0;stroke-width:1.5;stroke-linecap:round;stroke-dasharray:34;stroke-dashoffset:34;animation:lh-cruc 23s cubic-bezier(.4,0,.2,1) infinite}
+      .lh-word{fill:#bd9156;font-family:"Jost","Avenir Next",Futura,"Century Gothic","Helvetica Neue",system-ui,sans-serif;font-weight:500;letter-spacing:.22em;font-size:22px;opacity:0;animation:lh-word 23s cubic-bezier(.4,0,.2,1) infinite}
+      .lh-sub{fill:#f29a3d;font-family:"Jost","Avenir Next",Futura,"Century Gothic","Helvetica Neue",system-ui,sans-serif;font-weight:500;letter-spacing:.22em;font-size:14px;opacity:0;animation:lh-sub 23s cubic-bezier(.4,0,.2,1) infinite}
+      .lh-fbord{fill:none;stroke:#e86c2f;stroke-width:1.5;stroke-linecap:round;stroke-linejoin:round;opacity:0;animation:lh-fbord 23s cubic-bezier(.4,0,.2,1) infinite}
+      .lh-code{opacity:0;animation:lh-code 23s cubic-bezier(.4,0,.2,1) infinite}
+      @keyframes lh-core{
+        0%{opacity:0;transform:scale(.5)} 1.3%{opacity:.75;transform:scale(1.05)}
+        3.5%{opacity:.55;transform:scale(.95)} 5.2%{opacity:.35;transform:scale(.92)}
+        22%{opacity:.55;transform:scale(1)} 37%{opacity:.24;transform:scale(.88)}
+        53%{opacity:.55;transform:scale(1)} 68%{opacity:.24;transform:scale(.88)}
+        83%{opacity:.55;transform:scale(1)} 99%{opacity:.35;transform:scale(.92)} 100%{opacity:0}
+      }
+      @keyframes lh-flame{
+        0%,6.5%{opacity:0;transform:scale(.4)} 8.7%{opacity:.9;transform:scale(1)}
+        10.4%{opacity:1;transform:scale(1.05)} 12%{opacity:1;transform:scale(1)}
+        99%{opacity:1;transform:scale(1)} 100%{opacity:0;transform:scale(.4)}
+      }
+      @keyframes lh-fbord{
+        0%,11.7%{opacity:0} 14.3%{opacity:1} 99%{opacity:1} 100%{opacity:0}
+      }
+      @keyframes lh-code{
+        0%,15.2%{opacity:0} 17.4%{opacity:1} 99%{opacity:1} 100%{opacity:0}
+      }
+      @keyframes lh-cruc{
+        0%,18.3%{stroke-dashoffset:34;opacity:0} 18.8%{opacity:.8}
+        22.6%{stroke-dashoffset:0;opacity:.8} 99%{stroke-dashoffset:0;opacity:.8}
+        100%{stroke-dashoffset:34;opacity:0}
+      }
+      @keyframes lh-sub{
+        0%,19.6%{opacity:0} 23.9%{opacity:.85} 99%{opacity:.85} 100%{opacity:0}
+      }
+      @keyframes lh-word{
+        0%,20.4%{opacity:0} 25.2%{opacity:1} 99%{opacity:1} 100%{opacity:0}
+      }
+      @media (prefers-reduced-motion:reduce){
+        .lh-flame,.lh-fbord,.lh-core,.lh-cruc,.lh-word,.lh-sub,.lh-code{animation:none}
+        .lh-flame{opacity:1} .lh-fbord{opacity:1} .lh-code{opacity:1}
+        .lh-core{opacity:.35;transform:scale(.95);transform-origin:50% 100%;transform-box:fill-box}
+        .lh-cruc{opacity:.8;stroke-dashoffset:0} .lh-word{opacity:1} .lh-sub{opacity:.85}
+      }
+    ]]></style>
+  </defs>
+  <path class="lh-cruc" d="M 33 65 L 67 65"/>
+  <path class="lh-flame" d="M 50 62 C 33 57 33 37 50 23 C 67 37 67 57 50 62 Z"/>
+  <path class="lh-fbord" d="M 50 62 C 33 57 33 37 50 23 C 67 37 67 57 50 62 Z"/>
+  <path class="lh-core" d="M 50 57 C 41 52 41 40 50 32 C 59 40 59 52 50 57 Z"/>
+  <g class="lh-code" fill="none" stroke="#c25a1a" stroke-width="1.6" stroke-linecap="round" stroke-linejoin="round">
+    <polyline points="46,42 43,46 46,50"/>
+    <line x1="51" y1="43" x2="48" y2="49"/>
+    <polyline points="54,42 57,46 54,50"/>
+  </g>
+  <text class="lh-word" x="80" y="46" text-anchor="start">lian</text>
+  <text class="lh-sub" x="83" y="64" text-anchor="start">build</text>
+</svg>
diff --git a/assets/presentation/public/lian-v.svg b/assets/presentation/public/lian-v.svg
new file mode 100644
index 0000000..35f026c
--- /dev/null
+++ b/assets/presentation/public/lian-v.svg
@@ -0,0 +1,70 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="85 19 56 100" preserveAspectRatio="xMidYMid meet" role="img" aria-label="Lian Build">
+  <title>Lian Build</title>
+  <defs>
+    <linearGradient id="lv-fg" x1="0" y1="1" x2="0" y2="0">
+      <stop offset="0" stop-color="#e86c2f"/>
+      <stop offset="1" stop-color="#f5a623"/>
+    </linearGradient>
+    <linearGradient id="lv-cg" x1="0" y1="1" x2="0" y2="0">
+      <stop offset="0" stop-color="#f5a623"/>
+      <stop offset="1" stop-color="#fcd99e"/>
+    </linearGradient>
+    <style><![CDATA[
+      @import url('https://fonts.googleapis.com/css2?family=Jost:wght@400;500&display=swap');
+      .lv-flame{fill:url(#lv-fg);opacity:0;transform-origin:50% 100%;transform-box:fill-box;animation:lv-flame 23s cubic-bezier(.4,0,.2,1) infinite}
+      .lv-core{fill:url(#lv-cg);opacity:0;transform-origin:50% 100%;transform-box:fill-box;animation:lv-core 23s cubic-bezier(.4,0,.2,1) infinite}
+      .lv-cruc{fill:none;stroke:#a8a8a0;stroke-width:1.5;stroke-linecap:round;stroke-dasharray:36;stroke-dashoffset:36;animation:lv-cruc 23s cubic-bezier(.4,0,.2,1) infinite}
+      .lv-word{fill:#bd9156;font-family:"Jost","Avenir Next",Futura,"Century Gothic","Helvetica Neue",system-ui,sans-serif;font-weight:500;letter-spacing:.22em;font-size:22px;opacity:0;animation:lv-word 23s cubic-bezier(.4,0,.2,1) infinite}
+      .lv-sub{fill:#f29a3d;font-family:"Jost","Avenir Next",Futura,"Century Gothic","Helvetica Neue",system-ui,sans-serif;font-weight:500;letter-spacing:.22em;font-size:14px;opacity:0;animation:lv-sub 23s cubic-bezier(.4,0,.2,1) infinite}
+      .lv-fbord{fill:none;stroke:#e86c2f;stroke-width:1.5;stroke-linecap:round;stroke-linejoin:round;opacity:0;animation:lv-fbord 23s cubic-bezier(.4,0,.2,1) infinite}
+      .lv-code{opacity:0;animation:lv-code 23s cubic-bezier(.4,0,.2,1) infinite}
+      @keyframes lv-core{
+        0%{opacity:0;transform:scale(.5)} 1.3%{opacity:.75;transform:scale(1.05)}
+        3.5%{opacity:.55;transform:scale(.95)} 5.2%{opacity:.35;transform:scale(.92)}
+        22%{opacity:.55;transform:scale(1)} 37%{opacity:.24;transform:scale(.88)}
+        53%{opacity:.55;transform:scale(1)} 68%{opacity:.24;transform:scale(.88)}
+        83%{opacity:.55;transform:scale(1)} 99%{opacity:.35;transform:scale(.92)} 100%{opacity:0}
+      }
+      @keyframes lv-flame{
+        0%,6.5%{opacity:0;transform:scale(.4)} 8.7%{opacity:.9;transform:scale(1)}
+        10.4%{opacity:1;transform:scale(1.05)} 12%{opacity:1;transform:scale(1)}
+        99%{opacity:1;transform:scale(1)} 100%{opacity:0;transform:scale(.4)}
+      }
+      @keyframes lv-fbord{
+        0%,11.7%{opacity:0} 14.3%{opacity:1} 99%{opacity:1} 100%{opacity:0}
+      }
+      @keyframes lv-code{
+        0%,15.2%{opacity:0} 17.4%{opacity:1} 99%{opacity:1} 100%{opacity:0}
+      }
+      @keyframes lv-cruc{
+        0%,18.3%{stroke-dashoffset:36;opacity:0} 18.8%{opacity:.8}
+        22.6%{stroke-dashoffset:0;opacity:.8} 99%{stroke-dashoffset:0;opacity:.8}
+        100%{stroke-dashoffset:36;opacity:0}
+      }
+      @keyframes lv-sub{
+        0%,19.6%{opacity:0} 23.9%{opacity:.85} 99%{opacity:.85} 100%{opacity:0}
+      }
+      @keyframes lv-word{
+        0%,20.4%{opacity:0} 25.2%{opacity:1} 99%{opacity:1} 100%{opacity:0}
+      }
+      @media (prefers-reduced-motion:reduce){
+        .lv-flame,.lv-fbord,.lv-core,.lv-cruc,.lv-word,.lv-sub,.lv-code{animation:none}
+        .lv-flame{opacity:1} .lv-fbord{opacity:1} .lv-code{opacity:1}
+        .lv-core{opacity:.35;transform:scale(.95);transform-origin:50% 100%;transform-box:fill-box}
+        .lv-cruc{opacity:.8;stroke-dashoffset:0} .lv-word{opacity:1} .lv-sub{opacity:.85}
+      }
+    ]]></style>
+  </defs>
+  <path class="lv-cruc" d="M 92 66 L 128 66"/>
+  <path class="lv-flame" d="M 110 64 C 92 57 92 36 110 22 C 128 36 128 57 110 64 Z"/>
+  <path class="lv-fbord" d="M 110 64 C 92 57 92 36 110 22 C 128 36 128 57 110 64 Z"/>
+  <path class="lv-core" d="M 110 58 C 100 52 100 40 110 32 C 120 40 120 52 110 58 Z"/>
+  <g class="lv-code" fill="none" stroke="#c25a1a" stroke-width="1.6" stroke-linecap="round" stroke-linejoin="round">
+    <polyline points="106,42 103,46 106,50"/>
+    <line x1="112" y1="43" x2="109" y2="49"/>
+    <polyline points="115,42 118,46 115,50"/>
+  </g>
+  <text class="lv-word" x="113" y="93" text-anchor="middle">lian</text>
+  <text class="lv-sub" x="112" y="110" text-anchor="middle">build</text>
+</svg>
diff --git a/assets/presentation/run.sh b/assets/presentation/run.sh
index 6878963..469d92b 100755
--- a/assets/presentation/run.sh
+++ b/assets/presentation/run.sh
@@ -1,3 +1,5 @@
 #!/bin/bash
 
-pnpm run dev
+SLIDES_PATH=${1:-slides.md} 
+
+pnpm run dev $SLIDES_PATH
diff --git a/assets/presentation/work_group_info.md b/assets/presentation/work_group_info.md
new file mode 100644
index 0000000..7e793f2
--- /dev/null
+++ b/assets/presentation/work_group_info.md
@@ -0,0 +1,576 @@
+---
+theme: default
+title: Working Group — Ontología y Reflection
+titleTemplate: '%s - OntoRef'
+layout: cover
+keywords: Ontología,Reflection,Rust,Nickel,WorkingGroup,RustLasPalmas
+download: true
+exportFilename: ontoref-working-group
+monaco: true
+remoteAssets: true
+selectable: true
+record: true
+colorSchema: dark
+lineNumbers: false
+themeConfig:
+  primary: '#f74c00'
+  logoHeader: '/ferris.svg'
+fonts:
+  mono: 'Victor Mono'
+drawings:
+  enabled: true
+  persist: false
+  presenterOnly: false
+  syncAll: true
+scripts:
+  - setup/image-overlay.ts
+class: 'justify-center flex flex-cols'
+---
+
+<h1 class="font-medium text-3xl leading-tight">Working Group<br><span class="text-orange-400">Ontología y Reflection en tus proyectos</span></h1>
+
+<div class="flex justify-center mt-6 text-gray-400 text-lg italic">
+  Sesión de presentación · Rust-LasPalmas
+</div>
+
+<div class="flex justify-center mt-8">
+  <img class="w-40" src="/jesusperez_w.svg">
+</div>
+
+<style scoped>
+h1, h2, p { z-index: 10; }
+</style>
+
+<Footer />
+
+<!--
+Esto no es una charla. Es la sesión donde decidís si entráis a un grupo de trabajo cerrado.
+Saludo breve, agradecer la asistencia, y avisar: en 60 minutos saldrán con la decisión tomada.
+No hay seguimiento por correo, no hay "ya os mandaré info". Hoy entráis o no entráis.
+-->
+
+---
+layout: section
+---
+
+# Tres preguntas, antes de nada
+
+<!--
+Pasar al modo participativo desde el primer momento.
+No es retórica — pedir respuesta de verdad. Si nadie levanta la mano, esperar.
+El silencio funciona aquí.
+-->
+
+---
+layout: default
+---
+
+# Levantad la mano si esto os ha pasado
+
+<div class="mt-8 space-y-6">
+
+<v-click>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-2xl mt-1">1.</span>
+  <div class="text-gray-200 text-lg">
+    Postmortem donde la decisión que se violó <em>estaba escrita</em> en una ADR — pero nadie la recordaba.
+  </div>
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-2xl mt-1">2.</span>
+  <div class="text-gray-200 text-lg">
+    Onboarding de alguien nuevo (o tú mismo, después de tres meses) re-descubriendo <em>por qué</em> el código está así.
+  </div>
+</div>
+
+</v-click>
+
+<v-click>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-2xl mt-1">3.</span>
+  <div class="text-gray-200 text-lg">
+    Diagrama de arquitectura que <em>mintió</em> tres meses después de pintarlo.
+  </div>
+</div>
+
+</v-click>
+
+</div>
+
+<!--
+Click por click. Pausa entre cada uno.
+Si la sala levanta la mano en los tres, ya tienes audiencia. Si solo en uno, ese es el dolor a destacar.
+Reconocer en voz alta: "ya, lo conocemos todos."
+-->
+
+---
+layout: default
+---
+
+# El patrón
+
+<div class="mt-6 grid grid-cols-2 gap-8">
+
+<div>
+
+**Para el código exigimos**
+
+<div class="mt-4 space-y-3">
+
+<div class="flex items-start gap-2">
+  <span class="text-green-400 font-bold text-sm mt-1">✓</span>
+  <div class="text-gray-300">Tipos estrictos</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-green-400 font-bold text-sm mt-1">✓</span>
+  <div class="text-gray-300">Tests que validan invariantes</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-green-400 font-bold text-sm mt-1">✓</span>
+  <div class="text-gray-300">Compilador que rechaza incoherencias</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-green-400 font-bold text-sm mt-1">✓</span>
+  <div class="text-gray-300">Errores antes de llegar a producción</div>
+</div>
+
+</div>
+
+</div>
+
+<div>
+
+<v-click>
+
+**Para las decisiones aceptamos**
+
+<div class="mt-4 space-y-3">
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400 font-bold text-sm mt-1">✗</span>
+  <div class="text-gray-300">Markdown libre</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400 font-bold text-sm mt-1">✗</span>
+  <div class="text-gray-300">Diagramas que envejecen</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400 font-bold text-sm mt-1">✗</span>
+  <div class="text-gray-300">Convención y disciplina humana</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-red-400 font-bold text-sm mt-1">✗</span>
+  <div class="text-gray-300">Errores que aparecen en el postmortem</div>
+</div>
+
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<!--
+Esta es LA slide. La incoherencia entre las dos columnas es el insight completo.
+"Rust nos da tipos hasta el borde del binario. Después, texto libre."
+Aquí no se vende solución todavía.
+-->
+
+---
+layout: section
+---
+
+# ¿Y si las tratáramos igual?
+
+<div class="mt-8 text-gray-400 text-xl italic">
+  Tipos. Contratos. Validación al guardar, no en el postmortem.
+</div>
+
+<!--
+Pausa larga aquí.
+La pregunta queda en el aire. La siguiente slide es la respuesta práctica.
+-->
+
+---
+layout: default
+---
+
+# Lo que vamos a ver ahora
+
+<div class="mt-8 space-y-5 text-gray-300 text-lg">
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-xl mt-1">→</span>
+  <div>Un proyecto vacío</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-xl mt-1">→</span>
+  <div>Un axioma · una tensión · una decisión</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-xl mt-1">→</span>
+  <div>Un cambio que viola la decisión</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-xl mt-1">→</span>
+  <div>El typecheck rechazándolo</div>
+</div>
+
+</div>
+
+<div class="mt-12 text-center text-gray-500 text-sm italic">
+  8 líneas de código. Sin frameworks. Sin instalar nada raro.
+</div>
+
+<!--
+Slide-anchor para la mini-demo. Aquí salgo de las slides y voy a la terminal.
+~10 minutos de demo en vivo.
+La frase final ("8 líneas de código") es un compromiso público que la demo tiene que sostener.
+-->
+
+---
+layout: cover
+class: 'justify-center flex flex-cols'
+---
+
+<h1 class="font-medium text-3xl text-orange-400">Demo en vivo</h1>
+
+<div class="flex justify-center mt-6 text-gray-400 text-lg italic">
+  El terminal y un editor. Nada más.
+</div>
+
+<!--
+Salir de slidev. Terminal en pantalla compartida.
+Pre-cargado:
+  - directorio /tmp/demo-wg/ vacío
+  - dos buffers preparados pero no escritos
+  - ontoref CLI instalada y verificada
+  - nickel typecheck verificado
+Plan B si algo falla: tener un asciinema grabado que reproducir.
+Volver a las slides cuando suene la frase: "Esto que acabáis de ver es lo más pequeño que se puede tener."
+-->
+
+---
+layout: default
+---
+
+# Lo que acabáis de ver
+
+<div class="mt-6 grid grid-cols-2 gap-8">
+
+<div>
+
+**Lo construido en 8 minutos**
+
+<div class="mt-4 space-y-3 text-gray-300">
+
+<div class="flex items-start gap-2">
+  <span class="text-orange-400 font-bold text-sm mt-1">·</span>
+  <div>Un <code>core.ncl</code> con un axioma y una tensión</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-orange-400 font-bold text-sm mt-1">·</span>
+  <div>Una ADR como contrato vivo</div>
+</div>
+
+<div class="flex items-start gap-2">
+  <span class="text-orange-400 font-bold text-sm mt-1">·</span>
+  <div>Un cambio rechazado al guardar</div>
+</div>
+
+</div>
+
+</div>
+
+<div>
+
+<v-click>
+
+**Lo que esto se vuelve con tiempo**
+
+<div class="mt-4 text-gray-300 text-sm">
+Un proyecto real, después de meses:
+</div>
+
+<div class="mt-3 font-mono text-xs text-gray-400 bg-gray-900 p-3 rounded">
+.ontology/<br>
+├── core.ncl       <span class="text-gray-600">(20 nodos, 22 aristas)</span><br>
+├── state.ncl      <span class="text-gray-600">(5 dimensiones FSM)</span><br>
+├── gate.ncl       <span class="text-gray-600">(membranas)</span><br>
+└── manifest.ncl<br>
+adrs/              <span class="text-gray-600">(20 decisiones vivas)</span><br>
+reflection/modes/  <span class="text-gray-600">(rutinas tipadas)</span>
+</div>
+
+<div class="mt-3 text-gray-500 text-sm italic">
+La estructura es la misma. Solo crece el contenido.
+</div>
+
+</v-click>
+
+</div>
+
+</div>
+
+<!--
+Click reveal: lo que tienen ahora vs destino realista.
+NO es "tienes que llegar aquí". Es "esto crece a esto cuando le dedicas tiempo".
+Frase clave: "lo de antes y esto es el mismo schema. Lo importante es empezar."
+-->
+
+---
+layout: section
+---
+
+# La propuesta
+
+<div class="mt-6 text-gray-400 text-xl">
+  Working group cerrado. Cuatro o cinco personas.
+</div>
+
+<!--
+Cambio de tono. Ahora es la oferta concreta.
+-->
+
+---
+layout: default
+---
+
+# Cómo funciona
+
+<div class="mt-4 space-y-4 text-gray-300">
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-sm mt-1 w-32 flex-none">Formato</span>
+  <div>Presencial · coworking en Las Palmas</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-sm mt-1 w-32 flex-none">Tamaño</span>
+  <div>4–5 personas · cohorte cerrada</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-sm mt-1 w-32 flex-none">Sesiones</span>
+  <div>4 sesiones de ~2 horas · una cada 2 semanas</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-sm mt-1 w-32 flex-none">Entre sesiones</span>
+  <div>~1 hora de iteración sobre vuestro proyecto</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-sm mt-1 w-32 flex-none">Coste</span>
+  <div>Vuestro tiempo. Coworking compartido.</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-sm mt-1 w-32 flex-none">Después</span>
+  <div>Si funciona, abrimos un siguiente ciclo. No hay programa eterno.</div>
+</div>
+
+</div>
+
+<!--
+Honestidad sobre el compromiso real.
+"Cohorte cerrada" — no se incorpora gente a mitad. Razón: continuidad y lenguaje común.
+"4 sesiones" — no más. Si necesitan más, abren otro ciclo después.
+-->
+
+---
+layout: default
+---
+
+# Qué os lleváis
+
+<div class="mt-8 space-y-5 text-gray-300 text-lg">
+
+<div class="flex items-start gap-3">
+  <span class="text-green-400 font-bold text-xl mt-1">✓</span>
+  <div>Un <code>.ontology/</code> mínimo funcionando para un proyecto vuestro</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-green-400 font-bold text-xl mt-1">✓</span>
+  <div>Una ADR formalizada como constraint que protege ese proyecto</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-green-400 font-bold text-xl mt-1">✓</span>
+  <div>Una <code>reflection mode</code> propia para una rutina vuestra</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-green-400 font-bold text-xl mt-1">✓</span>
+  <div>Capacidad de seguir solos, sin depender del grupo</div>
+</div>
+
+</div>
+
+<div class="mt-12 text-gray-500 italic text-center">
+  No salís con un curso. Salís con código vuestro vivo.
+</div>
+
+<!--
+Outcome material. Cuatro cosas concretas, ninguna abstracta.
+La frase del cierre es la promesa.
+-->
+
+---
+layout: default
+---
+
+# Qué esto NO es
+
+<div class="mt-8 space-y-4 text-gray-300">
+
+<div class="flex items-start gap-3">
+  <span class="text-red-400 font-bold text-xl mt-1">✗</span>
+  <div>Un curso de Nickel desde cero</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-red-400 font-bold text-xl mt-1">✗</span>
+  <div>Una venta de OntoRef como producto</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-red-400 font-bold text-xl mt-1">✗</span>
+  <div>Soporte 24/7 entre sesiones</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-red-400 font-bold text-xl mt-1">✗</span>
+  <div>Asistencia opcional · si fallas dos sesiones, sales del ciclo</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-red-400 font-bold text-xl mt-1">✗</span>
+  <div>Una comunidad amplia · esto es trabajo concentrado, no networking</div>
+</div>
+
+</div>
+
+<!--
+Gestionar expectativas con honestidad.
+La línea de "fallas dos sesiones, sales del ciclo" parece dura — es deliberada.
+La continuidad es lo que define el grupo. Sin ella, no hay valor.
+Mejor decirlo aquí que en sesión 3 cuando alguien falló dos veces.
+-->
+
+---
+layout: default
+---
+
+# Lo que necesito de cada uno antes de la primera sesión
+
+<div class="mt-6 space-y-4 text-gray-300">
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-sm mt-1 w-8 flex-none">1.</span>
+  <div>Un proyecto vuestro · Rust o no, lo que tengáis vivo</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-sm mt-1 w-8 flex-none">2.</span>
+  <div>Nickel y la CLI de OntoRef instaladas (os paso la guía)</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-sm mt-1 w-8 flex-none">3.</span>
+  <div>Una decisión arquitectónica de ese proyecto que recordéis · cualquiera</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-orange-400 font-bold text-sm mt-1 w-8 flex-none">4.</span>
+  <div>Disposición a romper esa decisión en vivo y ver qué pasa</div>
+</div>
+
+</div>
+
+<div class="mt-10 text-gray-500 italic">
+  Si traéis esto, en sesión 1 ya estáis modelando.
+</div>
+
+<!--
+La preparación entre la decisión-de-entrar y la sesión 1 es lo que filtra a quien va en serio.
+Lista corta y concreta. Sin requisitos masivos.
+-->
+
+---
+layout: cover
+class: 'justify-center flex flex-cols'
+---
+
+<h1 class="font-medium text-3xl text-orange-400">¿Quién entra?</h1>
+
+<div class="flex justify-center mt-6 text-gray-300 text-lg">
+  Levantad la mano. No por correo después.
+</div>
+
+<div class="flex justify-center mt-12 text-gray-500 italic text-base">
+  Mejor tres comprometidos que cinco difusos.
+</div>
+
+<!--
+Esta es la slide del momento de decisión.
+NO seguir presentando después de esto. NO añadir "ya os pensáis si queréis".
+Pausa. Mirar a la sala. Esperar manos.
+Tomar nombres y mails ahí mismo. Crear el grupo de Telegram/Signal antes de salir.
+-->
+
+---
+layout: default
+---
+
+# Si no entras hoy
+
+<div class="mt-8 space-y-4 text-gray-300 text-lg">
+
+<div class="flex items-start gap-3">
+  <span class="text-gray-500 font-bold text-sm mt-1">·</span>
+  <div>El siguiente ciclo se abre cuando éste termine — sin fecha fijada todavía.</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-gray-500 font-bold text-sm mt-1">·</span>
+  <div>Si quieres aviso del próximo, déjalo dicho hoy.</div>
+</div>
+
+<div class="flex items-start gap-3">
+  <span class="text-gray-500 font-bold text-sm mt-1">·</span>
+  <div>El material del approach está abierto en Github · podéis explorar solos.</div>
+</div>
+
+</div>
+
+<div class="mt-12 text-center text-gray-400 italic">
+  Gracias por venir.
+</div>
+
+<Footer />
+
+<!--
+Slide para no dejar a nadie con la sensación de "rechazado".
+Quien no entró hoy tiene puerta abierta para el siguiente ciclo.
+Y el approach está públicamente disponible — la sesión les sirvió aunque no entren al WG.
+-->
diff --git a/assets/work-group-ore/README.md b/assets/work-group-ore/README.md
new file mode 100644
index 0000000..49787f5
--- /dev/null
+++ b/assets/work-group-ore/README.md
@@ -0,0 +1,33 @@
+# Working Group OntoRef — Material de Apoyo
+
+Material para la sesión de presentación + arranque del Working Group local
+(Rust-LasPalmas) sobre Ontología y Reflection.
+
+## Estructura
+
+```text
+work-group-ore/
+├── README.md                    Este archivo
+├── setup.md                     Guía de instalación para participantes
+└── demo/
+    ├── README.md                Cómo correr las demos
+    ├── script.md                Guion timing-by-timing de la mini-demo
+    ├── option-a-owner/          Demo principal: owner por servicio
+    ├── option-b-layers/         Demo alternativa: capas sin acople
+    └── recording/               Plan B — asciinema pre-grabada
+```
+
+## Slides asociadas
+
+`../presentation/work_group_info.md` — sesión de presentación en Slidev.
+
+## Flujo recomendado
+
+1. **Antes de la sesión presentación**: ensayar la demo en vivo con
+   `demo/option-a-owner/` siguiendo `demo/script.md`. Grabar la versión asciinema
+   como Plan B (`demo/recording/record.sh`).
+2. **Durante la sesión presentación**: usar las slides + cambiar a terminal en
+   la slide-anchor "Demo en vivo".
+3. **Tras la sesión presentación**: enviar `setup.md` a quienes se apuntaron.
+4. **Sesión 0 del WG**: cada participante llega con setup completo y un
+   proyecto suyo identificado.
diff --git a/assets/work-group-ore/demo/README.md b/assets/work-group-ore/demo/README.md
new file mode 100644
index 0000000..1c731c4
--- /dev/null
+++ b/assets/work-group-ore/demo/README.md
@@ -0,0 +1,59 @@
+# Mini-demo — Working Group OntoRef
+
+Dos opciones de mini-demo (~8-10 min) para la sesión de presentación. La
+demo se ejecuta en vivo desde una terminal cuando las slides llegan a la
+slide-anchor "Demo en vivo".
+
+## Cuál usar
+
+| Opción | Recomendada para | Complejidad |
+|---|---|---|
+| **A — Owner por servicio** | Audiencia mixta de developers (cualquier stack) | Baja, ~12 líneas |
+| **B — Capas sin acople** | Audiencia rustacean / backend con experiencia | Media, ~30 líneas |
+
+**Recomendación**: usar la opción A en la primera sesión. La opción B queda
+guardada por si en alguna repetición del WG la audiencia exige más
+sofisticación.
+
+## Estructura de cada opción
+
+Cada opción tiene tres archivos `.ncl` que representan los tres pasos de la
+demo:
+
+```text
+option-X/
+├── README.md                Descripción de la opción
+├── core.start.ncl           Paso 1: archivo mínimo sin contrato
+├── core.with-contract.ncl   Paso 2: contrato añadido, nickel export pasa
+└── core.broken.ncl          Paso 3: violación, nickel export falla
+```
+
+El presentador no copia estos archivos durante la demo — los va escribiendo
+en vivo. Los archivos sirven como:
+
+- Referencia de lo que tiene que terminar escribiendo
+- Plan B si se queda en blanco (`cp core.with-contract.ncl /tmp/demo-wg/core.ncl`)
+- Material de grabación asciinema para Plan B audiovisual
+
+## Ejecutar manualmente para ensayar
+
+```bash
+cd /tmp && rm -rf demo-wg && mkdir demo-wg && cd demo-wg
+
+# Paso 1
+cp /Users/Akasha/Development/ontoref/assets/work-group-ore/demo/option-a-owner/core.start.ncl ./core.ncl
+nickel export core.ncl    # exporta JSON
+
+# Paso 2
+cp /Users/Akasha/Development/ontoref/assets/work-group-ore/demo/option-a-owner/core.with-contract.ncl ./core.ncl
+nickel export core.ncl    # exporta JSON
+
+# Paso 3
+cp /Users/Akasha/Development/ontoref/assets/work-group-ore/demo/option-a-owner/core.broken.ncl ./core.ncl
+nickel export core.ncl    # falla con missing definition for 'owner'
+```
+
+## Plan B audiovisual
+
+Ver `recording/README.md` para la grabación con asciinema y cómo reproducirla
+si la demo en vivo falla.
diff --git a/assets/work-group-ore/demo/option-a-owner/README.md b/assets/work-group-ore/demo/option-a-owner/README.md
new file mode 100644
index 0000000..57db70a
--- /dev/null
+++ b/assets/work-group-ore/demo/option-a-owner/README.md
@@ -0,0 +1,38 @@
+# Opción A — Owner por servicio (recomendada)
+
+Demo principal de la sesión de presentación. Modela una decisión universal
+("cada servicio tiene un owner asignado") como contrato Nickel y muestra
+cómo el typecheck rechaza una violación de la misma.
+
+## Por qué esta opción
+
+- **Reconocible al instante** — cualquier developer ha vivido el postmortem
+  donde nadie sabe quién mantiene el servicio
+- **Modelable en pocas líneas** — el contrato es un record con dos campos
+- **Error de typecheck legible** — `missing field 'owner'` no necesita
+  explicación
+
+## Archivos
+
+| Archivo | Paso de la demo | Resultado de `nickel export` |
+|---|---|---|
+| `core.start.ncl` | Paso 2 — sin contrato | ✓ exporta JSON (pero no valida nada) |
+| `core.with-contract.ncl` | Paso 3 — contrato aplicado | ✓ exporta JSON (estructura validada) |
+| `core.broken.ncl` | Paso 4 — violación | ✗ error: `missing definition for 'owner'` |
+
+> **Por qué `nickel export` y no `nickel typecheck`**: los contratos
+> `| Array Service` se ejecutan durante la evaluación, no durante el chequeo
+> de tipos. `nickel typecheck core.broken.ncl` pasa silenciosamente porque
+> solo verifica tipos estáticos. `nickel export` evalúa el archivo entero y
+> aplica los contratos — eso es lo que activa el error.
+
+## Verificación rápida (ensayo)
+
+```bash
+cd /Users/Akasha/Development/ontoref/assets/work-group-ore/demo/option-a-owner
+nickel export core.start.ncl           # exporta JSON
+nickel export core.with-contract.ncl   # exporta JSON
+nickel export core.broken.ncl          # falla con error legible — esperado
+```
+
+Ver `../script.md` para el guion completo de la demo.
diff --git a/assets/work-group-ore/demo/option-a-owner/core.broken.ncl b/assets/work-group-ore/demo/option-a-owner/core.broken.ncl
new file mode 100644
index 0000000..117d9a3
--- /dev/null
+++ b/assets/work-group-ore/demo/option-a-owner/core.broken.ncl
@@ -0,0 +1,15 @@
+let Service = {
+  name | String,
+  owner | String,
+}
+in
+
+{
+  axiom = "Cada servicio en producción tiene un owner identificable",
+
+  services | Array Service = [
+    { name = "api", owner = "backend" },
+    { name = "frontend", owner = "frontend" },
+    { name = "metrics" },
+  ],
+}
diff --git a/assets/work-group-ore/demo/option-a-owner/core.start.ncl b/assets/work-group-ore/demo/option-a-owner/core.start.ncl
new file mode 100644
index 0000000..b79b1d7
--- /dev/null
+++ b/assets/work-group-ore/demo/option-a-owner/core.start.ncl
@@ -0,0 +1,8 @@
+{
+  axiom = "Cada servicio en producción tiene un owner identificable",
+
+  services = [
+    { name = "api", owner = "backend" },
+    { name = "frontend", owner = "frontend" },
+  ],
+}
diff --git a/assets/work-group-ore/demo/option-a-owner/core.with-contract.ncl b/assets/work-group-ore/demo/option-a-owner/core.with-contract.ncl
new file mode 100644
index 0000000..8b482a4
--- /dev/null
+++ b/assets/work-group-ore/demo/option-a-owner/core.with-contract.ncl
@@ -0,0 +1,14 @@
+let Service = {
+  name | String,
+  owner | String,
+}
+in
+
+{
+  axiom = "Cada servicio en producción tiene un owner identificable",
+
+  services | Array Service = [
+    { name = "api", owner = "backend" },
+    { name = "frontend", owner = "frontend" },
+  ],
+}
diff --git a/assets/work-group-ore/demo/option-b-layers/README.md b/assets/work-group-ore/demo/option-b-layers/README.md
new file mode 100644
index 0000000..623ea1a
--- /dev/null
+++ b/assets/work-group-ore/demo/option-b-layers/README.md
@@ -0,0 +1,68 @@
+# Opción B — Capas sin acople (alternativa rustacean)
+
+Demo alternativa, más sofisticada. Modela la regla de arquitectura hexagonal
+"el dominio no depende de infraestructura, la aplicación no depende de
+infraestructura" como contrato Nickel custom, y muestra el typecheck
+rechazando una dependencia ilegal.
+
+## Cuándo usar esta en vez de A
+
+- Audiencia con experiencia previa en arquitectura limpia / hexagonal
+- Cohorte que ya pasó por el WG y vuelve para repetir con tema más denso
+- Sesión más larga (15 min en vez de 8) donde da tiempo a explicar las capas
+
+**Para la primera sesión con audiencia mixta: usa la opción A.**
+
+## Archivos
+
+| Archivo | Paso de la demo | Resultado de `nickel export` |
+|---|---|---|
+| `core.start.ncl` | Paso 2 — sin contrato custom | ✓ exporta JSON (estructura básica) |
+| `core.with-contract.ncl` | Paso 3 — contrato `ValidLayerDependency` aplicado | ✓ exporta JSON (todas las dependencias son legales) |
+| `core.broken.ncl` | Paso 4 — dependencia ilegal | ✗ error: `Layer violation: user (Domain) cannot depend on db_pool (Infrastructure)` |
+
+> **Composición de contratos**: los dos contratos (`Dependency` para la forma
+> del record y `ValidLayerDependency` para la regla de capas) se aplican
+> encadenados con `| Array Dependency | Array ValidLayerDependency`. No se
+> mergean con `&` porque uno es record-shape y el otro es custom contract —
+> el `|` los aplica secuencialmente al mismo valor.
+
+## Concepto modelado
+
+Tres capas:
+
+- `'Domain` — lógica pura, no debe conocer infraestructura
+- `'Application` — orquesta dominio, no debe conocer infraestructura
+- `'Infrastructure` — adaptadores, puede depender de cualquier capa
+
+Reglas codificadas en `allowed_dependency`:
+
+| Source | Targets permitidos |
+|---|---|
+| `'Domain` | `'Domain` solamente |
+| `'Application` | `'Domain` y `'Application` |
+| `'Infrastructure` | cualquiera |
+
+## Verificación rápida (ensayo)
+
+```bash
+cd /Users/Akasha/Development/ontoref/assets/work-group-ore/demo/option-b-layers
+nickel export core.start.ncl           # exporta JSON
+nickel export core.with-contract.ncl   # exporta JSON
+nickel export core.broken.ncl          # falla — Layer violation: user (Domain) cannot depend on db_pool (Infrastructure)
+```
+
+## Adaptación del guion
+
+El guion en `../script.md` está escrito para opción A. Para opción B, los
+cambios:
+
+- **Paso 2**: tecleas la lista de dependencias sin contrato custom
+- **Paso 3**: tecleas el `let allowed_dependency = ...` y el
+  `ValidLayerDependency` con `std.contract.custom`
+- **Paso 4**: añades una dependencia `'Domain -> 'Infrastructure` y el typecheck
+  rechaza con el mensaje del custom contract
+- **Paso 5**: arreglarlo significa o cambiar la capa del módulo origen, o
+  refactorizar para introducir un puerto/adapter
+
+Tiempo: ~12-15 min en vez de 8-10.
diff --git a/assets/work-group-ore/demo/option-b-layers/core.broken.ncl b/assets/work-group-ore/demo/option-b-layers/core.broken.ncl
new file mode 100644
index 0000000..07b989d
--- /dev/null
+++ b/assets/work-group-ore/demo/option-b-layers/core.broken.ncl
@@ -0,0 +1,45 @@
+let Layer = [| 'Domain, 'Application, 'Infrastructure |]
+in
+
+let layer_to_string = fun l =>
+  l |> match {
+    'Domain => "Domain",
+    'Application => "Application",
+    'Infrastructure => "Infrastructure",
+  }
+in
+
+let allowed_dependency = fun source target =>
+  (source == 'Domain && target == 'Domain)
+  || (source == 'Application && (target == 'Domain || target == 'Application))
+  || (source == 'Infrastructure)
+in
+
+let Dependency = {
+  from_module | String,
+  from_layer | Layer,
+  to_module | String,
+  to_layer | Layer,
+}
+in
+
+let ValidLayerDependency = std.contract.custom (fun label =>
+  fun value =>
+    if allowed_dependency value.from_layer value.to_layer then
+      'Ok value
+    else
+      'Error {
+        message = "Layer violation: %{value.from_module} (%{layer_to_string value.from_layer}) cannot depend on %{value.to_module} (%{layer_to_string value.to_layer})",
+      }
+)
+in
+
+{
+  axiom = "El dominio no depende de infraestructura. La aplicación tampoco.",
+
+  dependencies | Array Dependency | Array ValidLayerDependency = [
+    { from_module = "user_service", from_layer = 'Application, to_module = "user", to_layer = 'Domain },
+    { from_module = "user_repository", from_layer = 'Infrastructure, to_module = "user", to_layer = 'Domain },
+    { from_module = "user", from_layer = 'Domain, to_module = "db_pool", to_layer = 'Infrastructure },
+  ],
+}
diff --git a/assets/work-group-ore/demo/option-b-layers/core.start.ncl b/assets/work-group-ore/demo/option-b-layers/core.start.ncl
new file mode 100644
index 0000000..3e757a5
--- /dev/null
+++ b/assets/work-group-ore/demo/option-b-layers/core.start.ncl
@@ -0,0 +1,8 @@
+{
+  axiom = "El dominio no depende de infraestructura. La aplicación tampoco.",
+
+  dependencies = [
+    { from_module = "user_service", from_layer = 'Application, to_module = "user", to_layer = 'Domain },
+    { from_module = "user_repository", from_layer = 'Infrastructure, to_module = "user", to_layer = 'Domain },
+  ],
+}
diff --git a/assets/work-group-ore/demo/option-b-layers/core.with-contract.ncl b/assets/work-group-ore/demo/option-b-layers/core.with-contract.ncl
new file mode 100644
index 0000000..b31a62a
--- /dev/null
+++ b/assets/work-group-ore/demo/option-b-layers/core.with-contract.ncl
@@ -0,0 +1,44 @@
+let Layer = [| 'Domain, 'Application, 'Infrastructure |]
+in
+
+let layer_to_string = fun l =>
+  l |> match {
+    'Domain => "Domain",
+    'Application => "Application",
+    'Infrastructure => "Infrastructure",
+  }
+in
+
+let allowed_dependency = fun source target =>
+  (source == 'Domain && target == 'Domain)
+  || (source == 'Application && (target == 'Domain || target == 'Application))
+  || (source == 'Infrastructure)
+in
+
+let Dependency = {
+  from_module | String,
+  from_layer | Layer,
+  to_module | String,
+  to_layer | Layer,
+}
+in
+
+let ValidLayerDependency = std.contract.custom (fun label =>
+  fun value =>
+    if allowed_dependency value.from_layer value.to_layer then
+      'Ok value
+    else
+      'Error {
+        message = "Layer violation: %{value.from_module} (%{layer_to_string value.from_layer}) cannot depend on %{value.to_module} (%{layer_to_string value.to_layer})",
+      }
+)
+in
+
+{
+  axiom = "El dominio no depende de infraestructura. La aplicación tampoco.",
+
+  dependencies | Array Dependency | Array ValidLayerDependency = [
+    { from_module = "user_service", from_layer = 'Application, to_module = "user", to_layer = 'Domain },
+    { from_module = "user_repository", from_layer = 'Infrastructure, to_module = "user", to_layer = 'Domain },
+  ],
+}
diff --git a/assets/work-group-ore/demo/recording/README.md b/assets/work-group-ore/demo/recording/README.md
new file mode 100644
index 0000000..2878527
--- /dev/null
+++ b/assets/work-group-ore/demo/recording/README.md
@@ -0,0 +1,50 @@
+# Plan B Audiovisual — Asciinema pre-grabada
+
+Si la demo en vivo falla durante la sesión (cuelgue de editor, problema con
+nickel, conexión de red rara), se reproduce esta grabación. La audiencia
+sigue viendo el approach funcionando — no se pierde el momento.
+
+## Grabar la demo
+
+Antes de la sesión, ejecutar:
+
+```bash
+cd /Users/Akasha/Development/ontoref/assets/work-group-ore/demo/recording
+./record.sh
+```
+
+El script:
+
+1. Crea un directorio temporal `/tmp/demo-wg-rec/`
+2. Inicia `asciinema rec demo.cast`
+3. Reproduce la demo paso a paso con timing visible
+4. Genera `demo.cast` en este directorio
+
+## Reproducir durante la sesión
+
+Si la demo en vivo se atasca:
+
+```bash
+asciinema play /Users/Akasha/Development/ontoref/assets/work-group-ore/demo/recording/demo.cast
+```
+
+Para reproducir a velocidad acelerada:
+
+```bash
+asciinema play -s 2 demo.cast
+```
+
+## Requisitos
+
+- `asciinema` instalado (`brew install asciinema` o `pip install asciinema`)
+- `nickel` 1.16.0+ accesible en `$PATH`
+
+## Verificar que la grabación funciona
+
+Después de grabar, reproducir desde el principio:
+
+```bash
+asciinema play demo.cast
+```
+
+Si se ve la demo entera reproduciéndose sin saltos raros, está lista.
diff --git a/assets/work-group-ore/demo/recording/record.sh b/assets/work-group-ore/demo/recording/record.sh
new file mode 100755
index 0000000..09a92f0
--- /dev/null
+++ b/assets/work-group-ore/demo/recording/record.sh
@@ -0,0 +1,107 @@
+#!/usr/bin/env bash
+# record.sh - Graba la demo "Owner por servicio" con asciinema.
+# Genera demo.cast reproducible como Plan B durante la sesión presentación.
+
+set -euo pipefail
+
+DEMO_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+SOURCE_DIR="${DEMO_DIR}/../option-a-owner"
+WORK_DIR="/tmp/demo-wg-rec"
+CAST_FILE="${DEMO_DIR}/demo.cast"
+
+if ! command -v asciinema >/dev/null 2>&1; then
+  echo "ERROR: asciinema no instalado. brew install asciinema" >&2
+  exit 1
+fi
+
+if ! command -v nickel >/dev/null 2>&1; then
+  echo "ERROR: nickel no instalado. brew install nickel-lang" >&2
+  exit 1
+fi
+
+cat <<EOF
+Va a grabar la demo en:
+  $CAST_FILE
+
+Pulsa enter para empezar la grabación. Después podrás reproducirla con:
+  asciinema play $CAST_FILE
+EOF
+read -r
+
+# Reset working directory
+rm -rf "$WORK_DIR"
+mkdir -p "$WORK_DIR"
+
+# Pequeño script Python que reproduce la demo dentro de asciinema
+PLAY_SCRIPT="$(mktemp)"
+cat > "$PLAY_SCRIPT" <<'PLAY'
+#!/usr/bin/env bash
+set -e
+
+CYAN='\033[1;36m'
+YELLOW='\033[1;33m'
+GREEN='\033[1;32m'
+RED='\033[1;31m'
+GRAY='\033[0;90m'
+NC='\033[0m'
+
+WORK_DIR="$1"
+SOURCE_DIR="$2"
+
+cd "$WORK_DIR"
+
+step() { echo -e "\n${YELLOW}━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━${NC}"; echo -e "${YELLOW}  $1${NC}"; echo -e "${YELLOW}━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━${NC}\n"; sleep 2; }
+say()  { echo -e "${GRAY}# $1${NC}"; sleep 2; }
+cmd()  { echo -e "${CYAN}\$ $1${NC}"; sleep 1; eval "$1"; sleep 2; }
+
+step "Paso 1 — Punto de partida"
+say "Esto es lo que cualquiera tiene antes de empezar."
+cmd "pwd"
+cmd "ls"
+
+step "Paso 2 — Escribir core.ncl mínimo"
+say "Defino lo que quiero que sea cierto. Sin contrato todavía."
+cmd "cat $SOURCE_DIR/core.start.ncl"
+cp "$SOURCE_DIR/core.start.ncl" core.ncl
+cmd "nickel export core.ncl"
+say "Exporta. Pero no valida nada todavía — la estructura es lo único garantizado."
+
+step "Paso 3 — Añadir el contrato"
+say "Defino qué es un Servicio. Esto es la ADR como código, no como markdown."
+cmd "cat $SOURCE_DIR/core.with-contract.ncl"
+cp "$SOURCE_DIR/core.with-contract.ncl" core.ncl
+cmd "nickel export core.ncl"
+say "Sigue pasando. Los servicios cumplen el contrato."
+
+step "Paso 4 — El rompimiento"
+say "Viernes por la tarde, prisa. Servicio nuevo de metrics, owner mañana."
+cmd "cat $SOURCE_DIR/core.broken.ncl"
+cp "$SOURCE_DIR/core.broken.ncl" core.ncl
+echo -e "${CYAN}\$ nickel export core.ncl${NC}"
+sleep 1
+nickel export core.ncl 2>&1 || true
+sleep 4
+say "El contrato no me dejó. No es un linter post-hoc — es validación al guardar."
+say "Si esto fuera un PR, el CI falla antes de pasar a review."
+
+step "Paso 5 — Arreglo y cierre"
+cp "$SOURCE_DIR/core.with-contract.ncl" core.ncl
+cmd "nickel export core.ncl"
+cmd "wc -l core.ncl"
+say "Doce líneas. Nickel solo. Sin daemon, sin frameworks."
+say "Esto se hace en cinco minutos en cualquier proyecto vuestro."
+
+sleep 3
+PLAY
+chmod +x "$PLAY_SCRIPT"
+
+# Run inside asciinema
+asciinema rec --command "$PLAY_SCRIPT '$WORK_DIR' '$SOURCE_DIR'" --title "OntoRef WG — mini-demo" "$CAST_FILE"
+
+rm -f "$PLAY_SCRIPT"
+rm -rf "$WORK_DIR"
+
+echo ""
+echo "Grabación completa: $CAST_FILE"
+echo "Para reproducir:"
+echo "  asciinema play $CAST_FILE"
diff --git a/assets/work-group-ore/demo/script.md b/assets/work-group-ore/demo/script.md
new file mode 100644
index 0000000..edc00bc
--- /dev/null
+++ b/assets/work-group-ore/demo/script.md
@@ -0,0 +1,235 @@
+# Guion mini-demo — Owner por servicio (Opción A)
+
+Tiempo total: 8-10 minutos. Lo que hay que teclear y lo que hay que decir,
+paso a paso. Para la opción B, el ritmo es similar pero con más densidad
+técnica — ver `option-b-layers/README.md`.
+
+## Pre-requisitos en la máquina del presentador
+
+- Terminal con tipografía visible desde la última fila del coworking
+- Editor con resaltado de sintaxis para Nickel (idealmente) o cualquier editor
+- Pre-cargado en una pestaña: el `core.ncl` final (`with-contract.ncl`) por si
+  hay bloqueo
+- `/tmp/demo-wg/` vacío antes de empezar
+
+## Pre-paso (cero, no contar como demo)
+
+Limpiar y posicionarse:
+
+```bash
+rm -rf /tmp/demo-wg && mkdir /tmp/demo-wg && cd /tmp/demo-wg
+clear
+```
+
+---
+
+## Paso 1 — directorio vacío (~30s)
+
+**Teclear:**
+
+```bash
+pwd
+ls
+```
+
+**Decir:**
+
+> Esto es lo que cualquiera de vosotros tiene antes de empezar. Sin framework,
+> sin scaffold, sin instalar nada. Vamos a montar el esqueleto del approach
+> con Nickel solo.
+
+---
+
+## Paso 2 — escribir `core.ncl` mínimo (~2 min)
+
+**Abrir editor:**
+
+```bash
+$EDITOR core.ncl
+```
+
+**Escribir** (despacio, comentando mientras se escribe):
+
+```nickel
+{
+  axiom = "Cada servicio en producción tiene un owner identificable",
+
+  services = [
+    { name = "api", owner = "backend" },
+    { name = "frontend", owner = "frontend" },
+  ],
+}
+```
+
+**Decir mientras tecleas:**
+
+> Voy a escribir lo que quiero que sea cierto en mi proyecto: que cada
+> servicio tenga un dueño identificable. Lo que pongo aquí es texto, todavía
+> no es contrato. Solo dice que la palabra `axiom` y la palabra `services`
+> existen.
+
+**Validar:**
+
+```bash
+nickel export core.ncl
+```
+
+**Esperar a ver:** el JSON evaluado, sin errores.
+
+**Decir:**
+
+> Vale, esto pasa. Pero ojo: aquí no he validado nada todavía. Es como tener
+> un YAML con sintaxis correcta. La estructura es lo único que se garantiza.
+> Si mañana añado un servicio sin owner, esto sigue exportando sin quejarse.
+> Vamos a arreglar eso.
+
+---
+
+## Paso 3 — añadir contrato (~3 min)
+
+**Editar para que quede así:**
+
+```nickel
+let Service = {
+  name | String,
+  owner | String,
+} in
+
+{
+  axiom = "Cada servicio en producción tiene un owner identificable",
+
+  services | Array Service = [
+    { name = "api", owner = "backend" },
+    { name = "frontend", owner = "frontend" },
+  ],
+}
+```
+
+**Decir mientras tecleas:**
+
+> Esto es el contrato. Defino qué es un Servicio: tiene un nombre y un owner,
+> ambos strings. Y le aplico el contrato a la lista — todos los elementos
+> tienen que cumplir la forma. Esto, en un proyecto real, es lo que en
+> markdown llamaríais una ADR. La diferencia es que aquí la ADR es código.
+
+**Validar:**
+
+```bash
+nickel export core.ncl
+```
+
+**Esperar a ver:** el JSON evaluado, sin errores.
+
+**Decir:**
+
+> Sigue pasando. Los servicios cumplen el contrato.
+
+---
+
+## Paso 4 — el rompimiento (~2 min)
+
+**Decir antes de tocar el código:**
+
+> Vale. Ahora viernes por la tarde, hay prisa, hay que meter un servicio nuevo
+> de métricas. Owner todavía no asignado, ya lo arreglo el lunes.
+
+**Editar para añadir el servicio sin owner:**
+
+```nickel
+  services | Array Service = [
+    { name = "api", owner = "backend" },
+    { name = "frontend", owner = "frontend" },
+    { name = "metrics" },
+  ],
+```
+
+**Validar:**
+
+```bash
+nickel export core.ncl
+```
+
+**Esperar a ver:** error que menciona `missing definition for 'owner'` y
+señala la línea del contrato y el record que viola.
+
+**LEER EL ERROR EN VOZ ALTA, despacio.**
+
+**Decir:**
+
+> El contrato no me dejó. No es un linter post-hoc, no es un test que tarda
+> cinco minutos en CI. Es la validación del archivo, ahora mismo, en mi
+> editor. Si esto fuera un PR, el CI falla antes de pasar a review. Tres
+> meses después, cuando metrics se rompa, no va a ser que nadie sepa quién
+> lo mantiene.
+
+**Pausa larga.** Dejar que la sala procese.
+
+---
+
+## Paso 5 — arreglar y cierre (~1.5 min)
+
+**Editar para añadir owner:**
+
+```nickel
+    { name = "metrics", owner = "platform" },
+```
+
+**Validar:**
+
+```bash
+nickel export core.ncl
+```
+
+**Esperar a ver:** el JSON con el servicio nuevo incluido.
+
+**Decir:**
+
+> Y arreglarlo es trivial. Pongo el owner, pasa. La fricción está en la
+> decisión, no en escribir el código.
+
+**Mostrar tamaño:**
+
+```bash
+wc -l core.ncl
+```
+
+**Esperar a ver:** ~12 líneas.
+
+**Decir:**
+
+> Doce líneas. Esto es Nickel solo. Sin daemon, sin CLI mía, sin proyecto
+> enorme. Lo de antes lo construyes en cinco minutos en cualquier proyecto
+> vuestro.
+
+---
+
+## Volver a slides
+
+Cierra el editor (no la terminal — quédate en `/tmp/demo-wg`). Cambia a
+slidev. La siguiente slide hace el contraste con un `.ontology/` ya crecido.
+
+## Plan B si algo se cuelga
+
+| Síntoma | Acción |
+|---|---|
+| Editor crashea al escribir | `cp` el archivo del paso correspondiente: `cp /Users/Akasha/Development/ontoref/assets/work-group-ore/demo/option-a-owner/core.with-contract.ncl ./core.ncl` |
+| `nickel` no responde | Plan B audiovisual: `asciinema play /Users/Akasha/Development/ontoref/assets/work-group-ore/demo/recording/demo.cast` |
+| Te quedas en blanco mid-demo | Pausa, di "voy a copiar la versión completa para no perder tiempo", `cp core.with-contract.ncl ./core.ncl`, sigue desde paso 4 |
+| Pregunta inesperada que para la demo | Apunta la pregunta en pizarra, di "lo retomamos después de la demo", sigue |
+
+> **Nota técnica**: usamos `nickel export` (no `nickel typecheck`) porque los
+> contratos `| Array Service` y `std.contract.custom` se evalúan en runtime,
+> no durante el chequeo de tipos estáticos. `nickel export` evalúa la
+> expresión a JSON y dispara los contratos en el camino — si alguno falla, el
+> error sale por `stderr`. Es el comando que mejor refleja "qué pasa cuando
+> tu CI valida el archivo".
+
+## Errores comunes durante la demo
+
+- **No esperar a que la sala lea el error**: la pausa larga después del fallo
+  del typecheck es donde aterriza el insight. Cuenta hasta cinco mentalmente.
+- **Hablar más rápido cuando uno está nervioso**: el tempo de la demo es lento
+  a propósito. Si te oyes acelerando, respira.
+- **Disculparse por la simpleza**: NO lo hagas. La simpleza ES el argumento.
+  Si dices "esto es muy simple, en producción es más complicado", pierdes la
+  fuerza del ejemplo.
diff --git a/assets/work-group-ore/setup.md b/assets/work-group-ore/setup.md
new file mode 100644
index 0000000..80ebc1b
--- /dev/null
+++ b/assets/work-group-ore/setup.md
@@ -0,0 +1,125 @@
+# Setup — Working Group OntoRef
+
+Lo que tenéis que tener instalado y preparado **antes** de la sesión 1 del
+working group. Tiempo estimado: 20-30 minutos si todo va bien, 1 hora si toca
+debuggear instalación.
+
+## Requisitos previos
+
+- macOS o Linux (ambos probados; Windows no soportado en este ciclo)
+- Acceso a terminal y editor de texto cualquiera (vim, neovim, vscode, zed,
+  helix, lo que uséis)
+- Conexión a internet en la sesión (para clonar repos)
+
+## 1. Instalar Nickel
+
+Nickel es el lenguaje de contratos. Se instala como un binario único.
+
+### macOS
+
+```bash
+brew install nickel-lang
+```
+
+Si no usáis brew:
+
+```bash
+curl -L https://github.com/tweag/nickel/releases/latest/download/nickel-x86_64-apple-darwin --output ~/.local/bin/nickel
+chmod +x ~/.local/bin/nickel
+```
+
+### Linux
+
+```bash
+curl -L https://github.com/tweag/nickel/releases/latest/download/nickel-x86_64-unknown-linux-gnu --output ~/.local/bin/nickel
+chmod +x ~/.local/bin/nickel
+```
+
+(Asegúrate de que `~/.local/bin` está en tu `$PATH`.)
+
+### Verificar
+
+```bash
+nickel --version
+```
+
+Debe mostrar `nickel 1.16.0` o superior.
+
+## 2. Instalar la CLI de OntoRef
+
+OntoRef es la herramienta que opera sobre `.ontology/` y se construye con
+`cargo`. Necesitáis tener Rust instalado (`rustup` si no lo tenéis).
+
+```bash
+git clone https://github.com/jesusperez-pro/ontoref.git ~/dev/ontoref
+cd ~/dev/ontoref
+cargo install --path . --bin ontoref
+```
+
+(El path del repo y la URL son orientativos — si Jesús os da uno distinto,
+usad el que os pase.)
+
+### Verificar
+
+```bash
+ontoref --version
+ontoref describe --help
+```
+
+Si `ontoref describe --help` devuelve la lista de subcomandos, está listo.
+
+## 3. Preparar lo que traéis a la sesión
+
+### Un proyecto vuestro
+
+Cualquier proyecto que tengáis vivo y conozcáis bien. **Rust o no, da igual.**
+Lo que importa es que sepáis hablar de él. Si tenéis varios, traed UNO — el
+que más os duela en términos de coherencia perdida.
+
+### Una decisión arquitectónica recordada
+
+Algo que decidisteis en su momento y que sigue vivo en el código:
+
+- "Usamos PostgreSQL en vez de SQLite porque…"
+- "Todas las APIs públicas tienen rate-limiting porque…"
+- "El módulo de auth no importa de billing porque…"
+- "Cualquier campo público de la API tiene que estar versionado porque…"
+
+No necesita ser sofisticada. Necesita ser **vuestra** y **viva**.
+
+### Disposición a romperla
+
+Vamos a modelar esa decisión como contrato y luego vamos a intentar violarla
+en vivo para ver el typecheck rechazándola. Si sois posesivos con vuestras
+decisiones, el ejercicio se hace cuesta arriba.
+
+## 4. Comprobaciones finales (24h antes)
+
+Ejecutad esto en una terminal limpia:
+
+```bash
+nickel --version       # debe responder
+ontoref --version      # debe responder
+mkdir -p ~/tmp/wg-test && cd ~/tmp/wg-test
+echo '{ x = 1 }' > test.ncl
+nickel export test.ncl    # debe imprimir {"x": 1}
+```
+
+Si los cuatro comandos pasan, estáis listos.
+
+## Si algo falla
+
+- **Antes de la sesión**: avisad por el grupo de Telegram. Es mejor 30 minutos
+  de troubleshooting compartido que perder 30 minutos de la sesión.
+- **Durante la sesión**: pareja con quien sí tiene el setup hasta que se
+  resuelva. La sesión no espera, pero tampoco se deja a nadie atrás.
+
+## Recursos opcionales (no necesarios para la sesión 1)
+
+- [Nickel Documentation](https://nickel-lang.org/) — para hojearlo si queréis
+  contexto
+- [OntoRef README](https://github.com/jesusperez-pro/ontoref) — la herramienta
+  por dentro
+
+No hace falta leerlos antes. Lo que se necesita aprender, se aprende en
+sesión.
diff --git a/crates/ontoref-daemon/Cargo.toml b/crates/ontoref-daemon/Cargo.toml
index f606afb..db41dab 100644
--- a/crates/ontoref-daemon/Cargo.toml
+++ b/crates/ontoref-daemon/Cargo.toml
@@ -28,6 +28,8 @@ uuid        = { workspace = true, optional = true }
 axum-server = { version = "0.7", features = ["tls-rustls"], optional = true }
 rmcp        = { version = "1", features = ["server", "transport-io", "transport-streamable-http-server"], optional = true }
 schemars    = { version = "1", optional = true }
+async-graphql     = { version = "=8.0.0-rc.5", optional = true }
+async-graphql-axum = { version = "=8.0.0-rc.5", optional = true }
 thiserror = { workspace = true }
 dashmap = { workspace = true }
 clap = { workspace = true }
@@ -44,12 +46,13 @@ ontoref-ontology = { path = "../ontoref-ontology", features = ["derive"] }
 libc = { workspace = true }
 
 [features]
-default = ["db", "nats", "ui", "mcp"]
+default = ["db", "nats", "ui", "mcp", "graphql"]
 db      = ["stratum-db/remote"]
 nats    = ["dep:platform-nats"]
 ui      = ["dep:tera", "dep:toml", "dep:uuid"]
 tls     = ["ui", "dep:axum-server"]
 mcp     = ["ui", "dep:rmcp", "dep:schemars"]
+graphql = ["dep:async-graphql", "dep:async-graphql-axum", "axum/ws"]
 
 [dev-dependencies]
 tokio-test = { workspace = true }
diff --git a/crates/ontoref-daemon/src/api.rs b/crates/ontoref-daemon/src/api.rs
index 06259a0..215bdd5 100644
--- a/crates/ontoref-daemon/src/api.rs
+++ b/crates/ontoref-daemon/src/api.rs
@@ -1,5 +1,7 @@
 use std::net::IpAddr;
 use std::path::PathBuf;
+#[cfg(any(feature = "mcp", feature = "graphql"))]
+use std::sync::atomic::AtomicBool;
 use std::sync::atomic::{AtomicU64, Ordering};
 use std::sync::Arc;
 use std::time::{Instant, SystemTime, UNIX_EPOCH};
@@ -92,6 +94,75 @@ fn is_uuid_v4(s: &str) -> bool {
             .all(|(i, &c)| matches!(i, 8 | 13 | 18 | 23) || c.is_ascii_hexdigit())
 }
 
+/// Runtime on/off flags for optional services compiled into the binary.
+///
+/// Toggles are `AtomicBool` so middleware reads are a single load instruction
+/// with no lock contention on the hot path.
+pub struct ServiceFlags {
+    #[cfg(feature = "mcp")]
+    pub mcp: AtomicBool,
+    #[cfg(feature = "graphql")]
+    pub graphql: AtomicBool,
+}
+
+impl ServiceFlags {
+    pub fn new() -> Arc<Self> {
+        Arc::new(Self {
+            #[cfg(feature = "mcp")]
+            mcp: AtomicBool::new(true),
+            #[cfg(feature = "graphql")]
+            graphql: AtomicBool::new(true),
+        })
+    }
+
+    /// Returns `true` if the named service is enabled (or not compiled in).
+    pub fn is_enabled(&self, service: &str) -> bool {
+        match service {
+            #[cfg(feature = "mcp")]
+            "mcp" => self.mcp.load(Ordering::Relaxed),
+            #[cfg(feature = "graphql")]
+            "graphql" => self.graphql.load(Ordering::Relaxed),
+            _ => false,
+        }
+    }
+
+    /// Toggle a named service. Returns the new state, or `None` if the service
+    /// is unknown or not compiled in.
+    #[cfg_attr(
+        not(any(feature = "mcp", feature = "graphql")),
+        allow(unused_variables)
+    )]
+    pub fn set(&self, service: &str, enabled: bool) -> Option<bool> {
+        match service {
+            #[cfg(feature = "mcp")]
+            "mcp" => {
+                self.mcp.store(enabled, Ordering::Relaxed);
+                Some(enabled)
+            }
+            #[cfg(feature = "graphql")]
+            "graphql" => {
+                self.graphql.store(enabled, Ordering::Relaxed);
+                Some(enabled)
+            }
+            _ => None,
+        }
+    }
+}
+
+// Cannot be derived: AtomicBool::default() is false, but services start
+// enabled.
+#[allow(clippy::derivable_impls)]
+impl Default for ServiceFlags {
+    fn default() -> Self {
+        Self {
+            #[cfg(feature = "mcp")]
+            mcp: AtomicBool::new(true),
+            #[cfg(feature = "graphql")]
+            graphql: AtomicBool::new(true),
+        }
+    }
+}
+
 /// Shared application state injected into handlers.
 #[derive(Clone)]
 pub struct AppState {
@@ -143,9 +214,36 @@ pub struct AppState {
     /// accepts this password and creates a "_daemon" session with Role::Admin.
     #[cfg(feature = "ui")]
     pub daemon_admin_hash: Option<String>,
+    /// Argon2id PHC hash of the GraphQL read token from
+    /// `config.graphql.read_token_hash`. `None` means GraphQL endpoints
+    /// require no dedicated token.
+    #[cfg(feature = "graphql")]
+    pub graphql_read_token_hash: Option<String>,
+    /// Runtime enable/disable flags for optional services (MCP, GraphQL).
+    /// Shared across all clones of AppState — toggling affects all connections.
+    pub service_flags: Arc<ServiceFlags>,
 }
 
 impl AppState {
+    /// Verify a raw token against `graphql_read_token_hash`.
+    ///
+    /// Returns `true` when:
+    /// - no hash is configured (`None`) — open access, or
+    /// - the token matches the configured argon2id hash.
+    #[cfg(feature = "graphql")]
+    pub fn verify_graphql_token(&self, token: &str) -> bool {
+        let Some(ref hash) = self.graphql_read_token_hash else {
+            return true;
+        };
+        use crate::registry::{verify_keys_list, KeyEntry, Role};
+        let synthetic = vec![KeyEntry {
+            role: Role::Viewer,
+            hash: hash.clone(),
+            label: "graphql".into(),
+        }];
+        verify_keys_list(&synthetic, token).is_some()
+    }
+
     fn touch_activity(&self) {
         let now = SystemTime::now()
             .duration_since(UNIX_EPOCH)
@@ -367,6 +465,11 @@ pub fn router(state: AppState) -> axum::Router {
             put(project_config_update),
         );
 
+    // Service toggle endpoints — daemon admin protected.
+    let app = app
+        .route("/services", get(services_status))
+        .route("/services/{service}", put(service_toggle));
+
     // Session endpoints — gated on ui feature (requires SessionStore).
     #[cfg(feature = "ui")]
     let app = app
@@ -407,18 +510,54 @@ pub fn router(state: AppState) -> axum::Router {
     };
 
     // MCP streamable-HTTP endpoint — stateless per-request factory.
+    // Toggle middleware checks service_flags.mcp before forwarding.
     #[cfg(feature = "mcp")]
     let app = {
         use rmcp::transport::streamable_http_server::{
             session::local::LocalSessionManager, StreamableHttpServerConfig, StreamableHttpService,
         };
+        let mcp_flags = Arc::clone(&state.service_flags);
         let mcp_state = state.clone();
         let mcp_svc = StreamableHttpService::new(
             move || Ok(crate::mcp::OntoreServer::new(mcp_state.clone())),
             std::sync::Arc::new(LocalSessionManager::default()),
             StreamableHttpServerConfig::default(),
         );
-        app.nest_service("/mcp", mcp_svc)
+        let mcp_router =
+            axum::Router::new()
+                .fallback_service(mcp_svc)
+                .layer(axum::middleware::from_fn(
+                    move |req: axum::extract::Request, next: axum::middleware::Next| {
+                        let flags = Arc::clone(&mcp_flags);
+                        async move {
+                            if !flags.mcp.load(Ordering::Relaxed) {
+                                return StatusCode::SERVICE_UNAVAILABLE.into_response();
+                            }
+                            next.run(req).await
+                        }
+                    },
+                ));
+        app.nest("/mcp", mcp_router)
+    };
+
+    // GraphQL endpoint — schema built once with AppState; GraphiQL at GET /graphql.
+    // Toggle middleware checks service_flags.graphql before forwarding.
+    #[cfg(feature = "graphql")]
+    let app = {
+        let gql_flags = Arc::clone(&state.service_flags);
+        let gql_router =
+            crate::graphql::graphql_router(&state).route_layer(axum::middleware::from_fn(
+                move |req: axum::extract::Request, next: axum::middleware::Next| {
+                    let flags = Arc::clone(&gql_flags);
+                    async move {
+                        if !flags.graphql.load(Ordering::Relaxed) {
+                            return StatusCode::SERVICE_UNAVAILABLE.into_response();
+                        }
+                        next.run(req).await
+                    }
+                },
+            ));
+        app.nest("/graphql", gql_router)
     };
 
     app
@@ -3504,6 +3643,130 @@ impl IntoResponse for crate::error::DaemonError {
     }
 }
 
+// ── Service toggles ──────────────────────────────────────────────────────────
+
+#[derive(Serialize)]
+struct ServiceEntry {
+    service: &'static str,
+    compiled: bool,
+    enabled: bool,
+}
+
+/// List all optional services with their compiled-in and runtime-enabled
+/// status.
+async fn services_status(State(state): State<AppState>) -> impl IntoResponse {
+    state.touch_activity();
+    let services = vec![
+        ServiceEntry {
+            service: "mcp",
+            #[cfg(feature = "mcp")]
+            compiled: true,
+            #[cfg(not(feature = "mcp"))]
+            compiled: false,
+            #[cfg(feature = "mcp")]
+            enabled: state.service_flags.mcp.load(Ordering::Relaxed),
+            #[cfg(not(feature = "mcp"))]
+            enabled: false,
+        },
+        ServiceEntry {
+            service: "graphql",
+            #[cfg(feature = "graphql")]
+            compiled: true,
+            #[cfg(not(feature = "graphql"))]
+            compiled: false,
+            #[cfg(feature = "graphql")]
+            enabled: state.service_flags.graphql.load(Ordering::Relaxed),
+            #[cfg(not(feature = "graphql"))]
+            enabled: false,
+        },
+    ];
+    Json(services)
+}
+
+#[derive(Deserialize)]
+struct ServiceToggleRequest {
+    // used in the #[cfg(feature = "ui")] handler; appears dead without that feature
+    #[cfg_attr(not(feature = "ui"), allow(dead_code))]
+    enabled: bool,
+}
+
+/// Enable or disable a named optional service at runtime. Daemon admin Bearer
+/// required.
+///
+/// Gated on `ui` because daemon admin auth (`daemon_admin_hash`) requires the
+/// session store compiled in with the `ui` feature. Without `ui`, the endpoint
+/// returns 501 — no auth surface exists to protect the toggle.
+#[cfg(feature = "ui")]
+async fn service_toggle(
+    State(state): State<AppState>,
+    headers: axum::http::HeaderMap,
+    Path(service): Path<String>,
+    Json(req): Json<ServiceToggleRequest>,
+) -> impl IntoResponse {
+    state.touch_activity();
+
+    let Some(ref admin_hash) = state.daemon_admin_hash else {
+        return (
+            StatusCode::UNAUTHORIZED,
+            Json(serde_json::json!({"error": "daemon admin auth not configured"})),
+        )
+            .into_response();
+    };
+    let bearer = headers
+        .get(axum::http::header::AUTHORIZATION)
+        .and_then(|v| v.to_str().ok())
+        .and_then(|v| v.strip_prefix("Bearer "))
+        .map(str::trim)
+        .filter(|s| !s.is_empty());
+    let Some(token) = bearer else {
+        return (
+            StatusCode::UNAUTHORIZED,
+            Json(serde_json::json!({"error": "Authorization: Bearer <admin-token> required"})),
+        )
+            .into_response();
+    };
+    let key_entry = crate::registry::KeyEntry {
+        role: crate::registry::Role::Admin,
+        hash: admin_hash.clone(),
+        label: "daemon-admin".to_string(),
+    };
+    if crate::registry::verify_keys_list(std::slice::from_ref(&key_entry), token).is_none() {
+        return (
+            StatusCode::UNAUTHORIZED,
+            Json(serde_json::json!({"error": "invalid admin credentials"})),
+        )
+            .into_response();
+    }
+
+    match state.service_flags.set(&service, req.enabled) {
+        Some(new_state) => {
+            tracing::info!(service = %service, enabled = new_state, "service toggled");
+            Json(serde_json::json!({"service": service, "enabled": new_state})).into_response()
+        }
+        None => (
+            StatusCode::NOT_FOUND,
+            Json(
+                serde_json::json!({"error": format!("unknown or not compiled service: {service}")}),
+            ),
+        )
+            .into_response(),
+    }
+}
+
+#[cfg(not(feature = "ui"))]
+async fn service_toggle(
+    State(_state): State<AppState>,
+    Path(_service): Path<String>,
+    Json(_req): Json<ServiceToggleRequest>,
+) -> impl IntoResponse {
+    (
+        StatusCode::NOT_IMPLEMENTED,
+        Json(
+            serde_json::json!({"error": "service toggle requires the ui feature (daemon admin auth)"}),
+        ),
+    )
+}
+
 #[cfg(test)]
 mod tests {
     use std::sync::atomic::AtomicU64;
@@ -3575,6 +3838,9 @@ mod tests {
             ncl_write_lock: Arc::new(crate::ui::ncl_write::NclWriteLock::new()),
             #[cfg(feature = "ui")]
             daemon_admin_hash: None,
+            #[cfg(feature = "graphql")]
+            graphql_read_token_hash: None,
+            service_flags: ServiceFlags::new(),
         }
     }
 
diff --git a/crates/ontoref-daemon/src/config.rs b/crates/ontoref-daemon/src/config.rs
index e9b502a..9c7332f 100644
--- a/crates/ontoref-daemon/src/config.rs
+++ b/crates/ontoref-daemon/src/config.rs
@@ -33,6 +33,9 @@ pub struct DaemonNclConfig {
     #[cfg(feature = "db")]
     #[serde(default)]
     pub db: DbConfig,
+    #[cfg(feature = "graphql")]
+    #[serde(default)]
+    pub graphql: GraphqlConfig,
 }
 
 /// `ui` section — template and asset paths, optional TLS cert overrides.
@@ -202,6 +205,19 @@ pub struct DaemonRuntimeConfig {
     pub notification_ack_required: Vec<String>,
 }
 
+/// `graphql` section — GraphQL endpoint access control (feature-gated).
+///
+/// `read_token_hash` is an argon2id PHC string produced by
+/// `ontoref-daemon --hash-password <token>`. Empty string disables auth.
+#[cfg(feature = "graphql")]
+#[derive(Debug, Deserialize, Default, ConfigFields)]
+#[config_section(id = "graphql", ncl_file = ".ontoref/config.ncl")]
+pub struct GraphqlConfig {
+    /// Argon2id PHC hash of the read token. Empty = no auth required.
+    #[serde(default)]
+    pub read_token_hash: String,
+}
+
 /// `db` section — SurrealDB connection (feature-gated).
 #[cfg(feature = "db")]
 #[derive(Debug, Deserialize, Default, ConfigFields)]
diff --git a/crates/ontoref-daemon/src/graphql/mod.rs b/crates/ontoref-daemon/src/graphql/mod.rs
new file mode 100644
index 0000000..22891b1
--- /dev/null
+++ b/crates/ontoref-daemon/src/graphql/mod.rs
@@ -0,0 +1,1097 @@
+/// GraphQL surface over the ontology graph, impact traversal, project state,
+/// backlog, ADRs, cross-project config, and notifications — with Apollo
+/// Federation v2 entity resolution and WebSocket subscriptions.
+use std::net::SocketAddr;
+use std::sync::Arc;
+
+use async_graphql::{
+    futures_util::{self, StreamExt as _},
+    http::ALL_WEBSOCKET_PROTOCOLS,
+    Context, Enum, Object, Schema, SimpleObject, Subscription,
+};
+use async_graphql_axum::{GraphQLProtocol, GraphQLRequest, GraphQLResponse, GraphQLWebSocket};
+use axum::{
+    extract::{ConnectInfo, Query, State, WebSocketUpgrade},
+    http::{header::AUTHORIZATION, HeaderMap, HeaderValue},
+    response::{Html, IntoResponse},
+    routing::get,
+    Extension, Router,
+};
+use tokio_stream::wrappers::BroadcastStream;
+
+use crate::{
+    api::{extract_bearer, AppState},
+    cache::NclCache,
+    federation::FederatedQuery,
+    notifications::{NotificationEvent, NotificationView},
+};
+
+pub type OntoSchema = Schema<QueryRoot, MutationRoot, SubscriptionRoot>;
+
+// ── Output types ─────────────────────────────────────────────────────────────
+
+#[derive(SimpleObject, Clone)]
+pub struct OntologyNode {
+    pub slug: String,
+    pub id: String,
+    pub name: String,
+    pub kind: String,
+    pub description: Option<String>,
+}
+
+#[derive(SimpleObject)]
+pub struct ImpactEntry {
+    pub slug: String,
+    pub node_id: String,
+    pub node_name: String,
+    pub depth: i32,
+    pub direction: String,
+    pub via: String,
+}
+
+#[derive(SimpleObject)]
+pub struct ProjectIdentity {
+    pub slug: String,
+    pub name: String,
+    pub kind: String,
+    pub description: Option<String>,
+}
+
+#[derive(SimpleObject)]
+pub struct BacklogItem {
+    pub id: String,
+    pub title: String,
+    pub kind: String,
+    pub priority: String,
+    pub status: String,
+}
+
+#[derive(SimpleObject)]
+pub struct AdrSummary {
+    pub id: String,
+    pub slug: String,
+    pub title: String,
+    pub status: String,
+    pub date: Option<String>,
+}
+
+#[derive(SimpleObject)]
+pub struct AdrRecord {
+    pub id: String,
+    pub slug: String,
+    pub title: String,
+    pub status: String,
+    pub date: Option<String>,
+    pub context: Option<String>,
+    pub decision: Option<String>,
+    pub consequences: Option<String>,
+}
+
+#[derive(SimpleObject)]
+pub struct CrossProjectSummary {
+    pub slug: String,
+    pub kind: String,
+    pub export_ok: bool,
+    pub section_count: i32,
+}
+
+#[derive(SimpleObject)]
+pub struct CrossProjectConflict {
+    pub section: String,
+    pub field: String,
+    pub project_a: String,
+    pub value_a: String,
+    pub project_b: String,
+    pub value_b: String,
+}
+
+#[derive(Enum, Clone, Copy, PartialEq, Eq)]
+pub enum NotificationEventGql {
+    OntologyChanged,
+    AdrChanged,
+    ReflectionChanged,
+    Custom,
+}
+
+impl From<NotificationEvent> for NotificationEventGql {
+    fn from(e: NotificationEvent) -> Self {
+        match e {
+            NotificationEvent::OntologyChanged => Self::OntologyChanged,
+            NotificationEvent::AdrChanged => Self::AdrChanged,
+            NotificationEvent::ReflectionChanged => Self::ReflectionChanged,
+            NotificationEvent::Custom => Self::Custom,
+        }
+    }
+}
+
+#[derive(SimpleObject, Clone)]
+pub struct NotificationGql {
+    pub id: i64,
+    pub project: String,
+    pub event: NotificationEventGql,
+    pub files: Vec<String>,
+    pub timestamp: i64,
+    pub source_actor: Option<String>,
+    pub custom_kind: Option<String>,
+    pub custom_title: Option<String>,
+}
+
+impl From<NotificationView> for NotificationGql {
+    fn from(n: NotificationView) -> Self {
+        Self {
+            id: n.id as i64,
+            project: n.project,
+            event: n.event.into(),
+            files: n.files,
+            timestamp: n.timestamp as i64,
+            source_actor: n.source_actor,
+            custom_kind: n.custom_kind,
+            custom_title: n.custom_title,
+        }
+    }
+}
+
+// ── Registry topology types ──────────────────────────────────────────────────
+
+#[derive(SimpleObject)]
+pub struct RegistryNamespacesGql {
+    pub own: Vec<String>,
+    pub replicate_to: Vec<String>,
+    pub mirror_from: Option<String>,
+    pub prefixes: Vec<String>,
+}
+
+#[derive(SimpleObject)]
+pub struct RegistryEntryGql {
+    pub id: String,
+    pub endpoint: String,
+    pub role: String,
+    pub tls: bool,
+    pub namespaces: RegistryNamespacesGql,
+}
+
+#[derive(SimpleObject)]
+pub struct RegistryTopologyGql {
+    pub participant: String,
+    pub default_id: Option<String>,
+    pub entries: Vec<RegistryEntryGql>,
+    /// Derived participant namespaces: ["domains/<participant>/",
+    /// "modes/<participant>/"]
+    pub namespaces: Vec<String>,
+}
+
+// ── Vault status types (ADR-017 + ADR-019) ──────────────────────────────────
+
+#[derive(SimpleObject)]
+pub struct RecipientGroupGql {
+    pub name: String,
+    pub keys: Vec<String>,
+}
+
+#[derive(SimpleObject)]
+pub struct RecipientRuleGql {
+    pub path: String,
+    pub groups: Vec<String>,
+}
+
+#[derive(SimpleObject)]
+pub struct ActorBindingGql {
+    pub actor: String,
+    pub role: String,
+}
+
+#[derive(SimpleObject)]
+pub struct VaultLastOpGql {
+    pub ts: String,
+    pub actor: String,
+    pub op: String,
+    pub registry_id: Option<String>,
+    pub detail: Option<String>,
+}
+
+#[derive(SimpleObject)]
+pub struct VaultStatusGql {
+    pub vault_id: String,
+    pub vault_dir: String,
+    pub backend: String,
+    pub registry_endpoint: Option<String>,
+    pub master_key: Option<String>,
+    pub access_sops_present: bool,
+    pub recipient_count: i32,
+    pub recipient_groups: Vec<RecipientGroupGql>,
+    pub recipient_rules: Vec<RecipientRuleGql>,
+    pub actor_key_bindings: Vec<ActorBindingGql>,
+    pub last_operation: Option<VaultLastOpGql>,
+}
+
+// ── Query root ───────────────────────────────────────────────────────────────
+
+pub struct QueryRoot;
+
+fn collect_conflicts(
+    index: &std::collections::BTreeMap<(String, String), Vec<(String, String)>>,
+) -> Vec<CrossProjectConflict> {
+    let mut conflicts = Vec::new();
+    for ((section, field), entries) in index {
+        for i in 0..entries.len() {
+            for j in (i + 1)..entries.len() {
+                if entries[i].1 != entries[j].1 {
+                    conflicts.push(CrossProjectConflict {
+                        section: section.clone(),
+                        field: field.clone(),
+                        project_a: entries[i].0.clone(),
+                        value_a: entries[i].1.clone(),
+                        project_b: entries[j].0.clone(),
+                        value_b: entries[j].1.clone(),
+                    });
+                }
+            }
+        }
+    }
+    conflicts
+}
+
+#[Object]
+impl QueryRoot {
+    /// Resolve a single ontology node by project slug and node id.
+    async fn ontology_node(
+        &self,
+        ctx: &Context<'_>,
+        slug: String,
+        id: String,
+    ) -> async_graphql::Result<Option<OntologyNode>> {
+        let state = ctx.data::<AppState>()?;
+        let fed = FederatedQuery::new(state.registry.clone());
+        Ok(fed.resolve(&slug, &id).await.map(|f| {
+            let name = str_field(&f.node, "name").unwrap_or_else(|| f.node_id.clone());
+            let kind = str_field(&f.node, "kind").unwrap_or_else(|| "unknown".to_owned());
+            let description = str_field(&f.node, "description");
+            OntologyNode {
+                slug: f.slug,
+                id: f.node_id,
+                name,
+                kind,
+                description,
+            }
+        }))
+    }
+
+    /// BFS impact graph from a node, with optional cross-project traversal.
+    async fn impact_graph(
+        &self,
+        ctx: &Context<'_>,
+        slug: String,
+        node_id: String,
+        max_depth: Option<i32>,
+        include_external: Option<bool>,
+    ) -> async_graphql::Result<Vec<ImpactEntry>> {
+        let state = ctx.data::<AppState>()?;
+        let fed = FederatedQuery::new(state.registry.clone());
+        let depth = max_depth.unwrap_or(3).max(0) as u32;
+        let entries = fed
+            .impact_graph(&slug, &node_id, depth, include_external.unwrap_or(false))
+            .await;
+        Ok(entries
+            .into_iter()
+            .map(|e| ImpactEntry {
+                slug: e.slug,
+                node_id: e.node_id,
+                node_name: e.node_name,
+                depth: e.depth as i32,
+                direction: e.direction,
+                via: e.via,
+            })
+            .collect())
+    }
+
+    /// Project identity from core.ncl (slug defaults to primary project).
+    async fn project(
+        &self,
+        ctx: &Context<'_>,
+        slug: Option<String>,
+    ) -> async_graphql::Result<Option<ProjectIdentity>> {
+        let state = ctx.data::<AppState>()?;
+        let (root, cache, import_path) = project_ctx(state, slug.as_deref());
+        let path = root.join(".ontology").join("core.ncl");
+        if !path.exists() {
+            return Ok(None);
+        }
+        let (json, _) = cache
+            .export(&path, import_path.as_deref())
+            .await
+            .map_err(|e| async_graphql::Error::new(e.to_string()))?;
+
+        let identity = json.get("identity").unwrap_or(&json);
+        let name = str_field(identity, "name")
+            .or_else(|| str_field(&json, "name"))
+            .unwrap_or_else(|| "unknown".to_owned());
+        let kind = str_field(identity, "kind")
+            .or_else(|| str_field(&json, "kind"))
+            .unwrap_or_else(|| "project".to_owned());
+        let description =
+            str_field(identity, "description").or_else(|| str_field(&json, "description"));
+        let resolved_slug = slug.unwrap_or_else(|| state.registry.primary_slug().to_owned());
+        Ok(Some(ProjectIdentity {
+            slug: resolved_slug,
+            name,
+            kind,
+            description,
+        }))
+    }
+
+    /// Backlog items from reflection/backlog.ncl (slug defaults to primary).
+    async fn backlog(
+        &self,
+        ctx: &Context<'_>,
+        slug: Option<String>,
+    ) -> async_graphql::Result<Vec<BacklogItem>> {
+        let state = ctx.data::<AppState>()?;
+        let (root, cache, import_path) = project_ctx(state, slug.as_deref());
+        let path = root.join("reflection").join("backlog.ncl");
+        if !path.exists() {
+            return Ok(vec![]);
+        }
+        let (json, _) = cache
+            .export(&path, import_path.as_deref())
+            .await
+            .map_err(|e| async_graphql::Error::new(e.to_string()))?;
+        let items = json
+            .get("items")
+            .and_then(|v| v.as_array())
+            .cloned()
+            .unwrap_or_default();
+        Ok(items
+            .iter()
+            .map(|item| BacklogItem {
+                id: str_field(item, "id").unwrap_or_default(),
+                title: str_field(item, "title").unwrap_or_default(),
+                kind: str_field(item, "kind").unwrap_or_default(),
+                priority: str_field(item, "priority").unwrap_or_default(),
+                status: str_field(item, "status").unwrap_or_default(),
+            })
+            .collect())
+    }
+
+    /// List ADR summaries for a project (slug defaults to primary).
+    async fn adrs(
+        &self,
+        ctx: &Context<'_>,
+        slug: Option<String>,
+    ) -> async_graphql::Result<Vec<AdrSummary>> {
+        let state = ctx.data::<AppState>()?;
+        let (root, cache, import_path) = project_ctx(state, slug.as_deref());
+        let adrs_dir = root.join("adrs");
+        let resolved_slug = slug.unwrap_or_else(|| state.registry.primary_slug().to_owned());
+
+        let entries = match std::fs::read_dir(&adrs_dir) {
+            Ok(e) => e,
+            Err(_) => return Ok(vec![]),
+        };
+
+        let mut summaries = Vec::new();
+        for entry in entries.flatten() {
+            let path = entry.path();
+            if path.extension().and_then(|e| e.to_str()) != Some("ncl") {
+                continue;
+            }
+            let file_id = path
+                .file_stem()
+                .and_then(|s| s.to_str())
+                .unwrap_or("")
+                .to_owned();
+            let Ok((json, _)) = cache.export(&path, import_path.as_deref()).await else {
+                continue;
+            };
+            summaries.push(AdrSummary {
+                id: str_field(&json, "id").unwrap_or_else(|| file_id.clone()),
+                slug: resolved_slug.clone(),
+                title: str_field(&json, "title").unwrap_or_else(|| file_id.clone()),
+                status: str_field(&json, "status").unwrap_or_else(|| "unknown".to_owned()),
+                date: str_field(&json, "date"),
+            });
+        }
+        summaries.sort_by(|a, b| a.id.cmp(&b.id));
+        Ok(summaries)
+    }
+
+    /// Full ADR record by id fragment (slug defaults to primary).
+    async fn adr(
+        &self,
+        ctx: &Context<'_>,
+        id: String,
+        slug: Option<String>,
+    ) -> async_graphql::Result<Option<AdrRecord>> {
+        let state = ctx.data::<AppState>()?;
+        let (root, cache, import_path) = project_ctx(state, slug.as_deref());
+        let resolved_slug = slug.unwrap_or_else(|| state.registry.primary_slug().to_owned());
+        let adrs_dir = root.join("adrs");
+
+        let entries = match std::fs::read_dir(&adrs_dir) {
+            Ok(e) => e,
+            Err(_) => return Ok(None),
+        };
+        for entry in entries.flatten() {
+            let path = entry.path();
+            if path.extension().and_then(|e| e.to_str()) != Some("ncl") {
+                continue;
+            }
+            let stem = path
+                .file_stem()
+                .and_then(|s| s.to_str())
+                .unwrap_or("")
+                .to_string();
+            if !stem.contains(id.as_str()) {
+                continue;
+            }
+            let (json, _) = cache
+                .export(&path, import_path.as_deref())
+                .await
+                .map_err(|e| async_graphql::Error::new(e.to_string()))?;
+            return Ok(Some(AdrRecord {
+                id: str_field(&json, "id").unwrap_or_else(|| stem.clone()),
+                slug: resolved_slug,
+                title: str_field(&json, "title").unwrap_or_else(|| stem.clone()),
+                status: str_field(&json, "status").unwrap_or_default(),
+                date: str_field(&json, "date"),
+                context: str_field(&json, "context"),
+                decision: str_field(&json, "decision"),
+                consequences: str_field(&json, "consequences"),
+            }));
+        }
+        Ok(None)
+    }
+
+    /// Project summaries for all registered projects with config surfaces.
+    async fn cross_project_summaries(
+        &self,
+        ctx: &Context<'_>,
+    ) -> async_graphql::Result<Vec<CrossProjectSummary>> {
+        let state = ctx.data::<AppState>()?;
+        let candidates: Vec<_> = state.registry.all();
+
+        let mut out = Vec::new();
+        for proj in candidates {
+            let Some(surface) = proj.config_surface.as_ref() else {
+                continue;
+            };
+            let entry = surface.entry_point_path(&proj.root);
+            let export_ok = proj
+                .cache
+                .export(&entry, proj.import_path.as_deref())
+                .await
+                .is_ok();
+            out.push(CrossProjectSummary {
+                slug: proj.slug.clone(),
+                kind: format!("{:?}", surface.kind),
+                export_ok,
+                section_count: surface.sections.len() as i32,
+            });
+        }
+        Ok(out)
+    }
+
+    /// Cross-project config conflicts — all pairs of projects with differing
+    /// values for the same (section, field).
+    async fn cross_project_conflicts(
+        &self,
+        ctx: &Context<'_>,
+    ) -> async_graphql::Result<Vec<CrossProjectConflict>> {
+        let state = ctx.data::<AppState>()?;
+        let candidates: Vec<_> = state.registry.all();
+
+        // (section, field) → Vec<(slug, value_string)>
+        let mut index: std::collections::BTreeMap<(String, String), Vec<(String, String)>> =
+            std::collections::BTreeMap::new();
+
+        for proj in &candidates {
+            let Some(surface) = proj.config_surface.as_ref() else {
+                continue;
+            };
+            let entry = surface.entry_point_path(&proj.root);
+            let Ok((json, _)) = proj.cache.export(&entry, proj.import_path.as_deref()).await else {
+                continue;
+            };
+            for section in &surface.sections {
+                let Some(sec_val) = json.get(&section.id).and_then(|v| v.as_object()) else {
+                    continue;
+                };
+                for (field, val) in sec_val {
+                    index
+                        .entry((section.id.clone(), field.clone()))
+                        .or_default()
+                        .push((proj.slug.clone(), val.to_string()));
+                }
+            }
+        }
+
+        let conflicts = collect_conflicts(&index);
+        Ok(conflicts)
+    }
+
+    /// Federation entity resolver — resolves OntologyNode by (slug, id).
+    #[graphql(entity)]
+    async fn find_ontology_node_by_slug_and_id(
+        &self,
+        ctx: &Context<'_>,
+        slug: String,
+        id: String,
+    ) -> async_graphql::Result<Option<OntologyNode>> {
+        self.ontology_node(ctx, slug, id).await
+    }
+
+    /// Registry topology declared in a project's manifest.ncl registry_provides
+    /// field.
+    async fn registry_topology(
+        &self,
+        ctx: &Context<'_>,
+        slug: Option<String>,
+    ) -> async_graphql::Result<Option<RegistryTopologyGql>> {
+        let state = ctx.data::<AppState>()?;
+        let (root, cache, import_path) = project_ctx(state, slug.as_deref());
+        let path = root.join(".ontology").join("manifest.ncl");
+        if !path.exists() {
+            return Ok(None);
+        }
+        let (json, _) = cache
+            .export(&path, import_path.as_deref())
+            .await
+            .map_err(|e| async_graphql::Error::new(e.to_string()))?;
+
+        let Some(provides) = json.get("registry_provides") else {
+            return Ok(None);
+        };
+        let participant = str_field(provides, "participant").unwrap_or_default();
+        if participant.is_empty() {
+            return Ok(None);
+        }
+
+        let regs = provides
+            .get("registries")
+            .and_then(|r| r.get("registries"))
+            .and_then(|r| r.as_array())
+            .cloned()
+            .unwrap_or_default();
+        let default_id = provides
+            .get("registries")
+            .and_then(|r| r.get("default"))
+            .and_then(|v| v.as_str())
+            .map(|s| s.to_owned());
+
+        let entries = regs
+            .iter()
+            .map(|e| {
+                let ns = e.get("namespaces").unwrap_or(&serde_json::Value::Null);
+                RegistryEntryGql {
+                    id: str_field(e, "id").unwrap_or_default(),
+                    endpoint: str_field(e, "endpoint").unwrap_or_default(),
+                    role: str_field(e, "role").unwrap_or_else(|| "dev".to_owned()),
+                    tls: e.get("tls").and_then(|v| v.as_bool()).unwrap_or(true),
+                    namespaces: RegistryNamespacesGql {
+                        own: string_array(ns, "own"),
+                        replicate_to: string_array(ns, "replicate_to"),
+                        mirror_from: str_field(ns, "mirror_from"),
+                        prefixes: string_array(ns, "prefixes"),
+                    },
+                }
+            })
+            .collect();
+
+        let namespaces = vec![
+            format!("domains/{participant}/"),
+            format!("modes/{participant}/"),
+        ];
+
+        Ok(Some(RegistryTopologyGql {
+            participant,
+            default_id,
+            entries,
+            namespaces,
+        }))
+    }
+
+    /// Canonical participant namespaces for a project id.
+    /// Returns ["domains/<project_id>/", "modes/<project_id>/"].
+    async fn participant_namespace(
+        &self,
+        _ctx: &Context<'_>,
+        project_id: String,
+    ) -> async_graphql::Result<Vec<String>> {
+        Ok(vec![
+            format!("domains/{project_id}/"),
+            format!("modes/{project_id}/"),
+        ])
+    }
+
+    /// Credential vault state declared in a project's
+    /// .ontoref/project.ncl::sops block (ADR-017 + ADR-019). Read-only —
+    /// never decrypts; only inspects declarative state and the local vault
+    /// dir.
+    async fn vault_status(
+        &self,
+        ctx: &Context<'_>,
+        slug: Option<String>,
+    ) -> async_graphql::Result<Option<VaultStatusGql>> {
+        let state = ctx.data::<AppState>()?;
+        let (root, cache, import_path) = project_ctx(state, slug.as_deref());
+        let project_ncl = root.join(".ontoref").join("project.ncl");
+        if !project_ncl.exists() {
+            return Ok(None);
+        }
+        let (project, _) = cache
+            .export(&project_ncl, import_path.as_deref())
+            .await
+            .map_err(|e| async_graphql::Error::new(e.to_string()))?;
+
+        let Some(sops) = project.get("sops").filter(|v| !v.is_null()) else {
+            return Ok(None);
+        };
+        if !sops
+            .get("enabled")
+            .and_then(|v| v.as_bool())
+            .unwrap_or(false)
+        {
+            return Ok(None);
+        }
+
+        let project_slug = str_field(&project, "slug").unwrap_or_default();
+        let vault_id = sops
+            .get("vault_id")
+            .and_then(|v| v.as_str())
+            .filter(|s| !s.is_empty())
+            .unwrap_or(&project_slug)
+            .to_string();
+        let backend = sops
+            .get("vault_backend")
+            .and_then(|v| v.as_str())
+            .unwrap_or("restic")
+            .to_string();
+        let registry_endpoint = sops
+            .get("registry_endpoint")
+            .and_then(|v| v.as_str())
+            .filter(|s| !s.is_empty())
+            .map(String::from);
+
+        // master_key: project override → global config.
+        let master_key = sops
+            .get("master_key_path")
+            .and_then(|v| v.as_str())
+            .filter(|s| !s.is_empty())
+            .map(String::from)
+            .or_else(|| {
+                let home = std::env::var_os("HOME").map(std::path::PathBuf::from)?;
+                let global = home.join(".config").join("ontoref").join("config.ncl");
+                if !global.exists() {
+                    return None;
+                }
+                let out = std::process::Command::new("nickel")
+                    .arg("export")
+                    .arg(&global)
+                    .output()
+                    .ok()?;
+                if !out.status.success() {
+                    return None;
+                }
+                let v: serde_json::Value = serde_json::from_slice(&out.stdout).ok()?;
+                v.get("vault")
+                    .and_then(|x| x.get("master_key_path"))
+                    .and_then(|s| s.as_str())
+                    .filter(|s| !s.is_empty())
+                    .map(String::from)
+            });
+
+        let home = std::env::var_os("HOME")
+            .map(std::path::PathBuf::from)
+            .ok_or_else(|| async_graphql::Error::new("HOME not resolvable"))?;
+        let vault_dir = home
+            .join(".config")
+            .join("ontoref")
+            .join("vaults")
+            .join(&vault_id);
+        let access_sops = vault_dir.join("access.sops.yaml");
+        let logs_jsonl = vault_dir.join("logs").join("access.jsonl");
+
+        let access_sops_present = access_sops.exists();
+        let recipient_count: i32 = if access_sops_present {
+            std::fs::read_to_string(&access_sops)
+                .map(|s| {
+                    s.lines()
+                        .filter(|l| l.trim_start().starts_with("- recipient: age1"))
+                        .count() as i32
+                })
+                .unwrap_or(0)
+        } else {
+            0
+        };
+
+        let recipient_groups: Vec<RecipientGroupGql> = sops
+            .get("recipient_groups")
+            .and_then(|v| v.as_object())
+            .map(|map| {
+                map.iter()
+                    .map(|(k, v)| RecipientGroupGql {
+                        name: k.clone(),
+                        keys: v
+                            .as_array()
+                            .map(|a| {
+                                a.iter()
+                                    .filter_map(|x| x.as_str().map(String::from))
+                                    .collect()
+                            })
+                            .unwrap_or_default(),
+                    })
+                    .collect()
+            })
+            .unwrap_or_default();
+
+        let recipient_rules: Vec<RecipientRuleGql> = sops
+            .get("recipient_rules")
+            .and_then(|v| v.as_array())
+            .map(|a| {
+                a.iter()
+                    .filter_map(|r| {
+                        Some(RecipientRuleGql {
+                            path: r.get("path")?.as_str()?.to_string(),
+                            groups: r
+                                .get("groups")?
+                                .as_array()?
+                                .iter()
+                                .filter_map(|g| g.as_str().map(String::from))
+                                .collect(),
+                        })
+                    })
+                    .collect()
+            })
+            .unwrap_or_default();
+
+        let actor_key_bindings: Vec<ActorBindingGql> = sops
+            .get("actor_key_bindings")
+            .and_then(|v| v.as_object())
+            .map(|m| {
+                m.iter()
+                    .filter_map(|(k, v)| {
+                        Some(ActorBindingGql {
+                            actor: k.clone(),
+                            role: v.as_str()?.to_string(),
+                        })
+                    })
+                    .collect()
+            })
+            .unwrap_or_default();
+
+        let last_operation = if logs_jsonl.exists() {
+            std::fs::read_to_string(&logs_jsonl)
+                .ok()
+                .and_then(|s| s.lines().last().map(String::from))
+                .and_then(|line| serde_json::from_str::<serde_json::Value>(&line).ok())
+                .and_then(|v| {
+                    Some(VaultLastOpGql {
+                        ts: v.get("ts")?.as_str()?.to_string(),
+                        actor: v.get("actor")?.as_str()?.to_string(),
+                        op: v.get("op")?.as_str()?.to_string(),
+                        registry_id: v
+                            .get("registry_id")
+                            .and_then(|x| x.as_str())
+                            .map(String::from),
+                        detail: v.get("detail").and_then(|x| x.as_str()).map(String::from),
+                    })
+                })
+        } else {
+            None
+        };
+
+        Ok(Some(VaultStatusGql {
+            vault_id,
+            vault_dir: vault_dir.display().to_string(),
+            backend,
+            registry_endpoint,
+            master_key,
+            access_sops_present,
+            recipient_count,
+            recipient_groups,
+            recipient_rules,
+            actor_key_bindings,
+            last_operation,
+        }))
+    }
+}
+
+// ── Mutation root ────────────────────────────────────────────────────────────
+
+pub struct MutationRoot;
+
+#[Object]
+impl MutationRoot {
+    /// Emit a custom notification on a project. Returns the new notification
+    /// id.
+    async fn emit_notification(
+        &self,
+        ctx: &Context<'_>,
+        kind: String,
+        title: String,
+        project: Option<String>,
+        payload_json: Option<String>,
+    ) -> async_graphql::Result<i64> {
+        let state = ctx.data::<AppState>()?;
+        let slug = project
+            .as_deref()
+            .unwrap_or_else(|| state.registry.primary_slug());
+        let payload = payload_json
+            .as_deref()
+            .map(serde_json::from_str)
+            .transpose()
+            .map_err(|e| async_graphql::Error::new(format!("payload_json: {e}")))?;
+        let id = state
+            .notifications
+            .push_custom(slug, kind, title, payload, None, None);
+        Ok(id as i64)
+    }
+
+    /// Propose a status change for a backlog item (creates a backlog_review
+    /// notification pending admin approval).
+    async fn propose_backlog_status(
+        &self,
+        ctx: &Context<'_>,
+        id: String,
+        proposed_status: String,
+        slug: Option<String>,
+    ) -> async_graphql::Result<bool> {
+        let state = ctx.data::<AppState>()?;
+        let project = slug
+            .as_deref()
+            .unwrap_or_else(|| state.registry.primary_slug());
+        state.notifications.push_custom(
+            project,
+            "backlog_review",
+            format!("Status proposal: {id} → {proposed_status}"),
+            Some(serde_json::json!({ "id": id, "proposed_status": proposed_status })),
+            None,
+            None,
+        );
+        Ok(true)
+    }
+
+    /// Acknowledge specific notifications by id list. Returns count acked.
+    async fn ack_notifications(
+        &self,
+        ctx: &Context<'_>,
+        ids: Vec<i64>,
+        actor_token: String,
+        project: Option<String>,
+    ) -> async_graphql::Result<i32> {
+        let state = ctx.data::<AppState>()?;
+        let slug = project
+            .as_deref()
+            .unwrap_or_else(|| state.registry.primary_slug());
+        let count = ids
+            .iter()
+            .filter(|&&i| state.notifications.ack_one(slug, &actor_token, i as u64))
+            .count();
+        Ok(count as i32)
+    }
+
+    /// Acknowledge all pending notifications for an actor. Returns count acked.
+    async fn ack_all_notifications(
+        &self,
+        ctx: &Context<'_>,
+        actor_token: String,
+        project: Option<String>,
+    ) -> async_graphql::Result<i32> {
+        let state = ctx.data::<AppState>()?;
+        let slug = project
+            .as_deref()
+            .unwrap_or_else(|| state.registry.primary_slug());
+        Ok(state.notifications.ack_all(slug, &actor_token) as i32)
+    }
+}
+
+// ── Subscription root ────────────────────────────────────────────────────────
+
+pub struct SubscriptionRoot;
+
+#[Subscription]
+impl SubscriptionRoot {
+    /// Real-time notification stream. Optionally filter by project slug.
+    async fn notifications(
+        &self,
+        ctx: &Context<'_>,
+        project: Option<String>,
+    ) -> async_graphql::Result<impl futures_util::Stream<Item = NotificationGql>> {
+        let state = ctx.data::<AppState>()?;
+        let rx = state.notifications.subscribe();
+        Ok(BroadcastStream::new(rx)
+            .filter_map(|r| std::future::ready(r.ok()))
+            .filter(move |notif| {
+                let keep = project
+                    .as_deref()
+                    .is_none_or(|p| p == notif.project.as_str());
+                std::future::ready(keep)
+            })
+            .map(NotificationGql::from))
+    }
+}
+
+// ── Helpers ──────────────────────────────────────────────────────────────────
+
+fn string_array(v: &serde_json::Value, key: &str) -> Vec<String> {
+    v.get(key)
+        .and_then(|a| a.as_array())
+        .map(|arr| {
+            arr.iter()
+                .filter_map(|s| s.as_str().map(|s| s.to_owned()))
+                .collect()
+        })
+        .unwrap_or_default()
+}
+
+fn str_field(v: &serde_json::Value, key: &str) -> Option<String> {
+    v.get(key)?.as_str().map(|s| s.to_owned())
+}
+
+fn project_ctx(
+    state: &AppState,
+    slug: Option<&str>,
+) -> (std::path::PathBuf, Arc<NclCache>, Option<String>) {
+    if let Some(s) = slug {
+        if let Some(ctx) = state.registry.get(s) {
+            return (ctx.root.clone(), ctx.cache.clone(), ctx.import_path.clone());
+        }
+    }
+    (
+        state.project_root.clone(),
+        state.cache.clone(),
+        state.nickel_import_path.clone(),
+    )
+}
+
+// ── Axum handlers ────────────────────────────────────────────────────────────
+
+#[derive(serde::Deserialize)]
+struct GraphiqlQuery {
+    token: Option<String>,
+}
+
+/// GET /graphql — serves GraphiQL IDE.
+///
+/// When the primary project has auth enabled, a `?token=<password>` query
+/// parameter is required. On success the token is pre-configured in GraphiQL
+/// as the default `Authorization: Bearer` header so queries run without manual
+/// setup. The same token is forwarded as a WebSocket `connectionParam` for
+/// subscriptions.
+async fn graphiql_handler(
+    Extension(app_state): Extension<AppState>,
+    ConnectInfo(addr): ConnectInfo<SocketAddr>,
+    Query(q): Query<GraphiqlQuery>,
+) -> impl IntoResponse {
+    // If a graphql read token is configured, require it as ?token=.
+    if app_state.graphql_read_token_hash.is_some() {
+        let provided = q.token.as_deref().unwrap_or("");
+        if !app_state.verify_graphql_token(provided) {
+            let mut headers = HeaderMap::new();
+            if let Ok(val) = HeaderValue::from_str(&format!("Bearer {}", provided)) {
+                headers.insert(AUTHORIZATION, val);
+            }
+            app_state.auth_rate_limiter.record_failure(addr.ip());
+            tracing::warn!(peer_ip = %addr.ip(), endpoint = "GET /graphql (IDE)", "unauthorized GraphQL access attempt");
+            return (
+                axum::http::StatusCode::UNAUTHORIZED,
+                axum::Json(serde_json::json!({"error": "invalid or missing GraphQL token — use ?token=<token>"})),
+            ).into_response();
+        }
+    }
+
+    let bearer_value = q.token.as_deref().map(|t| format!("Bearer {t}"));
+    let mut builder = async_graphql::http::GraphiQLSource::build()
+        .endpoint("/graphql")
+        .subscription_endpoint("/graphql/ws");
+    if let Some(ref bv) = bearer_value {
+        builder = builder
+            .header("Authorization", bv)
+            .ws_connection_param("Authorization", bv);
+    }
+    Html(builder.finish()).into_response()
+}
+
+async fn graphql_handler(
+    State(schema): State<OntoSchema>,
+    Extension(app_state): Extension<AppState>,
+    ConnectInfo(addr): ConnectInfo<SocketAddr>,
+    headers: HeaderMap,
+    req: GraphQLRequest,
+) -> impl IntoResponse {
+    if let Some(token) = extract_bearer(&headers) {
+        if !app_state.verify_graphql_token(token) {
+            app_state.auth_rate_limiter.record_failure(addr.ip());
+            tracing::warn!(peer_ip = %addr.ip(), endpoint = "POST /graphql", "unauthorized GraphQL access attempt");
+            return (
+                axum::http::StatusCode::UNAUTHORIZED,
+                axum::Json(serde_json::json!({"error": "invalid GraphQL token"})),
+            )
+                .into_response();
+        }
+    } else if app_state.graphql_read_token_hash.is_some() {
+        tracing::warn!(peer_ip = %addr.ip(), endpoint = "POST /graphql", "unauthenticated GraphQL request — token required");
+        return (
+            axum::http::StatusCode::UNAUTHORIZED,
+            axum::Json(serde_json::json!({"error": "Authorization: Bearer <token> required"})),
+        )
+            .into_response();
+    }
+    let gql: GraphQLResponse = schema.execute(req.into_inner()).await.into();
+    gql.into_response()
+}
+
+async fn graphql_ws_handler(
+    State(schema): State<OntoSchema>,
+    Extension(app_state): Extension<AppState>,
+    ConnectInfo(addr): ConnectInfo<SocketAddr>,
+    headers: HeaderMap,
+    protocol: GraphQLProtocol,
+    ws: WebSocketUpgrade,
+) -> impl IntoResponse {
+    // HTTP-level auth — verify graphql token from Bearer header.
+    if let Some(token) = extract_bearer(&headers) {
+        if !app_state.verify_graphql_token(token) {
+            app_state.auth_rate_limiter.record_failure(addr.ip());
+            tracing::warn!(peer_ip = %addr.ip(), endpoint = "GET /graphql/ws", "unauthorized GraphQL WS upgrade attempt");
+            return (
+                axum::http::StatusCode::UNAUTHORIZED,
+                axum::Json(serde_json::json!({"error": "invalid GraphQL token"})),
+            )
+                .into_response();
+        }
+    } else if app_state.graphql_read_token_hash.is_some() {
+        tracing::warn!(peer_ip = %addr.ip(), endpoint = "GET /graphql/ws", "unauthenticated GraphQL WS upgrade — token required");
+        return (
+            axum::http::StatusCode::UNAUTHORIZED,
+            axum::Json(serde_json::json!({"error": "Authorization: Bearer <token> required"})),
+        )
+            .into_response();
+    }
+    // Token forwarded into WS context for subscription resolvers if needed.
+    let bearer = extract_bearer(&headers).map(|s| s.to_owned());
+    ws.protocols(ALL_WEBSOCKET_PROTOCOLS)
+        .on_upgrade(move |stream| async move {
+            let mut data = async_graphql::Data::default();
+            if let Some(token) = bearer {
+                data.insert(token);
+            }
+            GraphQLWebSocket::new(stream, schema, protocol)
+                .with_data(data)
+                .serve()
+                .await;
+        })
+}
+
+// ── Router ───────────────────────────────────────────────────────────────────
+
+pub fn graphql_router(state: &AppState) -> Router {
+    let schema = Schema::build(QueryRoot, MutationRoot, SubscriptionRoot)
+        .data(state.clone())
+        .enable_federation()
+        .finish();
+    Router::new()
+        .route("/", get(graphiql_handler).post(graphql_handler))
+        .route("/ws", get(graphql_ws_handler))
+        .layer(Extension(state.clone()))
+        .with_state(schema)
+}
diff --git a/crates/ontoref-daemon/src/lib.rs b/crates/ontoref-daemon/src/lib.rs
index f66d67a..da8d5c7 100644
--- a/crates/ontoref-daemon/src/lib.rs
+++ b/crates/ontoref-daemon/src/lib.rs
@@ -18,6 +18,8 @@ pub mod config;
 pub mod config_coherence;
 pub mod error;
 pub mod federation;
+#[cfg(feature = "graphql")]
+pub mod graphql;
 #[cfg(feature = "mcp")]
 pub mod mcp;
 #[cfg(feature = "nats")]
diff --git a/crates/ontoref-daemon/src/main.rs b/crates/ontoref-daemon/src/main.rs
index 2014cf5..0efb7e3 100644
--- a/crates/ontoref-daemon/src/main.rs
+++ b/crates/ontoref-daemon/src/main.rs
@@ -868,6 +868,17 @@ async fn main() {
         }
     };
 
+    // Load GraphQL read token hash from config (feature-gated).
+    #[cfg(feature = "graphql")]
+    let graphql_read_token_hash: Option<String> = loaded_ncl_config.as_ref().and_then(|c| {
+        let h = c.graphql.read_token_hash.trim();
+        if h.is_empty() {
+            None
+        } else {
+            Some(h.to_owned())
+        }
+    });
+
     // Capture display values before they are moved into AppState.
     #[cfg(feature = "ui")]
     let project_root_str = project_root.display().to_string();
@@ -915,6 +926,9 @@ async fn main() {
                 ncl_write_lock: Arc::new(ontoref_daemon::ui::ncl_write::NclWriteLock::new()),
                 #[cfg(feature = "ui")]
                 daemon_admin_hash: daemon_admin_hash.clone(),
+                #[cfg(feature = "graphql")]
+                graphql_read_token_hash: graphql_read_token_hash.clone(),
+                service_flags: ontoref_daemon::api::ServiceFlags::new(),
             }
         }
         #[cfg(not(feature = "nats"))]
@@ -949,6 +963,9 @@ async fn main() {
                 ncl_write_lock: Arc::new(ontoref_daemon::ui::ncl_write::NclWriteLock::new()),
                 #[cfg(feature = "ui")]
                 daemon_admin_hash,
+                #[cfg(feature = "graphql")]
+                graphql_read_token_hash,
+                service_flags: ontoref_daemon::api::ServiceFlags::new(),
             }
         }
     };
@@ -1036,6 +1053,10 @@ async fn main() {
         return;
     }
 
+    // Capture before state is moved into the router.
+    #[cfg(feature = "graphql")]
+    let graphql_auth_protected = state.graphql_read_token_hash.is_some();
+
     let app = api::router(state).layer(TraceLayer::new_for_http().make_span_with(
         |req: &axum::http::Request<_>| {
             // Health-check endpoints are polled frequently — log at TRACE to
@@ -1072,6 +1093,51 @@ async fn main() {
 
     info!(addr = %addr, "listening");
 
+    // ── Active feature log ───────────────────────────────────────────────
+    {
+        #[allow(unused_mut)]
+        let mut features: Vec<&str> = Vec::new();
+        #[cfg(feature = "db")]
+        {
+            features.push("db");
+        }
+        #[cfg(feature = "nats")]
+        {
+            features.push("nats");
+        }
+        #[cfg(feature = "ui")]
+        {
+            features.push("ui");
+        }
+        #[cfg(feature = "mcp")]
+        {
+            features.push("mcp");
+        }
+        #[cfg(feature = "graphql")]
+        {
+            features.push("graphql");
+        }
+        #[cfg(feature = "tls")]
+        {
+            features.push("tls");
+        }
+        info!(features = ?features, "active features");
+    }
+
+    #[cfg(feature = "graphql")]
+    {
+        let auth_status = if graphql_auth_protected {
+            "token-protected"
+        } else {
+            "open (no token configured)"
+        };
+        info!(
+            endpoint = "/graphql",
+            auth = auth_status,
+            "GraphQL available"
+        );
+    }
+
     #[cfg(feature = "ui")]
     if let Some((ref tdir, ref public)) = ui_startup {
         #[cfg(feature = "tls")]
diff --git a/crates/ontoref-daemon/src/mcp/mod.rs b/crates/ontoref-daemon/src/mcp/mod.rs
index a28d701..5215b39 100644
--- a/crates/ontoref-daemon/src/mcp/mod.rs
+++ b/crates/ontoref-daemon/src/mcp/mod.rs
@@ -306,6 +306,7 @@ impl OntoreServer {
             .with_async_tool::<ConfigCoherenceTool>()
             .with_async_tool::<ConfigQuickrefTool>()
             .with_async_tool::<ConfigUpdateTool>()
+            .with_async_tool::<VaultStatusTool>()
     }
 
     fn project_ctx(&self, slug: Option<&str>) -> ProjectCtx {
@@ -346,9 +347,66 @@ impl OntoreServer {
     }
 }
 
+// ── MCP tool catalog (ADR-015)
+// ─────────────────────────────────────────────────
+
+/// All MCP tools registered via `#[ontoref_derive::onto_mcp_tool(...)]`,
+/// sorted by tool name. Reads the linker-built inventory at runtime — no
+/// allocation of catalog state, no daemon-state dependency.
+pub fn catalog() -> Vec<&'static ontoref_ontology::McpToolEntry> {
+    let mut tools: Vec<&'static ontoref_ontology::McpToolEntry> =
+        inventory::iter::<ontoref_ontology::McpToolEntry>().collect();
+    tools.sort_by_key(|t| t.name);
+    tools
+}
+
+/// Render a single catalog entry into the JSON shape returned by the
+/// `ontoref_help` tool. `constraint` is mapped to `required: bool` and an
+/// optional `default` field; non-empty `kind` is surfaced as a type hint.
+fn render_tool_for_help(t: &ontoref_ontology::McpToolEntry) -> serde_json::Value {
+    let params: Vec<serde_json::Value> = t
+        .params
+        .iter()
+        .map(|p| {
+            let mut entry = serde_json::Map::new();
+            entry.insert("name".into(), serde_json::Value::String(p.name.into()));
+            if !p.kind.is_empty() {
+                entry.insert("kind".into(), serde_json::Value::String(p.kind.into()));
+            }
+            entry.insert(
+                "required".into(),
+                serde_json::Value::Bool(p.constraint == "required"),
+            );
+            if let Some(default) = p.constraint.strip_prefix("default=") {
+                entry.insert("default".into(), serde_json::Value::String(default.into()));
+            }
+            if !p.description.is_empty() {
+                entry.insert(
+                    "description".into(),
+                    serde_json::Value::String(p.description.into()),
+                );
+            }
+            serde_json::Value::Object(entry)
+        })
+        .collect();
+
+    serde_json::json!({
+        "name": t.name,
+        "description": t.description,
+        "category": t.category,
+        "params": params,
+    })
+}
+
 // ── Tool: list_projects
 // ─────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_list_projects",
+    description = "List all available projects. In single-project mode returns the current \
+                   project name.",
+    category = "discovery"
+)]
 struct ListProjectsTool;
 
 impl ToolBase for ListProjectsTool {
@@ -390,6 +448,13 @@ impl AsyncTool<OntoreServer> for ListProjectsTool {
 // ── Tool: search
 // ────────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_search",
+    description = "Free-text search across nodes, ADRs, and modes. Returns kind+id for use with \
+                   ontoref_get.",
+    category = "discovery",
+    params = "query:string:required:Free-text search query; project:string:optional:Project slug"
+)]
 struct SearchTool;
 
 impl ToolBase for SearchTool {
@@ -435,6 +500,12 @@ impl AsyncTool<OntoreServer> for SearchTool {
 // ── Tool: describe_project
 // ──────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_describe",
+    description = "Architecture overview: README summary and manifest.",
+    category = "discovery",
+    params = "project:string:optional:Project slug"
+)]
 struct DescribeProjectTool;
 
 impl ToolBase for DescribeProjectTool {
@@ -496,6 +567,12 @@ impl AsyncTool<OntoreServer> for DescribeProjectTool {
 // ── Tool: list_adrs
 // ─────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_list_adrs",
+    description = "List all ADRs with id, title, status, constraint counts.",
+    category = "discovery",
+    params = "project:string:optional:Project slug"
+)]
 struct ListAdrsTool;
 
 impl ToolBase for ListAdrsTool {
@@ -580,6 +657,12 @@ impl AsyncTool<OntoreServer> for ListAdrsTool {
 // ── Tool: get_adr
 // ───────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_get_adr",
+    description = "Full ADR by id or partial stem (e.g. adr-001).",
+    category = "ontology",
+    params = "id:string:required:ADR id or partial stem; project:string:optional:Project slug"
+)]
 struct GetAdrTool;
 
 impl ToolBase for GetAdrTool {
@@ -642,6 +725,12 @@ impl AsyncTool<OntoreServer> for GetAdrTool {
 // ── Tool: list_ontology_extensions
 // ──────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_list_ontology_extensions",
+    description = "List extra .ontology/*.ncl files beyond core/state/gate.",
+    category = "discovery",
+    params = "project:string:optional:Project slug"
+)]
 struct ListOntologyExtensionsTool;
 
 impl ToolBase for ListOntologyExtensionsTool {
@@ -705,6 +794,12 @@ impl AsyncTool<OntoreServer> for ListOntologyExtensionsTool {
 // ── Tool: get_ontology_extension
 // ────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_get_ontology_extension",
+    description = "Export a project-defined .ontology extension by stem (e.g. career, personal).",
+    category = "ontology",
+    params = "id:string:required:Extension stem; project:string:optional:Project slug"
+)]
 struct GetOntologyExtensionTool;
 
 impl ToolBase for GetOntologyExtensionTool {
@@ -771,6 +866,12 @@ impl AsyncTool<OntoreServer> for GetOntologyExtensionTool {
 // ── Tool: list_modes
 // ────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_list_modes",
+    description = "List all reflection modes with id, trigger, step count.",
+    category = "discovery",
+    params = "project:string:optional:Project slug"
+)]
 struct ListModesTool;
 
 impl ToolBase for ListModesTool {
@@ -842,6 +943,12 @@ impl AsyncTool<OntoreServer> for ListModesTool {
 // ── Tool: get_mode
 // ──────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_get_mode",
+    description = "Full reflection mode including all steps and preconditions.",
+    category = "ontology",
+    params = "id:string:required:Mode id; project:string:optional:Project slug"
+)]
 struct GetModeTool;
 
 impl ToolBase for GetModeTool {
@@ -904,6 +1011,13 @@ impl AsyncTool<OntoreServer> for GetModeTool {
 // ── Tool: get_backlog
 // ───────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_backlog_list",
+    description = "List backlog items, optionally filtered by status.",
+    category = "knowledge",
+    params = "project:string:optional:Project slug; \
+              status:string:optional:Open|InProgress|Done|Cancelled"
+)]
 struct GetBacklogTool;
 
 impl ToolBase for GetBacklogTool {
@@ -972,6 +1086,12 @@ impl AsyncTool<OntoreServer> for GetBacklogTool {
 // ── Tool: get_node
 // ──────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_get_node",
+    description = "Full ontology node by id substring match.",
+    category = "ontology",
+    params = "id:string:required:Node id substring; project:string:optional:Project slug"
+)]
 struct GetNodeTool;
 
 impl ToolBase for GetNodeTool {
@@ -1036,6 +1156,13 @@ impl AsyncTool<OntoreServer> for GetNodeTool {
 
 // ── Tool: ontoref_get (unified dispatcher) ───────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_get",
+    description = "Retrieve full details of any item by kind+id (from ontoref_search results).",
+    category = "ontology",
+    params = "id:string:required:Item id from ontoref_search; kind:string:required:node|adr|mode; \
+              project:string:optional:Project slug"
+)]
 struct GetTool;
 
 impl ToolBase for GetTool {
@@ -1088,6 +1215,12 @@ impl AsyncTool<OntoreServer> for GetTool {
 // ── Tool: ontoref_help
 // ────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_help",
+    description = "List all available ontoref MCP tools with descriptions, parameters, and the \
+                   current active project.",
+    category = "discovery"
+)]
 struct HelpTool;
 
 impl ToolBase for HelpTool {
@@ -1126,104 +1259,8 @@ impl AsyncTool<OntoreServer> for HelpTool {
             .and_then(|g| g.clone());
         let available = service.available_projects();
 
-        let tools = serde_json::json!([
-            { "name": "ontoref_help",          "description": "This help. Lists all tools, parameters, and current project.", "params": [] },
-            { "name": "ontoref_list_projects",  "description": "List all registered projects.",                               "params": [] },
-            { "name": "ontoref_set_project",    "description": "Set the default project for this session.",
-              "params": [{"name": "project", "required": true, "description": "Project slug from ontoref_list_projects"}] },
-            { "name": "ontoref_status",         "description": "Full dashboard: actors, notifications, backlog stats, ADR/mode counts, manifest.",
-              "params": [{"name": "project", "required": false}] },
-            { "name": "ontoref_search",         "description": "Free-text search across nodes, ADRs, and modes. Returns kind+id for use with ontoref_get.",
-              "params": [{"name": "query", "required": true}, {"name": "project", "required": false}] },
-            { "name": "ontoref_get",            "description": "Retrieve full details of any item by kind+id (from ontoref_search results).",
-              "params": [{"name": "id", "required": true}, {"name": "kind", "required": true, "values": ["node", "adr", "mode"]}, {"name": "project", "required": false}] },
-            { "name": "ontoref_describe",       "description": "Architecture overview: README summary and manifest.",
-              "params": [{"name": "project", "required": false}] },
-            { "name": "ontoref_list_adrs",      "description": "List all ADRs with id, title, status, constraint counts.",
-              "params": [{"name": "project", "required": false}] },
-            { "name": "ontoref_get_adr",        "description": "Full ADR by id or partial stem (e.g. adr-001).",
-              "params": [{"name": "id", "required": true}, {"name": "project", "required": false}] },
-            { "name": "ontoref_list_ontology_extensions", "description": "List extra .ontology/*.ncl files beyond core/state/gate.",
-              "params": [{"name": "project", "required": false}] },
-            { "name": "ontoref_get_ontology_extension",  "description": "Export a project-defined .ontology extension by stem (e.g. career, personal).",
-              "params": [{"name": "id", "required": true}, {"name": "project", "required": false}] },
-            { "name": "ontoref_list_modes",     "description": "List all reflection modes with id, trigger, step count.",
-              "params": [{"name": "project", "required": false}] },
-            { "name": "ontoref_get_mode",       "description": "Full reflection mode including all steps and preconditions.",
-              "params": [{"name": "id", "required": true}, {"name": "project", "required": false}] },
-            { "name": "ontoref_get_node",       "description": "Full ontology node by id substring match.",
-              "params": [{"name": "id", "required": true}, {"name": "project", "required": false}] },
-            { "name": "ontoref_backlog_list",   "description": "List backlog items, optionally filtered by status.",
-              "params": [{"name": "project", "required": false}, {"name": "status", "required": false, "values": ["Open", "InProgress", "Done", "Cancelled"]}] },
-            { "name": "ontoref_backlog",        "description": "Mutate backlog: add a new item or update an item's status.",
-              "params": [
-                {"name": "operation", "required": true, "values": ["add", "update_status"]},
-                {"name": "project",  "required": false},
-                {"name": "id",       "required": false, "note": "required for update_status"},
-                {"name": "status",   "required": false, "note": "required for update_status", "values": ["Open", "InProgress", "Done", "Cancelled"]},
-                {"name": "title",    "required": false, "note": "required for add"},
-                {"name": "kind",     "required": false, "values": ["Feature", "Bug", "Chore", "Research"]},
-                {"name": "priority", "required": false, "values": ["Critical", "High", "Medium", "Low"]},
-                {"name": "detail",   "required": false}
-              ] },
-            { "name": "ontoref_constraints",   "description": "All hard and soft architectural constraints extracted from all ADRs.",
-              "params": [{"name": "project", "required": false}] },
-            { "name": "ontoref_qa_list",        "description": "List Q&A entries from reflection/qa.ncl. Optionally filter by question substring.",
-              "params": [{"name": "project", "required": false}, {"name": "filter", "required": false}] },
-            { "name": "ontoref_qa_add",         "description": "Add a Q&A entry to reflection/qa.ncl (persisted to disk). Use to record architectural knowledge.",
-              "params": [
-                {"name": "question", "required": true},
-                {"name": "answer",   "required": false},
-                {"name": "actor",    "required": false, "default": "agent"},
-                {"name": "tags",     "required": false, "note": "array of strings"},
-                {"name": "related",  "required": false, "note": "array of node/ADR ids"},
-                {"name": "project",  "required": false}
-              ] },
-            { "name": "ontoref_action_list",    "description": "List quick actions from .ontoref/config.ncl quick_actions array.",
-              "params": [{"name": "project", "required": false}] },
-            { "name": "ontoref_action_add",     "description": "Create a new reflection mode file in reflection/modes/<id>.ncl.",
-              "params": [
-                {"name": "id",       "required": true, "note": "short slug, e.g. gen-docs"},
-                {"name": "label",    "required": true},
-                {"name": "trigger",  "required": true},
-                {"name": "steps",    "required": true, "note": "array of step strings"},
-                {"name": "icon",     "required": false, "values": ["book-open", "refresh", "code", "bolt"]},
-                {"name": "category", "required": false, "values": ["docs", "sync", "analysis", "test"]},
-                {"name": "actors",   "required": false, "note": "array: developer | agent | ci"},
-                {"name": "project",  "required": false}
-              ] },
-            { "name": "ontoref_validate_adrs",  "description": "Run typed constraint checks for all ADRs. Returns pass/fail per constraint with detail.",
-              "params": [{"name": "project", "required": false}] },
-            { "name": "ontoref_validate",       "description": "Run the full project validation suite: ADR constraints, content assets, connections, gate consistency.",
-              "params": [{"name": "project", "required": false}] },
-            { "name": "ontoref_impact",         "description": "BFS impact graph from a node: what else is affected if this node changes.",
-              "params": [
-                {"name": "node_id",          "required": true},
-                {"name": "depth",            "required": false, "default": 2},
-                {"name": "include_external", "required": false, "note": "traverse cross-project connections"},
-                {"name": "project",          "required": false}
-              ] },
-            { "name": "ontoref_guides",         "description": "Complete project guide: identity, axioms, practices, constraints, gate state, modes, actor policy, content assets, connections. Canonical entry point for cold-start context.",
-              "params": [{"name": "project", "required": false}, {"name": "actor", "required": false, "values": ["developer", "agent", "ci", "admin"]}] },
-            { "name": "ontoref_bookmark_list",  "description": "List saved search bookmarks for the project.",
-              "params": [{"name": "project", "required": false}, {"name": "filter", "required": false}] },
-            { "name": "ontoref_bookmark_add",   "description": "Save a search bookmark (node_id + title + optional tags).",
-              "params": [
-                {"name": "node_id", "required": true},
-                {"name": "title",   "required": true},
-                {"name": "kind",    "required": false, "values": ["node", "adr", "mode"]},
-                {"name": "tags",    "required": false, "note": "array of strings"},
-                {"name": "project", "required": false}
-              ] },
-            { "name": "ontoref_api_catalog",    "description": "Annotated daemon API surface: all HTTP routes with method, path, auth, actors, params, tags.",
-              "params": [
-                {"name": "actor", "required": false, "values": ["developer", "agent", "ci", "admin"]},
-                {"name": "tag",   "required": false, "note": "e.g. ontology, projects, search"},
-                {"name": "auth",  "required": false, "values": ["none", "viewer", "admin"]}
-              ] },
-            { "name": "ontoref_file_versions",  "description": "Per-file reload counters for ontology files. Counter increments on each daemon reload of that file.",
-              "params": [{"name": "project", "required": false}] },
-        ]);
+        let tools: Vec<serde_json::Value> =
+            catalog().iter().map(|t| render_tool_for_help(t)).collect();
 
         Ok(serde_json::json!({
             "current_project": current_project,
@@ -1236,6 +1273,12 @@ impl AsyncTool<OntoreServer> for HelpTool {
 
 // ── Tool: set_project ────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_set_project",
+    description = "Set the default project for this session.",
+    category = "discovery",
+    params = "project:string:required:Project slug from ontoref_list_projects"
+)]
 struct SetProjectTool;
 
 impl ToolBase for SetProjectTool {
@@ -1297,6 +1340,13 @@ impl AsyncTool<OntoreServer> for SetProjectTool {
 
 // ── Tool: project_status ─────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_status",
+    description = "Full dashboard: actors, notifications, backlog stats, ADR/mode counts, \
+                   manifest.",
+    category = "discovery",
+    params = "project:string:optional:Project slug"
+)]
 struct ProjectStatusTool;
 
 impl ToolBase for ProjectStatusTool {
@@ -1416,6 +1466,51 @@ impl AsyncTool<OntoreServer> for ProjectStatusTool {
             serde_json::Value::Null
         };
 
+        // Extract structured registry topology from manifest so MCP consumers
+        // don't need to navigate the raw NCL JSON themselves.
+        let registry = {
+            let rp = manifest.get("registry_provides");
+            let participant = rp
+                .and_then(|v| v.get("participant"))
+                .and_then(|v| v.as_str())
+                .unwrap_or("")
+                .to_owned();
+            let registries = rp.and_then(|v| v.get("registries"));
+            let default_id = registries
+                .and_then(|v| v.get("default"))
+                .and_then(|v| v.as_str())
+                .unwrap_or("primary");
+            let entry = registries
+                .and_then(|v| v.get("registries"))
+                .and_then(|v| v.as_array())
+                .and_then(|arr| {
+                    arr.iter()
+                        .find(|e| e.get("id").and_then(|i| i.as_str()) == Some(default_id))
+                })
+                .cloned()
+                .unwrap_or(serde_json::Value::Null);
+            let namespaces = entry
+                .get("namespaces")
+                .and_then(|n| n.get("own"))
+                .and_then(|v| v.as_array())
+                .cloned()
+                .unwrap_or_default();
+            let endpoint = entry
+                .get("endpoint")
+                .and_then(|v| v.as_str())
+                .unwrap_or("")
+                .to_owned();
+            if participant.is_empty() {
+                serde_json::Value::Null
+            } else {
+                serde_json::json!({
+                    "participant": participant,
+                    "endpoint":    endpoint,
+                    "namespaces":  namespaces,
+                })
+            }
+        };
+
         Ok(serde_json::json!({
             "project":       ctx.root.file_name().and_then(|n| n.to_str()).unwrap_or("unknown"),
             "root":          ctx.root.display().to_string(),
@@ -1424,6 +1519,7 @@ impl AsyncTool<OntoreServer> for ProjectStatusTool {
             "backlog":       backlog_stats,
             "adrs":          adr_count,
             "modes":         mode_count,
+            "registry":      registry,
             "manifest":      manifest,
         }))
     }
@@ -1432,6 +1528,18 @@ impl AsyncTool<OntoreServer> for ProjectStatusTool {
 // ── Tool: backlog
 // ─────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_backlog",
+    description = "Mutate backlog: add a new item or update an item's status.",
+    category = "knowledge",
+    params = "operation:string:required:add|update_status; project:string:optional:Project slug; \
+              id:string:optional:required for update_status; \
+              status:string:optional:Open|InProgress|Done|Cancelled (required for update_status); \
+              title:string:optional:required for add; \
+              kind:string:optional:Feature|Bug|Chore|Research; \
+              priority:string:optional:Critical|High|Medium|Low; detail:string:optional:Free-text \
+              detail"
+)]
 struct BacklogOpTool;
 
 impl ToolBase for BacklogOpTool {
@@ -1607,6 +1715,12 @@ fn collect_constraints(
 // ── Tool: get_constraints
 // ───────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_constraints",
+    description = "All hard and soft architectural constraints extracted from all ADRs.",
+    category = "validation",
+    params = "project:string:optional:Project slug"
+)]
 struct GetConstraintsTool;
 
 impl ToolBase for GetConstraintsTool {
@@ -1678,6 +1792,14 @@ impl AsyncTool<OntoreServer> for GetConstraintsTool {
 // ── Tool: qa_list
 // ────────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_qa_list",
+    description = "List Q&A entries from reflection/qa.ncl. Optionally filter by question \
+                   substring.",
+    category = "knowledge",
+    params = "project:string:optional:Project slug; filter:string:optional:Question substring \
+              filter"
+)]
 struct QaListTool;
 
 impl ToolBase for QaListTool {
@@ -1745,6 +1867,15 @@ impl AsyncTool<OntoreServer> for QaListTool {
 // ── Tool: qa_add
 // ─────────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_qa_add",
+    description = "Add a Q&A entry to reflection/qa.ncl (persisted to disk). Use to record \
+                   architectural knowledge.",
+    category = "knowledge",
+    params = "question:string:required:Question text; answer:string:optional:Answer text; \
+              actor:string:optional:default=agent; tags:array:optional:Tag strings; \
+              related:array:optional:Related node or ADR ids; project:string:optional:Project slug"
+)]
 struct QaAddTool;
 
 impl ToolBase for QaAddTool {
@@ -1822,6 +1953,12 @@ impl AsyncTool<OntoreServer> for QaAddTool {
 // ── Tool: action_list
 // ────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_action_list",
+    description = "List quick actions from .ontoref/config.ncl quick_actions array.",
+    category = "knowledge",
+    params = "project:string:optional:Project slug"
+)]
 struct ActionListTool;
 
 impl ToolBase for ActionListTool {
@@ -1879,6 +2016,16 @@ impl AsyncTool<OntoreServer> for ActionListTool {
 // ── Tool: action_add
 // ─────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_action_add",
+    description = "Create a new reflection mode file in reflection/modes/<id>.ncl.",
+    category = "knowledge",
+    params = "id:string:required:Mode id slug; label:string:required:Display label; \
+              trigger:string:required:Trigger description; steps:array:required:Step strings; \
+              icon:string:optional:book-open|refresh|code|bolt; \
+              category:string:optional:docs|sync|analysis|test; \
+              actors:array:optional:developer|agent|ci; project:string:optional:Project slug"
+)]
 struct ActionAddTool;
 
 impl ToolBase for ActionAddTool {
@@ -1980,6 +2127,13 @@ impl AsyncTool<OntoreServer> for ActionAddTool {
 // ── Tool: validate_adrs
 // ──────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_validate_adrs",
+    description = "Run typed constraint checks for all ADRs. Returns pass/fail per constraint \
+                   with detail.",
+    category = "validation",
+    params = "project:string:optional:Project slug"
+)]
 struct ValidateAdrsTool;
 
 impl ToolBase for ValidateAdrsTool {
@@ -2049,6 +2203,14 @@ struct ImpactInput {
     include_external: Option<bool>,
 }
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_impact",
+    description = "BFS impact graph from a node: what else is affected if this node changes.",
+    category = "ontology",
+    params = "node_id:string:required:Source node id; depth:u32:optional:default=2; \
+              include_external:bool:optional:Traverse cross-project edges; \
+              project:string:optional:Project slug"
+)]
 struct ImpactTool;
 
 impl ToolBase for ImpactTool {
@@ -2106,6 +2268,13 @@ impl AsyncTool<OntoreServer> for ImpactTool {
 // ── Tool: validate_project
 // ────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_validate",
+    description = "Run the full project validation suite: ADR constraints, content assets, \
+                   connections, gate consistency.",
+    category = "validation",
+    params = "project:string:optional:Project slug"
+)]
 struct ValidateProjectTool;
 
 impl ToolBase for ValidateProjectTool {
@@ -2173,6 +2342,14 @@ impl AsyncTool<OntoreServer> for ValidateProjectTool {
 // ── Tool: guides
 // ──────────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_guides",
+    description = "Complete project guide: identity, axioms, practices, constraints, gate state, \
+                   modes, actor policy, content assets, connections. Canonical entry point for \
+                   cold-start context.",
+    category = "discovery",
+    params = "project:string:optional:Project slug; actor:string:optional:developer|agent|ci|admin"
+)]
 struct GuidesTool;
 
 impl ToolBase for GuidesTool {
@@ -2233,6 +2410,14 @@ impl AsyncTool<OntoreServer> for GuidesTool {
 // ── Tool: api_catalog
 // ────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_api_catalog",
+    description = "Annotated daemon API surface: all HTTP routes with method, path, auth, actors, \
+                   params, tags.",
+    category = "discovery",
+    params = "actor:string:optional:developer|agent|ci|admin; tag:string:optional:Semantic tag \
+              filter; auth:string:optional:none|viewer|admin"
+)]
 struct ApiCatalogTool;
 
 impl ToolBase for ApiCatalogTool {
@@ -2305,6 +2490,13 @@ impl AsyncTool<OntoreServer> for ApiCatalogTool {
 // ── Tool: file_versions
 // ──────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_file_versions",
+    description = "Per-file reload counters for ontology files. Counter increments on each daemon \
+                   reload of that file.",
+    category = "discovery",
+    params = "project:string:optional:Project slug"
+)]
 struct FileVersionsTool;
 
 impl ToolBase for FileVersionsTool {
@@ -2433,6 +2625,12 @@ impl ServerHandler for OntoreServer {
 // ── Tool: bookmark_list
 // ─────────────────────────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_bookmark_list",
+    description = "List saved search bookmarks for the project.",
+    category = "knowledge",
+    params = "project:string:optional:Project slug; filter:string:optional:Substring filter"
+)]
 struct BookmarkListTool;
 
 impl ToolBase for BookmarkListTool {
@@ -2507,6 +2705,14 @@ impl AsyncTool<OntoreServer> for BookmarkListTool {
 // ── Tool: bookmark_add
 // ────────────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_bookmark_add",
+    description = "Save a search bookmark (node_id + title + optional tags).",
+    category = "knowledge",
+    params = "node_id:string:required:Node id; title:string:required:Bookmark title; \
+              kind:string:optional:node|adr|mode; tags:array:optional:Tag strings; \
+              project:string:optional:Project slug"
+)]
 struct BookmarkAddTool;
 
 impl ToolBase for BookmarkAddTool {
@@ -2588,6 +2794,13 @@ impl AsyncTool<OntoreServer> for BookmarkAddTool {
 // ── Tool: project_config (read full export or single section)
 // ─────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_project_config",
+    description = "Read a project's config surface. Returns the full NCL export (merged with \
+                   active overrides) for all sections, or a single section when section is given.",
+    category = "config",
+    params = "project:string:optional:Project slug; section:string:optional:Section id"
+)]
 struct ProjectConfigTool;
 
 impl ToolBase for ProjectConfigTool {
@@ -2658,6 +2871,13 @@ impl AsyncTool<OntoreServer> for ProjectConfigTool {
 // ── Tool: config_coherence
 // ──────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_config_coherence",
+    description = "Run the multi-consumer coherence check for a project's config surface. Reports \
+                   unclaimed NCL fields and consumer fields missing from the NCL export.",
+    category = "config",
+    params = "project:string:optional:Project slug; section:string:optional:Section id"
+)]
 struct ConfigCoherenceTool;
 
 impl ToolBase for ConfigCoherenceTool {
@@ -2725,6 +2945,14 @@ impl AsyncTool<OntoreServer> for ConfigCoherenceTool {
 // ── Tool: config_quickref
 // ───────────────────────────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_config_quickref",
+    description = "Generate living config documentation for a project. Combines current NCL \
+                   values, manifest rationales, per-section meta records, override history, and \
+                   coherence status.",
+    category = "config",
+    params = "project:string:optional:Project slug; section:string:optional:Section id"
+)]
 struct ConfigQuickrefTool;
 
 impl ToolBase for ConfigQuickrefTool {
@@ -2790,6 +3018,16 @@ impl AsyncTool<OntoreServer> for ConfigQuickrefTool {
 // ── Tool: config_update (override layer mutation)
 // ────────────────────────────
 
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_config_update",
+    description = "Mutate a config section via the override layer. Generates a \
+                   {section}.overrides.ncl file merged after the original NCL. dry_run defaults \
+                   to true.",
+    category = "config",
+    params = "project:string:required:Project slug; section:string:required:Section id; \
+              values:json:required:JSON object with field updates; \
+              dry_run:bool:optional:default=true; reason:string:optional:Audit reason"
+)]
 struct ConfigUpdateTool;
 
 impl ToolBase for ConfigUpdateTool {
@@ -2901,3 +3139,235 @@ pub async fn serve_stdio(state: AppState) -> anyhow::Result<()> {
         .map_err(|e| anyhow::anyhow!("MCP wait error: {e}"))?;
     Ok(())
 }
+
+// ── Tool: vault_status (ADR-017 / ADR-019) ──────────────────────────────────
+
+#[ontoref_derive::onto_mcp_tool(
+    name = "ontoref_vault_status",
+    description = "Read-only credential vault state: vault_id, recipient count, recipient groups \
+                   + path rules, last logged operation, master_key resolution. No decryption.",
+    category = "credentials",
+    params = "project:string:optional:Project slug (defaults to current project)"
+)]
+struct VaultStatusTool;
+
+impl ToolBase for VaultStatusTool {
+    type Parameter = ProjectParam;
+    type Output = serde_json::Value;
+    type Error = ToolError;
+
+    fn name() -> Cow<'static, str> {
+        "ontoref_vault_status".into()
+    }
+
+    fn description() -> Option<Cow<'static, str>> {
+        Some(
+            "Structured credential vault state for the active project (ADR-017 + ADR-019). \
+             Returns: vault_id, vault_dir, backend, registry_endpoint, master_key (path or \
+             unresolved-error), access_sops_present (bool), recipient_count, recipient_groups \
+             (group→key list), recipient_rules (path→groups), actor_key_bindings, last_operation \
+             from logs/access.jsonl. NEVER decrypts credentials."
+                .into(),
+        )
+    }
+
+    fn output_schema() -> Option<Arc<JsonObject>> {
+        None
+    }
+}
+
+impl AsyncTool<OntoreServer> for VaultStatusTool {
+    async fn invoke(
+        service: &OntoreServer,
+        param: ProjectParam,
+    ) -> Result<serde_json::Value, ToolError> {
+        debug!(tool = "vault_status", project = ?param.project);
+        let ctx = service.project_ctx(param.project.as_deref());
+
+        let project_ncl = ctx.root.join(".ontoref").join("project.ncl");
+        if !project_ncl.exists() {
+            return Ok(serde_json::json!({
+                "ok": false,
+                "reason": format!("no .ontoref/project.ncl at {}", ctx.root.display()),
+            }));
+        }
+
+        // Evaluate project.ncl via the cache (consistent with other tools).
+        let project_json: serde_json::Value = ctx
+            .cache
+            .export(&project_ncl, ctx.import_path.as_deref())
+            .await
+            .map(|(v, _)| v)
+            .map_err(|e| ToolError(format!("project.ncl evaluation failed: {e}")))?;
+
+        let sops = match project_json.get("sops") {
+            Some(s) if !s.is_null() => s,
+            _ => {
+                return Ok(serde_json::json!({
+                    "ok": false,
+                    "reason": "project.ncl has no sops block — vault not configured",
+                }));
+            }
+        };
+        let enabled = sops
+            .get("enabled")
+            .and_then(|v| v.as_bool())
+            .unwrap_or(false);
+        if !enabled {
+            return Ok(serde_json::json!({
+                "ok": false,
+                "reason": "sops.enabled = false in project.ncl",
+            }));
+        }
+
+        let slug = project_json
+            .get("slug")
+            .and_then(|v| v.as_str())
+            .unwrap_or("");
+        let vault_id = sops
+            .get("vault_id")
+            .and_then(|v| v.as_str())
+            .filter(|s| !s.is_empty())
+            .unwrap_or(slug)
+            .to_string();
+        let vault_backend = sops
+            .get("vault_backend")
+            .and_then(|v| v.as_str())
+            .unwrap_or("restic")
+            .to_string();
+        let registry_endpoint = sops
+            .get("registry_endpoint")
+            .and_then(|v| v.as_str())
+            .unwrap_or("")
+            .to_string();
+        let actor_key_bindings = sops
+            .get("actor_key_bindings")
+            .cloned()
+            .unwrap_or(serde_json::json!({}));
+        let recipient_groups = sops
+            .get("recipient_groups")
+            .cloned()
+            .unwrap_or(serde_json::json!({}));
+        let recipient_rules = sops
+            .get("recipient_rules")
+            .cloned()
+            .unwrap_or(serde_json::json!([]));
+
+        // master_key path: project override → global config.
+        let master_key = sops
+            .get("master_key_path")
+            .and_then(|v| v.as_str())
+            .filter(|s| !s.is_empty())
+            .map(String::from)
+            .or_else(|| {
+                let home = std::env::var_os("HOME").map(std::path::PathBuf::from)?;
+                let global = home.join(".config").join("ontoref").join("config.ncl");
+                if !global.exists() {
+                    return None;
+                }
+                let out = std::process::Command::new("nickel")
+                    .arg("export")
+                    .arg(&global)
+                    .output()
+                    .ok()?;
+                if !out.status.success() {
+                    return None;
+                }
+                let v: serde_json::Value = serde_json::from_slice(&out.stdout).ok()?;
+                v.get("vault")
+                    .and_then(|x| x.get("master_key_path"))
+                    .and_then(|s| s.as_str())
+                    .filter(|s| !s.is_empty())
+                    .map(String::from)
+            });
+
+        let home = std::env::var_os("HOME")
+            .map(std::path::PathBuf::from)
+            .ok_or_else(|| ToolError("HOME not resolvable".into()))?;
+        let vault_dir = home
+            .join(".config")
+            .join("ontoref")
+            .join("vaults")
+            .join(&vault_id);
+        let access_sops = vault_dir.join("access.sops.yaml");
+        let logs_jsonl = vault_dir.join("logs").join("access.jsonl");
+
+        // Recipient count from access.sops.yaml's plaintext sops.age metadata.
+        // The file format places one '- recipient: age1...' entry per recipient
+        // in the sops metadata block; counting matching lines avoids needing a
+        // YAML parser dependency in the daemon.
+        let recipient_count: u64 = if access_sops.exists() {
+            std::fs::read_to_string(&access_sops)
+                .map(|s| {
+                    s.lines()
+                        .filter(|l| l.trim_start().starts_with("- recipient: age1"))
+                        .count() as u64
+                })
+                .unwrap_or(0)
+        } else {
+            0
+        };
+
+        // Last operation from access.jsonl.
+        let last_op = if logs_jsonl.exists() {
+            std::fs::read_to_string(&logs_jsonl)
+                .ok()
+                .and_then(|s| s.lines().last().map(String::from))
+                .and_then(|line| serde_json::from_str::<serde_json::Value>(&line).ok())
+        } else {
+            None
+        };
+
+        Ok(serde_json::json!({
+            "ok":                  true,
+            "vault_id":            vault_id,
+            "vault_dir":           vault_dir.display().to_string(),
+            "backend":             vault_backend,
+            "registry_endpoint":   registry_endpoint,
+            "master_key":          master_key,
+            "access_sops_present": access_sops.exists(),
+            "recipient_count":     recipient_count,
+            "recipient_groups":    recipient_groups,
+            "recipient_rules":     recipient_rules,
+            "actor_key_bindings":  actor_key_bindings,
+            "last_operation":      last_op,
+        }))
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn catalog_contains_known_tools_and_unique_names() {
+        let entries = catalog();
+        assert!(
+            entries.len() >= 30,
+            "expected at least 30 MCP tools in inventory, found {}",
+            entries.len()
+        );
+
+        let names: Vec<&str> = entries.iter().map(|t| t.name).collect();
+        assert!(names.contains(&"ontoref_help"));
+        assert!(names.contains(&"ontoref_guides"));
+        assert!(names.contains(&"ontoref_status"));
+        assert!(names.contains(&"ontoref_api_catalog"));
+
+        let mut sorted_unique = names.clone();
+        sorted_unique.sort_unstable();
+        sorted_unique.dedup();
+        assert_eq!(
+            sorted_unique.len(),
+            names.len(),
+            "duplicate tool names in inventory: {names:?}"
+        );
+
+        for name in &names {
+            assert!(
+                name.starts_with("ontoref_"),
+                "tool name '{name}' does not start with 'ontoref_'"
+            );
+        }
+    }
+}
diff --git a/crates/ontoref-daemon/src/registry.rs b/crates/ontoref-daemon/src/registry.rs
index 2379544..d82d241 100644
--- a/crates/ontoref-daemon/src/registry.rs
+++ b/crates/ontoref-daemon/src/registry.rs
@@ -442,9 +442,25 @@ impl ProjectRegistry {
 
 /// Resolve nickel_import_paths entries into a colon-separated
 /// NICKEL_IMPORT_PATH string. Relative paths are resolved against the project
-/// root.
+/// root. The user-global schemas directory `~/.config/ontoref/schemas/` is
+/// always appended so files importing `ontoref-project.ncl` (the canonical
+/// project schema) resolve correctly even when the project's local
+/// nickel_import_paths does not include it.
+fn global_schemas_path() -> Option<String> {
+    let home = std::env::var_os("HOME")?;
+    let p = Path::new(&home)
+        .join(".config")
+        .join("ontoref")
+        .join("schemas");
+    p.is_dir().then(|| p.display().to_string())
+}
+
 fn resolve_import_path(paths: &[String], root: &Path) -> Option<String> {
-    let joined = paths
+    resolve_import_path_inner(paths, root, global_schemas_path().as_deref())
+}
+
+fn resolve_import_path_inner(paths: &[String], root: &Path, extra: Option<&str>) -> Option<String> {
+    let mut entries: Vec<String> = paths
         .iter()
         .map(|p| {
             let candidate = Path::new(p);
@@ -454,8 +470,13 @@ fn resolve_import_path(paths: &[String], root: &Path) -> Option<String> {
                 root.join(candidate).display().to_string()
             }
         })
-        .collect::<Vec<_>>()
-        .join(":");
+        .collect();
+    if let Some(s) = extra {
+        if !entries.iter().any(|e| e == s) {
+            entries.push(s.to_string());
+        }
+    }
+    let joined = entries.join(":");
     if joined.is_empty() {
         None
     } else {
@@ -758,28 +779,55 @@ mod tests {
 
     #[test]
     fn empty_paths_returns_none() {
-        assert_eq!(resolve_import_path(&[], Path::new("/root")), None);
+        assert_eq!(
+            resolve_import_path_inner(&[], Path::new("/root"), None),
+            None
+        );
     }
 
     #[test]
     fn absolute_path_returned_as_is() {
-        let result = resolve_import_path(&["/abs/path".to_string()], Path::new("/root"));
+        let result =
+            resolve_import_path_inner(&["/abs/path".to_string()], Path::new("/root"), None);
         assert_eq!(result, Some("/abs/path".to_string()));
     }
 
     #[test]
     fn relative_path_joined_with_root() {
-        let result = resolve_import_path(&["lib".to_string()], Path::new("/root"));
+        let result = resolve_import_path_inner(&["lib".to_string()], Path::new("/root"), None);
         assert_eq!(result, Some("/root/lib".to_string()));
     }
 
     #[test]
     fn multiple_paths_colon_separated() {
-        let result =
-            resolve_import_path(&["/abs".to_string(), "rel".to_string()], Path::new("/root"));
+        let result = resolve_import_path_inner(
+            &["/abs".to_string(), "rel".to_string()],
+            Path::new("/root"),
+            None,
+        );
         assert_eq!(result, Some("/abs:/root/rel".to_string()));
     }
 
+    #[test]
+    fn extra_global_schema_appended() {
+        let result = resolve_import_path_inner(
+            &["/abs".to_string()],
+            Path::new("/root"),
+            Some("/global/schemas"),
+        );
+        assert_eq!(result, Some("/abs:/global/schemas".to_string()));
+    }
+
+    #[test]
+    fn extra_global_schema_not_duplicated() {
+        let result = resolve_import_path_inner(
+            &["/global/schemas".to_string()],
+            Path::new("/root"),
+            Some("/global/schemas"),
+        );
+        assert_eq!(result, Some("/global/schemas".to_string()));
+    }
+
     // ── with_primary slug collision ───────────────────────────────────────
 
     fn minimal_ctx(slug: &str) -> Arc<ProjectContext> {
diff --git a/crates/ontoref-daemon/src/ui/handlers.rs b/crates/ontoref-daemon/src/ui/handlers.rs
index cb197cc..35a703c 100644
--- a/crates/ontoref-daemon/src/ui/handlers.rs
+++ b/crates/ontoref-daemon/src/ui/handlers.rs
@@ -247,6 +247,141 @@ impl Default for DomainCtx {
     }
 }
 
+/// Extract registry participant, namespaces, and endpoint from manifest JSON.
+/// Returns empty strings/vec when `registry_provides` is absent.
+/// Pure projection of `.ontoref/project.ncl::sops` into vault basics.
+/// Returns `Some((vault_id, vault_mode))` when sops is enabled, `None`
+/// otherwise. vault_mode: "declarative" if `recipient_rules` is a non-empty
+/// array, "legacy" otherwise.
+fn vault_basics_from_json(project_json: Option<&serde_json::Value>) -> Option<(String, String)> {
+    let pj = project_json?;
+    let sops = pj.get("sops")?;
+    let enabled = sops
+        .get("enabled")
+        .and_then(|v| v.as_bool())
+        .unwrap_or(false);
+    if !enabled {
+        return None;
+    }
+    let slug = pj.get("slug").and_then(|v| v.as_str()).unwrap_or("");
+    let vault_id = sops
+        .get("vault_id")
+        .and_then(|v| v.as_str())
+        .filter(|s| !s.is_empty())
+        .unwrap_or(slug)
+        .to_string();
+    let has_rules = sops
+        .get("recipient_rules")
+        .and_then(|v| v.as_array())
+        .map(|a| !a.is_empty())
+        .unwrap_or(false);
+    let mode = if has_rules { "declarative" } else { "legacy" }.to_string();
+    Some((vault_id, mode))
+}
+
+/// Extract vault state for the UI from a project.ncl JSON
+/// (.ontoref/project.ncl::sops). Returns (has_vault, vault_id, vault_mode,
+/// recipient_count, last_op_summary). recipient_count is parsed from
+/// access.sops.yaml metadata when the file exists locally.
+fn extract_vault_ctx(
+    project_json: Option<&serde_json::Value>,
+) -> (bool, String, String, i32, String) {
+    let Some((vault_id, mode)) = vault_basics_from_json(project_json) else {
+        return (false, String::new(), String::new(), 0, String::new());
+    };
+
+    // Local files — best-effort, no failure on absence.
+    let home = std::env::var_os("HOME").map(std::path::PathBuf::from);
+    let (recipient_count, last_op_summary) = if let Some(home) = home {
+        let vault_dir = home
+            .join(".config")
+            .join("ontoref")
+            .join("vaults")
+            .join(&vault_id);
+        let access_sops = vault_dir.join("access.sops.yaml");
+        let count: i32 = if access_sops.exists() {
+            std::fs::read_to_string(&access_sops)
+                .map(|s| {
+                    // sops emits both YAML (`- recipient: age1...`) and JSON
+                    // (`"recipient": "age1..."`). Match the substring that
+                    // both share — any line containing `recipient` followed
+                    // by an `age1` token.
+                    s.lines()
+                        .filter(|l| l.contains("recipient") && l.contains("age1"))
+                        .count() as i32
+                })
+                .unwrap_or(0)
+        } else {
+            0
+        };
+        let logs = vault_dir.join("logs").join("access.jsonl");
+        let last = if logs.exists() {
+            std::fs::read_to_string(&logs)
+                .ok()
+                .and_then(|s| s.lines().last().map(String::from))
+                .and_then(|line| serde_json::from_str::<serde_json::Value>(&line).ok())
+                .map(|v| {
+                    let op = v.get("op").and_then(|x| x.as_str()).unwrap_or("?");
+                    let actor = v.get("actor").and_then(|x| x.as_str()).unwrap_or("?");
+                    let ts = v.get("ts").and_then(|x| x.as_str()).unwrap_or("");
+                    format!("{op} by {actor} @ {ts}")
+                })
+                .unwrap_or_default()
+        } else {
+            String::new()
+        };
+        (count, last)
+    } else {
+        (0, String::new())
+    };
+    (true, vault_id, mode, recipient_count, last_op_summary)
+}
+
+fn extract_registry_ctx(
+    manifest_json: Option<&serde_json::Value>,
+) -> (String, Vec<String>, String) {
+    let Some(mj) = manifest_json else {
+        return (String::new(), Vec::new(), String::new());
+    };
+    let rp = mj.get("registry_provides");
+    let participant = rp
+        .and_then(|v| v.get("participant"))
+        .and_then(|v| v.as_str())
+        .unwrap_or("")
+        .to_owned();
+    if participant.is_empty() {
+        return (String::new(), Vec::new(), String::new());
+    }
+    let registries = rp.and_then(|v| v.get("registries"));
+    let default_id = registries
+        .and_then(|v| v.get("default"))
+        .and_then(|v| v.as_str())
+        .unwrap_or("primary");
+    let entry = registries
+        .and_then(|v| v.get("registries"))
+        .and_then(|v| v.as_array())
+        .and_then(|arr| {
+            arr.iter()
+                .find(|e| e.get("id").and_then(|i| i.as_str()) == Some(default_id))
+        });
+    let namespaces = entry
+        .and_then(|e| e.get("namespaces"))
+        .and_then(|n| n.get("own"))
+        .and_then(|v| v.as_array())
+        .map(|arr| {
+            arr.iter()
+                .filter_map(|v| v.as_str().map(str::to_owned))
+                .collect()
+        })
+        .unwrap_or_default();
+    let endpoint = entry
+        .and_then(|e| e.get("endpoint"))
+        .and_then(|v| v.as_str())
+        .unwrap_or("")
+        .to_owned();
+    (participant, namespaces, endpoint)
+}
+
 /// Resolve domain_origin and domain_provides context from a manifest JSON
 /// value.
 async fn resolve_domain_ctx(
@@ -328,8 +463,8 @@ async fn resolve_domain_ctx(
 }
 
 /// Insert logo and MCP metadata into a Tera context.
-/// Logos are loaded from `.ontoref/config.ncl`; MCP availability is
-/// compile-time.
+/// Logos are loaded from `.ontoref/config.ncl`; service flags reflect runtime
+/// state from `AppState::service_flags`.
 pub(crate) async fn insert_brand_ctx(
     ctx: &mut Context,
     root: &std::path::Path,
@@ -337,11 +472,12 @@ pub(crate) async fn insert_brand_ctx(
     import_path: Option<&str>,
     base_url: &str,
     registry: &Arc<crate::registry::ProjectRegistry>,
+    flags: &crate::api::ServiceFlags,
 ) {
     let (logo, logo_dark) = load_logos(root, cache, import_path, base_url).await;
     ctx.insert("logo", &logo);
     ctx.insert("logo_dark", &logo_dark);
-    insert_mcp_ctx(ctx);
+    insert_mcp_ctx(ctx, flags);
     // Expose repo + docs URLs from card for file link generation in templates.
     let (repo_url, docs_url) = if let Some(json) = load_config_json(root, cache, import_path).await
     {
@@ -382,16 +518,20 @@ pub(crate) async fn insert_brand_ctx(
         .unwrap_or(false);
     ctx.insert("has_provisioning", &is_provisioning);
 
-    // Domain ontology layer — manifest.domain_origin and manifest.domain_provides.
-    // Surfaces the two-layer ontology distinction in the UI header and sidebar.
+    // Domain ontology layer + registry topology — both read from manifest.ncl.
     let manifest_path = root.join(".ontology").join("manifest.ncl");
-    let domain = if manifest_path.exists() {
+    let manifest_json: Option<serde_json::Value> = if manifest_path.exists() {
         match cache.export(&manifest_path, import_path).await {
-            Ok((mj, _)) => resolve_domain_ctx(&mj, registry).await,
-            Err(_) => DomainCtx::default(),
+            Ok((mj, _)) => Some(mj),
+            Err(_) => None,
         }
     } else {
-        DomainCtx::default()
+        None
+    };
+
+    let domain = match &manifest_json {
+        Some(mj) => resolve_domain_ctx(mj, registry).await,
+        None => DomainCtx::default(),
     };
     ctx.insert("has_domain_origin", &domain.has_origin);
     ctx.insert("domain_origin_id", &domain.origin_id);
@@ -404,14 +544,50 @@ pub(crate) async fn insert_brand_ctx(
     ctx.insert("domain_provides_name", &domain.provides_name);
     ctx.insert("domain_provides_kind", &domain.provides_kind);
     ctx.insert("domain_impl_projects", &domain.impl_projects);
+
+    // Registry topology — manifest.registry_provides (default registry entry)
+    let (registry_participant, registry_namespaces, registry_endpoint) =
+        extract_registry_ctx(manifest_json.as_ref());
+    ctx.insert("has_registry", &!registry_participant.is_empty());
+    ctx.insert("registry_participant", &registry_participant);
+    ctx.insert("registry_namespaces", &registry_namespaces);
+    ctx.insert("registry_endpoint", &registry_endpoint);
+
+    // Vault state — project.ncl::sops (ADR-017 + ADR-019)
+    let project_ncl_path = root.join(".ontoref").join("project.ncl");
+    let project_json: Option<serde_json::Value> = if project_ncl_path.exists() {
+        match cache.export(&project_ncl_path, import_path).await {
+            Ok((pj, _)) => Some(pj),
+            Err(_) => None,
+        }
+    } else {
+        None
+    };
+    let (has_vault, vault_id, vault_mode, vault_recipients, vault_last_op) =
+        extract_vault_ctx(project_json.as_ref());
+    ctx.insert("has_vault", &has_vault);
+    ctx.insert("vault_id", &vault_id);
+    ctx.insert("vault_mode", &vault_mode);
+    ctx.insert("vault_recipients", &vault_recipients);
+    ctx.insert("vault_last_op", &vault_last_op);
 }
 
-/// Insert MCP metadata and daemon version into a Tera context.
-pub(crate) fn insert_mcp_ctx(ctx: &mut Context) {
+/// Insert MCP and GraphQL runtime state and daemon version into a Tera context.
+pub(crate) fn insert_mcp_ctx(ctx: &mut Context, flags: &crate::api::ServiceFlags) {
     ctx.insert("daemon_version", env!("CARGO_PKG_VERSION"));
+    #[cfg(feature = "graphql")]
+    ctx.insert(
+        "graphql_active",
+        &flags.graphql.load(std::sync::atomic::Ordering::Relaxed),
+    );
+    #[cfg(not(feature = "graphql"))]
+    ctx.insert("graphql_active", &false);
     #[cfg(feature = "mcp")]
     {
-        ctx.insert("mcp_active", &true);
+        ctx.insert(
+            "mcp_active",
+            &flags.mcp.load(std::sync::atomic::Ordering::Relaxed),
+        );
         ctx.insert(
             "mcp_tools",
             &[
@@ -495,6 +671,7 @@ pub async fn dashboard(State(state): State<AppState>) -> Result<Html<String>, Ui
         state.nickel_import_path.as_deref(),
         "/ui",
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -527,6 +704,7 @@ pub async fn graph(State(state): State<AppState>) -> Result<Html<String>, UiErro
         state.nickel_import_path.as_deref(),
         "/ui",
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -587,6 +765,7 @@ pub async fn sessions(State(state): State<AppState>) -> Result<Html<String>, UiE
         state.nickel_import_path.as_deref(),
         "/ui",
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -636,6 +815,7 @@ pub async fn notifications_page(State(state): State<AppState>) -> Result<Html<St
         state.nickel_import_path.as_deref(),
         "/ui",
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -663,6 +843,7 @@ pub async fn search_page(State(state): State<AppState>) -> Result<Html<String>,
         state.nickel_import_path.as_deref(),
         "/ui",
         &state.registry,
+        &state.service_flags,
     )
     .await;
     render(tera, "pages/search.html", &ctx).await
@@ -694,6 +875,7 @@ pub async fn search_page_mp(
         ctx_ref.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
     render(tera, "pages/search.html", &ctx).await
@@ -764,6 +946,7 @@ pub async fn modes(State(state): State<AppState>) -> Result<Html<String>, UiErro
         state.nickel_import_path.as_deref(),
         "/ui",
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -854,48 +1037,77 @@ pub async fn project_picker(State(state): State<AppState>) -> Result<Html<String
             (vec![], 0)
         };
 
-        // Manifest — layers, operational_modes, default_mode, repo_kind
+        // Manifest — layers, operational_modes, default_mode, repo_kind, registry
         let manifest_path = proj.root.join(".ontology").join("manifest.ncl");
-        let (layers, op_modes, default_mode, repo_kind) = if manifest_path.exists() {
-            match proj
-                .cache
+        let manifest_json: Option<serde_json::Value> = if manifest_path.exists() {
+            proj.cache
                 .export(&manifest_path, proj.import_path.as_deref())
                 .await
-            {
-                Ok((json, _)) => {
-                    let layers: Vec<serde_json::Value> = json
-                        .get("layers")
-                        .and_then(|v| v.as_array())
-                        .map(|arr| arr.iter().map(|l| serde_json::json!({
-                            "id":          l.get("id").and_then(|v| v.as_str()).unwrap_or(""),
-                            "description": l.get("description").and_then(|v| v.as_str()).unwrap_or(""),
-                        })).collect())
-                        .unwrap_or_default();
-                    let op_modes: Vec<serde_json::Value> = json
-                        .get("operational_modes")
-                        .and_then(|v| v.as_array())
-                        .map(|arr| arr.iter().map(|m| serde_json::json!({
-                            "id":          m.get("id").and_then(|v| v.as_str()).unwrap_or(""),
-                            "description": m.get("description").and_then(|v| v.as_str()).unwrap_or(""),
-                        })).collect())
-                        .unwrap_or_default();
-                    let default_mode = json
-                        .get("default_mode")
-                        .and_then(|v| v.as_str())
-                        .unwrap_or("")
-                        .to_string();
-                    let repo_kind = json
-                        .get("repo_kind")
-                        .and_then(|v| v.as_str())
-                        .unwrap_or("")
-                        .to_string();
-                    (layers, op_modes, default_mode, repo_kind)
-                }
-                Err(_) => (vec![], vec![], String::new(), String::new()),
-            }
+                .ok()
+                .map(|(j, _)| j)
         } else {
-            (vec![], vec![], String::new(), String::new())
+            None
         };
+        let (layers, op_modes, default_mode, repo_kind) = match &manifest_json {
+            Some(json) => {
+                let layers: Vec<serde_json::Value> = json
+                    .get("layers")
+                    .and_then(|v| v.as_array())
+                    .map(|arr| {
+                        arr.iter()
+                            .map(|l| {
+                                serde_json::json!({
+                                    "id":          l.get("id").and_then(|v| v.as_str()).unwrap_or(""),
+                                    "description": l.get("description").and_then(|v| v.as_str()).unwrap_or(""),
+                                })
+                            })
+                            .collect()
+                    })
+                    .unwrap_or_default();
+                let op_modes: Vec<serde_json::Value> = json
+                    .get("operational_modes")
+                    .and_then(|v| v.as_array())
+                    .map(|arr| {
+                        arr.iter()
+                            .map(|m| {
+                                serde_json::json!({
+                                    "id":          m.get("id").and_then(|v| v.as_str()).unwrap_or(""),
+                                    "description": m.get("description").and_then(|v| v.as_str()).unwrap_or(""),
+                                })
+                            })
+                            .collect()
+                    })
+                    .unwrap_or_default();
+                let default_mode = json
+                    .get("default_mode")
+                    .and_then(|v| v.as_str())
+                    .unwrap_or("")
+                    .to_string();
+                let repo_kind = json
+                    .get("repo_kind")
+                    .and_then(|v| v.as_str())
+                    .unwrap_or("")
+                    .to_string();
+                (layers, op_modes, default_mode, repo_kind)
+            }
+            None => (vec![], vec![], String::new(), String::new()),
+        };
+        let (registry_participant, registry_namespaces, registry_endpoint) =
+            extract_registry_ctx(manifest_json.as_ref());
+
+        // Vault — read project.ncl::sops (ADR-017 + ADR-019)
+        let project_ncl_path = proj.root.join(".ontoref").join("project.ncl");
+        let project_json: Option<serde_json::Value> = if project_ncl_path.exists() {
+            proj.cache
+                .export(&project_ncl_path, proj.import_path.as_deref())
+                .await
+                .ok()
+                .map(|(j, _)| j)
+        } else {
+            None
+        };
+        let (has_vault, vault_id, vault_mode, vault_recipients, vault_last_op) =
+            extract_vault_ctx(project_json.as_ref());
 
         // card — loaded from `.ontoref/config.ncl` `card` field (which imports
         // ../card.ncl)
@@ -934,25 +1146,33 @@ pub async fn project_picker(State(state): State<AppState>) -> Result<Html<String
         let opmode = if proj.push_only { "push" } else { "daemon" };
 
         projects.push(serde_json::json!({
-            "slug":          proj.slug,
-            "root":          proj.root.display().to_string(),
-            "auth":          proj.auth_enabled(),
-            "card":          card,
-            "description":   description,
-            "default_mode":  default_mode,
-            "repo_kind":     repo_kind,
-            "repos":         repos,
-            "session_count": sessions.len(),
-            "sessions":      sessions,
-            "notif_count":   notifications.len(),
-            "notifications": notifications,
-            "backlog_open":  backlog_open,
-            "backlog_items": backlog_items,
-            "layers":        layers,
-            "op_modes":      op_modes,
-            "showcase":      showcase,
-            "generated":     generated,
-            "opmode":        opmode,
+            "slug":                 proj.slug,
+            "root":                 proj.root.display().to_string(),
+            "auth":                 proj.auth_enabled(),
+            "card":                 card,
+            "description":          description,
+            "default_mode":         default_mode,
+            "repo_kind":            repo_kind,
+            "repos":                repos,
+            "session_count":        sessions.len(),
+            "sessions":             sessions,
+            "notif_count":          notifications.len(),
+            "notifications":        notifications,
+            "backlog_open":         backlog_open,
+            "backlog_items":        backlog_items,
+            "layers":               layers,
+            "op_modes":             op_modes,
+            "showcase":             showcase,
+            "generated":            generated,
+            "opmode":               opmode,
+            "registry_participant": registry_participant,
+            "registry_namespaces":  registry_namespaces,
+            "registry_endpoint":    registry_endpoint,
+            "has_vault":            has_vault,
+            "vault_id":             vault_id,
+            "vault_mode":           vault_mode,
+            "vault_recipients":     vault_recipients,
+            "vault_last_op":        vault_last_op,
         }));
     }
 
@@ -960,7 +1180,7 @@ pub async fn project_picker(State(state): State<AppState>) -> Result<Html<String
     ctx.insert("projects", &projects);
     ctx.insert("base_url", "/ui");
     ctx.insert("hide_project_nav", &true);
-    insert_mcp_ctx(&mut ctx);
+    insert_mcp_ctx(&mut ctx, &state.service_flags);
     render(tera, "pages/project_picker.html", &ctx).await
 }
 
@@ -1228,6 +1448,7 @@ pub async fn dashboard_mp(
         ctx_ref.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -1347,6 +1568,7 @@ pub async fn api_catalog_page_mp(
         ctx_ref.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -1460,6 +1682,7 @@ pub async fn graph_mp(
         ctx_ref.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -1512,6 +1735,7 @@ pub async fn sessions_mp(
         ctx_ref.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -1597,6 +1821,7 @@ pub async fn notifications_mp(
         ctx_ref.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -2563,6 +2788,7 @@ pub async fn backlog_page(State(state): State<AppState>) -> Result<Html<String>,
         state.nickel_import_path.as_deref(),
         "/ui",
         &state.registry,
+        &state.service_flags,
     )
     .await;
     render(tera, "pages/backlog.html", &ctx).await
@@ -2599,6 +2825,7 @@ pub async fn backlog_page_mp(
         proj.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
     render(tera, "pages/backlog.html", &ctx).await
@@ -3140,10 +3367,48 @@ pub async fn manage_page(State(state): State<AppState>) -> Result<Html<String>,
     ctx.insert("hide_project_nav", &true);
     ctx.insert("error", &Option::<String>::None);
     ctx.insert("daemon_admin_enabled", &state.daemon_admin_hash.is_some());
-    insert_mcp_ctx(&mut ctx);
+    insert_mcp_ctx(&mut ctx, &state.service_flags);
     render(tera, "pages/manage.html", &ctx).await
 }
 
+/// HTMX endpoint — toggle a runtime service (mcp/graphql) on or off.
+/// AdminGuard ensures only authenticated daemon admins can call this.
+pub async fn service_toggle_ui(
+    State(state): State<AppState>,
+    _guard: super::auth::AdminGuard,
+    axum::extract::Path(service): axum::extract::Path<String>,
+) -> impl axum::response::IntoResponse {
+    use std::sync::atomic::Ordering;
+
+    use axum::http::StatusCode;
+
+    let new_state = state.service_flags.set(&service, {
+        // Toggle: read current, flip it.
+        match service.as_str() {
+            #[cfg(feature = "mcp")]
+            "mcp" => !state.service_flags.mcp.load(Ordering::Relaxed),
+            #[cfg(feature = "graphql")]
+            "graphql" => !state.service_flags.graphql.load(Ordering::Relaxed),
+            _ => return (StatusCode::NOT_FOUND, "unknown service").into_response(),
+        }
+    });
+
+    let Some(enabled) = new_state else {
+        return (StatusCode::NOT_FOUND, "unknown or not compiled service").into_response();
+    };
+
+    tracing::info!(service = %service, enabled = enabled, "service toggled via UI");
+
+    // Return an HTMX-friendly fragment: just the badge so the button can swap it.
+    let badge_class = if enabled {
+        "badge badge-success badge-sm"
+    } else {
+        "badge badge-error badge-sm"
+    };
+    let label = if enabled { "enabled" } else { "disabled" };
+    axum::response::Html(format!(r#"<span class="{badge_class}">{label}</span>"#)).into_response()
+}
+
 pub async fn manage_logout(
     State(state): State<AppState>,
     request: axum::extract::Request,
@@ -3348,6 +3613,7 @@ pub async fn modes_mp(
         ctx_ref.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -3439,6 +3705,7 @@ pub async fn compose_page_mp(
         ctx_ref.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -3648,6 +3915,7 @@ pub async fn actions_page(State(state): State<AppState>) -> Result<Html<String>,
         state.nickel_import_path.as_deref(),
         "/ui",
         &state.registry,
+        &state.service_flags,
     )
     .await;
     render(tera, "pages/actions.html", &ctx).await
@@ -3679,6 +3947,7 @@ pub async fn actions_page_mp(
         ctx_ref.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
     render(tera, "pages/actions.html", &ctx).await
@@ -3705,6 +3974,7 @@ pub async fn qa_page(State(state): State<AppState>) -> Result<Html<String>, UiEr
         state.nickel_import_path.as_deref(),
         "/ui",
         &state.registry,
+        &state.service_flags,
     )
     .await;
     render(tera, "pages/qa.html", &ctx).await
@@ -3736,6 +4006,7 @@ pub async fn qa_page_mp(
         ctx_ref.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
     render(tera, "pages/qa.html", &ctx).await
@@ -4230,6 +4501,7 @@ pub async fn config_page_mp(
         ctx_ref.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -4323,6 +4595,7 @@ pub async fn adrs_page_mp(
         ctx_ref.import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -4403,6 +4676,7 @@ pub async fn personal_page_mp(
         import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -4479,6 +4753,7 @@ pub async fn career_page_mp(
         import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
@@ -4547,8 +4822,92 @@ pub async fn provisioning_page_mp(
         import_path.as_deref(),
         &base_url,
         &state.registry,
+        &state.service_flags,
     )
     .await;
 
     render(tera, "pages/provisioning.html", &ctx).await
 }
+
+#[cfg(test)]
+mod tests {
+    use serde_json::json;
+
+    use super::*;
+
+    #[test]
+    fn vault_basics_returns_none_when_no_project_json() {
+        assert!(vault_basics_from_json(None).is_none());
+    }
+
+    #[test]
+    fn vault_basics_returns_none_when_sops_absent() {
+        let j = json!({ "slug": "demo" });
+        assert!(vault_basics_from_json(Some(&j)).is_none());
+    }
+
+    #[test]
+    fn vault_basics_returns_none_when_sops_disabled() {
+        let j = json!({ "slug": "demo", "sops": { "enabled": false } });
+        assert!(vault_basics_from_json(Some(&j)).is_none());
+    }
+
+    #[test]
+    fn vault_basics_legacy_mode_no_rules() {
+        let j = json!({
+            "slug": "demo",
+            "sops": { "enabled": true, "vault_id": "demo-vault" }
+        });
+        let (id, mode) = vault_basics_from_json(Some(&j)).expect("vault");
+        assert_eq!(id, "demo-vault");
+        assert_eq!(mode, "legacy");
+    }
+
+    #[test]
+    fn vault_basics_legacy_mode_empty_rules_array() {
+        let j = json!({
+            "slug": "demo",
+            "sops": { "enabled": true, "recipient_rules": [] }
+        });
+        let (id, mode) = vault_basics_from_json(Some(&j)).expect("vault");
+        assert_eq!(id, "demo");
+        assert_eq!(mode, "legacy");
+    }
+
+    #[test]
+    fn vault_basics_declarative_mode_with_rules() {
+        let j = json!({
+            "slug": "libre-wuji",
+            "sops": {
+                "enabled": true,
+                "vault_id": "libre-wuji",
+                "recipient_rules": [
+                    { "path": "access\\.sops\\.yaml$", "groups": ["admin"] }
+                ]
+            }
+        });
+        let (id, mode) = vault_basics_from_json(Some(&j)).expect("vault");
+        assert_eq!(id, "libre-wuji");
+        assert_eq!(mode, "declarative");
+    }
+
+    #[test]
+    fn vault_basics_falls_back_to_slug_when_vault_id_missing() {
+        let j = json!({
+            "slug": "fallback-slug",
+            "sops": { "enabled": true }
+        });
+        let (id, _mode) = vault_basics_from_json(Some(&j)).expect("vault");
+        assert_eq!(id, "fallback-slug");
+    }
+
+    #[test]
+    fn vault_basics_falls_back_to_slug_when_vault_id_empty_string() {
+        let j = json!({
+            "slug": "fallback-slug",
+            "sops": { "enabled": true, "vault_id": "" }
+        });
+        let (id, _mode) = vault_basics_from_json(Some(&j)).expect("vault");
+        assert_eq!(id, "fallback-slug");
+    }
+}
diff --git a/crates/ontoref-daemon/src/ui/login.rs b/crates/ontoref-daemon/src/ui/login.rs
index 536488d..b41ae4d 100644
--- a/crates/ontoref-daemon/src/ui/login.rs
+++ b/crates/ontoref-daemon/src/ui/login.rs
@@ -22,7 +22,7 @@ pub async fn login_page(
     ctx.insert("base_url", &base_url);
     ctx.insert("hide_project_nav", &true);
     ctx.insert("current_role", "");
-    insert_mcp_ctx(&mut ctx);
+    insert_mcp_ctx(&mut ctx, &state.service_flags);
     if let Some(proj) = state.registry.get(&slug) {
         insert_brand_ctx(
             &mut ctx,
@@ -31,6 +31,7 @@ pub async fn login_page(
             proj.import_path.as_deref(),
             &base_url,
             &state.registry,
+            &state.service_flags,
         )
         .await;
     }
@@ -78,7 +79,7 @@ pub async fn login_submit(
             tctx.insert("base_url", &base_url);
             tctx.insert("hide_project_nav", &true);
             tctx.insert("current_role", "");
-            insert_mcp_ctx(&mut tctx);
+            insert_mcp_ctx(&mut tctx, &state.service_flags);
             insert_brand_ctx(
                 &mut tctx,
                 &ctx.root,
@@ -86,6 +87,7 @@ pub async fn login_submit(
                 ctx.import_path.as_deref(),
                 &base_url,
                 &state.registry,
+                &state.service_flags,
             )
             .await;
             match render(tera, "pages/login.html", &tctx).await {
@@ -104,7 +106,7 @@ pub async fn manage_login_page(State(state): State<AppState>) -> Result<Html<Str
     ctx.insert("daemon_admin_enabled", &state.daemon_admin_hash.is_some());
     ctx.insert("hide_project_nav", &true);
     ctx.insert("current_role", "");
-    insert_mcp_ctx(&mut ctx);
+    insert_mcp_ctx(&mut ctx, &state.service_flags);
     render(tera, "pages/manage_login.html", &ctx).await
 }
 
@@ -145,7 +147,7 @@ pub async fn manage_login_submit(
         tctx.insert("daemon_admin_enabled", &true);
         tctx.insert("hide_project_nav", &true);
         tctx.insert("current_role", "");
-        insert_mcp_ctx(&mut tctx);
+        insert_mcp_ctx(&mut tctx, &state.service_flags);
         match render(tera, "pages/manage_login.html", &tctx).await {
             Ok(html) => html.into_response(),
             Err(_) => StatusCode::INTERNAL_SERVER_ERROR.into_response(),
diff --git a/crates/ontoref-daemon/src/ui/mod.rs b/crates/ontoref-daemon/src/ui/mod.rs
index 173fc6e..f9bb619 100644
--- a/crates/ontoref-daemon/src/ui/mod.rs
+++ b/crates/ontoref-daemon/src/ui/mod.rs
@@ -56,6 +56,10 @@ fn multi_router(state: AppState) -> axum::Router {
         .route("/manage", get(handlers::manage_page_guarded))
         .route("/manage/add", post(handlers::manage_add_guarded))
         .route("/manage/remove", post(handlers::manage_remove_guarded))
+        .route(
+            "/manage/services/{service}/toggle",
+            post(handlers::service_toggle_ui),
+        )
         .route(
             "/manage/login",
             get(login::manage_login_page).post(login::manage_login_submit),
diff --git a/crates/ontoref-daemon/templates/base.html b/crates/ontoref-daemon/templates/base.html
index 8b54051..fa8e5bd 100644
--- a/crates/ontoref-daemon/templates/base.html
+++ b/crates/ontoref-daemon/templates/base.html
@@ -505,8 +505,20 @@
                 <span class="inline-block w-1.5 h-1.5 rounded-full bg-success ml-1"></span>
               </span>
             </li>
-            <li class="divider my-0.5"></li>
             {% endif %}
+            {% if graphql_active %}
+            <li>
+              <a href="/graphql" target="_blank" class="gap-1.5 text-base-content/60">
+                <svg class="w-4 h-4 flex-shrink-0" fill="none" stroke="currentColor" viewBox="0 0 24 24">
+                  <path stroke-linecap="round" stroke-linejoin="round" stroke-width="2"
+                    d="M13 10V3L4 14h7v7l9-11h-7z"/>
+                </svg>
+                GraphQL
+                <span class="inline-block w-1.5 h-1.5 rounded-full bg-success ml-1"></span>
+              </a>
+            </li>
+            {% endif %}
+            <li class="divider my-0.5"></li>
             {% if slug and current_role == "viewer" %}
             <li class="px-2 py-0.5 text-xs text-base-content/40 font-mono">{{ slug }} · viewer</li>
             {% endif %}
@@ -1244,6 +1256,32 @@
           </div>
         </div>
         {% endif %}
+        {% if graphql_active %}
+        <a
+          href="/graphql"
+          target="_blank"
+          class="btn btn-xs btn-ghost gap-1"
+          title="GraphQL endpoint active"
+        >
+          <svg
+            class="w-3.5 h-3.5"
+            fill="none"
+            stroke="currentColor"
+            viewBox="0 0 24 24"
+          >
+            <path
+              stroke-linecap="round"
+              stroke-linejoin="round"
+              stroke-width="2"
+              d="M13 10V3L4 14h7v7l9-11h-7z"
+            />
+          </svg>
+          <span class="text-xs font-mono">GraphQL</span>
+          <span
+            class="inline-block w-1.5 h-1.5 rounded-full bg-success"
+          ></span>
+        </a>
+        {% endif %}
         <!-- Manage gear: only on main picker page (no project) -->
         {% if not slug %}
         <a
@@ -1382,6 +1420,45 @@
       </div>
     </div>
 
+    {% if (has_registry or has_vault) and slug %}
+    <div class="border-b border-base-content/8 bg-base-200/50 px-4 py-1 hidden sm:flex items-center gap-2 text-[11px] font-mono text-base-content/40 overflow-x-auto">
+      {% if has_registry %}
+      <svg class="w-3 h-3 flex-shrink-0 text-accent/50" fill="none" stroke="currentColor" viewBox="0 0 24 24">
+        <path stroke-linecap="round" stroke-linejoin="round" stroke-width="2"
+          d="M5 12h14M5 12a2 2 0 01-2-2V6a2 2 0 012-2h14a2 2 0 012 2v4a2 2 0 01-2 2M5 12a2 2 0 00-2 2v4a2 2 0 002 2h14a2 2 0 002-2v-4a2 2 0 00-2-2m-2-4h.01M17 16h.01"/>
+      </svg>
+      <span class="text-base-content/30">registry</span>
+      <span class="text-accent/70 font-semibold">{{ registry_participant }}</span>
+      {% if registry_namespaces | length > 0 %}
+      <span class="text-base-content/20">·</span>
+      {% for ns in registry_namespaces %}
+      <span class="text-base-content/50">{{ ns }}</span>
+      {% if not loop.last %}<span class="text-base-content/20">·</span>{% endif %}
+      {% endfor %}
+      {% endif %}
+      {% if registry_endpoint %}
+      <span class="text-base-content/20">·</span>
+      <span class="text-base-content/30 truncate max-w-xs" title="{{ registry_endpoint }}">{{ registry_endpoint }}</span>
+      {% endif %}
+      {% endif %}
+      {% if has_vault %}
+      {% if has_registry %}<span class="text-base-content/20">·</span>{% endif %}
+      <span class="text-base-content/30">vault</span>
+      <span class="text-accent/70 font-semibold" title="vault id">{{ vault_id }}</span>
+      <span class="text-base-content/20">·</span>
+      <span class="px-1.5 py-0.5 rounded {% if vault_mode == 'declarative' %}bg-success/15 text-success{% else %}bg-warning/15 text-warning{% endif %} text-[10px] font-semibold uppercase tracking-wide" title="vault mode (declarative = recipient_rules; legacy = SOPS_AGE_RECIPIENTS)">{{ vault_mode }}</span>
+      {% if vault_recipients > 0 %}
+      <span class="text-base-content/20">·</span>
+      <span class="text-base-content/40" title="recipients in access.sops.yaml">{{ vault_recipients }} recip</span>
+      {% endif %}
+      {% if vault_last_op %}
+      <span class="text-base-content/20">·</span>
+      <span class="text-base-content/30 truncate max-w-xs" title="last audit log entry">{{ vault_last_op }}</span>
+      {% endif %}
+      {% endif %}
+    </div>
+    {% endif %}
+
     <main class="{% block main_class %}container mx-auto px-4 py-6 max-w-7xl{% endblock main_class %}">
       {% block content %}{% endblock content %}
     </main>
diff --git a/crates/ontoref-daemon/templates/pages/manage.html b/crates/ontoref-daemon/templates/pages/manage.html
index aa98e3f..f69a7ac 100644
--- a/crates/ontoref-daemon/templates/pages/manage.html
+++ b/crates/ontoref-daemon/templates/pages/manage.html
@@ -21,6 +21,52 @@
   {% if error %}<span>{{ error }}</span>{% endif %}
 </div>
 
+<!-- Runtime Services -->
+<div class="card bg-base-200 border border-base-content/10 mb-6">
+  <div class="card-body p-5">
+    <h2 class="card-title text-sm font-semibold">Runtime Services</h2>
+    <p class="text-xs text-base-content/40 -mt-1">Toggle optional services without restarting the daemon.</p>
+    <div class="flex flex-wrap gap-3 mt-3">
+      {% if mcp_active is defined %}
+      <div class="flex items-center gap-2 bg-base-300 rounded-lg px-3 py-2">
+        <span class="text-xs font-mono font-semibold">MCP</span>
+        <span id="mcp-badge" class="{% if mcp_active %}badge badge-success badge-sm{% else %}badge badge-error badge-sm{% endif %}">
+          {% if mcp_active %}enabled{% else %}disabled{% endif %}
+        </span>
+        <button
+          class="btn btn-xs btn-ghost ml-1"
+          hx-post="/ui/manage/services/mcp/toggle"
+          hx-target="#mcp-badge"
+          hx-swap="outerHTML"
+          title="Toggle MCP">
+          <svg class="w-3.5 h-3.5" fill="none" stroke="currentColor" viewBox="0 0 24 24">
+            <path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M4 4v5h.582m15.356 2A8.001 8.001 0 004.582 9m0 0H9m11 11v-5h-.581m0 0a8.003 8.003 0 01-15.357-2m15.357 2H15"/>
+          </svg>
+        </button>
+      </div>
+      {% endif %}
+      {% if graphql_active is defined %}
+      <div class="flex items-center gap-2 bg-base-300 rounded-lg px-3 py-2">
+        <span class="text-xs font-mono font-semibold">GraphQL</span>
+        <span id="graphql-badge" class="{% if graphql_active %}badge badge-success badge-sm{% else %}badge badge-error badge-sm{% endif %}">
+          {% if graphql_active %}enabled{% else %}disabled{% endif %}
+        </span>
+        <button
+          class="btn btn-xs btn-ghost ml-1"
+          hx-post="/ui/manage/services/graphql/toggle"
+          hx-target="#graphql-badge"
+          hx-swap="outerHTML"
+          title="Toggle GraphQL">
+          <svg class="w-3.5 h-3.5" fill="none" stroke="currentColor" viewBox="0 0 24 24">
+            <path stroke-linecap="round" stroke-linejoin="round" stroke-width="2" d="M4 4v5h.582m15.356 2A8.001 8.001 0 004.582 9m0 0H9m11 11v-5h-.581m0 0a8.003 8.003 0 01-15.357-2m15.357 2H15"/>
+          </svg>
+        </button>
+      </div>
+      {% endif %}
+    </div>
+  </div>
+</div>
+
 <!-- Registered Projects -->
 <div class="mb-6" id="manage-projects-section">
   {% include "partials/manage_projects_section.html" %}
diff --git a/crates/ontoref-daemon/templates/pages/project_picker.html b/crates/ontoref-daemon/templates/pages/project_picker.html
index 2462f54..dec77e6 100644
--- a/crates/ontoref-daemon/templates/pages/project_picker.html
+++ b/crates/ontoref-daemon/templates/pages/project_picker.html
@@ -89,6 +89,16 @@
           {% if p.repo_kind %}
           <span class="badge badge-outline badge-xs flex-shrink-0 text-base-content/40">{{ p.repo_kind }}</span>
           {% endif %}
+          {% if p.registry_participant %}
+          <span class="badge badge-outline badge-xs flex-shrink-0 font-mono text-accent/70" title="Registry participant: {{ p.registry_participant }}">
+            ⟳ {{ p.registry_participant }}
+          </span>
+          {% endif %}
+          {% if p.has_vault %}
+          <span class="badge badge-outline badge-xs flex-shrink-0 font-mono {% if p.vault_mode == 'declarative' %}text-success{% else %}text-warning{% endif %}" title="Vault {{ p.vault_id }} · {{ p.vault_mode }} mode · {{ p.vault_recipients }} recipients">
+            ⛁ {{ p.vault_id }}{% if p.vault_recipients > 0 %} · {{ p.vault_recipients }}{% endif %}
+          </span>
+          {% endif %}
         </div>
         <!-- Quick-access shortcut icons -->
         <div class="flex items-center gap-0.5 flex-shrink-0">
@@ -205,6 +215,51 @@
              hx-trigger="revealed"
              hx-swap="innerHTML"></div>
 
+        {% if p.registry_participant %}
+        <details class="collapse collapse-arrow bg-base-300/40 rounded-lg">
+          <summary class="collapse-title text-xs font-semibold py-2 min-h-0 px-3 cursor-pointer">
+            Registry · <span class="font-mono font-normal text-accent/80">{{ p.registry_participant }}</span>
+          </summary>
+          <div class="collapse-content px-3 pb-3">
+            {% if p.registry_endpoint %}
+            <p class="text-xs font-medium text-base-content/40 uppercase tracking-wider mb-1">Endpoint</p>
+            <p class="font-mono text-xs text-base-content/60 mb-2 break-all">{{ p.registry_endpoint }}</p>
+            {% endif %}
+            {% if p.registry_namespaces %}
+            <p class="text-xs font-medium text-base-content/40 uppercase tracking-wider mb-1">Namespaces</p>
+            <div class="space-y-0.5">
+              {% for ns in p.registry_namespaces %}
+              <p class="font-mono text-xs text-base-content/70">{{ ns }}</p>
+              {% endfor %}
+            </div>
+            {% endif %}
+          </div>
+        </details>
+        {% endif %}
+
+        {% if p.has_vault %}
+        <details class="collapse collapse-arrow bg-base-300/40 rounded-lg">
+          <summary class="collapse-title text-xs font-semibold py-2 min-h-0 px-3 cursor-pointer">
+            Vault · <span class="font-mono font-normal {% if p.vault_mode == 'declarative' %}text-success{% else %}text-warning{% endif %}">{{ p.vault_id }}</span>
+            <span class="ml-2 px-1.5 py-0.5 rounded text-[10px] font-semibold uppercase tracking-wide {% if p.vault_mode == 'declarative' %}bg-success/15 text-success{% else %}bg-warning/15 text-warning{% endif %}">{{ p.vault_mode }}</span>
+          </summary>
+          <div class="collapse-content px-3 pb-3">
+            <p class="text-xs font-medium text-base-content/40 uppercase tracking-wider mb-1">Mode</p>
+            <p class="font-mono text-xs text-base-content/70 mb-2">
+              {% if p.vault_mode == 'declarative' %}per-file recipient routing (ADR-019){% else %}legacy uniform recipients (ADR-017){% endif %}
+            </p>
+            {% if p.vault_recipients > 0 %}
+            <p class="text-xs font-medium text-base-content/40 uppercase tracking-wider mb-1">Recipients in access.sops.yaml</p>
+            <p class="font-mono text-xs text-base-content/70 mb-2">{{ p.vault_recipients }}</p>
+            {% endif %}
+            {% if p.vault_last_op %}
+            <p class="text-xs font-medium text-base-content/40 uppercase tracking-wider mb-1">Last audit entry</p>
+            <p class="font-mono text-xs text-base-content/70 break-all">{{ p.vault_last_op }}</p>
+            {% endif %}
+          </div>
+        </details>
+        {% endif %}
+
         {% if p.layers or p.op_modes %}
         <details class="collapse collapse-arrow bg-base-300/40 rounded-lg">
           <summary class="collapse-title text-xs font-semibold py-2 min-h-0 px-3 cursor-pointer">
diff --git a/crates/ontoref-derive/src/lib.rs b/crates/ontoref-derive/src/lib.rs
index 78f9dae..e863538 100644
--- a/crates/ontoref-derive/src/lib.rs
+++ b/crates/ontoref-derive/src/lib.rs
@@ -266,6 +266,143 @@ fn emit_onto_api(attr: OntoApiAttr, item: proc_macro2::TokenStream) -> proc_macr
     }
 }
 
+// ── #[onto_mcp_tool(...)]
+// ──────────────────────────────────────────────────
+
+/// Attribute macro for MCP tool unit-structs in ontoref-daemon.
+///
+/// Registers the tool in the MCP catalog at link time via
+/// `inventory::submit!(McpToolEntry{...})`. The annotated item is emitted
+/// unchanged — `ToolBase` and `AsyncTool` impls below the struct continue to
+/// own the executable behaviour (ADR-015).
+///
+/// # Required keys
+/// - `name = "ontoref_xxx"` — must equal `ToolBase::name()` for the same struct
+/// - `description = "..."` — one-line agent-facing description
+///
+/// # Optional keys
+/// - `category = "discovery"` — semantic grouping (discovery | ontology |
+///   knowledge | validation | config). Empty by default.
+/// - `params = "name:type:constraint:desc; ..."` — same grammar as
+///   `#[onto_api(params = ...)]`; semicolon-separated, four-field entries.
+///
+/// # Example
+/// ```ignore
+/// #[onto_mcp_tool(
+///     name = "ontoref_search",
+///     description = "Free-text search across nodes, ADRs, and modes.",
+///     category = "discovery",
+///     params = "query:string:required:Search term; project:string:optional:Project slug",
+/// )]
+/// struct SearchTool;
+/// ```
+#[proc_macro_attribute]
+pub fn onto_mcp_tool(args: TokenStream, input: TokenStream) -> TokenStream {
+    match expand_onto_mcp_tool(args, input) {
+        Ok(ts) => ts.into(),
+        Err(err) => err.to_compile_error().into(),
+    }
+}
+
+fn expand_onto_mcp_tool(
+    args: TokenStream,
+    input: TokenStream,
+) -> syn::Result<proc_macro2::TokenStream> {
+    let item = proc_macro2::TokenStream::from(input);
+
+    let kv_args = syn::parse::Parser::parse(
+        Punctuated::<MetaNameValue, Token![,]>::parse_terminated,
+        args,
+    )?;
+
+    let mut name: Option<String> = None;
+    let mut description: Option<String> = None;
+    let mut category = String::new();
+    let mut params_raw: Option<String> = None;
+
+    for kv in &kv_args {
+        let key = kv
+            .path
+            .get_ident()
+            .ok_or_else(|| syn::Error::new_spanned(&kv.path, "expected identifier"))?
+            .to_string();
+        let val = lit_str(&kv.value)
+            .ok_or_else(|| syn::Error::new_spanned(&kv.value, "expected string literal"))?;
+
+        match key.as_str() {
+            "name" => name = Some(val),
+            "description" => description = Some(val),
+            "category" => category = val,
+            "params" => params_raw = Some(val),
+            other => {
+                return Err(syn::Error::new_spanned(
+                    &kv.path,
+                    format!(
+                        "unknown onto_mcp_tool key: {other}; expected name, description, \
+                         category, params"
+                    ),
+                ))
+            }
+        }
+    }
+
+    let name = name.ok_or_else(|| {
+        syn::Error::new(
+            Span::call_site(),
+            "#[onto_mcp_tool] requires name = \"ontoref_xxx\"",
+        )
+    })?;
+    let description = description.ok_or_else(|| {
+        syn::Error::new(
+            Span::call_site(),
+            "#[onto_mcp_tool] requires description = \"...\"",
+        )
+    })?;
+
+    let params = parse_params(params_raw.as_deref().unwrap_or(""))?;
+
+    let name_lit = LitStr::new(&name, Span::call_site());
+    let desc_lit = LitStr::new(&description, Span::call_site());
+    let category_lit = LitStr::new(&category, Span::call_site());
+
+    let param_exprs: Vec<_> = params
+        .iter()
+        .map(|p| {
+            let n = LitStr::new(&p.name, Span::call_site());
+            let k = LitStr::new(&p.kind, Span::call_site());
+            let c = LitStr::new(&p.constraint, Span::call_site());
+            let d = LitStr::new(&p.description, Span::call_site());
+            quote! {
+                ::ontoref_ontology::ApiParam { name: #n, kind: #k, constraint: #c, description: #d }
+            }
+        })
+        .collect();
+
+    let unique = name
+        .bytes()
+        .fold(5381u64, |h, b| h.wrapping_mul(33).wrapping_add(b as u64));
+    let static_ident =
+        syn::Ident::new(&format!("__ONTOREF_MCP_TOOL_{unique:x}"), Span::call_site());
+
+    Ok(quote! {
+        ::inventory::submit! {
+            ::ontoref_ontology::McpToolEntry {
+                name:        #name_lit,
+                description: #desc_lit,
+                category:    #category_lit,
+                params:      &[#(#param_exprs),*],
+                source_file: file!(),
+            }
+        }
+
+        #[doc(hidden)]
+        #[allow(non_upper_case_globals, dead_code)]
+        static #static_ident: () = ();
+
+        #item
+    })
+}
+
 // ── Attribute parsing
 // ─────────────────────────────────────────────────────────
 
diff --git a/crates/ontoref-ontology/src/lib.rs b/crates/ontoref-ontology/src/lib.rs
index 316c94a..924be5d 100644
--- a/crates/ontoref-ontology/src/lib.rs
+++ b/crates/ontoref-ontology/src/lib.rs
@@ -18,16 +18,21 @@ pub mod contrib;
 #[cfg(feature = "derive")]
 pub mod api;
 
+#[cfg(feature = "derive")]
+pub mod mcp;
+
 #[cfg(feature = "derive")]
 pub use api::{ApiParam, ApiRouteEntry};
 #[cfg(feature = "derive")]
 pub use contrib::{ConfigFieldsEntry, NodeContribution, TestCoverage};
 pub use error::OntologyError;
+#[cfg(feature = "derive")]
+pub use mcp::McpToolEntry;
 pub use ontology::{Core, Gate, Ontology, State};
 // Re-export the proc-macro crate so consumers only need to depend on
 // `ontoref-ontology` with the `derive` feature — no separate ontoref-derive dep.
 #[cfg(feature = "derive")]
-pub use ontoref_derive::{onto_validates, ConfigFields, OntologyNode};
+pub use ontoref_derive::{onto_mcp_tool, onto_validates, ConfigFields, OntologyNode};
 pub use types::{
     AbstractionLevel, CoreConfig, Coupling, Dimension, DimensionState, Duration, Edge, EdgeType,
     GateConfig, Horizon, Membrane, Node, OpeningCondition, Permeability, Pole, Protocol,
diff --git a/crates/ontoref-ontology/src/mcp.rs b/crates/ontoref-ontology/src/mcp.rs
new file mode 100644
index 0000000..3a3c5d6
--- /dev/null
+++ b/crates/ontoref-ontology/src/mcp.rs
@@ -0,0 +1,24 @@
+use crate::api::ApiParam;
+
+/// Static metadata for an MCP tool exposed by ontoref-daemon.
+///
+/// Registered at link time via `inventory::submit!` — generated by
+/// `#[onto_mcp_tool(...)]` proc-macro attribute on each tool struct.
+/// Collected via `inventory::iter::<McpToolEntry>()` (ADR-015).
+///
+/// `params` reuses [`ApiParam`] — both API routes and MCP tools are protocol
+/// surfaces and share the same `name:type:constraint:description` shape.
+#[derive(serde::Serialize, Clone)]
+pub struct McpToolEntry {
+    pub name: &'static str,
+    pub description: &'static str,
+    /// Semantic grouping (e.g. "discovery", "ontology", "knowledge",
+    /// "validation", "config"). Used by future tiering work; safe to leave
+    /// empty.
+    pub category: &'static str,
+    pub params: &'static [ApiParam],
+    /// Source file path captured via `file!()` at the macro call site.
+    pub source_file: &'static str,
+}
+
+inventory::collect!(McpToolEntry);
diff --git a/domains/provisioning/commands.nu b/domains/provisioning/commands.nu
index bfa731d..335960c 100644
--- a/domains/provisioning/commands.nu
+++ b/domains/provisioning/commands.nu
@@ -175,8 +175,416 @@ export def "backlog show" [id: string] {
     }
 }
 
+# ─── install (Mixed/platform only) ───────────────────────────────────────────
+
+def detect-platform []: nothing -> string {
+    if (do { ^which docker } | complete | get exit_code) == 0 { "docker" }
+    else if (do { ^which podman } | complete | get exit_code) == 0 { "podman" }
+    else if (do { ^which kubectl } | complete | get exit_code) == 0 { "kubernetes" }
+    else { "" }
+}
+
+export def "install" [
+    --mode: string = "solo"     # Deployment mode: solo|multi-user|cicd|enterprise
+    --platform: string = ""     # Container platform: docker|podman|kubernetes (auto-detected)
+    --dry-run                   # Print what would be run without executing
+]: nothing -> nothing {
+    let root = (project-root)
+
+    let installer_bin = [$root, "platform", "target", "release", "provisioning-installer"] | path join
+    let installer_src = [$root, "platform", "installer"] | path join
+    let just_available = (do { ^which just } | complete | get exit_code) == 0
+
+    let resolved_platform = if ($platform | is-not-empty) { $platform } else { detect-platform }
+
+    print $"Platform install  mode=($mode)  platform=(if ($resolved_platform | is-empty) { '?' } else { $resolved_platform })"
+    print ""
+
+    if ($resolved_platform | is-empty) {
+        print "No supported container platform found (docker, podman, kubectl)."
+        print "Install one and retry."
+        return
+    }
+
+    if ($installer_bin | path exists) {
+        print $"installer binary: ($installer_bin)"
+        if $dry_run {
+            print $"[dry-run] would run: ($installer_bin) --headless --mode ($mode) --platform ($resolved_platform) --yes"
+            return
+        }
+        do { ^$installer_bin --headless --mode $mode --platform $resolved_platform --yes } | complete | ignore
+        return
+    }
+
+    if ($installer_src | path exists) {
+        print "installer binary not built — building from source..."
+        if $dry_run {
+            print $"[dry-run] would run: cargo build --release --manifest-path ($installer_src)/Cargo.toml"
+            return
+        }
+        let r = do { ^cargo build --release --manifest-path $"($installer_src)/Cargo.toml" } | complete
+        if $r.exit_code != 0 {
+            error make { msg: $"installer build failed:\n($r.stderr)" }
+        }
+        do { ^$installer_bin --headless --mode $mode --platform $resolved_platform --yes } | complete | ignore
+        return
+    }
+
+    if $just_available {
+        print "No installer binary or source found — delegating to: just install"
+        if $dry_run {
+            print $"[dry-run] would run: just install mode=($mode)"
+            return
+        }
+        let r = do { ^just install $"mode=($mode)" } | complete
+        if $r.exit_code != 0 {
+            error make { msg: $"just install failed:\n($r.stderr)" }
+        }
+        return
+    }
+
+    print "installer not found and 'just' is not available."
+    print $"Build the installer first: cargo build --release --manifest-path ($installer_src)/Cargo.toml"
+}
+
+# ─── ops (ADR-037: NATS ops contract) ────────────────────────────────────────
+
+def workspace-id []: nothing -> string {
+    let card = load-card
+    if ($card | is-empty) { "unknown" } else { $card.id }
+}
+
+def nats-available []: nothing -> bool {
+    (do { ^which nats } | complete | get exit_code) == 0
+}
+
+def keeper-cli-available []: nothing -> bool {
+    (do { ^which keeper-cli } | complete | get exit_code) == 0
+}
+
+def rad-available []: nothing -> bool {
+    (do { ^which rad } | complete | get exit_code) == 0
+}
+
+export def "ops list" []: nothing -> nothing {
+    let ws = workspace-id
+    if not (nats-available) {
+        print "nats CLI not found — install nats-io/natscli"
+        return
+    }
+    let stream = $"OPS_PENDING_(($ws | str upcase))"
+    let r = do { ^nats stream info $stream --json } | complete
+    if $r.exit_code != 0 {
+        print $"Stream ($stream) not found or NATS unreachable: ($r.stderr | str trim)"
+        return
+    }
+    let info = $r.stdout | from json
+    print $"stream:    ($stream)"
+    print $"messages:  ($info.state.messages)"
+    print $"first_seq: ($info.state.first_seq)"
+    print $"last_seq:  ($info.state.last_seq)"
+}
+
+export def "ops describe" [id: string]: nothing -> nothing {
+    let ws = workspace-id
+    if not (nats-available) { print "nats CLI not found"; return }
+    let subject = $"ops.pending.($ws).>"
+    let r = do { ^nats req $subject "" --count 50 --json } | complete
+    if $r.exit_code != 0 { print $"NATS request failed: ($r.stderr | str trim)"; return }
+    let msgs = $r.stdout | lines | each { |l| $l | from json | get data? | default {} }
+    let found = $msgs | where { |m| ($m.jti? | default "") == $id }
+    if ($found | is-empty) { print $"Op not found: ($id)"; return }
+    let op = $found | first
+    print $"jti:                ($op.jti)"
+    print $"op_type:            ($op.op_type? | default 'unknown')"
+    print $"target:             ($op.target? | default 'unknown')"
+    print $"expected_version:   ($op.expected_state_version? | default 'none')"
+    print $"scopes:             ($op.scopes? | default [] | str join ', ')"
+    print $"issued_at:          ($op.iat? | default 'unknown')"
+}
+
+export def "ops sign" [id: string]: nothing -> nothing {
+    if not (keeper-cli-available) { print "keeper-cli not found — build from platform/crates/ops-keeper"; return }
+    let ws = workspace-id
+    let r = do { ^keeper-cli --workspace $ws sign $id } | complete
+    if $r.exit_code != 0 {
+        error make { msg: $"keeper-cli sign failed: ($r.stderr | str trim)" }
+    }
+    print $r.stdout
+}
+
+export def "ops history" [--workspace: string = ""]: nothing -> nothing {
+    let ws = if ($workspace | is-empty) { workspace-id } else { $workspace }
+    let repo_path = $env.HOME | path join ".local/share/radicle/repos" $"state-($ws)"
+    if not ($repo_path | path exists) {
+        print $"Radicle state repo not found at ($repo_path)"
+        print "Run: rad clone rad:<rid> state-($ws)"
+        return
+    }
+    let r = do { ^git -C $repo_path log --oneline --no-decorate -20 } | complete
+    if $r.exit_code != 0 { print $"git log failed: ($r.stderr | str trim)"; return }
+    $r.stdout | lines | each { |l|
+        let parts = $l | split row " " | enumerate
+        let hash = $parts | where index == 0 | get item.0
+        let msg  = $parts | skip 1 | get item | str join " "
+        { commit: $hash, message: $msg }
+    }
+}
+
+# ─── playbook (extensions/playbooks/) ────────────────────────────────────────
+
+def playbooks-root []: nothing -> string {
+    let root = project-root
+    [$root, "..", "..", "provisioning", "extensions", "playbooks"] | path join | path expand
+}
+
+export def "playbook list" []: nothing -> nothing {
+    let root = playbooks-root
+    if not ($root | path exists) { print "No extensions/playbooks/ directory found."; return }
+    ^ls $root | where type == dir | get name | each { |d|
+        let playbook_path = [$d, "playbook.ncl"] | path join
+        let has_run = ([$d, "run.nu"] | path join | path exists)
+        { name: ($d | path basename), has_run: $has_run, has_playbook: ($playbook_path | path exists) }
+    }
+}
+
+export def "playbook describe" [name: string]: nothing -> nothing {
+    let root = playbooks-root
+    let playbook_path = [$root, $name, "playbook.ncl"] | path join
+    if not ($playbook_path | path exists) {
+        print $"Playbook not found: ($name)"
+        print $"Expected at: ($playbook_path)"
+        return
+    }
+    let r = do { ^nickel export $playbook_path } | complete
+    if $r.exit_code != 0 { print $"Failed to load playbook: ($r.stderr | str trim)"; return }
+    let p = $r.stdout | from json
+    print $"name:         ($p.name)"
+    print $"description:  ($p.description)"
+    print $"version:      ($p.version)"
+    print ""
+    print "preconditions:"
+    $p.preconditions? | default [] | each { |c| print $"  - ($c)" }
+    print ""
+    print "steps:"
+    $p.steps? | default [] | enumerate | each { |s|
+        print $"  ($s.index + 1). ($s.item.id): ($s.item.description)"
+    }
+    if ($p.rollback_strategy? | default "none") != "none" {
+        print $"\nrollback: ($p.rollback_strategy)"
+    }
+}
+
+export def "playbook run" [name: string, --dry-run]: nothing -> nothing {
+    let root = playbooks-root
+    let run_path = [$root, $name, "run.nu"] | path join
+    if not ($run_path | path exists) {
+        print $"run.nu not found for playbook: ($name)"
+        return
+    }
+    let ws = workspace-id
+    if $dry_run {
+        let dry_path = [$root, $name, "tests", "dry_run.nu"] | path join
+        if ($dry_path | path exists) {
+            let r = do { ^nu $dry_path --workspace $ws } | complete
+            if $r.exit_code != 0 { error make { msg: $"dry-run failed: ($r.stderr | str trim)" } }
+            print $r.stdout
+        } else {
+            print $"[dry-run] would execute: nu ($run_path) --workspace ($ws)"
+            print "No dry_run.nu found — simulating step listing only."
+            playbook describe $name
+        }
+        return
+    }
+    let r = do { ^nu $run_path --workspace $ws } | complete
+    if $r.exit_code != 0 {
+        error make { msg: $"playbook run failed: ($r.stderr | str trim)" }
+    }
+    print $r.stdout
+}
+
+export def "playbook history" []: nothing -> nothing {
+    let ws = workspace-id
+    let repo_path = $env.HOME | path join ".local/share/radicle/repos" $"state-($ws)"
+    if not ($repo_path | path exists) {
+        print $"Radicle state repo not found at ($repo_path)"
+        return
+    }
+    let r = do { ^git -C $repo_path log --oneline --no-decorate --grep="playbook:" -20 } | complete
+    if $r.exit_code != 0 { print $"git log failed: ($r.stderr | str trim)"; return }
+    if ($r.stdout | is-empty) { print "No playbook audit entries found."; return }
+    $r.stdout | lines | each { |l|
+        let parts = $l | split row " " | enumerate
+        let hash = $parts | where index == 0 | get item.0
+        let msg  = $parts | skip 1 | get item | str join " "
+        { commit: $hash, message: $msg }
+    }
+}
+
+# ─── keeper (ADR-037 dual-mode signing) ──────────────────────────────────────
+
+export def "keeper status" []: nothing -> nothing {
+    let ws = workspace-id
+    if not (keeper-cli-available) { print "keeper-cli not found"; return }
+    let r = do { ^keeper-cli --workspace $ws status } | complete
+    if $r.exit_code != 0 {
+        print $"keeper-cli unavailable: ($r.stderr | str trim)"
+        return
+    }
+    print $r.stdout
+}
+
+export def "keeper policy" []: nothing -> nothing {
+    let ws = workspace-id
+    let policy_repo = $env.HOME | path join ".local/share/radicle/repos" $"policy-($ws)"
+    if not ($policy_repo | path exists) {
+        print $"Policy repo not cloned locally: policy-($ws)"
+        print "Run: rad clone rad:<rid> policy-($ws)"
+        return
+    }
+    let policy_path = [$policy_repo, "policy.ncl"] | path join
+    if not ($policy_path | path exists) { print "policy.ncl not found in policy repo"; return }
+    let r = do { ^nickel export $policy_path } | complete
+    if $r.exit_code != 0 { print $"Failed to load policy: ($r.stderr | str trim)"; return }
+    let p = $r.stdout | from json
+    print $"workspace:     ($p.workspace)"
+    print $"version:       ($p.version)"
+    print ""
+    print "auto_sign rules:"
+    $p.auto_sign? | default [] | each { |rule|
+        print $"  op_type=($rule.op_type) target=($rule.target? | default '*') scope=($rule.scope? | default 'any')"
+    }
+    print ""
+    print "require_manual rules:"
+    $p.require_manual? | default [] | each { |rule|
+        print $"  op_type=($rule.op_type) target=($rule.target? | default '*')"
+    }
+}
+
+export def "keeper switch" [mode: string]: nothing -> nothing {
+    let valid_modes = ["auto", "operator-only"]
+    if not ($valid_modes | any { |m| $m == $mode }) {
+        error make { msg: $"Invalid mode '($mode)'. Valid modes: ($valid_modes | str join ', ')" }
+    }
+    let playbook = if $mode == "auto" { "switch_to_vm_ops" } else { "switch_to_operator_only" }
+    print $"Delegating to playbook: ($playbook)"
+    playbook run $playbook
+}
+
+# ─── governance (ADR-038 Radicle delegation) ─────────────────────────────────
+
+export def "governance delegations" []: nothing -> nothing {
+    let ws = workspace-id
+    let repos = ["policy", "desired", "state"] | each { |r| $"($r)-($ws)" }
+    $repos | each { |repo_name|
+        let r = do { ^rad id --repo $repo_name --json } | complete
+        if $r.exit_code != 0 {
+            { repo: $repo_name, error: ($r.stderr | str trim), delegates: [] }
+        } else {
+            let info = $r.stdout | from json
+            { repo: $repo_name, threshold: ($info.threshold? | default 1), delegates: ($info.delegates? | default []) }
+        }
+    }
+}
+
+export def "governance signers" []: nothing -> nothing {
+    let ws = workspace-id
+    if not (nats-available) { print "nats CLI not found"; return }
+    let r = do { ^nats auth account ls --json } | complete
+    if $r.exit_code != 0 {
+        print $"NATS auth query failed: ($r.stderr | str trim)"
+        return
+    }
+    let accounts = $r.stdout | from json
+    let ops_accounts = $accounts | where { |a| ($a.name? | default "") | str contains $ws }
+    if ($ops_accounts | is-empty) { print $"No NATS accounts found for workspace: ($ws)"; return }
+    $ops_accounts | each { |a|
+        { account: $a.name, subject: ($a.subject? | default ""), signers: ($a.signing_keys? | default []) }
+    }
+}
+
 # ─── help + main ──────────────────────────────────────────────────────────────
 
+# ─── registry (i sub-domain) ──────────────────────────────────────────────────
+
+const DEFAULT_REGISTRY = "reg.librecloud.online"
+
+def _find-capabilities-ncl []: nothing -> string {
+    let ws_path = ($env | get -o PROVISIONING_WORKSPACE_PATH | default "")
+    if ($ws_path | is-not-empty) {
+        let candidates = (do { glob $"($ws_path)/infra/*/capabilities.ncl" } | default [])
+        if ($candidates | is-not-empty) { return ($candidates | first) }
+    }
+    mut dir = $env.PWD
+    for _ in 0..6 {
+        let cur = $dir   # immutable copy for closure capture in glob
+        let candidates = (do { glob $"($cur)/infra/*/capabilities.ncl" } | default [])
+        if ($candidates | is-not-empty) { return ($candidates | first) }
+        let parent = ($dir | path dirname)
+        if $parent == $dir { break }
+        $dir = $parent
+    }
+    ""
+}
+
+export def "i resolve-registry" []: nothing -> string {
+    let from_env = ($env | get -o PROVISIONING_REGISTRY | default "")
+    if ($from_env | is-not-empty) { return $from_env }
+    let caps = (_find-capabilities-ncl)
+    if ($caps | is-not-empty) {
+        let ip  = ($env.NICKEL_IMPORT_PATH? | default ($env.PROVISIONING? | default ""))
+        let ncl = (do { ^nickel export --import-path $ip $caps } | complete)
+        if $ncl.exit_code == 0 {
+            let data        = ($ncl.stdout | from json)
+            let reg_default = ($data | get -o provides.registries.default | default "")
+            let registries  = ($data | get -o provides.registries.registries | default [])
+            if ($reg_default | is-not-empty) and ($registries | is-not-empty) {
+                let entry = ($registries | where { |e| $e.id == $reg_default } | first | default null)
+                if ($entry != null) and ($entry.endpoint? | default "" | is-not-empty) {
+                    return $entry.endpoint
+                }
+            }
+        }
+    }
+    $DEFAULT_REGISTRY
+}
+
+export def "i list" [--live]: nothing -> nothing {
+    let reg = (i resolve-registry)
+    if not $live {
+        print $"registry : ($reg)"
+        print "Use --live to query the live catalog."
+        return
+    }
+    let url = $"https://($reg)/v2/_catalog"
+    let body = try { http get $url } catch {
+        print $"Registry catalog unavailable at ($url)"
+        return
+    }
+    let repos = ($body | get -o repositories | default [])
+
+    let domains = ($repos | where { |r| $r | str starts-with "domains/" } | each { |r|
+        let parts = ($r | str replace "domains/" "" | split row "/")
+        { kind: "domain", participant: ($parts | first), id: ($parts | skip 1 | str join "/"), ref: $"($reg)/($r)" }
+    })
+    let modes = ($repos | where { |r| $r | str starts-with "modes/" } | each { |r|
+        let parts = ($r | str replace "modes/" "" | split row "/")
+        { kind: "mode", participant: ($parts | first), id: ($parts | skip 1 | str join "/"), ref: $"($reg)/($r)" }
+    })
+
+    let all = ($domains | append $modes)
+    if ($all | is-empty) {
+        print $"No domains/ or modes/ artifacts found in ($reg)"
+        return
+    }
+    print $"registry: ($reg)  artifacts: ($all | length)"
+    print ""
+    $all | each { |a|
+        print $"  ($a.kind)  ($a.participant)/($a.id)  → ($a.ref)"
+    }
+    null
+}
+
 def show-help [] {
     print "Provisioning — ontoref domain extension"
     print ""
@@ -197,11 +605,42 @@ def show-help [] {
     print "COMMANDS  (Mixed/platform only)"
     print "  backlog [--priority]          Backlog items filtered by High|Medium|Low"
     print "  backlog show <id>             Full detail of a backlog item"
+    print "  install [--mode] [--platform] Install provisioning platform services"
+    print ""
+    print "COMMANDS  (DevWorkspace — ops contract, ADR-037)"
+    print "  ops list                      Pending ops queue depth for this workspace"
+    print "  ops describe <id>             Full op detail: JWT claims, scopes, expected_version"
+    print "  ops sign <id>                 Operator signs a pending op via keeper-cli"
+    print "  ops history [<workspace>]     Applied ops from <workspace>-state Radicle ledger"
+    print ""
+    print "COMMANDS  (DevWorkspace — playbooks)"
+    print "  playbook list                 Available playbooks for this workspace"
+    print "  playbook describe <name>      Steps, params, preconditions of a playbook"
+    print "  playbook run <name>           Execute a playbook (use --dry-run first)"
+    print "  playbook history              Past playbook executions from audit ledger"
+    print ""
+    print "COMMANDS  (DevWorkspace — keeper daemon, ADR-037)"
+    print "  keeper status                 Keeper mode (auto|operator-only|down) + policy version"
+    print "  keeper policy                 Active keeper policy from policy-<workspace> Radicle repo"
+    print "  keeper switch <mode>          Switch keeper mode (delegates to switch_to_<mode> playbook)"
+    print ""
+    print "COMMANDS  (DevWorkspace — governance, ADR-038)"
+    print "  governance delegations        Radicle delegation sets for policy/desired/state repos"
+    print "  governance signers            Active NATS JWT signers and M-of-N quorum status"
+    print ""
+    print "COMMANDS  (all repo_kinds — registry)"
+    print "  i resolve-registry            Resolved OCI registry endpoint (env → capabilities.ncl → default)"
+    print "  i list                        Show resolved registry endpoint"
+    print "  i list --live                 Query _catalog and list live domains/ and modes/ artifacts"
 }
 
 def main [
     ...args: string
     --priority: string = ""
+    --mode: string = "solo"
+    --platform: string = ""
+    --dry-run
+    --live
 ] {
     if ($args | is-empty) or ($args | first) == "help" { show-help; return }
     let sub = ($args | str join " ")
@@ -212,9 +651,28 @@ def main [
         "gates"        => { gates }
         "card"         => { card }
         "capabilities" => { capabilities }
+        "install"      => { if $dry_run { install --mode $mode --platform $platform --dry-run } else { install --mode $mode --platform $platform } }
         "backlog"      => { backlog --priority $priority }
-        _ if ($sub | str starts-with "backlog show ")  => { backlog show ($args | get 2) }
-        _ if ($sub | str starts-with "validate ")      => { validate ($args | skip 1 | str join " ") }
+        "ops list"                      => { ops list }
+        "ops history"                   => { ops history }
+        "keeper status"                 => { keeper status }
+        "keeper policy"                 => { keeper policy }
+        "governance delegations"        => { governance delegations }
+        "governance signers"            => { governance signers }
+        "playbook list"                 => { playbook list }
+        "playbook history"              => { playbook history }
+        _ if ($sub | str starts-with "backlog show ")       => { backlog show ($args | get 2) }
+        _ if ($sub | str starts-with "validate ")           => { validate ($args | skip 1 | str join " ") }
+        _ if ($sub | str starts-with "ops describe ")       => { ops describe ($args | get 2) }
+        _ if ($sub | str starts-with "ops sign ")           => { ops sign ($args | get 2) }
+        _ if ($sub | str starts-with "playbook describe ")  => { playbook describe ($args | get 2) }
+        _ if ($sub | str starts-with "playbook run ")       => {
+            if $dry_run { playbook run ($args | get 2) --dry-run } else { playbook run ($args | get 2) }
+        }
+        _ if ($sub | str starts-with "keeper switch ")      => { keeper switch ($args | get 2) }
+        _ if ($sub | str starts-with "ops history ")        => { ops history --workspace ($args | get 2) }
+        "i resolve-registry"                                 => { i resolve-registry | print }
+        "i list"                                             => { if $live { i list --live } else { i list } }
         _              => { print $"Unknown command: ($sub)"; show-help }
     }
 }
diff --git a/domains/provisioning/domain.ncl b/domains/provisioning/domain.ncl
index 238ed93..7e89e34 100644
--- a/domains/provisioning/domain.ncl
+++ b/domains/provisioning/domain.ncl
@@ -18,6 +18,33 @@ s.Domain & {
     { id = "capabilities",                  description = "Platform capabilities from manifest" },
     { id = "backlog [--priority]",          description = "Backlog items filtered by High|Medium|Low (platform only)" },
     { id = "backlog show <id>",             description = "Full detail of a backlog item" },
+    { id = "install [--mode] [--platform]", description = "Install provisioning platform services (platform only)" },
+
+    # ── Ops contract (adr-037) — pending queue, signed commands, audit ──────
+    { id = "ops list",                      description = "Pending and recent ops for the workspace (NATS ops.pending and ops.cmd subjects)" },
+    { id = "ops describe <id>",             description = "Full detail of a single op including JWT claims, scopes, expected_state_version" },
+    { id = "ops sign <id>",                 description = "Operator signs a pending op via keeper-cli (requires authorized signer key)" },
+    { id = "ops history [--workspace]",     description = "Audit log of applied ops from <workspace>-state Radicle ledger" },
+
+    # ── Playbooks — executable operational artifacts ────────────────────────
+    { id = "playbook list",                 description = "Available playbooks for this workspace's repo_kind" },
+    { id = "playbook describe <name>",      description = "Full definition (steps, params, preconditions) of a playbook" },
+    { id = "playbook run <name>",           description = "Execute a playbook (use --dry-run first to simulate without side effects)" },
+    { id = "playbook history",              description = "Past playbook executions, params, outcomes from audit ledger" },
+
+    # ── Keeper (signing daemon) status and control ──────────────────────────
+    { id = "keeper status",                 description = "Current keeper mode (auto | operator-only | down) and policy version" },
+    { id = "keeper policy",                 description = "Show active keeper policy for this workspace from policy-<workspace> Radicle repo" },
+    { id = "keeper switch <mode>",          description = "Switch keeper mode (delegates to switch_to_<mode> playbook)" },
+
+    # ── Governance (adr-038) — Radicle delegation queries ────────────────────
+    { id = "governance delegations",        description = "Current Radicle delegation set for the workspace's policy/desired/state repos" },
+    { id = "governance signers",            description = "Active signers for ops and policy (M-of-N quorum status)" },
+
+    # ── Registry (OCI / zot) ─────────────────────────────────────────────────
+    { id = "i resolve-registry",            description = "Resolved OCI registry endpoint: PROVISIONING_REGISTRY env → capabilities.ncl → constant fallback" },
+    { id = "i list",                        description = "Show resolved registry endpoint without querying live catalog" },
+    { id = "i list --live",                 description = "Query _catalog on the resolved registry; list live domains/<participant>/<id> and modes/<participant>/<id>" },
   ],
 
   pages = [
diff --git a/install/install.nu b/install/install.nu
index 13cf621..6f3a1ea 100755
--- a/install/install.nu
+++ b/install/install.nu
@@ -52,7 +52,8 @@ def main [] {
   mkdir $bin_dir
 
   # ── 1. Binary → ontoref-daemon.bin ────────────────────────────────────────
-  let bin_src  = $"($repo_root)/target/release/ontoref-daemon"
+  let target_dir = (cd $repo_root; ^cargo metadata --format-version 1 --no-deps | from json | get target_directory)
+  let bin_src  = $"($target_dir)/release/ontoref-daemon"
   let bin_dest = $"($bin_dir)/ontoref-daemon.bin"
 
   if not ($bin_src | path exists) {
diff --git a/install/resources/config.ncl b/install/resources/config.ncl
index eb37147..442c1b9 100644
--- a/install/resources/config.ncl
+++ b/install/resources/config.ncl
@@ -12,13 +12,28 @@ let Port = std.contract.from_predicate
   (fun p => std.is_number p && p > 1024 && p < 65536)
   in
 
-let LogLevel  = [| 'trace, 'debug, 'info, 'warn, 'error |] in
-let Rotation  = [| 'daily, 'hourly, 'never |] in
-let Actor     = [| 'developer, 'agent, 'ci, 'admin |] in
-let Severity  = [| 'Hard, 'Soft |] in
+let LogLevel     = [| 'trace, 'debug, 'info, 'warn, 'error |] in
+let Rotation     = [| 'daily, 'hourly, 'never |] in
+let Actor        = [| 'developer, 'agent, 'ci, 'admin |] in
+let Severity     = [| 'Hard, 'Soft |] in
+let VaultBackend = [| 'restic, 'kopia |] in
 
 let KeyRole   = [| 'admin, 'viewer |] in
 let KeyEntry  = { role | KeyRole, hash | String } in
+let ActorKeyBindings = { _ | String } in
+let RecipientGroup   = { _ | Array String } in
+let RecipientRule    = { path | String, groups | Array String } in
+let SopsConfig = {
+  enabled | Bool | default = false,
+  vault_id | String | optional,
+  vault_backend | VaultBackend | default = 'restic,
+  registry_endpoint | String | optional,
+  age_key_paths | Array String | default = [],
+  master_key_path | String | optional,
+  actor_key_bindings | ActorKeyBindings | default = {},
+  recipient_groups | RecipientGroup    | default = {},
+  recipient_rules  | Array RecipientRule | default = [],
+} in
 let ProjectEntry = {
   slug                | String,
   root                | String,
@@ -26,6 +41,7 @@ let ProjectEntry = {
   keys                | Array KeyEntry | default = [],
   remote_url          | String | default = "",
   push_only           | Bool   | default = false,
+  sops                | SopsConfig | optional,
 } in
 
 {
@@ -75,6 +91,31 @@ let ProjectEntry = {
     path | String = "",
   },
 
+  # Default vault settings (ADR-017). Per-project override in <project>/.ontoref/project.ncl::sops.
+  vault = {
+    # Path to the master age private key (.kage) used to open project vaults.
+    # Empty here = each project must declare sops.master_key_path explicitly.
+    master_key_path | String = "",
+    # Default vault backend tool. Switching requires re-initializing local repos.
+    backend | VaultBackend = 'restic,
+    # Cosign keypair for signing/verifying src-vault OCI artifacts.
+    # Empty paths = cosign operations error explicitly. No env-var fallback.
+    cosign = {
+      key_path | String = "",   # cosign.key — used by `secrets push` to sign
+      pub_path | String = "",   # cosign.pub — used by `secrets sync` and verify
+      # Upload signature metadata to Rekor (sigstore public transparency log).
+      # Default false (private self-hosted model). Set true only if you publish
+      # artifacts that need globally verifiable, tamper-evident audit trail.
+      # Verification side propagates: false → cosign verify uses --insecure-ignore-tlog.
+      tlog | Bool = false,
+      # Cosign 2+ requires a --signing-config without Rekor URLs when tlog=false.
+      # Generate once with:
+      #   curl https://raw.githubusercontent.com/sigstore/root-signing/refs/heads/main/targets/signing_config.v0.2.json \
+      #     | jq 'del(.rekorTlogUrls) | del(.rekorTlogConfig)' > <path>
+      signing_config_path | String = "",
+    },
+  },
+
   ui = {
     # Resolved from platform data dir when empty:
     #   macOS: ~/Library/Application Support/ontoref/templates
diff --git a/install/resources/schemas/ontoref-project.ncl b/install/resources/schemas/ontoref-project.ncl
index 6f96499..7e9d7a8 100644
--- a/install/resources/schemas/ontoref-project.ncl
+++ b/install/resources/schemas/ontoref-project.ncl
@@ -9,6 +9,45 @@ let KeyEntry = {
   label | String | default = "",
 } in
 
+# Maps ONTOREF_ACTOR value → role name declared in scopes/<role>.ncl.
+# Example: { developer = "developer", ci = "cdci", ontoref = "ontoref" }
+let actor_key_bindings_type = { _ | String } in
+
+let VaultBackend = [| 'restic, 'kopia |] in
+
+# Per-file recipient routing (optional). When `recipient_rules` is declared,
+# bootstrap generates <vault_dir>/.sops.yaml from these definitions and sops
+# encrypts each file with the union of recipient groups matching its path.
+# Use case: multi-tenant projects where different clients/agents/teams
+# decrypt different credential files within the same vault.
+let recipient_group_type = { _ | Array String } in   # group → list of age1... pubkeys
+
+let recipient_rule_type = {
+  path   | String,                                   # path_regex relative to vault dir
+  groups | Array String,                             # union of these recipient_groups
+} in
+
+let SopsConfig = {
+  enabled | Bool | default = false,
+  # Slug used for ~/.config/ontoref/vaults/<vault_id>/. Defaults to project slug when absent.
+  vault_id | String | optional,
+  # Restic or kopia. Switching requires re-initializing the local repo.
+  vault_backend | VaultBackend | default = 'restic,
+  # ZOT registry endpoint where src-vault/<vault_id>:latest is pushed/pulled.
+  registry_endpoint | String | optional,
+  # Paths to age public key files for each declared recipient (relative or absolute).
+  age_key_paths | Array String | default = [],
+  # Path to the master age private key (.kage). Overrides the global path in config.ncl.
+  # Never store this inside the vault directory or the project root.
+  master_key_path | String | optional,
+  # ONTOREF_ACTOR → role name binding. Used by credential resolution to select the correct
+  # sops-encrypted file in scopes/<role>.ncl before calling oras.
+  actor_key_bindings | actor_key_bindings_type | default = {},
+  # Per-file recipient routing — empty = use SOPS_AGE_RECIPIENTS env (legacy mode).
+  recipient_groups | recipient_group_type    | default = {},
+  recipient_rules  | Array recipient_rule_type | default = [],
+} in
+
 let ProjectEntry = {
   slug                | String,
   # Absolute local path to the project root. Empty string for push_only projects.
@@ -21,6 +60,8 @@ let ProjectEntry = {
   remote_url          | String | default = "",
   # When true: no local file watch, no NCL import. Project pushes ontology via POST /sync.
   push_only           | Bool   | default = false,
+  # Registry credential vault configuration. Absent = no registry credentials declared.
+  sops                | SopsConfig | optional,
 } in
 
 {
diff --git a/install/resources/templates/integration/README.md b/install/resources/templates/integration/README.md
new file mode 100644
index 0000000..b5a5919
--- /dev/null
+++ b/install/resources/templates/integration/README.md
@@ -0,0 +1,78 @@
+# Integration Templates (ADR-015 / federated OCI artifacts)
+
+Three starting templates for integration. Pick the role you want to play; you
+can be both a producer and a consumer.
+
+## Pick the template that matches what you want to do
+
+| Template | Role | Outcome |
+|---|---|---|
+| `domain-producer/` | Define a typed contract for structured data | Push `domains/<participant>/<id>:<version>` to ZOT |
+| `mode-producer/` | Author an orchestration mode that consumes domains | Push `modes/<participant>/<id>:<version>` to ZOT |
+| `mode-consumer/` | Adopt someone else's published mode | Cabling file in `infra/<ws>/integrations/` |
+
+## Layout
+
+```text
+domain-producer/
+  contract.ncl.template       ← schema (Nickel types) for the structured data
+  example.json.template       ← reference instance matching the contract
+  manifest.ncl.template       ← DomainArtifact descriptor
+
+mode-producer/
+  provisioning.ncl.template   ← IntegrationMode declaration
+  domains.lock.ncl.template   ← pinned versions of consumed domains
+  manifest.ncl.template       ← ModeArtifact descriptor
+
+mode-consumer/
+  cabling.ncl.template        ← binds mode params to workspace values
+```
+
+## Workflow
+
+### Producer (one-time per artifact)
+
+```bash
+# 1. Copy template to your catalog dir
+cp -r install/resources/templates/integration/domain-producer/  catalog/domains/my-domain/
+mv catalog/domains/my-domain/contract.ncl.template  catalog/domains/my-domain/contract.ncl
+mv catalog/domains/my-domain/example.json.template  catalog/domains/my-domain/example.json
+mv catalog/domains/my-domain/manifest.ncl.template  catalog/domains/my-domain/manifest.ncl
+
+# 2. Edit the files: replace <placeholders>, define your types
+
+# 3. Publish
+ore secrets bootstrap                                    # one-time, vault setup
+prvng integration domain publish catalog/domains/my-domain <participant>
+prvng integration domain verify <participant>/my-domain 0.1.0
+```
+
+### Consumer (one-time per mode you adopt)
+
+```bash
+# 1. Subscribe — pulls the mode + its domain dependencies, scaffolds cabling
+prvng integration subscribe <mode-id> \
+    --mode-file infra/modes/<mode-id>.ncl \
+    --workspace-dir .
+
+# 2. Edit the generated cabling file
+$EDITOR infra/<ws>/integrations/<mode-id>.ncl
+
+# 3. Validate that all bindings resolve
+prvng integration validate <mode-id> --workspace-dir .
+
+# 4. Invoke the mode
+prvng integration invoke <mode-id> --binary <name>
+```
+
+## Without templates: minimal viable
+
+The artifact format is documented in `provisioning/schemas/lib/integration/oci_artifact_format.ncl`.
+You can write `contract.ncl`, `manifest.ncl`, `provisioning.ncl` from scratch following
+the type definitions there. Templates only save copy-paste time.
+
+## See also
+
+- `reflection/qa.ncl::integration-*` — FAQ with diagrams
+- `provisioning/schemas/lib/integration/oci_artifact_format.ncl` — types
+- `prvng integration --help` — command surface
diff --git a/install/resources/templates/integration/domain-producer/contract.ncl.template b/install/resources/templates/integration/domain-producer/contract.ncl.template
new file mode 100644
index 0000000..30d7ba1
--- /dev/null
+++ b/install/resources/templates/integration/domain-producer/contract.ncl.template
@@ -0,0 +1,32 @@
+# <my-domain> contract v0.1.0
+# Domain contract: typed shape for <describe what kind of data>.
+# Consumers receive the schema via OCI pull and bind values via cabling.
+
+# Helper enums and sub-records.
+let _<TypeName> = [| 'value_one, 'value_two |] in
+
+let _<NestedRecord> = {
+  field_a | String,
+  field_b | _<TypeName>,
+} in
+
+# Top-level context type — this is what consumers receive at the binding point.
+# Convention: name it <DomainName>Context.
+let _<MyDomain>Context = {
+  schema_version | String | default = "0.1.0",
+  workspace      | String,                            # provided by the consumer
+  mode_id        | String,                            # provided by the mode invoking this binding
+
+  # ── Replace below with your domain-specific fields ──
+  endpoint       | String
+                 | doc "Description of what this field means.",
+  options        | _<NestedRecord> | optional
+                 | doc "Optional structured config.",
+  flags          | Array String | default = [],
+} in
+
+{
+  <TypeName>          = _<TypeName>,
+  <NestedRecord>      = _<NestedRecord>,
+  <MyDomain>Context  = _<MyDomain>Context,
+}
diff --git a/install/resources/templates/integration/domain-producer/example.json.template b/install/resources/templates/integration/domain-producer/example.json.template
new file mode 100644
index 0000000..126e8d4
--- /dev/null
+++ b/install/resources/templates/integration/domain-producer/example.json.template
@@ -0,0 +1,11 @@
+{
+  "schema_version": "0.1.0",
+  "workspace": "<example-workspace>",
+  "mode_id": "<example-mode-using-this-domain>",
+  "endpoint": "https://example.com/api",
+  "options": {
+    "field_a": "sample",
+    "field_b": "value_one"
+  },
+  "flags": []
+}
diff --git a/install/resources/templates/integration/domain-producer/manifest.ncl.template b/install/resources/templates/integration/domain-producer/manifest.ncl.template
new file mode 100644
index 0000000..22d4210
--- /dev/null
+++ b/install/resources/templates/integration/domain-producer/manifest.ncl.template
@@ -0,0 +1,27 @@
+let oci = import "schemas/lib/integration/oci_artifact_format.ncl" in
+
+{
+  artifact = {
+    media_type  = "application/vnd.ontoref.domain.v1",
+    id          = "<my-domain-id>",
+    version     = "0.1.0",
+    description = "<one-line summary of what this domain represents>",
+    layers = [
+      {
+        media_type  = "application/vnd.ontoref.domain.contract.v1",
+        description = "contract.ncl — typed schema for <my-domain-id>",
+        required    = true,
+      },
+      {
+        media_type  = "application/vnd.ontoref.domain.example.v1",
+        description = "example.json — reference instance",
+        required    = false,
+      },
+    ],
+    # ADR-017 G2 — declare which RegistryEntry credentials this artifact depends
+    # on. Used by `ore secrets close` impact analysis to identify which artifacts
+    # are affected when credentials change. Empty if the domain does not consume
+    # registry credentials at runtime.
+    uses_registry = "primary",
+  } | oci.DomainArtifact,
+}
diff --git a/install/resources/templates/integration/mode-consumer/cabling.ncl.template b/install/resources/templates/integration/mode-consumer/cabling.ncl.template
new file mode 100644
index 0000000..d40941c
--- /dev/null
+++ b/install/resources/templates/integration/mode-consumer/cabling.ncl.template
@@ -0,0 +1,32 @@
+# <mode-id> cabling — bindings for this workspace.
+# Auto-scaffolded by `prvng integration subscribe`. Edit each binding to
+# resolve the placeholders against your workspace's values, secrets, manifest.
+
+let mode = import "<path-to-pulled-mode>/provisioning.ncl" in
+
+{
+  mode_id   = mode.id,
+
+  # One binding per domain in mode.domains_used. The shape MUST match the
+  # corresponding contract.ncl::<DomainName>Context. Run
+  #   prvng integration validate <mode-id> --workspace-dir .
+  # to confirm every field resolves and typechecks.
+  bindings = {
+
+    "secret-delivery" = {
+      schema_version = "0.1.0",
+      workspace      = "<your-workspace-slug>",
+      mode_id        = mode.id,
+      # ── Domain-specific fields below — see the secret-delivery contract ──
+      # secrets = [...]
+    },
+
+    "<other-domain-id>" = {
+      schema_version = "0.1.0",
+      workspace      = "<your-workspace-slug>",
+      mode_id        = mode.id,
+      # ── Fill in per the domain's contract.ncl ──
+    },
+
+  },
+}
diff --git a/install/resources/templates/integration/mode-producer/domains.lock.ncl.template b/install/resources/templates/integration/mode-producer/domains.lock.ncl.template
new file mode 100644
index 0000000..d3beda2
--- /dev/null
+++ b/install/resources/templates/integration/mode-producer/domains.lock.ncl.template
@@ -0,0 +1,18 @@
+let oci = import "schemas/lib/integration/oci_artifact_format.ncl" in
+
+# Pinned versions + digests of the domains this mode consumes. Auto-populated
+# by `prvng integration subscribe`; producer hand-writes the initial version
+# matching the contracts.ncl set referenced in provisioning.ncl::domains_used.
+
+{
+  schema_version = "0.1.0",
+  domains = {
+    "secret-delivery" = {
+      version    = "0.1.0",
+      digest     = "sha256:<populate-after-publish>",
+      pulled_at  = "<set-on-pull>",
+      media_type = "application/vnd.ontoref.domain.v1",
+    } | oci.DomainLockEntry,
+    # Add one entry per domain in provisioning.ncl::domains_used.
+  },
+} | oci.DomainLock
diff --git a/install/resources/templates/integration/mode-producer/manifest.ncl.template b/install/resources/templates/integration/mode-producer/manifest.ncl.template
new file mode 100644
index 0000000..3ef8b73
--- /dev/null
+++ b/install/resources/templates/integration/mode-producer/manifest.ncl.template
@@ -0,0 +1,24 @@
+let oci = import "schemas/lib/integration/oci_artifact_format.ncl" in
+
+{
+  artifact = {
+    media_type  = "application/vnd.ontoref.mode.v1",
+    id          = "<my-mode-id>",
+    version     = "0.1.0",
+    description = "<one-line summary of the mode's orchestration>",
+    participant = "<my-participant>",
+    layers = [
+      {
+        media_type  = "application/vnd.ontoref.mode.declaration.v1",
+        description = "provisioning.ncl — IntegrationMode declaration",
+        required    = true,
+      },
+      {
+        media_type  = "application/vnd.ontoref.mode.lock.v1",
+        description = "domains.lock.ncl — pinned domain versions",
+        required    = true,
+      },
+    ],
+    uses_registry = "primary",
+  } | oci.ModeArtifact,
+}
diff --git a/install/resources/templates/integration/mode-producer/provisioning.ncl.template b/install/resources/templates/integration/mode-producer/provisioning.ncl.template
new file mode 100644
index 0000000..c464ffb
--- /dev/null
+++ b/install/resources/templates/integration/mode-producer/provisioning.ncl.template
@@ -0,0 +1,43 @@
+let oci = import "schemas/lib/integration/oci_artifact_format.ncl" in
+
+# IntegrationMode declaration — the orchestration this artifact represents.
+# Consumers subscribe via `prvng integration subscribe <mode-id>` which pulls
+# this declaration and the domains_used dependencies, then scaffolds a
+# cabling file for them to bind workspace-specific values.
+
+{
+  id          = "<my-mode-id>",
+  participant = "<my-participant>",
+  version     = "0.1.0",
+  description = "<one-line summary>",
+  direction   = 'outbound,            # outbound | inbound | bidirectional
+  trigger     = 'manual,              # manual | scheduled | event
+
+  # Domains this mode consumes. Each entry pins a version + media type.
+  # The cabling file consumers write will provide values matching each domain's
+  # contract.ncl shape.
+  domains_used = [
+    "secret-delivery:0.1.0",
+    "<your-other-domain>:0.1.0",
+  ],
+
+  # Steps the mode executes. Each step receives the assembled context (bindings
+  # from the cabling file) via stdin or argv.
+  steps = [
+    {
+      id          = "step-1",
+      description = "<what this step does>",
+      invocation  = {
+        method = 'stdin_context,
+        binary = {
+          source = 'cargo_install,
+          name   = "<binary-name>",
+          version = "0.1.0",
+          cargo_crate = "<crate-name>",
+        },
+        args = [],
+        env  = {},
+      } | oci.Invocation,
+    },
+  ],
+}
diff --git a/install/resources/templates/sops/README.md b/install/resources/templates/sops/README.md
new file mode 100644
index 0000000..f9ea878
--- /dev/null
+++ b/install/resources/templates/sops/README.md
@@ -0,0 +1,65 @@
+# Credential Vault Templates (ADR-017)
+
+Three starting templates for adopting the credential vault model. Copy the snippet
+that matches your tenancy needs into your project's `.ontoref/project.ncl` and
+`.ontology/manifest.ncl`. None is mandatory — the schema fields are optional and
+projects can declare them ad-hoc.
+
+## Pick one
+
+| Template | When to use |
+|---|---|
+| **single-team** | One team, no tenant separation. CI uses same credentials as developers. |
+| **multi-tenant** | Multiple clients/projects share the same vault but must not see each other's credentials. |
+| **agent-first** | AI agents (MCP) consume credentials with strict RO scope; humans operate normally. |
+
+## Layout
+
+```text
+single-team/
+  project.ncl.snippet     ← paste into <your-project>/.ontoref/project.ncl
+  manifest.ncl.snippet    ← paste into <your-project>/.ontology/manifest.ncl
+multi-tenant/
+  project.ncl.snippet
+  manifest.ncl.snippet
+agent-first/
+  project.ncl.snippet
+  manifest.ncl.snippet
+```
+
+## Snippet conventions
+
+- `<placeholder>` — replace with project-specific value
+- `age1...` — generate via `age-keygen -o ~/.age/keys/<name>.key.txt`, copy
+  the public key from the line `# public key: ...`
+- All paths in `credential_sops` / `credential_sops_rw` are relative to the
+  src-vault root (i.e. `~/.config/ontoref/vaults/<vault_id>/src-vault/`)
+
+## Migration from one template to another
+
+Templates only differ in `recipient_groups` + `recipient_rules` and the
+`credential_sops` paths. To switch:
+
+1. Edit `project.ncl::sops.recipient_groups` + `recipient_rules`
+2. Edit `manifest.ncl::registry_provides.registries[].credential_sops*`
+3. `ore secrets rekey` — regenerates `<vault>/.sops.yaml` and re-encrypts
+4. `ore secrets close` — pushes the updated state
+
+## No template: legacy minimal
+
+If none of the templates fits, the minimal viable path is:
+
+```ncl
+sops = {
+  enabled            = true,
+  vault_id           = "<your-slug>",
+  master_key_path    = "/path/to/.kage",
+  vault_backend      = 'restic,
+  registry_endpoint  = "<your-zot-host>",
+  actor_key_bindings = { developer = "developer" },
+}
+```
+
+`recipient_groups` + `recipient_rules` left absent. The bootstrap will require
+`SOPS_AGE_RECIPIENTS` env var listing all recipient public keys. Same recipients
+across all encrypted files. Path A in the FAQ.
diff --git a/install/resources/templates/sops/agent-first/manifest.ncl.snippet b/install/resources/templates/sops/agent-first/manifest.ncl.snippet
new file mode 100644
index 0000000..21a4b37
--- /dev/null
+++ b/install/resources/templates/sops/agent-first/manifest.ncl.snippet
@@ -0,0 +1,37 @@
+# ───── Path C: agent-first manifest snippet ─────
+# Two RegistryEntries: 'primary' for humans (full RW), 'agent-ro' for agents (RO).
+# Agent role's scope file should declare ops = ['pull, 'verify, 'list] and
+# namespaces restricted to what agents may read.
+
+  registry_provides = m.make_registry_provides {
+    participant = "<your-slug>",
+    registries  = m.make_registries_config {
+      default    = "primary",
+      registries = [
+        m.make_registry_entry {
+          id                 = "primary",
+          endpoint           = "<your-zot-host>",
+          role               = 'primary,
+          tls                = true,
+          namespaces = {
+            own      = ["domains/<your-slug>/", "modes/<your-slug>/"],
+            prefixes = ["domains/<your-slug>/", "modes/<your-slug>/"],
+          },
+          credential_sops    = "registry/developer-ro.sops.yaml",
+          credential_sops_rw = "registry/admin-rw.sops.yaml",
+        },
+        m.make_registry_entry {
+          id                 = "agent-ro",
+          endpoint           = "<your-zot-host>",
+          role               = 'dev,
+          tls                = true,
+          namespaces = {
+            own      = [],
+            prefixes = ["domains/<your-slug>/"],
+          },
+          credential_sops    = "registry/agent-readonly.sops.yaml",
+          # No credential_sops_rw — agents cannot push, full stop.
+        },
+      ],
+    },
+  },
diff --git a/install/resources/templates/sops/agent-first/project.ncl.snippet b/install/resources/templates/sops/agent-first/project.ncl.snippet
new file mode 100644
index 0000000..7518572
--- /dev/null
+++ b/install/resources/templates/sops/agent-first/project.ncl.snippet
@@ -0,0 +1,30 @@
+# ───── Path C: agent-first ─────
+# AI agents (MCP) read a single agent-readonly credential file. Humans operate
+# normally with admin/developer roles. Paste inside s.make_project { ... }.
+
+  sops = {
+    enabled            = true,
+    vault_id           = "<your-slug>",
+    vault_backend      = 'restic,
+    registry_endpoint  = "<your-zot-host>",
+
+    actor_key_bindings = {
+      developer = "developer",
+      ci        = "cdci",
+      agent     = "agent-ro",       # mapped to RO scope file
+      admin     = "admin",
+    },
+
+    recipient_groups = {
+      admin     = ["age1admin..."],
+      developer = ["age1developer..."],
+      agents    = ["age1agent..."],   # one age key per running agent instance
+    },
+
+    recipient_rules = [
+      { path = "access\\.sops\\.yaml$",                   groups = ["admin", "developer"] },
+      { path = "registry/admin-.*\\.sops\\.yaml$",        groups = ["admin"] },
+      { path = "registry/developer-.*\\.sops\\.yaml$",    groups = ["admin", "developer"] },
+      { path = "registry/agent-readonly\\.sops\\.yaml$",  groups = ["admin", "agents"] },
+    ],
+  },
diff --git a/install/resources/templates/sops/multi-tenant/manifest.ncl.snippet b/install/resources/templates/sops/multi-tenant/manifest.ncl.snippet
new file mode 100644
index 0000000..a42c836
--- /dev/null
+++ b/install/resources/templates/sops/multi-tenant/manifest.ncl.snippet
@@ -0,0 +1,49 @@
+# ───── Path B: multi-tenant manifest snippet ─────
+# One RegistryEntry per tenant — credential_sops paths must match the
+# recipient_rules patterns declared in project.ncl.
+
+  registry_provides = m.make_registry_provides {
+    participant = "<your-slug>",
+    registries  = m.make_registries_config {
+      default    = "primary",
+      registries = [
+        m.make_registry_entry {
+          id                 = "primary",
+          endpoint           = "<your-zot-host>",
+          role               = 'primary,
+          tls                = true,
+          namespaces = {
+            own      = ["domains/<your-slug>/", "modes/<your-slug>/"],
+            prefixes = ["domains/<your-slug>/", "modes/<your-slug>/"],
+          },
+          # Default tenant — administrative fallback.
+          credential_sops    = "registry/ro.sops.yaml",
+          credential_sops_rw = "registry/rw.sops.yaml",
+        },
+        m.make_registry_entry {
+          id                 = "clientA",
+          endpoint           = "<your-zot-host>",
+          role               = 'dev,
+          tls                = true,
+          namespaces = {
+            own      = ["domains/clientA/", "modes/clientA/"],
+            prefixes = ["domains/clientA/"],
+          },
+          credential_sops    = "registry/clientA-ro.sops.yaml",
+          credential_sops_rw = "registry/clientA-rw.sops.yaml",
+        },
+        m.make_registry_entry {
+          id                 = "clientB",
+          endpoint           = "<your-zot-host>",
+          role               = 'dev,
+          tls                = true,
+          namespaces = {
+            own      = ["domains/clientB/", "modes/clientB/"],
+            prefixes = ["domains/clientB/"],
+          },
+          credential_sops    = "registry/clientB-ro.sops.yaml",
+          credential_sops_rw = "registry/clientB-rw.sops.yaml",
+        },
+      ],
+    },
+  },
diff --git a/install/resources/templates/sops/multi-tenant/project.ncl.snippet b/install/resources/templates/sops/multi-tenant/project.ncl.snippet
new file mode 100644
index 0000000..ce35448
--- /dev/null
+++ b/install/resources/templates/sops/multi-tenant/project.ncl.snippet
@@ -0,0 +1,34 @@
+# ───── Path B: multi-tenant ─────
+# Multiple clients share the vault; their credentials are encrypted to disjoint
+# recipient sets. Paste inside s.make_project { ... }. Adapt groups + rules.
+
+  sops = {
+    enabled            = true,
+    vault_id           = "<your-slug>",
+    vault_backend      = 'restic,
+    registry_endpoint  = "<your-zot-host>",
+
+    actor_key_bindings = {
+      developer = "developer",
+      ci        = "cdci",
+      agent     = "ontoref",
+      admin     = "admin",
+    },
+
+    # Recipient groups — each is a list of age public keys. Use empty [] for
+    # placeholder groups you populate later.
+    recipient_groups = {
+      admin   = ["age1admin..."],
+      ops     = ["age1ops..."],
+      clientA = ["age1clientA-lead..."],
+      clientB = ["age1clientB-lead..."],
+    },
+
+    # Path-to-recipient-set rules. Each *.sops.yaml created in the vault is
+    # encrypted with the union of its rule's groups. Order matters — first match wins.
+    recipient_rules = [
+      { path = "access\\.sops\\.yaml$",                   groups = ["admin", "ops"] },
+      { path = "registry/clientA-.*\\.sops\\.yaml$",      groups = ["admin", "clientA"] },
+      { path = "registry/clientB-.*\\.sops\\.yaml$",      groups = ["admin", "clientB"] },
+    ],
+  },
diff --git a/install/resources/templates/sops/single-team/manifest.ncl.snippet b/install/resources/templates/sops/single-team/manifest.ncl.snippet
new file mode 100644
index 0000000..02f6049
--- /dev/null
+++ b/install/resources/templates/sops/single-team/manifest.ncl.snippet
@@ -0,0 +1,23 @@
+# ───── Path A: single-team manifest snippet ─────
+# Paste inside your m.make_manifest { ... } block. Replace <placeholders>.
+
+  registry_provides = m.make_registry_provides {
+    participant = "<your-slug>",
+    registries  = m.make_registries_config {
+      default    = "primary",
+      registries = [
+        m.make_registry_entry {
+          id                 = "primary",
+          endpoint           = "<your-zot-host>",
+          role               = 'primary,
+          tls                = true,
+          namespaces = {
+            own      = ["domains/<your-slug>/", "modes/<your-slug>/"],
+            prefixes = ["domains/<your-slug>/", "modes/<your-slug>/"],
+          },
+          credential_sops    = "registry/ro.sops.yaml",   # RO — pull/list
+          credential_sops_rw = "registry/rw.sops.yaml",   # RW — push
+        },
+      ],
+    },
+  },
diff --git a/install/resources/templates/sops/single-team/project.ncl.snippet b/install/resources/templates/sops/single-team/project.ncl.snippet
new file mode 100644
index 0000000..677fa37
--- /dev/null
+++ b/install/resources/templates/sops/single-team/project.ncl.snippet
@@ -0,0 +1,20 @@
+# ───── Path A: single-team ─────
+# Paste inside your s.make_project { ... } block. Replace <placeholders>.
+# Bootstrap requires SOPS_AGE_RECIPIENTS env var (comma-separated age public keys)
+# because no recipient_groups are declared in this template.
+
+  sops = {
+    enabled            = true,
+    vault_id           = "<your-slug>",
+    vault_backend      = 'restic,
+    registry_endpoint  = "<your-zot-host>",
+    # master_key_path absent → resolves from ~/.config/ontoref/config.ncl::vault.master_key_path.
+    # Override per-project only when this project requires a different key.
+
+    actor_key_bindings = {
+      developer = "developer",
+      ci        = "cdci",
+      agent     = "ontoref",
+      admin     = "admin",
+    },
+  },
diff --git a/justfile b/justfile
index 5f7874b..4a9688c 100644
--- a/justfile
+++ b/justfile
@@ -3,6 +3,7 @@ import 'justfiles/test.just'
 import 'justfiles/dev.just'
 import 'justfiles/ci.just'
 import 'justfiles/assets.just'
+import 'justfiles/secrets.just'
 
 project_root := justfile_directory()
 
diff --git a/justfiles/_secrets_lib.sh b/justfiles/_secrets_lib.sh
new file mode 100644
index 0000000..3c8b09d
--- /dev/null
+++ b/justfiles/_secrets_lib.sh
@@ -0,0 +1,245 @@
+#!/usr/bin/env bash
+# justfiles/_secrets_lib.sh — shared loader for secrets-* recipes (ADR-017).
+# Sourced from each recipe via: source "{{justfile_directory()}}/justfiles/_secrets_lib.sh"
+#
+# After sourcing, the following variables are set and validated:
+#   PROJECT_ROOT, PROJECT_NCL, IMPORT_PATH
+#   PROJECT_JSON       — exported project.ncl as JSON (jq-able)
+#   GLOBAL_JSON        — exported ~/.config/ontoref/config.ncl as JSON
+#   SLUG               — project.ncl::slug
+#   VAULT_ID           — project.ncl::sops.vault_id, fallback to SLUG
+#   VAULT_TOOL         — project.ncl::sops.vault_backend (restic|kopia)
+#   REGISTRY           — project.ncl::sops.registry_endpoint
+#   MASTER_KEY         — project override → ~/.config/ontoref/config.ncl::vault.master_key_path
+#   COSIGN_KEY_PATH    — global vault.cosign.key_path (signing)
+#   COSIGN_PUB_PATH    — global vault.cosign.pub_path (verification)
+#   VAULT_DIR          — ~/.config/ontoref/vaults/<vault_id>
+#   ACCESS_SOPS        — $VAULT_DIR/access.sops.yaml
+
+set -euo pipefail
+
+PROJECT_ROOT="${ONTOREF_PROJECT_ROOT:-$(pwd)}"
+PROJECT_NCL="${PROJECT_ROOT}/.ontoref/project.ncl"
+[[ -f "$PROJECT_NCL" ]] || { echo "No .ontoref/project.ncl found at $PROJECT_ROOT"; exit 1; }
+
+IMPORT_PATH="${NICKEL_IMPORT_PATH:-${PROVISIONING:-}}"
+PROJECT_JSON=$(nickel export --import-path "$IMPORT_PATH" "$PROJECT_NCL")
+GLOBAL_JSON=$(nickel export --import-path "$IMPORT_PATH" "$HOME/.config/ontoref/config.ncl" 2>/dev/null || echo '{}')
+
+SLUG=$(echo "$PROJECT_JSON" | jq -r '.slug // empty')
+VAULT_ID=$(echo "$PROJECT_JSON" | jq -r '.sops.vault_id // empty')
+VAULT_ID="${VAULT_ID:-$SLUG}"
+VAULT_TOOL=$(echo "$PROJECT_JSON" | jq -r '.sops.vault_backend // "restic"')
+REGISTRY=$(echo "$PROJECT_JSON" | jq -r '.sops.registry_endpoint // empty')
+
+MASTER_KEY=$(echo "$PROJECT_JSON" | jq -r '.sops.master_key_path // empty')
+if [[ -z "$MASTER_KEY" ]]; then
+    MASTER_KEY=$(echo "$GLOBAL_JSON" | jq -r '.vault.master_key_path // empty')
+fi
+
+COSIGN_KEY_PATH=$(echo "$GLOBAL_JSON" | jq -r '.vault.cosign.key_path // empty')
+COSIGN_PUB_PATH=$(echo "$GLOBAL_JSON" | jq -r '.vault.cosign.pub_path // empty')
+COSIGN_TLOG=$(echo "$GLOBAL_JSON" | jq -r '.vault.cosign.tlog // false')
+COSIGN_SIGNING_CONFIG=$(echo "$GLOBAL_JSON" | jq -r '.vault.cosign.signing_config_path // empty')
+
+# Sign/verify flags derived from tlog policy. When tlog=false we sign without
+# Rekor (cosign 2+ requires --signing-config that excludes rekorTlogUrls/Config)
+# and verify with --insecure-ignore-tlog so missing entries do not fail.
+# Symmetric — sign and verify must agree.
+if [[ "$COSIGN_TLOG" == "true" ]]; then
+    COSIGN_SIGN_FLAGS=""
+    COSIGN_VERIFY_FLAGS=""
+else
+    if [[ -z "$COSIGN_SIGNING_CONFIG" ]]; then
+        echo "vault.cosign.tlog=false but signing_config_path not set — generate one and declare it" >&2
+        exit 1
+    fi
+    [[ -f "$COSIGN_SIGNING_CONFIG" ]] || {
+        echo "Cosign signing-config not found at $COSIGN_SIGNING_CONFIG" >&2
+        exit 1
+    }
+    COSIGN_SIGN_FLAGS="--signing-config $COSIGN_SIGNING_CONFIG"
+    COSIGN_VERIFY_FLAGS="--insecure-ignore-tlog"
+fi
+
+VAULT_DIR="$HOME/.config/ontoref/vaults/$VAULT_ID"
+ACCESS_SOPS="$VAULT_DIR/access.sops.yaml"
+
+[[ -n "$VAULT_ID" ]]   || { echo "Cannot determine vault_id (slug empty and sops.vault_id absent)"; exit 1; }
+[[ -n "$MASTER_KEY" ]] || { echo "master_key_path not declared (project.ncl::sops or ~/.config/ontoref/config.ncl::vault)"; exit 1; }
+[[ -f "$MASTER_KEY" ]] || { echo "Master key (.kage) not found at $MASTER_KEY"; exit 1; }
+
+# Auto-rotate audit log on every recipe invocation (call at end of this file
+# after vault_rotate_audit_log is defined — bash does not hoist functions).
+
+# Vault lock TTL (minutes). Past this, locks are considered abandoned and
+# admin can force-close. Per ADR-017 G1 — cooperative lock atop ZOT ACL.
+VAULT_LOCK_TTL_MIN=60
+VAULT_LOCK_TAG="lock"   # OCI ref: src-vault/<id>:lock
+
+# Audit log rotation thresholds. access.jsonl is rotated when EITHER applies:
+#   size:  current file > VAULT_AUDIT_ROTATE_BYTES (default 10MB)
+#   age:   current file mtime older than VAULT_AUDIT_ROTATE_DAYS (default 90)
+# Rotated to access.jsonl.<UTC-ISO>.gz inside the same logs/ dir.
+VAULT_AUDIT_ROTATE_BYTES=${VAULT_AUDIT_ROTATE_BYTES:-10485760}
+VAULT_AUDIT_ROTATE_DAYS=${VAULT_AUDIT_ROTATE_DAYS:-90}
+
+# Rotate $VAULT_DIR/logs/access.jsonl when size or age exceeds thresholds.
+# Idempotent: noop when file absent or under both thresholds. Safe to call at
+# the start of every secrets-* operation.
+vault_rotate_audit_log() {
+    local log="$VAULT_DIR/logs/access.jsonl"
+    [[ -f "$log" ]] || return 0
+
+    local size_bytes age_days now_epoch mtime_epoch should_rotate=0
+    size_bytes=$(stat -f %z "$log" 2>/dev/null || stat -c %s "$log" 2>/dev/null || echo 0)
+    [[ "$size_bytes" -gt "$VAULT_AUDIT_ROTATE_BYTES" ]] && should_rotate=1
+
+    if [[ "$should_rotate" -eq 0 ]]; then
+        now_epoch=$(date -u +%s)
+        mtime_epoch=$(stat -f %m "$log" 2>/dev/null || stat -c %Y "$log" 2>/dev/null || echo "$now_epoch")
+        age_days=$(( (now_epoch - mtime_epoch) / 86400 ))
+        [[ "$age_days" -gt "$VAULT_AUDIT_ROTATE_DAYS" ]] && should_rotate=1
+    fi
+
+    if [[ "$should_rotate" -eq 1 ]]; then
+        local stamp rotated
+        stamp=$(date -u +%Y-%m-%dT%H-%M-%SZ)
+        rotated="$VAULT_DIR/logs/access.jsonl.${stamp}.gz"
+        gzip -c "$log" > "$rotated" && : > "$log"
+        # Append a marker entry recording the rotation so audit chain stays continuous.
+        printf '{"ts":"%s","actor":"%s","op":"audit-log-rotated","detail":"previous=%s size=%s"}\n' \
+            "$(date -u +%Y-%m-%dT%H:%M:%SZ)" "${ONTOREF_ACTOR:-system}" \
+            "access.jsonl.${stamp}.gz" "$size_bytes" >> "$log"
+    fi
+}
+
+# Fetch current lock state from ZOT. Returns JSON {locked_by,actor,since,expires}
+# or empty string if no lock or fetch fails. ZOT credentials must be in $TMPDIR_CFG/config.json.
+# Caller provides $TMPDIR_CFG before invoking.
+vault_lock_fetch() {
+    local out
+    out=$(DOCKER_CONFIG="$TMPDIR_CFG" oras manifest fetch \
+        "${REGISTRY}/src-vault/${VAULT_ID}:${VAULT_LOCK_TAG}" 2>/dev/null) || return 0
+    # Lock metadata is in the manifest annotations — extract our keys.
+    echo "$out" | jq -c '{
+        locked_by:       (.annotations."ontoref.vault.locked_by"       // ""),
+        actor:           (.annotations."ontoref.vault.actor"           // ""),
+        since:           (.annotations."ontoref.vault.since"           // ""),
+        expires:         (.annotations."ontoref.vault.expires"         // ""),
+        force_closed_by: (.annotations."ontoref.vault.force_closed_by" // "")
+    }'
+}
+
+# Push a lock artifact with the given metadata. Requires $TMPDIR_CFG with ZOT creds.
+# Args: locked_by actor since expires
+vault_lock_push() {
+    local locked_by="$1" actor="$2" since="$3" expires="$4"
+    # oras rejects absolute paths — work in a tmp dir with relative filename.
+    local lock_dir lock_file rc
+    lock_dir=$(mktemp -d)
+    lock_file="lock.json"
+    # Lock metadata is in OCI annotations; the blob just needs ≥ 1 byte —
+    # ZOT rejects empty blobs (returns 400 on the PUT). Use a minimal JSON
+    # placeholder so the file is self-describing if pulled.
+    printf '{"kind":"vault-lock"}\n' > "$lock_dir/$lock_file"
+    set +e
+    ( cd "$lock_dir" && DOCKER_CONFIG="$TMPDIR_CFG" oras push \
+        "${REGISTRY}/src-vault/${VAULT_ID}:${VAULT_LOCK_TAG}" \
+        --artifact-type "application/vnd.ontoref.vault.lock.v1" \
+        --annotation "ontoref.vault.locked_by=${locked_by}" \
+        --annotation "ontoref.vault.actor=${actor}" \
+        --annotation "ontoref.vault.since=${since}" \
+        --annotation "ontoref.vault.expires=${expires}" \
+        "${lock_file}:application/vnd.ontoref.vault.lock.metadata.v1" ) >/dev/null 2>&1
+    rc=$?
+    set -e
+    rm -rf "$lock_dir"
+    return $rc
+}
+
+# Delete the lock artifact from ZOT. Requires $TMPDIR_CFG.
+vault_lock_release() {
+    DOCKER_CONFIG="$TMPDIR_CFG" oras manifest delete -f \
+        "${REGISTRY}/src-vault/${VAULT_ID}:${VAULT_LOCK_TAG}" >/dev/null 2>&1 || true
+}
+
+# Build a $TMPDIR_CFG from the decrypted access.sops.yaml. Sets the global
+# TMPDIR_CFG variable; caller is responsible for `rm -rf "$TMPDIR_CFG"` later.
+vault_zot_config_open() {
+    [[ -f "$ACCESS_SOPS" ]] || { echo "[lock] access.sops.yaml absent — bootstrap or sync first"; return 1; }
+    local creds zu zp
+    creds=$(SOPS_AGE_KEY_FILE="$MASTER_KEY" sops --decrypt "$ACCESS_SOPS")
+    zu=$(echo "$creds" | grep '^zot_username:' | awk '{print $2}')
+    zp=$(echo "$creds" | grep '^zot_password:' | awk '{print $2}')
+    TMPDIR_CFG=$(mktemp -d)
+    local b64
+    b64=$(printf '%s:%s' "$zu" "$zp" | base64)
+    printf '{"auths":{"%s":{"auth":"%s"}}}' "$REGISTRY" "$b64" > "$TMPDIR_CFG/config.json"
+    unset creds zu zp b64
+}
+
+# ── ADR-017 G2 — Impact analysis ───────────────────────────────────────────────
+# vault_impact_report: list relative paths of *.sops.yaml files in the vault
+# that have changed since the last restic snapshot. Empty list = no credential
+# changes. Output one path per line, relative to $VAULT_DIR.
+vault_impact_changed_creds() {
+    local repo="$VAULT_DIR/repo"
+    [[ -d "$repo" ]] || return 0
+    local vk
+    vk=$(SOPS_AGE_KEY_FILE="$MASTER_KEY" sops --decrypt --extract '["vault_key"]' "$ACCESS_SOPS" 2>/dev/null)
+    [[ -n "$vk" ]] || return 0
+
+    local snaps
+    snaps=$(RESTIC_PASSWORD="$vk" restic -r "$repo" snapshots --json 2>/dev/null) || return 0
+    local last_id
+    last_id=$(echo "$snaps" | jq -r 'last | .id // empty' 2>/dev/null)
+    [[ -n "$last_id" ]] || return 0
+
+    # Take an in-memory snapshot of current files (no restic backup — too invasive
+    # for a dry diff). Use a hash comparison instead: hash each *.sops.yaml in the
+    # last snapshot via 'restic dump' and compare to current hash.
+    local cur_dir snap_dir tmp_diff
+    cur_dir="$VAULT_DIR"
+    tmp_diff=$(mktemp)
+    while IFS= read -r rel; do
+        [[ -z "$rel" ]] && continue
+        local cur_hash="" snap_hash=""
+        if [[ -f "$cur_dir/$rel" ]]; then
+            cur_hash=$(shasum -a 256 "$cur_dir/$rel" | awk '{print $1}')
+        fi
+        snap_hash=$(RESTIC_PASSWORD="$vk" restic -r "$repo" dump "$last_id" "$rel" 2>/dev/null \
+            | shasum -a 256 | awk '{print $1}')
+        if [[ "$cur_hash" != "$snap_hash" ]]; then
+            echo "$rel" >> "$tmp_diff"
+        fi
+    done < <(find "$VAULT_DIR" -name "*.sops.yaml" -type f 2>/dev/null \
+             | while read -r f; do echo "${f#${VAULT_DIR}/}"; done)
+    cat "$tmp_diff"
+    rm -f "$tmp_diff"
+}
+
+# vault_impact_registries: given a list of changed sops paths (relative), find
+# which RegistryEntry IDs in the project manifest reference them via
+# credential_sops or credential_sops_rw. Output: <entry_id>\t<sops_path>\t<field>
+vault_impact_registries() {
+    local manifest_ncl="$PROJECT_ROOT/.ontology/manifest.ncl"
+    [[ -f "$manifest_ncl" ]] || return 0
+    local manifest_json
+    manifest_json=$(nickel export --import-path "$IMPORT_PATH" "$manifest_ncl" 2>/dev/null) || return 0
+    local entries
+    entries=$(echo "$manifest_json" | jq -c '.registry_provides.registries.registries // []' 2>/dev/null)
+    [[ "$entries" != "null" && -n "$entries" ]] || return 0
+
+    while IFS= read -r changed_full; do
+        # Strip leading 'src-vault/' since credential_sops paths are relative to src-vault/.
+        local rel="${changed_full#src-vault/}"
+        echo "$entries" | jq -r --arg rel "$rel" '
+            .[] | select(.credential_sops == $rel or .credential_sops_rw == $rel)
+                | "\(.id)\t\($rel)\t\(if .credential_sops == $rel then "credential_sops" else "credential_sops_rw" end)"
+        '
+    done
+}
+
+# Final action of this lib: rotate audit log if needed. Idempotent.
+vault_rotate_audit_log
diff --git a/justfiles/build.just b/justfiles/build.just
index e94bf78..7862fc4 100644
--- a/justfiles/build.just
+++ b/justfiles/build.just
@@ -8,7 +8,7 @@ build-all:
 
 # Build ontoref-daemon in release mode and install
 build-daemon: build-css
-    cargo build --release -p ontoref-daemon
+    cargo build --release -p ontoref-daemon --features "db,nats,ui,mcp,graphql"
     nu install/install.nu
 
 # Build UnoCSS bundle for the daemon UI
@@ -23,3 +23,29 @@ build-standalone:
 # Check all crates compile (fast feedback, no codegen)
 check:
     cargo check --all-targets --all-features
+
+# Generate a new GraphQL read token and print its Argon2id hash.
+# Copy the TOKEN value to use in API calls; copy the HASH into config.ncl graphql.read_token_hash.
+generate-graphql-token:
+    #!/usr/bin/env bash
+    set -euo pipefail
+    TOKEN=$(openssl rand -base64 18 | tr -d '=/+' | head -c 24)
+    HASH=$(ontoref-daemon --hash-password "$TOKEN")
+    echo "TOKEN: $TOKEN"
+    echo "HASH:  $HASH"
+    echo ""
+    echo "Add to ~/.config/ontoref/config.ncl:"
+    echo "  graphql = { read_token_hash | String = \"$HASH\" },"
+
+# Generate a new admin key-pair and print the Argon2id hash for config.ncl.
+# Use ontoref-daemon --gen-keys for full key-pair generation with role assignment.
+generate-admin-key:
+    #!/usr/bin/env bash
+    set -euo pipefail
+    TOKEN=$(openssl rand -base64 18 | tr -d '=/+' | head -c 24)
+    HASH=$(ontoref-daemon --hash-password "$TOKEN")
+    echo "TOKEN: $TOKEN"
+    echo "HASH:  $HASH"
+    echo ""
+    echo "Add to your project keys[] in config.ncl:"
+    echo "  { role = 'admin, hash = \"$HASH\", label = \"admin\" },"
diff --git a/justfiles/ci.just b/justfiles/ci.just
index 2034f7b..505d5fa 100644
--- a/justfiles/ci.just
+++ b/justfiles/ci.just
@@ -32,7 +32,7 @@ help:
     @echo "  just clean            - Clean build artifacts"
 
 # Run all CI checks
-ci-full: ci-lint-rust ci-fmt-toml ci-lint-toml ci-lint-nickel ci-lint-markdown ci-lint-prose ci-test ci-audit ci-check-config-sync ci-docs
+ci-full: ci-lint-rust ci-fmt-toml ci-lint-toml ci-lint-nickel ci-lint-markdown ci-lint-prose ci-test ci-secrets-error-fixtures ci-secrets-smoke ci-vault-lock-smoke ci-audit ci-check-config-sync ci-docs
     @echo "✅ All CI checks passed!"
 
 # ==============================================================================
@@ -119,6 +119,38 @@ ci-test-coverage:
     @echo "📊 Running tests with coverage..."
     cargo llvm-cov --all-features --lcov --output-path lcov.info
 
+# End-to-end credential chain smoke (ADR-017 + ADR-019).
+# Pushes a throwaway vault to ZOT, pulls it back from an isolated HOME, decrypts.
+# Skipped automatically when ZOT credentials are not exported — keeps `ci-full`
+# green on CI workers without registry access. To run it: export
+# RECOVERY_ZOT_USER and RECOVERY_ZOT_PASS, then `just ci-secrets-smoke`.
+ci-secrets-smoke:
+    #!/usr/bin/env bash
+    set -euo pipefail
+    if [[ -z "${RECOVERY_ZOT_USER:-}${ZOT_USER:-}" ]] || [[ -z "${RECOVERY_ZOT_PASS:-}${ZOT_PASS:-}" ]]; then
+        echo "🔐 ci-secrets-smoke: skipped (RECOVERY_ZOT_USER/PASS not set)"
+        exit 0
+    fi
+    echo "🔐 Running credential chain smoke..."
+    bash reflection/tests/test_credential_chain.sh
+
+# Vault lock concurrency smoke (ADR-017 G1). Same skip-when-unauth behavior.
+ci-vault-lock-smoke:
+    #!/usr/bin/env bash
+    set -euo pipefail
+    if [[ -z "${RECOVERY_ZOT_USER:-}${ZOT_USER:-}" ]] || [[ -z "${RECOVERY_ZOT_PASS:-}${ZOT_PASS:-}" ]]; then
+        echo "🔐 ci-vault-lock-smoke: skipped (RECOVERY_ZOT_USER/PASS not set)"
+        exit 0
+    fi
+    echo "🔐 Running vault lock concurrency smoke..."
+    bash reflection/tests/test_vault_lock.sh
+
+# Named-error fixture coverage (15 [error-tag] cases — protocol baseline).
+# Always runs — the fixtures are local, no registry needed.
+ci-secrets-error-fixtures:
+    @echo "🔐 Running named-error fixture coverage..."
+    nu reflection/tests/test_secrets.nu
+
 
 
 # ==============================================================================
@@ -177,10 +209,15 @@ hooks-run-all:
 # Install
 # ==============================================================================
 
-# Build and install daemon (alias — prefer build-daemon)
+# Build and install ontoref locally (binary + CLI wrapper + data + nulib + config dir).
+# Despite the name, this installs the full ontoref distribution — see install/install.nu.
 install-daemon:
     just build-daemon
 
+# Alias for install-daemon — clearer name for the full local install.
+install:
+    just install-daemon
+
 # ==============================================================================
 # Utility Commands
 # ==============================================================================
diff --git a/justfiles/secrets.just b/justfiles/secrets.just
new file mode 100644
index 0000000..b492cdf
--- /dev/null
+++ b/justfiles/secrets.just
@@ -0,0 +1,770 @@
+# Secrets Module — src-vault credential management (ADR-017)
+# =============================================================
+# All operations read credentials from ~/.config/ontoref/vaults/<vault_id>/access.sops.yaml,
+# decrypted by the .kage master key resolved from project.ncl::sops.master_key_path
+# (per-project override) or ~/.config/ontoref/config.ncl::vault.master_key_path (global).
+# Cosign keypair for vault signing/verification lives in ~/.config/ontoref/config.ncl::vault.cosign.
+# Direct restic/kopia or sops invocations bypass lock and audit — use these recipes instead.
+#
+# Common variable loading lives in justfiles/_secrets_lib.sh. Each recipe sources it via
+# `{{source_directory()}}/_secrets_lib.sh` to get a consistent, validated env.
+
+# Generate a new age keypair for a role. Prints public + private to stdout.
+secrets-gen-key role:
+    #!/usr/bin/env bash
+    set -euo pipefail
+    which age-keygen > /dev/null || { echo "age-keygen not found — install age"; exit 1; }
+    echo "Generating age keypair for role: {{role}}"
+    age-keygen
+
+# Bootstrap the vault for a new project. Admin-only.
+# Collects all recipient public keys, creates access.sops.yaml encrypted for all recipients,
+# initializes the local restic/kopia repo. Run secrets-push afterward to upload to ZOT.
+#
+# Prerequisites:
+#   - All recipient age public keys collected (one per role)
+#   - Admin .kage available at master_key_path
+#   - SOPS_AGE_RECIPIENTS set to comma-separated list of all recipient age public keys
+secrets-bootstrap:
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+    mkdir -p "$VAULT_DIR/logs"
+
+    # Per-file recipient routing — when project declares recipient_rules, generate
+    # <vault_dir>/.sops.yaml so sops uses different recipient unions per file path.
+    # Otherwise, fall back to SOPS_AGE_RECIPIENTS env var (legacy single-set mode).
+    RULES_JSON=$(echo "$PROJECT_JSON" | jq -c '.sops.recipient_rules // []')
+    GROUPS_JSON=$(echo "$PROJECT_JSON" | jq -c '.sops.recipient_groups // {}')
+    HAS_RULES=$(echo "$RULES_JSON" | jq 'length > 0')
+    SOPS_CFG_PATH="$VAULT_DIR/.sops.yaml"
+
+    if [[ "$HAS_RULES" == "true" ]]; then
+        # ADR-019 coherency check — every group referenced by a rule MUST exist
+        # as a key in recipient_groups, AND every recipient_group MUST resolve to
+        # at least one non-empty age public key. A typo or empty group produces
+        # an unreadable .sops.yaml — better to fail loud at bootstrap.
+        UNKNOWN_GROUPS=$(echo "$RULES_JSON" | jq -r --argjson groups "$GROUPS_JSON" '
+            [.[] | .groups[] | . as $g | select($groups | has($g) | not)] | unique | join(", ")
+        ')
+        if [[ -n "$UNKNOWN_GROUPS" ]]; then
+            echo "[recipient-rules] groups referenced in recipient_rules but missing from recipient_groups: $UNKNOWN_GROUPS" >&2
+            echo "  → declare them in project.ncl::sops.recipient_groups, or fix the typo in recipient_rules[].groups" >&2
+            exit 1
+        fi
+        EMPTY_GROUPS=$(echo "$GROUPS_JSON" | jq -r '
+            to_entries | map(select(.value | length == 0) | .key) | join(", ")
+        ')
+        if [[ -n "$EMPTY_GROUPS" ]]; then
+            echo "[recipient-rules] recipient_groups with no age keys: $EMPTY_GROUPS" >&2
+            echo "  → populate them in project.ncl::sops.recipient_groups, or remove the unused group(s)" >&2
+            exit 1
+        fi
+        UNREFERENCED_GROUPS=$(jq -nrc \
+            --argjson rules "$RULES_JSON" --argjson groups "$GROUPS_JSON" '
+            ($groups | keys) - ($rules | map(.groups[]) | unique) | join(", ")
+        ')
+        if [[ -n "$UNREFERENCED_GROUPS" ]]; then
+            echo "[recipient-rules] WARN: recipient_groups declared but never used by any rule: $UNREFERENCED_GROUPS" >&2
+        fi
+
+        # Resolve groups → age list per rule, write .sops.yaml
+        echo "creation_rules:" > "$SOPS_CFG_PATH"
+        # Use YAML single-quoted strings for path_regex so backslash escapes
+        # like \\. pass through to sops without YAML interpreting them.
+        echo "$RULES_JSON" | jq -r --argjson groups "$GROUPS_JSON" '
+            .[] | "  - path_regex: '"'"'\(.path)'"'"'\n    age: " +
+            ([.groups[] | $groups[.] // []] | flatten | unique | join(","))
+        ' >> "$SOPS_CFG_PATH"
+        echo "Generated $SOPS_CFG_PATH with $(echo "$RULES_JSON" | jq length) rule(s)."
+        # When sops sees --config, it uses it; we don't need --age flag.
+        SOPS_ENC_FLAGS=(--config "$SOPS_CFG_PATH")
+    else
+        [[ -n "${SOPS_AGE_RECIPIENTS:-}" ]] || {
+            echo "SOPS_AGE_RECIPIENTS not set — export age public keys, OR declare sops.recipient_groups+rules in project.ncl"
+            exit 1
+        }
+        SOPS_ENC_FLAGS=(--age "$SOPS_AGE_RECIPIENTS")
+    fi
+
+    if [[ -f "$ACCESS_SOPS" ]]; then
+        echo "access.sops.yaml already exists at $ACCESS_SOPS — skipping creation"
+    else
+        echo "Creating access.sops.yaml. Required inputs:"
+        echo "  zot_username + zot_password : ZOT registry credentials for src-vault/<id> namespace"
+        echo "  vault_key                    : restic/kopia repo password"
+        echo "  cosign_password              : password for vault.cosign.key_path (used by 'secrets push' to sign)"
+        echo
+        read -r -p "zot_username: " ZOT_USER
+        read -r -s -p "zot_password: " ZOT_PASS; echo
+        read -r -s -p "vault_key (restic/kopia password): " VAULT_KEY; echo
+        read -r -s -p "cosign_password (for cosign.key): " COSIGN_PWD; echo
+
+        printf 'zot_username: %s\nzot_password: %s\nvault_key: %s\ncosign_password: %s\n' \
+            "$ZOT_USER" "$ZOT_PASS" "$VAULT_KEY" "$COSIGN_PWD" \
+            | SOPS_AGE_KEY_FILE="$MASTER_KEY" sops "${SOPS_ENC_FLAGS[@]}" \
+                --filename-override "$ACCESS_SOPS" \
+                --input-type yaml --encrypt /dev/stdin > "$ACCESS_SOPS"
+        unset ZOT_USER ZOT_PASS VAULT_KEY COSIGN_PWD
+        echo "access.sops.yaml created and encrypted for all recipients"
+    fi
+
+    REPO_PATH="$VAULT_DIR/repo"
+    if [[ -d "$REPO_PATH" ]]; then
+        echo "Vault repo already exists at $REPO_PATH — skipping init"
+    else
+        CREDS=$(SOPS_AGE_KEY_FILE="$MASTER_KEY" sops --decrypt "$ACCESS_SOPS")
+        VAULT_KEY_VAL=$(echo "$CREDS" | grep vault_key | awk '{print $2}')
+        case "$VAULT_TOOL" in
+            restic) RESTIC_PASSWORD="$VAULT_KEY_VAL" restic -r "$REPO_PATH" init ;;
+            kopia)  KOPIA_PASSWORD="$VAULT_KEY_VAL" kopia repository create filesystem --path "$REPO_PATH" ;;
+        esac
+        unset VAULT_KEY_VAL CREDS
+        echo "Vault repo initialized at $REPO_PATH"
+    fi
+
+    # Initial audit entry — ensures access.jsonl exists for the first secrets-push.
+    # Subsequent operations append.
+    if [[ ! -f "$VAULT_DIR/logs/access.jsonl" ]]; then
+        printf '{"ts":"%s","actor":"%s","op":"bootstrap","detail":"vault initialized"}\n' \
+            "$(date -u +%Y-%m-%dT%H:%M:%SZ)" "${ONTOREF_ACTOR:-unknown}" \
+            > "$VAULT_DIR/logs/access.jsonl"
+    fi
+
+    # ── Layer 2 (ADR-017) — scopes + registry credentials inside src-vault/ ──
+    # The src-vault/ subtree IS the OCI artifact contents. domain_client.nu reads
+    # registry/{ro,rw}.sops.yaml via credential_sops/_rw paths from the manifest.
+    SRC_VAULT_DIR="$VAULT_DIR/src-vault"
+    mkdir -p "$SRC_VAULT_DIR/scopes" "$SRC_VAULT_DIR/registry"
+
+    # Generate scope files for each unique role declared in actor_key_bindings.
+    # Default scopes by role-name convention; edit later via 'ore secrets open'.
+    ROLES=$(echo "$PROJECT_JSON" | jq -r '[.sops.actor_key_bindings | .[]] | unique | .[]')
+    for ROLE in $ROLES; do
+        SCOPE_FILE="$SRC_VAULT_DIR/scopes/${ROLE}.ncl"
+        if [[ -f "$SCOPE_FILE" ]]; then continue; fi
+        # RW for human-operator roles, RO for automation roles. Convention only.
+        case "$ROLE" in
+            admin|developer|jesus|supervisor)
+                ACCESS="'rw"
+                OPS="['pull, 'push, 'verify, 'list]"
+                ;;
+            *)
+                ACCESS="'ro"
+                OPS="['pull, 'verify, 'list]"
+                ;;
+        esac
+        printf '{\n  role        = "%s",\n  access      = %s,\n  bound_actor = [],\n  namespaces  = ["domains/*/", "modes/*/"],\n  ops         = %s,\n}\n' \
+            "$ROLE" "$ACCESS" "$OPS" > "$SCOPE_FILE"
+        echo "Created scope: scopes/${ROLE}.ncl"
+    done
+
+    # Default Layer 2 credentials: registry/ro.sops.yaml + registry/rw.sops.yaml.
+    # Skipped in declarative mode (recipient_rules declared) — the operator names
+    # files per their tenancy scheme (e.g. clientA-ro, agent-readonly) and
+    # populates them via 'ore secrets open' after bootstrap. Legacy mode keeps
+    # the default ro/rw pair seeded with the same zot credentials for both.
+    if [[ "$HAS_RULES" == "true" ]]; then
+        echo "Skipping default registry/ro|rw.sops.yaml — declarative recipient_rules in effect."
+        echo "Populate registry/<file>.sops.yaml entries that match your declared rules."
+    else
+        L2_CREDS=$(SOPS_AGE_KEY_FILE="$MASTER_KEY" sops --decrypt "$ACCESS_SOPS")
+        L2_USER=$(echo "$L2_CREDS" | grep '^zot_username:' | awk '{print $2}')
+        L2_PASS=$(echo "$L2_CREDS" | grep '^zot_password:' | awk '{print $2}')
+        for KIND in ro rw; do
+            SOPS_FILE="$SRC_VAULT_DIR/registry/${KIND}.sops.yaml"
+            if [[ -f "$SOPS_FILE" ]]; then continue; fi
+            printf 'username: %s\npassword: %s\n' "$L2_USER" "$L2_PASS" \
+                | SOPS_AGE_KEY_FILE="$MASTER_KEY" sops "${SOPS_ENC_FLAGS[@]}" \
+                    --input-type yaml --encrypt /dev/stdin > "$SOPS_FILE"
+            echo "Created registry/${KIND}.sops.yaml"
+        done
+        unset L2_CREDS L2_USER L2_PASS
+    fi
+
+    echo "Bootstrap complete. Run 'ore secrets push' to upload the vault to ZOT."
+
+# Pull the latest src-vault OCI artifact from ZOT and restore to local vault dir.
+secrets-sync:
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+    [[ -n "$REGISTRY" ]] || { echo "sops.registry_endpoint not declared in project.ncl"; exit 1; }
+    mkdir -p "$VAULT_DIR/logs"
+
+    [[ -f "$ACCESS_SOPS" ]] || { echo "No access.sops.yaml at $VAULT_DIR — first sync needs admin to push the vault"; exit 1; }
+    CREDS=$(SOPS_AGE_KEY_FILE="$MASTER_KEY" sops --decrypt "$ACCESS_SOPS")
+    ZOT_USER=$(echo "$CREDS" | grep zot_username | awk '{print $2}')
+    ZOT_PASS=$(echo "$CREDS" | grep zot_password | awk '{print $2}')
+
+    TMPDIR_CFG=$(mktemp -d)
+    AUTH_B64=$(printf '%s:%s' "$ZOT_USER" "$ZOT_PASS" | base64)
+    printf '{"auths":{"%s":{"auth":"%s"}}}' "${REGISTRY}" "$AUTH_B64" > "$TMPDIR_CFG/config.json"
+
+    # Hard constraint (ADR-017 src-vault-cosign-signed): verify before trusting artifact.
+    [[ -n "$COSIGN_PUB_PATH" ]] || { rm -rf "$TMPDIR_CFG"; echo "vault.cosign.pub_path not set in ~/.config/ontoref/config.ncl"; exit 1; }
+    [[ -f "$COSIGN_PUB_PATH" ]] || { rm -rf "$TMPDIR_CFG"; echo "Cosign pubkey not found at $COSIGN_PUB_PATH"; exit 1; }
+    echo "Verifying src-vault/$VAULT_ID:latest signature (tlog=$COSIGN_TLOG)..."
+    cosign verify $COSIGN_VERIFY_FLAGS --key "$COSIGN_PUB_PATH" "${REGISTRY}/src-vault/${VAULT_ID}:latest" \
+        || { rm -rf "$TMPDIR_CFG"; echo "Vault artifact signature FAILED — rejecting pull"; exit 1; }
+
+    DOCKER_CONFIG="$TMPDIR_CFG" oras pull "${REGISTRY}/src-vault/${VAULT_ID}:latest" --output "$VAULT_DIR"
+
+    rm -rf "$TMPDIR_CFG"
+    unset ZOT_USER ZOT_PASS CREDS AUTH_B64
+    echo "Vault synced to $VAULT_DIR"
+
+    LOG_ENTRY=$(printf '{"ts":"%s","actor":"%s","op":"pull","detail":"secrets-sync"}' \
+        "$(date -u +%Y-%m-%dT%H:%M:%SZ)" "${ONTOREF_ACTOR:-unknown}")
+    echo "$LOG_ENTRY" >> "$VAULT_DIR/logs/access.jsonl"
+
+# Push the local vault snapshot to ZOT as an OCI artifact (cosign-signed).
+secrets-push:
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+    [[ -n "$REGISTRY" ]] || { echo "sops.registry_endpoint not declared in project.ncl"; exit 1; }
+    [[ -f "$ACCESS_SOPS" ]] || { echo "No access.sops.yaml — run secrets-bootstrap first"; exit 1; }
+
+    # Hard constraint (ADR-017 src-vault-cosign-signed): sign required.
+    [[ -n "$COSIGN_KEY_PATH" ]] || { echo "vault.cosign.key_path not set in ~/.config/ontoref/config.ncl"; exit 1; }
+    [[ -f "$COSIGN_KEY_PATH" ]] || { echo "Cosign key not found at $COSIGN_KEY_PATH"; exit 1; }
+
+    CREDS=$(SOPS_AGE_KEY_FILE="$MASTER_KEY" sops --decrypt "$ACCESS_SOPS")
+    ZOT_USER=$(echo "$CREDS" | grep '^zot_username:' | awk '{print $2}')
+    ZOT_PASS=$(echo "$CREDS" | grep '^zot_password:' | awk '{print $2}')
+    VAULT_KEY_VAL=$(echo "$CREDS" | grep '^vault_key:' | awk '{print $2}')
+    # cosign_password optional for back-compat with vaults bootstrapped before
+    # the field existed; fallback to existing $COSIGN_PASSWORD env if unset.
+    COSIGN_PWD_VAL=$(echo "$CREDS" | grep '^cosign_password:' | awk '{print $2}')
+
+    echo "Creating vault snapshot..."
+    case "$VAULT_TOOL" in
+        restic) RESTIC_PASSWORD="$VAULT_KEY_VAL" restic -r "$VAULT_DIR/repo" backup "$VAULT_DIR" --tag "secrets-push" ;;
+        kopia)  KOPIA_PASSWORD="$VAULT_KEY_VAL" kopia --config-file "$VAULT_DIR/repo/kopia.config" snapshot create "$VAULT_DIR" ;;
+    esac
+    unset VAULT_KEY_VAL
+
+    TMPDIR_CFG=$(mktemp -d)
+    AUTH_B64=$(printf '%s:%s' "$ZOT_USER" "$ZOT_PASS" | base64)
+    printf '{"auths":{"%s":{"auth":"%s"}}}' "${REGISTRY}" "$AUTH_B64" > "$TMPDIR_CFG/config.json"
+    unset ZOT_USER ZOT_PASS CREDS AUTH_B64
+
+    [[ -f "$VAULT_DIR/logs/access.jsonl" ]] || {
+        echo "access.jsonl missing — re-run 'ore secrets bootstrap' to initialize"
+        rm -rf "$TMPDIR_CFG"
+        exit 1
+    }
+
+    echo "Pushing src-vault/${VAULT_ID}:latest to $REGISTRY..."
+    # oras requires relative paths for layer files — cd into the vault dir.
+    # Walk src-vault/ to add scope files (.ncl), encrypted credentials (.sops.yaml),
+    # and recipient public keys (.age.pub) as typed layers.
+    LAYERS=("access.sops.yaml:application/vnd.ontoref.vault.access.v1"
+            "logs/access.jsonl:application/vnd.ontoref.vault.audit.v1")
+    if [[ -d "$VAULT_DIR/src-vault" ]]; then
+        while IFS= read -r f; do
+            REL="${f#${VAULT_DIR}/}"
+            case "$REL" in
+                src-vault/scopes/*.ncl)         LAYERS+=("$REL:application/vnd.ontoref.vault.scope.v1") ;;
+                src-vault/registry/*.sops.yaml) LAYERS+=("$REL:application/vnd.ontoref.vault.credential.v1") ;;
+            esac
+        done < <(find "$VAULT_DIR/src-vault" -type f \( -name "*.ncl" -o -name "*.sops.yaml" \))
+    fi
+
+    PUSH_LOG=$(mktemp)
+    set +e
+    ( cd "$VAULT_DIR" && DOCKER_CONFIG="$TMPDIR_CFG" oras push "${REGISTRY}/src-vault/${VAULT_ID}:latest" \
+        --artifact-type "application/vnd.ontoref.vault.v1" \
+        "${LAYERS[@]}" ) > "$PUSH_LOG" 2>&1
+    PUSH_EXIT=$?
+    set -e
+    cat "$PUSH_LOG"
+    if [[ $PUSH_EXIT -ne 0 ]]; then
+        rm -rf "$TMPDIR_CFG" "$PUSH_LOG"
+        echo "oras push failed (exit $PUSH_EXIT)"
+        exit 1
+    fi
+
+    VAULT_DIGEST=$(grep -oE 'sha256:[a-f0-9]{64}' "$PUSH_LOG" | head -1)
+    rm -f "$PUSH_LOG"
+    if [[ -z "$VAULT_DIGEST" ]]; then
+        rm -rf "$TMPDIR_CFG"
+        echo "Could not extract digest from oras push output — cannot sign"
+        exit 1
+    fi
+
+    IMAGE_REF="${REGISTRY}/src-vault/${VAULT_ID}@${VAULT_DIGEST}"
+    echo "Signing $IMAGE_REF (tlog=$COSIGN_TLOG)..."
+    # cosign needs DOCKER_CONFIG too — it HEADs the manifest before signing.
+    # COSIGN_PASSWORD: from access.sops.yaml when present, else fall back to env
+    # (back-compat with vaults bootstrapped before the field existed).
+    if [[ -n "$COSIGN_PWD_VAL" ]]; then
+        DOCKER_CONFIG="$TMPDIR_CFG" COSIGN_PASSWORD="$COSIGN_PWD_VAL" \
+            cosign sign $COSIGN_SIGN_FLAGS --key "$COSIGN_KEY_PATH" --yes "$IMAGE_REF"
+    else
+        DOCKER_CONFIG="$TMPDIR_CFG" \
+            cosign sign $COSIGN_SIGN_FLAGS --key "$COSIGN_KEY_PATH" --yes "$IMAGE_REF"
+    fi
+    unset COSIGN_PWD_VAL
+    echo "Vault artifact signed"
+
+    rm -rf "$TMPDIR_CFG"
+
+    LOG_ENTRY=$(printf '{"ts":"%s","actor":"%s","op":"push","detail":"secrets-push digest=%s"}' \
+        "$(date -u +%Y-%m-%dT%H:%M:%SZ)" "${ONTOREF_ACTOR:-unknown}" "$VAULT_DIGEST")
+    echo "$LOG_ENTRY" >> "$VAULT_DIR/logs/access.jsonl"
+    echo "Vault pushed and signed: $REGISTRY/src-vault/$VAULT_ID:latest"
+
+# Open the vault for editing. Acquires an OCI lock at src-vault/<id>:lock with
+# TTL (default 60 min). Conflicts: if another lock is active and not expired,
+# error with the holder's identity. After editing, run 'ore secrets close'.
+secrets-open:
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+    [[ -f "$ACCESS_SOPS" ]] || { echo "No access.sops.yaml at $VAULT_DIR — run secrets-sync first"; exit 1; }
+    [[ -n "$REGISTRY" ]]    || { echo "sops.registry_endpoint not declared in project.ncl"; exit 1; }
+
+    vault_zot_config_open || exit 1
+    trap 'rm -rf "$TMPDIR_CFG"' EXIT
+
+    NOW_TS=$(date -u +%Y-%m-%dT%H:%M:%SZ)
+    NOW_EPOCH=$(date -u +%s)
+    EXPIRES_EPOCH=$(( NOW_EPOCH + VAULT_LOCK_TTL_MIN * 60 ))
+    EXPIRES_TS=$(date -u -r "$EXPIRES_EPOCH" +%Y-%m-%dT%H:%M:%SZ 2>/dev/null \
+        || date -u -d "@$EXPIRES_EPOCH" +%Y-%m-%dT%H:%M:%SZ 2>/dev/null \
+        || echo "")
+    HOLDER="${USER:-$(whoami)}"
+    ACTOR="${ONTOREF_ACTOR:-developer}"
+
+    EXISTING=$(vault_lock_fetch)
+    if [[ -n "$EXISTING" ]]; then
+        EXIST_BY=$(echo "$EXISTING"     | jq -r .locked_by)
+        EXIST_ACTOR=$(echo "$EXISTING"  | jq -r .actor)
+        EXIST_SINCE=$(echo "$EXISTING"  | jq -r .since)
+        EXIST_EXP=$(echo "$EXISTING"    | jq -r .expires)
+        # Compare expires_epoch with now to detect abandoned lock.
+        EXIST_EXP_EPOCH=$(date -u -j -f "%Y-%m-%dT%H:%M:%SZ" "$EXIST_EXP" +%s 2>/dev/null \
+            || date -u -d "$EXIST_EXP" +%s 2>/dev/null || echo 0)
+        if [[ "$EXIST_BY" == "$HOLDER" ]]; then
+            echo "Vault already locked by you ($HOLDER, since $EXIST_SINCE). Refreshing TTL."
+        elif [[ "$EXIST_EXP_EPOCH" -gt "$NOW_EPOCH" ]]; then
+            echo "Vault locked by $EXIST_BY (actor=$EXIST_ACTOR) since $EXIST_SINCE, expires $EXIST_EXP."
+            echo "If lock is abandoned and you have admin rights, run: ore secrets force-unlock"
+            exit 1
+        else
+            echo "Lock expired ($EXIST_EXP < $NOW_TS). Force-closing abandoned lock from $EXIST_BY."
+        fi
+    fi
+
+    vault_lock_push "$HOLDER" "$ACTOR" "$NOW_TS" "$EXPIRES_TS" \
+        || { echo "Failed to push lock artifact (registry ACL may not allow)"; exit 1; }
+    echo "Lock acquired: $HOLDER until $EXPIRES_TS"
+
+    echo "Opening vault $VAULT_ID for editing..."
+    SOPS_AGE_KEY_FILE="$MASTER_KEY" sops "$ACCESS_SOPS"
+    echo "Edit complete. Run 'ore secrets close' to push and release the lock."
+
+# Close the vault: impact report → push updated state → release lock.
+# ADR-017 G2: list which credential files changed and which RegistryEntries
+# they affect. Operator confirms before the push proceeds.
+secrets-close:
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+
+    # Pre-flight impact analysis.
+    CHANGED=$(vault_impact_changed_creds)
+    if [[ -n "$CHANGED" ]]; then
+        echo "──────────────────────────────────────────────────────────────────"
+        echo "  Credential changes detected since last snapshot:"
+        echo "$CHANGED" | sed 's/^/    /'
+        echo
+        IMPACT=$(echo "$CHANGED" | vault_impact_registries)
+        if [[ -n "$IMPACT" ]]; then
+            echo "  Affected RegistryEntry references:"
+            echo "$IMPACT" | awk -F'\t' '{ printf "    %s  via %s (%s)\n", $1, $2, $3 }'
+            echo
+            echo "  Domains/modes declaring uses_registry to any of the entries above"
+            echo "  will resolve fresh credentials on next pull. Audit consumers if any"
+            echo "  service uses cached credentials in ~/.cache/ontoref/."
+        else
+            echo "  No RegistryEntry in this project's manifest references the changed file(s)."
+            echo "  Consumers in OTHER projects may still depend on these credentials."
+        fi
+        echo "──────────────────────────────────────────────────────────────────"
+        if [[ -z "${ONTOREF_SECRETS_YES:-}" ]]; then
+            read -r -p "  Proceed with push? [y/N] " ans
+            [[ "$ans" =~ ^[Yy] ]] || { echo "Aborted. Lock remains held."; exit 1; }
+        fi
+    else
+        echo "No credential changes since last snapshot."
+    fi
+
+    just secrets-push
+    vault_zot_config_open || { echo "[close] cannot resolve ZOT config — lock NOT released"; exit 1; }
+    trap 'rm -rf "$TMPDIR_CFG"' EXIT
+    EXISTING=$(vault_lock_fetch)
+    if [[ -z "$EXISTING" ]]; then
+        echo "No lock to release."
+        exit 0
+    fi
+    EXIST_BY=$(echo "$EXISTING" | jq -r .locked_by)
+    HOLDER="${USER:-$(whoami)}"
+    if [[ "$EXIST_BY" != "$HOLDER" ]]; then
+        echo "Lock held by $EXIST_BY, not $HOLDER. Refusing to release. Use force-unlock if needed."
+        exit 1
+    fi
+    vault_lock_release
+    echo "Lock released."
+
+# Force-release an abandoned or stuck lock (admin only).
+# Lock metadata records who force-closed for auditability.
+secrets-force-unlock:
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+    [[ -n "$REGISTRY" ]] || { echo "sops.registry_endpoint not declared"; exit 1; }
+    vault_zot_config_open || exit 1
+    trap 'rm -rf "$TMPDIR_CFG"' EXIT
+    EXISTING=$(vault_lock_fetch)
+    if [[ -z "$EXISTING" ]]; then
+        echo "No lock to release."
+        exit 0
+    fi
+    EXIST_BY=$(echo "$EXISTING" | jq -r .locked_by)
+    EXIST_SINCE=$(echo "$EXISTING" | jq -r .since)
+    HOLDER="${USER:-$(whoami)}"
+    NOW_TS=$(date -u +%Y-%m-%dT%H:%M:%SZ)
+    echo "Force-unlocking vault $VAULT_ID."
+    echo "  was held by:    $EXIST_BY since $EXIST_SINCE"
+    echo "  force-closed by: $HOLDER at $NOW_TS"
+    # Emit a tombstone-tagged audit before delete so the action is traceable.
+    LOG_ENTRY=$(printf '{"ts":"%s","actor":"%s","op":"force-unlock","detail":"was-held-by=%s since=%s"}' \
+        "$NOW_TS" "${ONTOREF_ACTOR:-developer}" "$EXIST_BY" "$EXIST_SINCE")
+    echo "$LOG_ENTRY" >> "$VAULT_DIR/logs/access.jsonl"
+    vault_lock_release
+    echo "Lock released."
+
+# Add a recipient. Two modes:
+# - DECLARATIVE (project declares recipient_rules): error with guidance to edit
+#   project.ncl::sops.recipient_groups and run 'secrets rekey' — that is the
+#   canonical source of truth. Mutating sops files directly would diverge from
+#   the declared state and the next rekey would re-clobber the change.
+# - LEGACY (no rules declared): apply --add-age across every *.sops.yaml in the vault.
+secrets-add-key AGE_PUBKEY ROLE:
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+    HAS_RULES=$(echo "$PROJECT_JSON" | jq -r '(.sops.recipient_rules // []) | length > 0')
+    if [[ "$HAS_RULES" == "true" ]]; then
+        echo "This vault uses declarative recipient_rules — direct add-key would drift."
+        echo "Edit project.ncl: add the pubkey to sops.recipient_groups[\"{{ROLE}}\"]"
+        echo "Then: ore secrets rekey"
+        exit 1
+    fi
+    echo "Adding recipient (role: {{ROLE}}) to vault $VAULT_ID..."
+    while IFS= read -r f; do
+        echo "  + ${f#${VAULT_DIR}/}"
+        SOPS_AGE_KEY_FILE="$MASTER_KEY" sops updatekeys --add-age "{{AGE_PUBKEY}}" "$f" >/dev/null
+    done < <(find "$VAULT_DIR" -name "*.sops.yaml")
+    echo "Recipient added to $(find "$VAULT_DIR" -name "*.sops.yaml" | wc -l | tr -d ' ') file(s)."
+    echo "Run 'ore secrets push' to distribute the updated vault."
+
+# Remove a recipient. Same DECLARATIVE/LEGACY split as add-key.
+secrets-remove-key AGE_PUBKEY:
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+    HAS_RULES=$(echo "$PROJECT_JSON" | jq -r '(.sops.recipient_rules // []) | length > 0')
+    if [[ "$HAS_RULES" == "true" ]]; then
+        echo "This vault uses declarative recipient_rules — direct remove-key would drift."
+        echo "Edit project.ncl: remove the pubkey from sops.recipient_groups[*]"
+        echo "Then: ore secrets rekey"
+        exit 1
+    fi
+    echo "Removing recipient from vault $VAULT_ID..."
+    while IFS= read -r f; do
+        echo "  - ${f#${VAULT_DIR}/}"
+        SOPS_AGE_KEY_FILE="$MASTER_KEY" sops updatekeys --remove-age "{{AGE_PUBKEY}}" "$f" >/dev/null
+    done < <(find "$VAULT_DIR" -name "*.sops.yaml")
+    echo "Recipient removed from $(find "$VAULT_DIR" -name "*.sops.yaml" | wc -l | tr -d ' ') file(s)."
+    echo "Rotate registry tokens if the removed key was suspected of compromise."
+
+# Re-encrypt all sops files. When project declares recipient_rules, regenerate
+# <vault_dir>/.sops.yaml from project.ncl FIRST so updatekeys reads the fresh
+# recipient unions per path. Otherwise, updatekeys uses whatever .sops.yaml
+# already exists (legacy mode).
+secrets-rekey:
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+    RULES_JSON=$(echo "$PROJECT_JSON" | jq -c '.sops.recipient_rules // []')
+    GROUPS_JSON=$(echo "$PROJECT_JSON" | jq -c '.sops.recipient_groups // {}')
+    HAS_RULES=$(echo "$RULES_JSON" | jq 'length > 0')
+
+    if [[ "$HAS_RULES" == "true" ]]; then
+        SOPS_CFG_PATH="$VAULT_DIR/.sops.yaml"
+        echo "creation_rules:" > "$SOPS_CFG_PATH"
+        # Use YAML single-quoted strings for path_regex so backslash escapes
+        # like \\. pass through to sops without YAML interpreting them.
+        echo "$RULES_JSON" | jq -r --argjson groups "$GROUPS_JSON" '
+            .[] | "  - path_regex: '"'"'\(.path)'"'"'\n    age: " +
+            ([.groups[] | $groups[.] // []] | flatten | unique | join(","))
+        ' >> "$SOPS_CFG_PATH"
+        echo "Regenerated $SOPS_CFG_PATH from project.ncl ($(echo "$RULES_JSON" | jq length) rule(s))."
+    fi
+
+    echo "Re-keying all sops files in $VAULT_DIR..."
+    find "$VAULT_DIR" -name "*.sops.yaml" | while read -r f; do
+        echo "  updatekeys: $f"
+        SOPS_AGE_KEY_FILE="$MASTER_KEY" sops updatekeys "$f"
+    done
+    echo "Re-key complete. Run 'ore secrets push' to distribute."
+
+# Add or update the cosign_password field inside access.sops.yaml without
+# opening the editor. Targets vaults bootstrapped before the field existed.
+# Idempotent: if the field is already present, value is replaced.
+# Recover a corrupted/missing access.sops.yaml by pulling the src-vault OCI
+# artifact from ZOT. Verified disaster-recovery flow (qa.ncl::credential-vault-
+# disaster-recovery). Requires the project.ncl::sops.registry_endpoint and
+# valid ZOT credentials in your environment via DOCKER_CONFIG (or admin
+# credentials passed inline).
+secrets-recover:
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+    [[ -n "$REGISTRY" ]] || { echo "sops.registry_endpoint not declared"; exit 1; }
+
+    # When access.sops.yaml is corrupt we cannot decrypt it to obtain credentials.
+    # Caller must provide ZOT credentials via env (RECOVERY_ZOT_USER + _PASS) OR
+    # have a pre-existing valid DOCKER_CONFIG pointing at this registry.
+    if [[ -n "${RECOVERY_ZOT_USER:-}" && -n "${RECOVERY_ZOT_PASS:-}" ]]; then
+        TMPCFG=$(mktemp -d)
+        AUTH=$(printf '%s:%s' "$RECOVERY_ZOT_USER" "$RECOVERY_ZOT_PASS" | base64)
+        printf '{"auths":{"%s":{"auth":"%s"}}}' "$REGISTRY" "$AUTH" > "$TMPCFG/config.json"
+        export DOCKER_CONFIG="$TMPCFG"
+        trap 'rm -rf "$TMPCFG"' EXIT
+        echo "Using RECOVERY_ZOT_USER credentials."
+    elif [[ -n "${DOCKER_CONFIG:-}" ]]; then
+        echo "Using existing DOCKER_CONFIG=$DOCKER_CONFIG"
+    else
+        echo "No credentials available. Set RECOVERY_ZOT_USER + RECOVERY_ZOT_PASS"
+        echo "or DOCKER_CONFIG before invoking."
+        exit 1
+    fi
+
+    RECOVERY=$(mktemp -d)
+    echo "Pulling src-vault/${VAULT_ID}:latest from $REGISTRY..."
+    if ! oras pull "${REGISTRY}/src-vault/${VAULT_ID}:latest" --output "$RECOVERY"; then
+        rm -rf "$RECOVERY"
+        echo "oras pull failed — verify registry and credentials"
+        exit 1
+    fi
+
+    [[ -f "$RECOVERY/access.sops.yaml" ]] || {
+        rm -rf "$RECOVERY"
+        echo "Recovered artifact does not contain access.sops.yaml"
+        exit 1
+    }
+
+    mkdir -p "$VAULT_DIR" "$VAULT_DIR/logs" "$VAULT_DIR/src-vault"
+    cp "$RECOVERY/access.sops.yaml" "$ACCESS_SOPS"
+    [[ -d "$RECOVERY/logs"      ]] && cp -R "$RECOVERY/logs/."      "$VAULT_DIR/logs/"      2>/dev/null || true
+    [[ -d "$RECOVERY/src-vault" ]] && cp -R "$RECOVERY/src-vault/." "$VAULT_DIR/src-vault/" 2>/dev/null || true
+    rm -rf "$RECOVERY"
+
+    # Verify decrypts cleanly with the configured master key.
+    if SOPS_AGE_KEY_FILE="$MASTER_KEY" sops --decrypt "$ACCESS_SOPS" >/dev/null 2>&1; then
+        echo "Recovered. access.sops.yaml decrypts with $MASTER_KEY"
+        printf '{"ts":"%s","actor":"%s","op":"recover","detail":"access.sops.yaml restored from %s"}\n' \
+            "$(date -u +%Y-%m-%dT%H:%M:%SZ)" "${ONTOREF_ACTOR:-developer}" \
+            "${REGISTRY}/src-vault/${VAULT_ID}:latest" >> "$VAULT_DIR/logs/access.jsonl"
+    else
+        echo "Restored but cannot decrypt — your .kage may not be a recipient"
+        exit 1
+    fi
+
+secrets-set-cosign-password:
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+    [[ -f "$ACCESS_SOPS" ]] || { echo "access.sops.yaml absent — bootstrap first"; exit 1; }
+
+    read -r -s -p "cosign_password (for cosign.key): " COSIGN_PWD; echo
+    [[ -n "$COSIGN_PWD" ]] || { echo "empty password rejected"; exit 1; }
+
+    # `sops set` updates a single field in-place, preserving the original
+    # recipient set (which lives in the file's sops metadata block). No
+    # decrypt + re-encrypt cycle, no leak of plaintext to disk.
+    SOPS_AGE_KEY_FILE="$MASTER_KEY" sops set "$ACCESS_SOPS" '["cosign_password"]' "\"$COSIGN_PWD\""
+    unset COSIGN_PWD
+
+    # Verify the field is present after the update.
+    if SOPS_AGE_KEY_FILE="$MASTER_KEY" sops --decrypt --extract '["cosign_password"]' "$ACCESS_SOPS" >/dev/null 2>&1; then
+        echo "  cosign_password set in $ACCESS_SOPS"
+        echo "  Run 'ore secrets push' to upload the updated vault."
+    else
+        echo "  set succeeded but field still not extractable — investigate"
+        exit 1
+    fi
+
+# Audit vault constraints from ADR-017:
+#   --check bootstrap-credentials   access.sops.yaml can be decrypted by master_key
+#   --check no-credential-env       manifest does not declare credential_env
+#   --check recipients              every .sops.yaml has ≥ 2 recipients (multi-recipient mandatory)
+secrets-audit check="":
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+
+    CHECK="${check:-all}"
+    PASS=0
+
+    if [[ "$CHECK" == "bootstrap-credentials" ]] || [[ "$CHECK" == "all" ]]; then
+        echo -n "  bootstrap-credentials: "
+        if [[ ! -f "$ACCESS_SOPS" ]]; then
+            echo "FAIL (access.sops.yaml not found at $ACCESS_SOPS)"
+            PASS=1
+        else
+            DECRYPT_CHECK=$(SOPS_AGE_KEY_FILE="$MASTER_KEY" sops --decrypt "$ACCESS_SOPS" 2>&1)
+            if [[ $? -eq 0 ]]; then echo "OK"; else echo "FAIL (decrypt error)"; PASS=1; fi
+        fi
+    fi
+
+    if [[ "$CHECK" == "no-credential-env" ]] || [[ "$CHECK" == "all" ]]; then
+        echo -n "  no-credential-env: "
+        if grep -r "credential_env" "${PROJECT_ROOT}/.ontology/manifest.ncl" 2>/dev/null | grep -q .; then
+            echo "FAIL (credential_env found in manifest.ncl — excluded by schema)"
+            PASS=1
+        else
+            echo "OK"
+        fi
+    fi
+
+    if [[ "$CHECK" == "recipients" ]] || [[ "$CHECK" == "all" ]]; then
+        echo -n "  multi-recipient-mandatory: "
+        if [[ -f "$ACCESS_SOPS" ]]; then
+            RECIPIENT_COUNT=$(grep -c "age1" "$ACCESS_SOPS" 2>/dev/null || echo 0)
+            if [[ "$RECIPIENT_COUNT" -ge 2 ]]; then echo "OK ($RECIPIENT_COUNT recipients)";
+            else echo "FAIL (only $RECIPIENT_COUNT recipient — multi-recipient required)"; PASS=1; fi
+        else
+            echo "SKIP (no access.sops.yaml)"
+        fi
+    fi
+
+    # ADR-019 — recipient routing constraints (skipped in legacy mode).
+    RULES=$(echo "$PROJECT_JSON" | jq -c '.sops.recipient_rules // []')
+    GROUPS=$(echo "$PROJECT_JSON" | jq -c '.sops.recipient_groups // {}')
+    HAS_RULES=$(echo "$RULES" | jq 'length > 0')
+
+    if [[ "$CHECK" == "recipient-routing-coherent" ]] || [[ "$CHECK" == "all" ]]; then
+        echo -n "  recipient-routing-coherent: "
+        if [[ "$HAS_RULES" != "true" ]]; then
+            echo "SKIP (legacy single-set mode)"
+        else
+            # Constraint 1: every group referenced by a rule must be declared in groups
+            UNDECLARED=$(echo "$RULES" | jq -r --argjson g "$GROUPS" '
+                [.[] | .groups[] | select(($g | has(.)) | not)] | unique | join(",")
+            ')
+            # Constraint 2: every rule must resolve to ≥ 1 recipient
+            EMPTY_RULES=$(echo "$RULES" | jq -r --argjson g "$GROUPS" '
+                [.[] | select((.groups | map($g[.] // []) | flatten | unique | length) == 0) | .path] | join(",")
+            ')
+            if [[ -n "$UNDECLARED" ]]; then
+                echo "FAIL (undeclared groups referenced: $UNDECLARED)"
+                PASS=1
+            elif [[ -n "$EMPTY_RULES" ]]; then
+                echo "FAIL (rules resolve to zero recipients: $EMPTY_RULES)"
+                PASS=1
+            else
+                echo "OK ($(echo "$RULES" | jq length) rule(s), $(echo "$GROUPS" | jq 'keys | length') group(s))"
+            fi
+        fi
+    fi
+
+    if [[ "$CHECK" == "recipient-routing-coverage" ]] || [[ "$CHECK" == "all" ]]; then
+        echo -n "  recipient-routing-coverage: "
+        if [[ "$HAS_RULES" != "true" ]]; then
+            echo "SKIP (legacy single-set mode)"
+        elif [[ ! -d "$VAULT_DIR" ]]; then
+            echo "SKIP (vault not bootstrapped)"
+        else
+            # For each *.sops.yaml in the vault, verify at least one rule.path matches.
+            UNCOVERED=""
+            while IFS= read -r f; do
+                rel="${f#${VAULT_DIR}/}"
+                # Try each rule's regex; rule order doesn't matter for coverage.
+                MATCHED=$(echo "$RULES" | jq -r --arg path "$rel" '
+                    [.[] | select($path | test(.path))] | length
+                ')
+                [[ "$MATCHED" == "0" ]] && UNCOVERED="$UNCOVERED $rel"
+            done < <(find "$VAULT_DIR" -name "*.sops.yaml" 2>/dev/null)
+            if [[ -n "$UNCOVERED" ]]; then
+                echo "FAIL (paths without matching rule:$UNCOVERED)"
+                PASS=1
+            else
+                echo "OK"
+            fi
+        fi
+    fi
+
+    if [[ "$CHECK" == "no-multi-vault" ]] || [[ "$CHECK" == "all" ]]; then
+        echo -n "  no-multi-vault: "
+        if grep -q "sops\.vaults *=" "$PROJECT_NCL" 2>/dev/null; then
+            echo "FAIL (sops.vaults declaration found — multi-vault is unimplemented per ADR-019)"
+            PASS=1
+        else
+            echo "OK"
+        fi
+    fi
+
+    exit $PASS
+
+# Read existing access.sops.yaml recipients and emit a recipient_groups +
+# recipient_rules stanza ready to paste into project.ncl::sops. Read-only —
+# does not mutate project.ncl. The default policy maps every existing recipient
+# to BOTH `admin` and `developer` groups (preserves current behaviour exactly)
+# and emits two rules: one for access.sops.yaml (admin+developer), one for
+# everything else (developer-only). Operator edits the stanza before pasting
+# to refine which keys belong to which group.
+secrets-migrate-to-declarative:
+    #!/usr/bin/env bash
+    source "{{source_directory()}}/_secrets_lib.sh"
+    [[ -f "$ACCESS_SOPS" ]] || { echo "No access.sops.yaml at $VAULT_DIR — bootstrap or sync first"; exit 1; }
+
+    # Already declarative? Refuse to overwrite — operator must rekey from project.ncl.
+    HAS_RULES=$(echo "$PROJECT_JSON" | jq -r '.sops.recipient_rules // [] | length > 0')
+    if [[ "$HAS_RULES" == "true" ]]; then
+        echo "[migrate] project.ncl already declares recipient_rules — nothing to migrate."
+        echo "  Run 'just secrets-rekey' if you changed recipient_groups and need to re-encrypt."
+        exit 0
+    fi
+
+    # Extract recipients from the existing access.sops.yaml metadata. sops
+    # records them under sops.age[].recipient as age1... entries — file format
+    # may be YAML or JSON depending on input-type, so use a generic regex.
+    RECIPIENTS=$(grep -oE 'age1[a-z0-9]+' "$ACCESS_SOPS" | sort -u)
+    [[ -n "$RECIPIENTS" ]] || { echo "[migrate] no age recipients found in $ACCESS_SOPS"; exit 1; }
+    COUNT=$(echo "$RECIPIENTS" | wc -l | tr -d ' ')
+
+    echo "# ── recipient routing migration helper (read-only) ────────────────────"
+    echo "# Detected $COUNT recipient(s) in $ACCESS_SOPS:"
+    echo "$RECIPIENTS" | sed 's/^/#   - /'
+    echo "#"
+    echo "# Default policy: all recipients in both 'admin' and 'developer' groups."
+    echo "# This preserves current behaviour exactly (no key loses access). Edit"
+    echo "# the groups below to scope down before pasting into project.ncl::sops."
+    echo "#"
+    echo "# Paste THIS into your project.ncl::sops record (replacing any existing"
+    echo "# recipient_groups / recipient_rules / SOPS_AGE_RECIPIENTS env reliance):"
+    echo "# ──────────────────────────────────────────────────────────────────────"
+    echo
+    AGE_LIST=$(echo "$RECIPIENTS" | awk '{printf "      \"%s\",\n", $0}')
+    printf '  recipient_groups = {\n'
+    printf '    admin = [\n%s\n    ],\n' "$AGE_LIST"
+    printf '    developer = [\n%s\n    ],\n' "$AGE_LIST"
+    printf '  },\n'
+    printf '  recipient_rules = [\n'
+    printf '    { path = "access\\\\.sops\\\\.yaml$",          groups = ["admin", "developer"] },\n'
+    printf '    { path = "registry/.*\\\\.sops\\\\.yaml$",     groups = ["developer"] },\n'
+    printf '    { path = "scopes/.*\\\\.ncl$",                 groups = ["admin"] },\n'
+    printf '  ],\n'
+    echo
+    echo "# ──────────────────────────────────────────────────────────────────────"
+    echo "# Next steps:"
+    echo "#   1. Edit project.ncl::sops to insert the stanza above."
+    echo "#   2. Refine groups (move keys between admin/developer to scope access)."
+    echo "#   3. Refine recipient_rules paths to match your repo layout."
+    echo "#   4. Run: just secrets-rekey   (re-encrypts every *.sops.yaml using"
+    echo "#      the new declared rules — same recipients = idempotent no-op)."
+    echo "#   5. Verify: just secrets-audit   (runs ADR-019 coherency checks)."
diff --git a/lian-build/build_directives.ncl b/lian-build/build_directives.ncl
new file mode 100644
index 0000000..696389e
--- /dev/null
+++ b/lian-build/build_directives.ncl
@@ -0,0 +1,41 @@
+let d       = import "defaults/build_directives.ncl" in
+let project = import "../provisioning/project.ncl" in
+
+d.make_build_directives {
+  workspace = project.workspace_id,
+
+  artifacts = [
+    d.make_artifact {
+      image      = "%{project.image_base}:%{project.image_tag}",
+      dockerfile = "Dockerfile",
+      context    = ".",
+      platforms  = ["linux/arm64"],
+      build_args = {
+        IMAGE_BASE   = project.lamina_base,
+        RUST_VERSION = project.rust_version,
+      },
+      extra_contexts = {
+        stratumiops = "../stratumiops",
+      },
+    },
+  ],
+
+  adapter = d.make_adapter_ref {
+    kind         = 'hcloud,
+    location     = "fsn1",
+    ssh_key_ref  = "orchestrator-buildkit",
+    network_ref  = "wuwei",
+    firewall_ref = "libre-wuji-fw",
+    internal_dns = "10.0.8.20",
+  },
+
+  registry = d.default_zot_registry "reg.librecloud.online" project.build_secrets.registry_push,
+
+  cache = d.ci_cache_policy project.workspace_id,
+
+  sccache = {
+    bucket     = "sccache",
+    endpoint   = "https://fsn1.your-objectstorage.com",
+    credential = project.build_secrets.sccache,
+  },
+}
diff --git a/lian-build/ctx-test.nu b/lian-build/ctx-test.nu
new file mode 100644
index 0000000..2d05fc6
--- /dev/null
+++ b/lian-build/ctx-test.nu
@@ -0,0 +1,134 @@
+#!/usr/bin/env nu
+# Local build test for ontoref-daemon.
+# Reads build_directives.ncl for IMAGE_BASE / RUST_VERSION, assembles the
+# build context via rsync, and invokes docker build with --build-context stratumiops.
+#
+# Usage:
+#   nu lian-build/ctx-test.nu                             # prepare ctx only
+#   nu lian-build/ctx-test.nu --stage builder             # build builder stage
+#   nu lian-build/ctx-test.nu --run                       # build final image
+#
+# Local lamina layers (after running lamina ctx-test.nu --layer rust/nickel):
+#   nu lian-build/ctx-test.nu --run --image-base lamina --rust-version local
+#
+# ARG overrides (bypass NCL):
+#   nu lian-build/ctx-test.nu --run --rust-version 1.90
+
+def read_build_args [script_dir: string] {
+    let directives = ($script_dir | path join "build_directives.ncl")
+
+    if not ($directives | path exists) {
+        print $"warning: ($directives) not found — using flag defaults"
+        return {}
+    }
+
+    if (which nickel | is-empty) {
+        print "warning: nickel not in PATH — using flag defaults"
+        return {}
+    }
+
+    let lian_root = if "LIAN_BUILD_ROOT" in $env {
+        $env.LIAN_BUILD_ROOT
+    } else {
+        $script_dir | path join ".." ".." "lian-build" | path expand
+    }
+
+    if not ($lian_root | path exists) {
+        print $"warning: lian-build root not found at ($lian_root) — set LIAN_BUILD_ROOT"
+        return {}
+    }
+
+    let result = do { ^nickel export --import-path $lian_root $directives } | complete
+    if $result.exit_code != 0 {
+        print $"warning: nickel eval failed — ($result.stderr)"
+        return {}
+    }
+    $result.stdout | from json | get artifacts.0.build_args
+}
+
+def resolve [flag: string, ncl: record, key: string, fallback: string] {
+    if not ($flag | is-empty) { $flag }
+    else { $ncl | get -i $key | default $fallback }
+}
+
+def main [
+    --stage: string    = ""
+    --run
+    --keep
+    --platform: string = "linux/arm64"
+    --rust-version: string = ""
+    --image-base: string   = ""
+] {
+    let script_dir       = $env.FILE_PWD
+    let project_root     = ($script_dir | path join "..") | path expand
+    let stratumiops_root = ($project_root | path join ".." "stratumiops") | path expand
+
+    if not ($stratumiops_root | path exists) {
+        error make { msg: $"stratumiops not found at ($stratumiops_root)" }
+    }
+
+    let ncl          = read_build_args $script_dir
+    let rust_version = resolve $rust_version $ncl "RUST_VERSION" "1.89"
+    let image_base   = resolve $image_base   $ncl "IMAGE_BASE"   "reg.librecloud.online/lamina"
+
+    let ctx = (^mktemp -d | str trim)
+
+    print $"ctx:          ($ctx)"
+    print $"project:      ($project_root)"
+    print $"stratumiops:  ($stratumiops_root)"
+    print $"IMAGE_BASE:   ($image_base)"
+    print $"RUST_VERSION: ($rust_version)"
+
+    ^rsync -a --delete ...[
+        "--exclude=.git/"
+        "--exclude=target/"
+        "--exclude=.coder/"
+        "--exclude=logs/"
+        "--exclude=logs-archive/"
+        "--exclude=lian-build/ctx-*"
+        $"($project_root)/"
+        $"($ctx)/"
+    ]
+
+    print $"context ready: ($ctx)"
+
+    let dockerfile = ($project_root | path join "Dockerfile")
+
+    let build_args = [
+        "--build-arg"     $"RUST_VERSION=($rust_version)"
+        "--build-arg"     $"IMAGE_BASE=($image_base)"
+        "--build-context" $"stratumiops=($stratumiops_root)"
+    ]
+
+    if $run {
+        let tag = "ontoref-daemon:local-test"
+        print $"\nbuilding final image: ($tag)"
+        ^docker build ...([
+            "--platform" $platform
+            "-t"         $tag
+            "-f"         $dockerfile
+        ] ++ $build_args ++ [$ctx])
+        if not $keep { ^rm -rf $ctx }
+        print "\nsmoke test"
+        ^docker run --rm -p "7891:7891" $tag
+        return
+    }
+
+    if ($stage | is-empty) {
+        print $"\nctx ready at: ($ctx)"
+        print "pass --stage <name> to build a specific stage, or --run for the final image"
+        if not $keep { ^rm -rf $ctx }
+        return
+    }
+
+    print $"\nbuilding stage: ($stage)"
+    ^docker build ...([
+        "--platform" $platform
+        "--target"   $stage
+        "-t"         $"ontoref-($stage):test"
+        "-f"         $dockerfile
+    ] ++ $build_args ++ [$ctx])
+
+    if $keep { print $"ctx preserved: ($ctx)" } else { ^rm -rf $ctx }
+    print $"\nimage: ontoref-($stage):test"
+}
diff --git a/ontology/defaults/decomposition.ncl b/ontology/defaults/decomposition.ncl
new file mode 100644
index 0000000..d399b2a
--- /dev/null
+++ b/ontology/defaults/decomposition.ncl
@@ -0,0 +1,58 @@
+let d = import "core.ncl" in
+
+{
+  lift_out_practice = d.make_node {
+    id          = "lift-out-practice",
+    name        = "Component Lift-Out",
+    pole        = 'Yang,
+    level       = 'Practice,
+    description = "A component developed inside a host project is extracted as an independent peer project when it passes the four-criterion test: orthogonal concern, plurality of consumers, divergent release cadence, and path-agnostic config. The extracted project registers in ontoref before any code moves.",
+    adrs        = ["adr-016-component-lift-out-pattern"],
+  },
+
+  lift_readiness_practice = d.make_node {
+    id          = "lift-readiness",
+    name        = "Lift Readiness Evaluation",
+    pole        = 'Spiral,
+    level       = 'Practice,
+    description = "Before extracting a component, four criteria are evaluated: (1) orthogonal concern — the component's domain is not the host's domain; (2) consumer plurality — at least two distinct callers exist or are imminent; (3) release cadence divergence — the component's lifecycle does not match the host's; (4) config path-agnostic — the component can receive its config from any caller without importing host infrastructure.",
+    adrs        = ["adr-016-component-lift-out-pattern"],
+  },
+
+  premature_decomposition_tension = d.make_node {
+    id          = "premature-decomposition",
+    name        = "Premature Decomposition Tension",
+    pole        = 'Yin,
+    level       = 'Tension,
+    description = "Extracting a component before it has proven orthogonality and multiple consumers creates coordination overhead without autonomy benefits. The four-criterion gate is the safeguard.",
+  },
+
+  host_coupling_tension = d.make_node {
+    id          = "host-coupling",
+    name        = "Host Coupling Tension",
+    pole        = 'Yin,
+    level       = 'Tension,
+    description = "A component that imports host-project infrastructure (config loaders, schema registries, workspace crates) cannot be an independent peer. Host coupling is the primary blocker for lift-out and must be severed at extraction time.",
+  },
+
+  edges = [
+    d.make_edge {
+      from   = "lift-readiness",
+      to     = "lift-out-practice",
+      kind   = 'FlowsTo,
+      note   = "readiness evaluation gates the extraction decision",
+    },
+    d.make_edge {
+      from   = "premature-decomposition",
+      to     = "lift-out-practice",
+      kind   = 'TensionWith,
+      note   = "extracting too early defeats the purpose",
+    },
+    d.make_edge {
+      from   = "host-coupling",
+      to     = "lift-readiness",
+      kind   = 'LimitedBy,
+      note   = "coupling to host infrastructure fails criterion 4",
+    },
+  ],
+}
diff --git a/ontology/defaults/manifest.ncl b/ontology/defaults/manifest.ncl
index 769c668..454ca2e 100644
--- a/ontology/defaults/manifest.ncl
+++ b/ontology/defaults/manifest.ncl
@@ -47,6 +47,15 @@ let c = import "content.ncl" in
   Requirement       = s.Requirement,
   CriticalDep       = s.CriticalDep,
   make_capability   = fun data => s.Capability   & data,
-  make_requirement  = fun data => s.Requirement  & data,
-  make_critical_dep = fun data => s.CriticalDep  & data,
+  make_requirement      = fun data => s.Requirement      & data,
+  make_critical_dep     = fun data => s.CriticalDep      & data,
+  # Registry topology builders
+  RegistryRole          = s.RegistryRole,
+  RegistryNamespaces    = s.RegistryNamespaces,
+  RegistryEntry         = s.RegistryEntry,
+  RegistriesConfig      = s.RegistriesConfig,
+  RegistryProvides      = s.RegistryProvides,
+  make_registry_entry   = fun data => s.RegistryEntry    & data,
+  make_registries_config = fun data => s.RegistriesConfig & data,
+  make_registry_provides = fun data => s.RegistryProvides & data,
 }
diff --git a/ontology/schemas/manifest.ncl b/ontology/schemas/manifest.ncl
index 97f5a8f..5b995ef 100644
--- a/ontology/schemas/manifest.ncl
+++ b/ontology/schemas/manifest.ncl
@@ -285,6 +285,53 @@ let critical_dep_type = {
   mitigation     | String | default = "",
 } in
 
+# ── Registry topology ────────────────────────────────────────────────────────
+# Declares the multi-registry topology for coordination: which registries
+# exist, their roles, and which participant namespaces this project owns.
+# Mirrors capabilities.ncl RegistriesConfig from the provisioning schema so
+# ontoref consumer projects can declare registry intent without a dep on the
+# provisioning crate. Consumed by resolve-registry, describe workspace, and
+# the GraphQL registry_topology resolver.
+
+let registry_role_type = [| 'primary, 'build, 'dev, 'mirror |] in
+
+let registry_namespaces_type = {
+  own          | Array String | default = [],
+  replicate_to | Array String | default = [],
+  mirror_from  | String       | optional,
+  prefixes     | Array String | default = [],
+} in
+
+let registry_entry_type = {
+  id                 | String,
+  endpoint           | String,
+  role               | registry_role_type       | default = 'dev,
+  tls                | Bool                     | default = true,
+  namespaces         | registry_namespaces_type | default = {},
+  # Credential references — credential_env is intentionally absent (see ADR-017).
+  # credential_sops paths are relative to the synced src-vault root, NOT the project
+  # root. The src-vault is the OCI artifact restored to ~/.config/ontoref/vaults/<vault_id>/src-vault/
+  # and contains encrypted credential files (typically registry/ro.sops.yaml, registry/rw.sops.yaml).
+  credential_sops    | String | optional,   # RO token — for pull/list operations
+  credential_sops_rw | String | optional,   # RW token — for push operations
+  credential_oidc    | String | optional,   # OIDC audience (future)
+  # Explicit dependency declaration for impact analysis (ADR-017 constraint)
+  uses_registry      | String | optional,   # references another RegistryEntry.id
+} in
+
+let registries_config_type = {
+  registries | Array registry_entry_type | default = [],
+  default    | String                    | optional,
+} in
+
+# Declared by a project that is a participant in the registry ecosystem:
+# owns domain/mode namespaces and can push artifacts.
+let registry_provides_type = {
+  # Identifier used to scope namespaces: domains/<participant>/, modes/<participant>/
+  participant | String,
+  registries  | registries_config_type | default = {},
+} in
+
 # ── Domain ontology layer ───────────────────────────────────────────────────
 # A project may have two distinct ontological layers:
 #   1. Self-ontology  (.ontology/) — the project describing itself: state, ADRs, practices.
@@ -335,6 +382,28 @@ let domain_origin_type = {
   integrated  | Bool             | default = true,
 } in
 
+# ── Level identity (ADR-018) ────────────────────────────────────────────────
+# Declares where a project sits in the ontoref specialization hierarchy.
+# 'Base = 1 (ontoref itself — generic protocol), 'Domain = 2 (project-type domain,
+# e.g. provisioning), 'Instance = 3 (concrete workspace/infra derived from a domain).
+# Absent at 'Base (base is always Override by definition; no parent exists).
+# Required at 'Domain and 'Instance to enable observable traversal (ore mode resolve).
+
+let level_index_type = [| 'Base, 'Domain, 'Instance |] in
+
+let level_type = {
+  # 'Base=1, 'Domain=2, 'Instance=3 — position in the specialization hierarchy.
+  index  | level_index_type,
+  # Stable identifier for this level (e.g. "ontoref-base", "provisioning-domain").
+  name   | String,
+  # Name of the upstream project this level extends. Absent at 'Base.
+  parent | String | optional,
+  # Absolute path to the parent project on disk.
+  # When set, ore validate modes --check delegate-chain verifies that each Delegate
+  # mode exists in {parent_path}/reflection/modes/ — Hard failure if not found there.
+  parent_path | String | optional,
+} in
+
 # ── Root manifest ───────────────────────────────────────────────────────────
 
 let manifest_type = {
@@ -378,6 +447,12 @@ let manifest_type = {
   domain_provides      | domain_provides_type           | optional,
   # Set when this project IS an implementation extending an upstream domain framework.
   domain_origin        | domain_origin_type             | optional,
+  # Declares this project's participant identity and registry topology.
+  # participant scopes artifact namespaces: domains/<participant>/ modes/<participant>/
+  registry_provides    | registry_provides_type         | optional,
+  # Level identity in the three-level hierarchy (ADR-018).
+  # Required at 'Domain and 'Instance. Absent at 'Base (ontoref itself).
+  level                | level_type                     | optional,
 } in
 
 {
@@ -410,4 +485,11 @@ let manifest_type = {
   DomainOntologyKind  = domain_ontology_kind_type,
   DomainProvides      = domain_provides_type,
   DomainOrigin        = domain_origin_type,
+  RegistryRole        = registry_role_type,
+  RegistryNamespaces  = registry_namespaces_type,
+  RegistryEntry       = registry_entry_type,
+  RegistriesConfig    = registries_config_type,
+  RegistryProvides    = registry_provides_type,
+  LevelIndex          = level_index_type,
+  Level               = level_type,
 }
diff --git a/provisioning/project.ncl b/provisioning/project.ncl
new file mode 100644
index 0000000..678656a
--- /dev/null
+++ b/provisioning/project.ncl
@@ -0,0 +1,14 @@
+{
+  name         = "ontoref",
+  workspace_id = "ontoref",
+  image_base   = "reg.librecloud.online/ontoref-daemon",
+  image_tag    = "latest",
+  rust_version = "1.89",
+  lamina_base  = "reg.librecloud.online/lamina",
+  port         = 7891,
+
+  build_secrets = {
+    registry_push = { kind = 'docker_config, path = "infra/libre-wuji/secrets/registry-push.sops.yaml" },
+    sccache       = { kind = 's3,            path = "infra/libre-wuji/secrets/hetzner-s3.sops.yaml" },
+  },
+}
diff --git a/reflection/backlog.ncl b/reflection/backlog.ncl
index 256f4d6..3bf4c0c 100644
--- a/reflection/backlog.ncl
+++ b/reflection/backlog.ncl
@@ -75,5 +75,422 @@ let s = import "schemas/backlog.ncl" in
       created  = "2026-03-13",
     },
 
+    {
+      id       = "bl-007",
+      title    = "ADR `'Grep` constraint check needs path-exclusion expressivity",
+      kind     = 'Wish,
+      priority = 'Low,
+      status   = 'Open,
+      detail   = "Surfaced by an adopting project (lian-build) writing adr-003 (`nickel-via-subprocess`). The ADR wants to enforce: 'no Rust file in src/ may call Command::new(\"nickel\") EXCEPT src/directives.rs which is the canonical dispatch site.' The current `'Grep` check shape (pattern, paths, must_be_empty) cannot express 'paths_exclude'. Workaround in lian-build: must_be_empty=false (positive presence) plus check_hint explaining the rule socially. Auditable but does not catch a second copy of the dispatch in another module. Alternatives: (A) add `paths_exclude: Array String` to `'Grep`; (B) new `'GrepFiltered` tag with include/exclude lists; (C) recommend `'NuCmd` for these cases (works today, opaque to static review). Friction site: lian-build/adrs/adr-003-nickel-via-subprocess.ncl::nickel-export-centralized. Mirror lives at lian-build/reflection/backlog.ncl::bl-002.",
+      related_adrs  = [],
+      graduates_to  = 'PrItem,
+      created  = "2026-05-03",
+    },
+
+    {
+      id       = "bl-010",
+      title    = "Canonicalize ontoref-instance directory layout (manifest path + federated-peer catalog path)",
+      kind     = 'Debt,
+      priority = 'Medium,
+      status   = 'Open,
+      detail   = m%"
+        SURFACED BY  adr-020 layer-2-biconditional check applied across the
+                     three ontoref-onboarded projects (2026-05-03 audit).
+
+        DRIFT OBSERVED — three projects, three conventions
+
+          ontoref       domains/, modes/                  (TOP-LEVEL — no catalog/)
+                        .ontology/manifest.ncl            (under .ontology/)
+                        registry_provides   declared
+
+          lian-build    catalog/domains/, catalog/modes/  (under catalog/)
+                        manifest.ncl                      (at root)
+                        .ontology/manifest.ncl            (pointer to root)
+                        registry_provides   declared
+
+          provisioning  catalog/domains/, catalog/components/, catalog/providers/
+                        .ontology/manifest.ncl            (under .ontology/)
+                        registry_provides   ABSENT despite federated-peer
+                                            catalog/domains/
+
+        THREE INDEPENDENT QUESTIONS
+
+          Q1 — federated-peer catalog path
+            Options:
+              (a) `catalog/domains/` and `catalog/modes/` (lian-build / provisioning)
+              (b) `domains/` and `modes/` at top-level (ontoref)
+              (c) Allow either, normalize at check time
+            Trade-off: (a) groups all federation artifacts under one prefix;
+            (b) gives them top-level visibility; (c) avoids migration but
+            makes the layer-2-biconditional check more elaborate.
+
+          Q2 — manifest.ncl canonical location
+            Options:
+              (a) Project root: `manifest.ncl` (lian-build)
+              (b) Self-ontology: `.ontology/manifest.ncl` (ontoref, provisioning)
+              (c) Both, with the non-canonical one as a pointer (lian-build does
+                  this — root is canonical, .ontology/manifest.ncl imports it)
+            Trade-off: (a) joins the top-level doc surface; (b) groups all
+            self-ontology files under one directory; (c) accommodates both
+            but adds one indirection.
+
+          Q3 — provisioning's missing registry_provides
+            Independent of Q1 / Q2: provisioning has catalog/domains/ but no
+            registry_provides declaration. Either it should add one (as a
+            federated peer publishing the protocol's shared domains under
+            its participant namespace), OR the protocol's shared domains
+            should move to a participant-scoped location somewhere else.
+            See bl-007 for related schema work.
+
+        WHEN THIS GRADUATES
+          Decision yields an ADR (or amendment to adr-020) committing the
+          canonical paths. Migration follows in three coordinated changes:
+          one per project. After migration, adr-020 promotes from 'Proposed
+          to 'Accepted and the layer-2-biconditional check passes ecosystem-
+          wide.
+
+        REFERENCES
+          - adrs/adr-020-three-layer-ontoref-instance-model.ncl
+          - reflection/backlog.ncl::bl-009                graduation parent
+          - reflection/qa.ncl::ontoref-three-layer-model  protocol-level
+                                                          explanation
+        "%,
+      related_adrs  = ["adr-020"],
+      graduates_to  = 'Adr,
+      created  = "2026-05-03",
+    },
+
+    {
+      id       = "bl-012",
+      title    = "ontoref runner + template bugs surfaced during dao-discipline rollout",
+      kind     = 'Bug,
+      priority = 'High,
+      status   = 'Done,
+      updated  = "2026-05-03",
+      detail   = m%"
+        SURFACED BY  adr-020 + lian-build adr-002/adr-003 smoke tests + the
+                     full update_ontoref end-to-end run during dao-discipline
+                     rollout (2026-05-03). What started as 3 validate.nu
+                     dispatcher bugs grew into 5 root-cause fixes when each
+                     fix exposed the next failure layer. All 5 fixed in this
+                     session; 4 of 5 propagated via just install + 1 via
+                     direct template edit.
+
+        FINAL BUG INVENTORY (all closed)
+
+        BUG #1 — port drift: client default 7891, daemon 7890
+          File:  reflection/modules/store.nu (daemon-url)
+          Cosmetic fix attempted first: changed literal 7891→7890 in 10 sites.
+          Real fix applied: single source of truth — daemon-url in store.nu
+          reads ~/.config/ontoref/config.ncl::daemon.port. The 9 consumer
+          sites import daemon-url from store.nu and call (daemon-url)
+          instead of duplicating the env-or-default pattern. Future port
+          changes are 1 edit in config.ncl, not 10.
+
+        BUG #2a — run-nucmd doesn't substitute placeholders
+          File:  reflection/nulib/modes.nu (exec-step-cmd)
+          Symptom: cmd "test -f {project_dir}/.ontology/manifest.ncl || ..."
+          ran with literal {project_dir}, hitting filesystem paths that
+          don't exist. Every mode step using placeholders silently failed.
+          Fix: build-cmd-context resolves {project_dir} / {ontoref_dir} /
+          {project_name} from env + .ontoref/project.ncl::slug; substitution
+          happens before bash/nu dispatch.
+
+        BUG #2b — paths with spaces break --import-path A:B:C
+          File:  reflection/modes/update_ontoref.ncl (validate-* steps)
+          Surfaced after Bug #2a fix exposed the second-order failure on
+          macOS where ontoref installs to ~/Library/Application Support/.
+          Fix: rewrote 4 cmds (add-manifest, add-connections, validate-
+          manifest, validate-connections) to quote {ontoref_dir} placeholder
+          and use --import-path repeated per directory instead of colon-
+          separated. Also future-proofs for any project path containing
+          spaces.
+
+        BUG #3 — template connections.ncl uses bare import "connections"
+          File:  templates/ontology/connections.ncl
+          Symptom: template generated files that didn't typecheck because
+          NICKEL_IMPORT_PATH didn't include a directory where bare
+          "connections" resolves. Existing onboarded projects worked
+          because they used import "reflection/schemas/connections.ncl"
+          (path-qualified) from older onboarding flows.
+          Fix: template aligned with convention.
+
+        BUG #4 — template manifest.ncl missing required consumption_modes
+          File:  templates/ontology/manifest.ncl
+          Symptom: generated manifest.ncl failed contract validation because
+          ProjectManifest schema requires consumption_modes (no default).
+          Surfaced after Bug #2b fix unblocked validate-manifest step.
+          Fix: template adds default 'Developer consumer scaffold with
+          comment marking it for replacement.
+
+        BUG #1 (original report; now superseded by inventory above) — run-nucmd doesn't cd to project root
+          File:     reflection/modules/validate.nu line ~89
+          Symptom:  NuCmd checks with relative paths silently mis-resolve.
+                    The cmd executes from validate.nu's invocation cwd, not
+                    from (adr-root). Concrete case: adr-020's
+                    layer-2-biconditional check evaluates to FALSE for both
+                    sides (no catalog/, no manifest.ncl) when invoked from
+                    reflection/modules/, producing a spurious 'biconditional
+                    holds' PASS even though the project violates it.
+          Fix:      cd (adr-root) before nu -c $check.cmd in run-nucmd.
+                    One-line patch.
+          Validates: re-run validate check-adr adr-020 from arbitrary cwd
+                    and confirm result matches direct-from-root invocation.
+
+        BUG #2 — run-cargo assumes workspace layout (crates/<name>/Cargo.toml)
+          File:     reflection/modules/validate.nu (run-cargo)
+          Symptom:  For single-crate projects (lian-build), the dispatcher
+                    looks at crates/lian-build/Cargo.toml which doesn't
+                    exist; the check fails with 'Cargo.toml not found'.
+          Concrete case: adr-003's no-nickel-rust-deps fails on lian-build
+                    despite Cargo.toml at project root containing no
+                    nickel-lang dep.
+          Fix:      Try Cargo.toml at root first; fall back to crates/<crate>/
+                    only if root Cargo.toml doesn't declare the named crate.
+                    Or: make the check accept an explicit `manifest_path`
+                    field instead of inferring from `crate` name.
+          Validates: re-run validate check-adr adr-003 from lian-build root;
+                    no-nickel-rust-deps should PASS.
+
+        BUG #3 — daemon-export-safe failure on some ADRs
+          File:     reflection/modules/store.nu line 200
+          Symptom:  `validate check-adr adr-001` against lian-build returns
+                    eval_block_with_input error from store.nu:200 ($result.stdout
+                    | from json with empty stdout because nickel export
+                    failed silently).
+          Hypothesis: daemon-export-safe doesn't pass the four
+                    nickel_import_paths from .ontoref/project.ncl, so ADRs
+                    that import from those paths fail to export. This was
+                    the root cause when I drafted lian-build adrs/ — they
+                    all import ../../../Development/ontoref/adrs/defaults.ncl.
+          Fix:      Read .ontoref/project.ncl::nickel_import_paths and pass
+                    them to nickel export as --import-path. Already partially
+                    done in some sites; needs auditing across daemon-export-safe.
+          Validates: re-run validate check-adr adr-001 from lian-build root;
+                    should not error from store.nu.
+
+        WHY HIGH PRIORITY
+          Without these fixes, the constraint discipline doesn't deliver its
+          value: false PASSes (Bug #1) silently mask real violations; false
+          FAILs (Bugs #2, #3) train readers to ignore the runner. Either
+          failure mode erodes trust in the validation surface.
+
+        REFERENCES
+          - reflection/modules/validate.nu     (file with all three bugs)
+          - reflection/modules/store.nu:200    (Bug #3 site)
+          - adrs/adr-020-three-layer-ontoref-instance-model.ncl
+                                                (Bug #1 surfaced here)
+          - lian-build/adrs/adr-002-cli-subcommand-discipline.ncl
+                                                (independent: my own NuCmd
+                                                 used bash `2>&1` instead of
+                                                 nu `out+err>`; fixed in
+                                                 same audit)
+          - lian-build/adrs/adr-003-nickel-via-subprocess.ncl
+                                                (Bug #2 surfaced here)
+        "%,
+      related_adrs  = ["adr-020"],
+      graduates_to  = 'PrItem,
+      created  = "2026-05-03",
+    },
+
+    {
+      id       = "bl-011",
+      title    = "Cross-project constraint application — protocol ADRs need a way to validate consumer projects",
+      kind     = 'Idea,
+      priority = 'Medium,
+      status   = 'Open,
+      detail   = m%"
+        SURFACED BY  adr-020 drafting (2026-05-03). The constraint runner in
+                     reflection/modules/validate.nu globs `adrs/adr-*.ncl`
+                     from the target project's root only. Protocol-level
+                     ADRs (currently adr-020, future adr-021+) live in
+                     <ontoref>/adrs/, so they do NOT apply when validating
+                     a consumer project (lian-build, provisioning).
+
+        EFFECT
+          adr-020's six constraints are runnable against ontoref itself, but
+          a consumer project's CI cannot use the same constraints to verify
+          its own Layer 1 / Layer 2 compliance. Each consumer would have to
+          either:
+            (a) copy adr-020 into its own adrs/ (cargo-cult, drift hazard)
+            (b) write its own adr-NNN that references the model (ceremony)
+            (c) wait for a protocol-aware validate runner
+
+        ALTERNATIVES (no decision committed)
+
+          A. Inherit by reference — `.ontoref/project.ncl` declares
+             `protocol_adrs = ["adr-020"]`, and validate.nu's adr-files
+             helper unions the project's own adrs/ with the named
+             protocol adrs from <ontoref-root>/adrs/. Uses the existing
+             ADR-018 level hierarchy machinery (parent-traversal).
+
+          B. Separate validate command — `validate check-protocol`
+             explicitly runs a fixed set of protocol-level ADRs against
+             the current target. Makes the protocol-vs-self distinction
+             explicit at the CLI; less magic than (A).
+
+          C. Cargo-cult copy on setup — `ontoref setup` writes a sentinel
+             file at `adrs/_protocol/adr-020.ncl` that pulls
+             `<ontoref>/adrs/adr-020-*.ncl` via NCL import. Lighter than
+             a full copy; brittle if the protocol path changes.
+
+          D. Status quo — protocol ADRs apply only to ontoref itself;
+             consumer projects implement their own self-mirror.
+
+        TRADE-OFF SUMMARY
+          - (A) extends the existing level-hierarchy plumbing; couples
+            validate.nu to the level resolution logic. Cleanest model.
+          - (B) is the smallest CLI change but requires every consumer
+            to remember to run it.
+          - (C) is the smallest setup change but creates phantom files.
+          - (D) is the least work and the least value.
+
+        WHEN THIS GRADUATES
+          When a second consumer project (beyond lian-build) onboards and
+          asks "how do I know my project conforms to adr-020?" Today
+          lian-build can be checked manually (and was, in 2026-05-03's
+          audit). At >=2 consumers the manual approach doesn't scale.
+
+        REFERENCES
+          - reflection/modules/validate.nu::adr-files                current root-only globbing
+          - reflection/modules/validate.nu::run-fileexists / run-nucmd  constraint runners (already exist)
+          - adr-018-level-hierarchy-mode-resolution-strategy           parent-traversal precedent
+          - adr-020-three-layer-ontoref-instance-model                 the ADR whose effectiveness this gates
+          - bl-010                                                     directory drift (independent of this gap)
+        "%,
+      related_adrs  = ["adr-020", "adr-018-level-hierarchy-mode-resolution-strategy"],
+      graduates_to  = 'Adr,
+      created  = "2026-05-03",
+    },
+
+    {
+      id       = "bl-008",
+      title    = "lian-build is a candidate satisfier of bl-002 (first external consuming project)",
+      kind     = 'Idea,
+      priority = 'Medium,
+      status   = 'Open,
+      detail   = "bl-002 ('First external project consuming the protocol') gates the protocol-maturity dimension's transition to 'protocol-stable'. lian-build now has its own ontoref instance (.ontology, reflection/{qa,backlog,modes}, adrs, manifest.ncl, catalog/{domains,modes}) and adopts the schema fully. Open question: does lian-build count as 'external' given that it's authored by the same maintainer set, or does the criterion require organizationally-distinct adoption? If it counts, this graduates bl-002. If not, lian-build is at minimum strong evidence the protocol is consumable. Decision should be recorded as an explicit acceptance criterion update on bl-002, not buried in a chat. Cross-references: lian-build's reflection/qa.ncl exemplifies a project's qa instance; reflection/backlog.ncl exemplifies project-level backlog; adrs/adr-001..003 exemplify project ADRs with constraint-bearing checks.",
+      related_dim   = "protocol-maturity",
+      graduates_to  = 'StateTransition,
+      created  = "2026-05-03",
+    },
+
+    {
+      id       = "bl-009",
+      title    = "Codify the three-layer ontoref model as a protocol-level concept",
+      kind     = 'Idea,
+      priority = 'Medium,
+      status   = 'InProgress,
+      updated  = "2026-05-03",
+      detail   = m%"
+        SURFACED BY  lian-build adoption (.coder/2026-05-03 conversation,
+                     reflection/qa.ncl::lian-build-what-and-why).
+        WHY HERE     when an adopting project sees the protocol, the layering
+                     of its ontoref instance must be discoverable from protocol
+                     documentation, not re-derived per project. Today the model
+                     is implicit — projects accidentally mix layers, and the
+                     'don't mix' rule lives in chat threads.
+
+        THE THREE LAYERS as observed across ontoref + lian-build + provisioning
+
+          LAYER 1 — Self-management ontoref (about the project itself)
+            paths:    .ontology/  reflection/  adrs/
+            audience: this project's developers and maintainers
+            purpose:  describe the project to itself — axioms, FSM dimensions,
+                      binding decisions, open questions
+            present:  EVERY ontoref-onboarded project
+
+          LAYER 2 — Specialized domain/mode ontoref (the integration surface)
+            paths:    schemas/  catalog/{domains,modes}/  manifest.ncl::registry_provides
+            audience: OTHER projects that want to integrate this project
+            purpose:  the contract surface other projects bind to —
+                      domain artifacts (typed contracts), mode artifacts
+                      (orchestration shapes), schemas, registry-namespace
+                      claim
+            present:  projects that publish federated artifacts (lian-build
+                      yes; some projects no — Layer 2 is OPTIONAL)
+
+          LAYER 3 — Caller-side implementations (NOT in this project)
+            paths:    <caller>/extensions/<this-project>/
+                      <workspace>/infra/<ws>/integrations/
+                      <caller>/catalog/components/<...>/  (when consuming)
+            audience: operators and CI of caller projects
+            purpose:  wiring this project into a specific workspace or pipeline
+            present:  PER CALLER, NEVER in this project's repo
+
+        WHY THE LAYERING MATTERS
+
+          The three layers collapse if mixed. Symptoms when mixing happens:
+          - Layer 3 content in Layer 1 → project FAQ becomes a how-to-deploy
+            guide for one specific caller; other callers get nothing.
+          - Layer 2 content in Layer 1 → schemas drift from the binary
+            because they're treated as 'project notes' instead of contracts.
+          - Layer 1 content in Layer 2 → caller pulls a contract artifact
+            and gets architectural rationale instead of a typed shape.
+          - Cross-project Layer 1 sharing → ontoref's qa.ncl would carry
+            project-specific entries that adopters can't filter out.
+
+          Each layer has a different lifecycle, audience, and validation
+          rule. Treating them as one violates the no-mixing convention
+          repeatedly observed in practice.
+
+        WHAT AN ADR WOULD COMMIT (constraint candidates — pick a subset)
+
+          (a) Layer 1 mandatory: every ontoref-onboarded project has
+              .ontology/core.ncl AND reflection/qa.ncl AND reflection/backlog.ncl
+              AND adrs/.
+              Check: 'FileExists per path.
+
+          (b) Layer 2 optional but well-formed when present: a project with
+              catalog/ MUST also have manifest.ncl::registry_provides
+              (and vice versa).
+              Check: 'NuCmd that asserts the biconditional.
+
+          (c) Layer 3 isolation: a project's src/ + Cargo.toml + .ontology/
+              + reflection/ MUST NOT contain caller-side cabling files
+              (cabling lives in callers).
+              Check: 'Grep with paths_exclude (depends on bl-007).
+
+          (d) Cross-layer reference convention: a project's qa entry that
+              touches another layer must be tagged accordingly
+              ("layer-1", "layer-2", "layer-3-boundary").
+              Check: 'NuCmd validating tag presence on cross-layer entries.
+
+        CANDIDATES TO ADD TO `ontoref setup`
+
+          - Generate reflection/qa.ncl scaffold (currently only backlog.ncl
+            and qa.ncl appear after manual setup).
+          - Generate manifest.ncl with registry_provides commented out (Layer 2
+            opt-in cue).
+          - Generate adrs/_template.ncl pre-filled with constraint-bearing
+            example shape (already done in ontoref itself — propagate).
+
+        OPEN QUESTIONS
+
+          - Naming: 'self-management' / 'integration-surface' / 'caller-side'
+            vs. 'Layer 1/2/3' vs. some other vocabulary. Numbered is precise
+            but flat; named is descriptive but verbose.
+          - Should Layer 2's `manifest.ncl::registry_provides` be in
+            `.ontology/` instead of project root? Today it lives at root in
+            both lian-build and ontoref; the alternative is .ontology/manifest.ncl
+            (which exists as a pointer in lian-build).
+          - How does this layering interact with ADR-018 (level-hierarchy:
+            Base / Domain / Instance)? Are the three layers orthogonal to
+            the three levels (a 3x3 matrix), or do they collapse onto each
+            other in some way?
+
+        REFERENCES
+          - lian-build/reflection/qa.ncl::lian-build-what-and-why  worked example
+          - lian-build/manifest.ncl                                Layer 2 example
+          - adr-018-level-hierarchy-mode-resolution-strategy       open interaction
+          - bl-002 / bl-008                                         protocol-maturity
+                                                                    transition gate
+        "%,
+      related_adrs  = ["adr-018-level-hierarchy-mode-resolution-strategy"],
+      related_dim   = "protocol-maturity",
+      graduates_to  = 'Adr,
+      created  = "2026-05-03",
+    },
+
   ],
 } | s.BacklogConfig
diff --git a/reflection/bin/ontoref.nu b/reflection/bin/ontoref.nu
index 851d20f..0d3c47a 100755
--- a/reflection/bin/ontoref.nu
+++ b/reflection/bin/ontoref.nu
@@ -25,6 +25,8 @@ use ../modules/graph.nu *
 use ../modules/validate.nu *
 use ../modules/migrate.nu *
 use ../modules/workflow.nu *
+use ../modules/secrets.nu *
+use ../modules/vault.nu *
 
 use ../nulib/fmt.nu *
 use ../nulib/shared.nu *
@@ -145,6 +147,8 @@ def "main help" [...args: string] {
   fmt-cmd $"($cmd) help search" "ontology search + bookmarks (NCL-persisted)"
   fmt-cmd $"($cmd) help qa" "Q&A knowledge base: query, add, list"
   fmt-cmd $"($cmd) help log" "action audit trail, follow, filter"
+  fmt-cmd $"($cmd) help secrets" "registry credential vault — bootstrap, sync, push (ADR-017)"
+  fmt-cmd $"($cmd) help vault" "vault snapshots, integrity check, status"
 
   # Active domain extension (conditional on project repo_kind)
   let proj_root = ($env.ONTOREF_PROJECT_ROOT? | default "")
@@ -305,6 +309,11 @@ def "main mode run" [id: string, --dry-run (-n), --yes (-y)] {
   run-mode $id --dry-run=$dry_run --yes=$yes
 }
 
+def "main mode resolve" [id: string, --fmt (-f): string = "text"] {
+  log-action $"mode resolve ($id)" "read"
+  mode resolve $id --fmt $fmt
+}
+
 # ── Run / Step ────────────────────────────────────────────────────────────────
 
 def "main run start" [mode: string, --task (-t): string = "", --fmt (-f): string = ""] {
@@ -355,6 +364,58 @@ def "main mg"   [action?: string] { missing-target "migrate" $action }
 def "main mg l" [--fmt (-f): string = ""] { log-action "migrate list" "read";    migrate list    --fmt $fmt }
 def "main mg p" [--fmt (-f): string = ""] { log-action "migrate pending" "read"; migrate pending --fmt $fmt }
 
+# ── Domain extensions ─────────────────────────────────────────────────────────
+# Static dispatch for declared domain extensions. The help renderer at startup
+# already displays the entry conditional on repo_kind match — execution is
+# wired here so `./ontoref <domain-id> <cmd> [args...]` reaches commands.nu.
+
+def --wrapped "main provisioning" [...args: string] {
+  log-action $"provisioning ($args | str join ' ')" "read"
+  let cmd_file = $"($env.ONTOREF_ROOT)/domains/provisioning/commands.nu"
+  ^nu $cmd_file ...$args
+}
+def --wrapped "main prov" [...args: string] { main provisioning ...$args }
+
+# ── Secrets ───────────────────────────────────────────────────────────────────
+
+def "main secrets" [action?: string] { missing-target "secrets" $action }
+def "main secrets help"        [] { help-group "secrets" }
+def "main secrets bootstrap"   [] { log-action "secrets bootstrap"  "write"; secrets bootstrap }
+def "main secrets sync"        [] { log-action "secrets sync"       "read";  secrets sync }
+def "main secrets push"        [] { log-action "secrets push"       "write"; secrets push }
+def "main secrets open"        [] { log-action "secrets open"       "write"; secrets open }
+def "main secrets close"       [] { log-action "secrets close"      "write"; secrets close }
+def "main secrets force-unlock" [] { log-action "secrets force-unlock" "write"; secrets force-unlock }
+def "main secrets set-cosign-password" [] { log-action "secrets set-cosign-password" "write"; secrets set-cosign-password }
+def "main secrets recover" [] { log-action "secrets recover" "write"; secrets recover }
+def "main secrets rekey"       [] { log-action "secrets rekey"      "write"; secrets rekey }
+def "main secrets status"      [] { log-action "secrets status"     "read";  secrets status }
+def "main secrets describe"    [] { log-action "secrets describe"   "read";  secrets describe }
+def "main secrets gen-key"     [role: string] {
+  log-action $"secrets gen-key ($role)" "read"
+  secrets gen-key $role
+}
+def "main secrets add-key"     [pubkey: string, role: string] {
+  log-action $"secrets add-key ($role)" "write"
+  secrets add-key $pubkey $role
+}
+def "main secrets remove-key"  [pubkey: string] {
+  log-action "secrets remove-key" "write"
+  secrets remove-key $pubkey
+}
+def "main secrets audit" [--check: string = ""] {
+  log-action $"secrets audit ($check)" "read"
+  secrets audit --check $check
+}
+
+# ── Vault ─────────────────────────────────────────────────────────────────────
+
+def "main vault" [action?: string] { missing-target "vault" $action }
+def "main vault help"      [] { help-group "vault" }
+def "main vault snapshots" [] { log-action "vault snapshots" "read"; vault snapshots }
+def "main vault check"     [] { log-action "vault check"     "read"; vault check }
+def "main vault status"    [] { log-action "vault status"    "read"; vault status }
+
 # ── Workflow ──────────────────────────────────────────────────────────────────
 
 def "main workflow" [action?: string] { missing-target "workflow" $action }
@@ -404,6 +465,14 @@ def "main validate justfile" [--fmt (-f): string = ""] {
   log-action "validate justfile" "read"
   validate justfile --fmt $fmt
 }
+def "main validate modes" [--check (-c): string = "all", --fmt (-f): string = "text"] {
+  log-action $"validate modes --check ($check)" "read"
+  validate modes --check $check --fmt $fmt
+}
+def "main validate modes --self-test" [--fmt (-f): string = "text"] {
+  log-action "validate modes --self-test" "read"
+  validate modes --self-test --fmt $fmt
+}
 
 # ── ADR ───────────────────────────────────────────────────────────────────────
 
@@ -515,6 +584,11 @@ def "main sync manifest-check" [--strict] {
   log-action "sync manifest-check" "read"
   sync manifest-check --strict=$strict
 }
+def "main sync protocol-files" [--apply, --fmt (-f): string = "", --actor: string = ""] {
+  log-action $"sync protocol-files (if $apply { 'apply' } else { 'dry-run' })" (if $apply { "write" } else { "read" })
+  let f = (resolve-fmt $fmt [text json])
+  sync protocol-files --apply=$apply --fmt $f --actor $actor
+}
 
 # ── Coder ─────────────────────────────────────────────────────────────────────
 
@@ -593,6 +667,10 @@ def "main describe project" [--fmt (-f): string = "", --actor: string = ""] {
   log-action "describe project" "read"
   let f = (resolve-fmt $fmt [text table json yaml toml]); describe project --fmt $f --actor $actor
 }
+def "main describe vault" [--fmt (-f): string = "", --actor: string = ""] {
+  log-action "describe vault" "read"
+  let f = (resolve-fmt $fmt [text table json yaml toml]); describe vault --fmt $f --actor $actor
+}
 def "main describe capabilities" [--fmt (-f): string = "", --actor: string = ""] {
   log-action "describe capabilities" "read"
   let f = (resolve-fmt $fmt [text table json yaml toml]); describe capabilities --fmt $f --actor $actor
@@ -1448,7 +1526,7 @@ def notify-daemon-project-add [project_path: string] {
     print $"  (ansi yellow)warn(ansi reset): daemon not notified — could not export project.ncl"
     return
   }
-  let url = $"($env.ONTOREF_DAEMON_URL? | default "http://127.0.0.1:7891")/projects"
+  let url = $"(daemon-url)/projects"
   let auth = (bearer-args)
   let r = do { ^curl -sf -X POST -H "Content-Type: application/json" ...$auth -d ($entry | to json) $url } | complete
   if $r.exit_code == 0 {
@@ -1462,7 +1540,7 @@ def notify-daemon-project-add [project_path: string] {
 # Silently skips when slug is empty or daemon is not reachable.
 def notify-daemon-project-remove [slug: string] {
   if ($slug | is-empty) or not (daemon-available) { return }
-  let url = $"($env.ONTOREF_DAEMON_URL? | default "http://127.0.0.1:7891")/projects/($slug)"
+  let url = $"(daemon-url)/projects/($slug)"
   let auth = (bearer-args)
   let r = do { ^curl -sf -X DELETE ...$auth $url } | complete
   if $r.exit_code == 0 {
diff --git a/reflection/defaults/workflow.ncl b/reflection/defaults/workflow.ncl
index 3864329..ecade57 100644
--- a/reflection/defaults/workflow.ncl
+++ b/reflection/defaults/workflow.ncl
@@ -170,7 +170,7 @@
       args          = ["build", "--release", "--workspace"],
       cargo_profile = "release",
       target        = "",
-      artifacts     = ["target/release/"],
+      artifacts     = ["{TARGET_DIR}/release/"],
     },
 
     release-musl-x86 = {
@@ -180,7 +180,7 @@
       args          = ["build", "--target", "x86_64-unknown-linux-musl", "--release"],
       cargo_profile = "release",
       target        = "x86_64-unknown-linux-musl",
-      artifacts     = ["target/x86_64-unknown-linux-musl/release/"],
+      artifacts     = ["{TARGET_DIR}/x86_64-unknown-linux-musl/release/"],
     },
 
     sbom = {
@@ -190,7 +190,7 @@
       args          = [],
       cargo_profile = "",
       target        = "",
-      artifacts     = ["sbom.json"],
+      artifacts     = ["{ROOT}/sbom.json"],
     },
 
     docs-html = {
@@ -200,7 +200,7 @@
       args          = ["doc", "--no-deps", "--workspace"],
       cargo_profile = "release",
       target        = "",
-      artifacts     = ["target/doc/"],
+      artifacts     = ["{TARGET_DIR}/doc/"],
     },
   },
 
diff --git a/reflection/migrations/0003-manifest-self-interrogation.ncl b/reflection/migrations/0003-manifest-self-interrogation.ncl
index 2aa2b4e..5b98b3e 100644
--- a/reflection/migrations/0003-manifest-self-interrogation.ncl
+++ b/reflection/migrations/0003-manifest-self-interrogation.ncl
@@ -3,9 +3,10 @@
   slug        = "manifest-self-interrogation",
   description = "Populate capabilities[], requirements[], and critical_deps[] in manifest.ncl",
   check       = {
-    tag         = "NuCmd",
-    cmd         = "let f = $\"($env.ONTOREF_PROJECT_ROOT)/.ontology/manifest.ncl\"; if not ($f | path exists) { exit 1 }; let r = (do { ^nickel export --format json --import-path $env.NICKEL_IMPORT_PATH $f } | complete); if $r.exit_code != 0 { exit 1 }; let m = ($r.stdout | from json); if (($m.capabilities? | default [] | length) > 0) and (($m.requirements? | default [] | length) > 0) { exit 0 } else { exit 1 }",
-    expect_exit = 0,
+    tag           = "Grep",
+    pattern       = "make_capability",
+    paths         = [".ontology/manifest.ncl"],
+    must_be_empty = false,
   },
   instructions = "
 Open .ontology/manifest.ncl and populate three arrays.
diff --git a/reflection/migrations/0010-manifest-capability-completeness.ncl b/reflection/migrations/0010-manifest-capability-completeness.ncl
index 3523da7..2c5e6c2 100644
--- a/reflection/migrations/0010-manifest-capability-completeness.ncl
+++ b/reflection/migrations/0010-manifest-capability-completeness.ncl
@@ -3,9 +3,10 @@
   slug        = "manifest-capability-completeness",
   description = "Ensure manifest.ncl declares capabilities[] covering all project functionality — not just ontology structure. Without capabilities, agents see a partial view of the project via describe capabilities.",
   check       = {
-    tag         = "NuCmd",
-    cmd         = "let root = $env.ONTOREF_PROJECT_ROOT; let manifest = $\"($root)/.ontology/manifest.ncl\"; if not ($manifest | path exists) { exit 1 }; let caps = (nickel export --format json $manifest | from json | get capabilities? | default []); if ($caps | length) >= 3 { exit 0 } else { exit 1 }",
-    expect_exit = 0,
+    tag           = "Grep",
+    pattern       = "make_capability",
+    paths         = [".ontology/manifest.ncl"],
+    must_be_empty = false,
   },
   instructions = "
 Your project's .ontology/manifest.ncl must declare capabilities[] that describe what the project
diff --git a/reflection/migrations/0014-workflow-layer-model.ncl b/reflection/migrations/0014-workflow-layer-model.ncl
index e6293bf..cb7f311 100644
--- a/reflection/migrations/0014-workflow-layer-model.ncl
+++ b/reflection/migrations/0014-workflow-layer-model.ncl
@@ -4,7 +4,7 @@
   description = "Workflow layer model: .ontology/workflow.ncl declares independent CI/build/distribution layers. Cargo profiles ci-test (pre-commit, debug=0) and ci (CI pipeline, line-tables-only) must exist. Nextest profiles ci-test and ci must exist. Generator: ore workflow generate.",
   check       = {
     tag         = "NuCmd",
-    cmd         = "let root = ($env.ONTOREF_PROJECT_ROOT? | default ($env.ONTOREF_ROOT? | default (pwd | path expand))); let wf = $\"($root)/.ontology/workflow.ncl\"; if not ($wf | path exists) { exit 1 }; let cargo_cfg = $\"($root)/.cargo/config.toml\"; if not ($cargo_cfg | path exists) { exit 1 }; let cargo_txt = (open --raw $cargo_cfg); let has_ci_test = ($cargo_txt | str contains \"[profile.ci-test]\"); let has_ci = ($cargo_txt | str contains \"[profile.ci]\"); let nextest_cfg = $\"($root)/.config/nextest.toml\"; if not ($nextest_cfg | path exists) { exit 1 }; let nextest_txt = (open --raw $nextest_cfg); let nx_ci_test = ($nextest_txt | str contains \"[profile.ci-test]\"); let nx_ci = ($nextest_txt | str contains \"[profile.ci]\"); if $has_ci_test and $has_ci and $nx_ci_test and $nx_ci { exit 0 } else { exit 1 }",
+    cmd         = "let root = ($env.ONTOREF_PROJECT_ROOT? | default ($env.ONTOREF_ROOT? | default (pwd | path expand))); let wf = $\"($root)/.ontology/workflow.ncl\"; if not ($wf | path exists) { exit 1 }; let has_rust = ($\"($root)/Cargo.toml\" | path exists); if not $has_rust { exit 0 }; let cargo_cfg = $\"($root)/.cargo/config.toml\"; if not ($cargo_cfg | path exists) { exit 1 }; let cargo_txt = (open --raw $cargo_cfg); let has_ci_test = ($cargo_txt | str contains \"[profile.ci-test]\"); let has_ci = ($cargo_txt | str contains \"[profile.ci]\"); let nextest_cfg = $\"($root)/.config/nextest.toml\"; if not ($nextest_cfg | path exists) { exit 1 }; let nextest_txt = (open --raw $nextest_cfg); let nx_ci_test = ($nextest_txt | str contains \"[profile.ci-test]\"); let nx_ci = ($nextest_txt | str contains \"[profile.ci]\"); if $has_ci_test and $has_ci and $nx_ci_test and $nx_ci { exit 0 } else { exit 1 }",
     expect_exit = 0,
   },
   instructions = "
diff --git a/reflection/migrations/0015-registry-topology.ncl b/reflection/migrations/0015-registry-topology.ncl
new file mode 100644
index 0000000..9393273
--- /dev/null
+++ b/reflection/migrations/0015-registry-topology.ncl
@@ -0,0 +1,62 @@
+{
+  id          = "0015",
+  slug        = "registry-topology",
+  description = "Declare registry_provides in .ontology/manifest.ncl with participant id and namespace topology. Domain artifacts now use domains/<participant>/<id>:<version> instead of the flat domains/<id>:<version> namespace.",
+
+  check = {
+    tag     = "Grep",
+    paths   = [".ontology/manifest.ncl"],
+    pattern = "registry_provides",
+  },
+
+  instructions = "
+## Registry topology declaration (migration 0015)
+
+The provisioning registry now uses participant-scoped namespaces for all artifacts:
+  domains/<participant>/<id>:<version>
+  modes/<participant>/<id>:<version>  (already correct if you used mode-push)
+
+### Step 1 — Declare registry_provides in .ontology/manifest.ncl
+
+Add inside your make_manifest { ... }:
+
+  registry_provides = m.make_registry_provides {
+    participant = \"<your-project-id>\",
+    registries  = m.make_registries_config {
+      default    = \"primary\",
+      registries = [
+        m.make_registry_entry {
+          id         = \"primary\",
+          endpoint   = \"\",
+          role       = 'primary,
+          tls        = true,
+          namespaces = {
+            own      = [\"domains/<your-project-id>/\", \"modes/<your-project-id>/\"],
+            prefixes = [\"domains/<your-project-id>/\", \"modes/<your-project-id>/\"],
+          },
+        },
+      ],
+    },
+  },
+
+Set endpoint = \"\" and configure PROVISIONING_REGISTRY at runtime, or fill it in
+once the registry hostname is stable.
+
+### Step 2 — Update domain-push calls (participant is now positional arg 2)
+
+  # Before
+  domain-push <path> --registry <reg>
+  # After
+  domain-push <path> <your-project-id> --registry <reg>
+
+### Step 3 — Update domain-pull / describe / verify (use participant/id form)
+
+  # Before: domain-pull <id> <version>
+  # After:  domain-pull <participant> <id> <version>
+  #         or via: prvng i domain pull <participant>/<id> <version>
+
+### Step 4 — Verify live state
+
+  PROVISIONING_REGISTRY=<host> ore provisioning i list --live
+",
+}
diff --git a/reflection/migrations/0016-registry-credential-vault.ncl b/reflection/migrations/0016-registry-credential-vault.ncl
new file mode 100644
index 0000000..4243d16
--- /dev/null
+++ b/reflection/migrations/0016-registry-credential-vault.ncl
@@ -0,0 +1,129 @@
+{
+  id          = "0016",
+  slug        = "registry-credential-vault",
+  description = "Add sops vault config to project.ncl and credential_sops fields to registry entries (ADR-017). Enables per-project multi-recipient age-encrypted credential vaults stored as OCI artifacts in ZOT, with isolated DOCKER_CONFIG for every oras invocation.",
+
+  check = {
+    tag     = "Grep",
+    paths   = [".ontoref/project.ncl"],
+    pattern = "sops",
+  },
+
+  instructions = m%"
+## Registry credential vault (migration 0016)
+
+Implements ADR-017: per-project sops multi-recipient vaults stored as OCI artifacts
+in ZOT. Credentials are never in env vars, never in ambient docker config, never
+plaintext on disk. Every oras invocation runs with an isolated DOCKER_CONFIG.
+
+### Reference material
+
+- **Helper contract**: `reflection/modules/secrets.nu` header docstring — defines
+  `resolve-vault-access` (Layer 0), `resolve-registry-credential` (Layer 2),
+  `assert-actor-authorized`, plus the 13 named errors callers must handle.
+- **Working example**: ontoref's own `.ontoref/project.ncl` and
+  `.ontology/manifest.ncl` — copy this shape into your project.
+- **CLI surface**: `ore secrets {bootstrap|sync|push|open|close|status|audit}` and
+  `ore vault {status|snapshots|check}` (P2 dispatcher wireup).
+
+### Step 1 — Set the global master key path (one-time per developer)
+
+In `~/.config/ontoref/config.ncl`:
+
+  vault = {
+    master_key_path = "/path/to/your/.kage",
+    backend         = 'restic,
+  },
+
+This becomes the default for all projects. A project may override it in its own
+`.ontoref/project.ncl::sops.master_key_path` if it requires a different key.
+
+### Step 2 — Add the sops block to .ontoref/project.ncl
+
+Inside the `make_project { ... }` call:
+
+  sops = {
+    enabled            = true,
+    vault_id           = "<your-project-slug>",
+    vault_backend      = 'restic,
+    registry_endpoint  = "<your-zot-host>",
+    actor_key_bindings = {
+      developer = "developer",
+      ci        = "cdci",
+      agent     = "ontoref",
+      admin     = "admin",
+    },
+    # master_key_path absent → resolves from ~/.config/ontoref/config.ncl::vault.master_key_path
+  },
+
+### Step 3 — Add registry_provides + credential_sops to .ontology/manifest.ncl
+
+  registry_provides = m.make_registry_provides {
+    participant = "<your-project-slug>",
+    registries  = m.make_registries_config {
+      default    = "primary",
+      registries = [
+        m.make_registry_entry {
+          id                 = "primary",
+          endpoint           = "<your-zot-host>",
+          role               = 'primary,
+          tls                = true,
+          namespaces = {
+            own      = ["domains/<your-slug>/", "modes/<your-slug>/"],
+            prefixes = ["domains/<your-slug>/", "modes/<your-slug>/"],
+          },
+          credential_sops    = "registry/ro.sops.yaml",   # RO — relative to src-vault root
+          credential_sops_rw = "registry/rw.sops.yaml",   # RW — relative to src-vault root
+        },
+      ],
+    },
+  },
+
+NOTE: `credential_sops` paths are relative to the src-vault root
+(`~/.config/ontoref/vaults/<vault_id>/src-vault/`), NOT to the project root.
+
+### Step 4 — Declare uses_registry in domains/modes that call oras
+
+Any domain or mode NCL file that pushes or pulls artifacts must add:
+
+  uses_registry = "<registry-entry-id>",   # id from make_registry_entry
+
+Enables impact analysis on `ore secrets close` to find affected services.
+
+### Step 5 — Bootstrap the vault (admin only, one-time per project)
+
+  export SOPS_AGE_RECIPIENTS="age1<developer>,age1<ci>,age1<admin>,..."
+  ore secrets bootstrap
+
+This creates `~/.config/ontoref/vaults/<vault_id>/access.sops.yaml`, initializes
+the local restic/kopia repo, and (next step) pushes the first src-vault artifact.
+
+### Step 6 — Push the bootstrapped vault to ZOT
+
+  export COSIGN_KEY_PATH=/path/to/cosign.key   # required by ADR-017
+  ore secrets push
+
+The artifact is signed with cosign — unsigned vault pushes are rejected by the
+`src-vault-cosign-signed` hard constraint.
+
+### Step 7 — Non-admin actors: sync the vault
+
+  ore secrets sync
+
+Pulls `src-vault/<vault_id>:latest` from ZOT and decrypts access.sops.yaml.
+
+### Step 8 — Verify ADR-017 constraints
+
+  ore secrets audit
+
+Runs all checks: bootstrap-credentials, no-credential-env, multi-recipient
+mandatory. Failures here block downstream operations.
+
+### Step 9 — Verify visible state
+
+  ore secrets status     # vault dir + last access entry + master_key resolution
+  ore vault   status     # backend + repo presence + last snapshot
+
+Both should report a healthy state with no errors.
+"%,
+}
diff --git a/reflection/migrations/0017-level-hierarchy-strategy.ncl b/reflection/migrations/0017-level-hierarchy-strategy.ncl
new file mode 100644
index 0000000..d2cd840
--- /dev/null
+++ b/reflection/migrations/0017-level-hierarchy-strategy.ncl
@@ -0,0 +1,84 @@
+{
+  id          = "0017",
+  slug        = "level-hierarchy-strategy",
+  description = "Declare level identity in manifest.ncl and strategy field in reflection modes (ADR-018). Required at Domain (level 2) and Instance (level 3) to enable observable traversal via ore mode resolve.",
+
+  check = {
+    tag     = "Grep",
+    paths   = [".ontology/manifest.ncl"],
+    pattern = "level\\s*=\\s*\\{",
+  },
+
+  instructions = "
+## Level hierarchy and mode resolution strategy (migration 0017)
+
+Implements ADR-018: three-level hierarchy with per-mode resolution strategy.
+This makes domain boundaries explicit and mode traversal observable.
+
+### Step 1 — Declare level identity in .ontology/manifest.ncl
+
+Inside the manifest record, add a `level` field. The value depends on what
+this project IS in the hierarchy:
+
+  # Level 2 — project domain (e.g. provisioning, personal):
+  level = {
+    index  = 'Domain,
+    name   = \"<your-domain-name>\",    # e.g. \"provisioning-domain\"
+    parent = \"ontoref-base\",
+  },
+
+  # Level 3 — domain instance (e.g. a concrete workspace or infra project):
+  level = {
+    index  = 'Instance,
+    name   = \"<your-instance-name>\",  # e.g. \"myorg-provisioning\"
+    parent = \"provisioning-domain\",   # matches the domain's level.name
+  },
+
+ontoref itself declares no level field — 'Base is the implicit default.
+
+### Step 2 — Add strategy to reflection modes in reflection/modes/*.ncl
+
+Each mode at level 2+ must declare how it relates to the parent level's mode.
+
+  # Mode is fully implemented here — traversal stops:
+  strategy = 'Override,
+
+  # Mode is not implemented here — traverse to parent for execution:
+  strategy = 'Delegate,
+
+  # Mode accumulates contributions from all levels (lower level wins conflicts):
+  strategy = 'Merge,
+
+  # Mode inherits specific steps from the parent, overrides the rest:
+  strategy = 'Compose,
+  extends  = [\"step-id-from-parent\", \"other-step-id\"],
+
+Implicit absence (no strategy field at level 2+) is treated as 'Delegate with
+a Soft validation warning from ore validate modes --check strategy-declared.
+
+### Step 3 — Verify (once ore validate modes is available)
+
+  ore validate modes --check level-declared
+  ore validate modes --check strategy-declared
+  ore validate modes --check delegate-chain
+
+### Step 4 — For strategy transitions: bind to state.ncl
+
+When a mode's strategy is expected to change as the project matures, declare
+a dimension in state.ncl. Example — a mode currently delegating that will
+eventually override:
+
+  {
+    id            = \"build-docs-strategy\",
+    description   = \"build-docs mode resolution strategy: Delegate → Override\",
+    current_state = \"Delegate\",
+    desired_state = \"Override\",
+    catalyst      = \"implement project-specific doc generation\",
+    blocker       = \"doc generation not yet implemented for this domain\",
+  },
+
+When the implementation is complete: update current_state = \"Override\" in state.ncl
+AND update strategy = 'Override in the mode NCL. This makes the transition an
+observable FSM event (tracked in git, visible via ore describe state).
+",
+}
diff --git a/reflection/migrations/0018-recipient-routing.ncl b/reflection/migrations/0018-recipient-routing.ncl
new file mode 100644
index 0000000..d9bbabe
--- /dev/null
+++ b/reflection/migrations/0018-recipient-routing.ncl
@@ -0,0 +1,136 @@
+{
+  id          = "0018",
+  slug        = "recipient-routing",
+  description = "Adopt per-file recipient routing for tenant isolation (ADR-019). Optional: only projects requiring multi-tenant credential separation need to apply. Adds recipient_groups + recipient_rules to project.ncl::sops; bootstrap then generates <vault_dir>/.sops.yaml from declarations and sops encrypts each file with the union of declared groups.",
+
+  # Adoption is opt-in. Projects without multi-tenant needs MAY skip.
+  # Mark as 'applied' when the project explicitly chose a path:
+  #   - Path A (single-team / legacy): project.ncl has sops but no recipient_groups
+  #   - Path B/C: project.ncl declares both recipient_groups and recipient_rules
+  # Either path satisfies the check below.
+  check = {
+    tag     = "Grep",
+    paths   = [".ontoref/project.ncl"],
+    pattern = "sops",
+  },
+
+  instructions = m%"
+## Per-file recipient routing for tenant isolation (migration 0018, ADR-019)
+
+This migration is **OPTIONAL**. Apply only if your project requires multi-tenant
+credential separation, AI-agent restricted access, or developer/admin role mixing
+that the single-recipient-set model from migration 0016 cannot express.
+
+### Decide your path
+
+  ┌────────────────────────────────────────────────────────────────────┐
+  │ Will multiple actors (clients/agents/teams) operate this project?  │
+  └─────────────────────────────────┬──────────────────────────────────┘
+                                    │
+                  ┌─────── No ──────┴─────── Yes ──────────────┐
+                  ▼                                             ▼
+  ┌──────────────────────────┐         ┌──────────────────────────────────┐
+  │ Path A — Single team     │         │ Will those actors need DISTINCT  │
+  │ Migration 0016 sufficient│         │ credential sets per actor?       │
+  │ — no further action      │         └──────────────────┬───────────────┘
+  └──────────────────────────┘                            │
+                                  ┌─────── No ────────────┴──── Yes ──────┐
+                                  ▼                                        ▼
+                  ┌─────────────────────────────────┐    ┌────────────────────────┐
+                  │ Stay with migration 0016         │    │ Path B/C: this 0018    │
+                  │ Re-evaluate when needs change    │    │ Apply steps below      │
+                  └──────────────────────────────────┘    └────────────────────────┘
+
+### Step 1 — Templates
+
+Three templates ship under install/resources/templates/sops/:
+
+  single-team/    — Path A reference (no migration needed)
+  multi-tenant/   — Path B reference (clients/services/teams isolated)
+  agent-first/    — Path C reference (AI agent has narrow read-only access)
+
+Pick one, copy its project.ncl.snippet and manifest.ncl.snippet into the
+matching files in your project. Adapt placeholders to your tenancy.
+
+### Step 2 — Declare groups and rules
+
+Inside .ontoref/project.ncl::sops, add:
+
+  recipient_groups = {
+    admin   = [\"age1...\"],            # admins (full access)
+    ops     = [\"age1...\"],            # ops/observability
+    clientA = [\"age1clientA-lead...\"], # one group per client/team/agent
+    clientB = [\"age1clientB-lead...\"],
+    agents  = [\"age1agent...\"],       # AI agents (typically RO)
+  },
+  recipient_rules = [
+    { path = \"access\\\\.sops\\\\.yaml$\",                  groups = [\"admin\", \"ops\"] },
+    { path = \"registry/clientA-.*\\\\.sops\\\\.yaml$\",     groups = [\"admin\", \"clientA\"] },
+    { path = \"registry/clientB-.*\\\\.sops\\\\.yaml$\",     groups = [\"admin\", \"clientB\"] },
+    { path = \"registry/agent-readonly\\\\.sops\\\\.yaml$\", groups = [\"admin\", \"agents\"] },
+  ],
+
+### Step 3 — Add per-tenant RegistryEntry items in manifest.ncl
+
+For each rule, declare a corresponding RegistryEntry whose credential_sops
+matches the rule's path pattern:
+
+  m.make_registry_entry {
+    id                 = \"clientA\",
+    endpoint           = \"<your-zot-host>\",
+    credential_sops    = \"registry/clientA-ro.sops.yaml\",
+    credential_sops_rw = \"registry/clientA-rw.sops.yaml\",
+    namespaces = {
+      own      = [\"domains/clientA/\", \"modes/clientA/\"],
+      prefixes = [\"domains/clientA/\"],
+    },
+  }
+
+### Step 4 — Bootstrap or migrate
+
+  NEW PROJECT       ore secrets bootstrap
+                    (generates <vault_dir>/.sops.yaml + access.sops.yaml,
+                     skips default registry/ro|rw.sops.yaml — operator
+                     populates per-tenant files via 'ore secrets open')
+
+  EXISTING PROJECT  ore secrets rekey
+                    (regenerates <vault_dir>/.sops.yaml from new declarations,
+                     re-encrypts every existing *.sops.yaml with the matching rule)
+
+### Step 5 — Verify
+
+  ore secrets describe          # see groups, rules, recipients per file
+  ore secrets audit             # 6 checks pass:
+                                #   bootstrap-credentials      OK
+                                #   no-credential-env          OK
+                                #   multi-recipient-mandatory  OK
+                                #   recipient-routing-coherent OK
+                                #   recipient-routing-coverage OK
+                                #   no-multi-vault             OK
+
+### What does NOT change
+
+- vault_id remains one per project (multi-vault is explicitly out of scope)
+- Master .kage continues per-developer
+- Restic/Kopia repo and OCI artifact remain unique per project
+- Vault lock semantics unchanged
+- All ADR-017 invariants preserved
+
+### When to escalate to multi-vault (not implemented)
+
+- Hard isolation: separate master keys per environment (prod vs staging)
+- Hard isolation: separate restic repos to prevent any cross-pull visibility
+- Compliance grade: regulator requires environments cannot share vault
+  contents even when encrypted
+
+When such a case appears, capture it in a new ADR superseding/extending
+ADR-019 before adopting multi-vault.
+
+### References
+
+  ADR-019                                          decision + 6 checkable constraints
+  reflection/qa.ncl::credential-vault-templates    three-path overview
+  install/resources/templates/sops/                ready-to-copy snippets
+  reflection/migrations/0016                       Layer 0/1/2 prerequisites
+"%,
+}
diff --git a/reflection/modes/update_ontoref.ncl b/reflection/modes/update_ontoref.ncl
index e64397b..a02a193 100644
--- a/reflection/modes/update_ontoref.ncl
+++ b/reflection/modes/update_ontoref.ncl
@@ -82,7 +82,7 @@ let s = import "../schema.ncl" in
       id         = "detect-daemon-api",
       action     = "Check whether daemon exposes /api/catalog (v2 capability)",
       actor      = 'Agent,
-      cmd        = "curl -sf ${ONTOREF_DAEMON_URL:-http://127.0.0.1:7891}/api/catalog > /dev/null && echo 'daemon: v2 api catalog available' || echo 'daemon: not reachable or pre-v2 (start/restart the daemon)'",
+      cmd        = "curl -sf ${ONTOREF_DAEMON_URL:-http://127.0.0.1:7890}/api/catalog > /dev/null && echo 'daemon: v2 api catalog available' || echo 'daemon: not reachable or pre-v2 (start/restart the daemon)'",
       depends_on = [],
       on_error   = { strategy = 'Continue },
       note       = "Non-blocking check. Daemon must be restarted to expose new API catalog endpoint.",
@@ -94,7 +94,7 @@ let s = import "../schema.ncl" in
       id         = "add-manifest",
       action     = "Create .ontology/manifest.ncl stub if missing",
       actor      = 'Agent,
-      cmd        = "test -f {project_dir}/.ontology/manifest.ncl || sed 's/{{ project_name }}/{project_name}/g' {ontoref_dir}/templates/ontology/manifest.ncl > {project_dir}/.ontology/manifest.ncl",
+      cmd        = "test -f \"{project_dir}/.ontology/manifest.ncl\" || sed 's/{{ project_name }}/{project_name}/g' \"{ontoref_dir}/templates/ontology/manifest.ncl\" > \"{project_dir}/.ontology/manifest.ncl\"",
       depends_on = [{ step = "detect-manifest", kind = 'Always }],
       on_error   = { strategy = 'Continue },
       note       = "Adds content asset and template declarations. Skipped if file already exists.",
@@ -104,7 +104,7 @@ let s = import "../schema.ncl" in
       id         = "add-connections",
       action     = "Create .ontology/connections.ncl stub if missing",
       actor      = 'Agent,
-      cmd        = "test -f {project_dir}/.ontology/connections.ncl || sed 's/{{ project_name }}/{project_name}/g' {ontoref_dir}/templates/ontology/connections.ncl > {project_dir}/.ontology/connections.ncl",
+      cmd        = "test -f \"{project_dir}/.ontology/connections.ncl\" || sed 's/{{ project_name }}/{project_name}/g' \"{ontoref_dir}/templates/ontology/connections.ncl\" > \"{project_dir}/.ontology/connections.ncl\"",
       depends_on = [{ step = "detect-connections", kind = 'Always }],
       on_error   = { strategy = 'Continue },
       note       = "Adds cross-project federation stub. Skipped if file already exists.",
@@ -116,20 +116,20 @@ let s = import "../schema.ncl" in
       id         = "validate-manifest",
       action     = "Nickel typecheck .ontology/manifest.ncl",
       actor      = 'Agent,
-      cmd        = "cd {project_dir} && nickel export --import-path {ontoref_dir}/ontology:{ontoref_dir}/ontology/schemas:{ontoref_dir}/ontology/defaults:{ontoref_dir} .ontology/manifest.ncl > /dev/null",
+      cmd        = "cd \"{project_dir}\" && nickel export --import-path \"{ontoref_dir}/ontology\" --import-path \"{ontoref_dir}/ontology/schemas\" --import-path \"{ontoref_dir}/ontology/defaults\" --import-path \"{ontoref_dir}\" .ontology/manifest.ncl > /dev/null",
       depends_on = [{ step = "add-manifest", kind = 'OnSuccess }],
       on_error   = { strategy = 'Continue },
-      note       = "Confirms manifest.ncl parses under the current content schema.",
+      note       = "Confirms manifest.ncl parses under the current content schema. Quotes around {ontoref_dir} survive the macOS install path that contains a space (~/Library/Application Support/ontoref); --import-path repeated per directory instead of colon-separated to avoid splitting on the space when bash word-splits.",
     },
 
     {
       id         = "validate-connections",
       action     = "Nickel typecheck .ontology/connections.ncl",
       actor      = 'Agent,
-      cmd        = "cd {project_dir} && nickel export --import-path {ontoref_dir}/reflection/schemas:{ontoref_dir} .ontology/connections.ncl > /dev/null",
+      cmd        = "cd \"{project_dir}\" && nickel export --import-path \"{ontoref_dir}/reflection/schemas\" --import-path \"{ontoref_dir}\" .ontology/connections.ncl > /dev/null",
       depends_on = [{ step = "add-connections", kind = 'OnSuccess }],
       on_error   = { strategy = 'Continue },
-      note       = "Confirms connections.ncl parses under the connections schema.",
+      note       = "Confirms connections.ncl parses under the connections schema. Quoted paths + repeated --import-path flag for the same reason as validate-manifest.",
     },
 
     # ── REPORT (aggregate — depends on all previous steps) ────────────────────
diff --git a/reflection/modules/adr.nu b/reflection/modules/adr.nu
index 4fb3ef4..ce3b95a 100644
--- a/reflection/modules/adr.nu
+++ b/reflection/modules/adr.nu
@@ -57,7 +57,14 @@ def adr-to-md []: record -> string {
     $lines = ($lines | append $"### ($c.id) [($c.severity)]")
     $lines = ($lines | append $c.claim)
     $lines = ($lines | append $"- Scope: ($c.scope)")
-    $lines = ($lines | append $"- Check: `($c.check_hint)`")
+    let check_display = if ($c | get -o check_hint | is-not-empty) {
+      $c.check_hint
+    } else if ($c | get -o check | is-not-empty) {
+      $c.check | to json --raw
+    } else {
+      "(no check)"
+    }
+    $lines = ($lines | append $"- Check: `($check_display)`")
     $lines = ($lines | append $"- Rationale: ($c.rationale)")
     $lines = ($lines | append $"")
   }
diff --git a/reflection/modules/backlog.nu b/reflection/modules/backlog.nu
index 066e8ea..2c09914 100644
--- a/reflection/modules/backlog.nu
+++ b/reflection/modules/backlog.nu
@@ -61,7 +61,13 @@ export def "backlog add" [
 
   let items    = (backlog-load)
   let next_num = if ($items | is-empty) { 1 } else {
-    ($items | each { |i| $i.id | str replace "bl-" "" | into int } | sort | last) + 1
+    let nums = ($items.id
+      | each { |id| $id | parse --regex '^bl-(?P<n>[0-9]+)$' }
+      | flatten
+      | get n
+      | each { into int }
+      | sort)
+    if ($nums | is-empty) { 1 } else { ($nums | last) + 1 }
   }
   let new_id = $"bl-($next_num | fill --alignment right --width 3 --character '0')"
   let today  = (date now | format date "%Y-%m-%d")
diff --git a/reflection/modules/describe.nu b/reflection/modules/describe.nu
index 9cb0df0..57b818b 100644
--- a/reflection/modules/describe.nu
+++ b/reflection/modules/describe.nu
@@ -39,7 +39,7 @@ export def nickel-import-path [root: string]: nothing -> string {
   }
 }
 
-use ../modules/store.nu [daemon-export-safe]
+use ../modules/store.nu [daemon-export-safe, daemon-url]
 use ./vcs.nu
 
 # Centralized output dispatcher for all describe commands.
@@ -89,6 +89,129 @@ export def "describe project" [
   emit-output $data $f { || render-project-text $data $a $root }
 }
 
+# ── describe vault ──────────────────────────────────────────────────────────────
+# "What is the credential vault state? Which mode (declarative vs legacy)?
+#  Recipient groups, rules, last audit entry. Mirrors ontoref_vault_status MCP
+#  tool and GraphQL vaultStatus resolver — same data, CLI surface."
+
+export def "describe vault" [
+  --fmt: string = "",
+  --actor: string = "",
+]: nothing -> nothing {
+  let root = (project-root)
+  let a = if ($actor | is-not-empty) { $actor } else { (actor-default) }
+  let f = if ($fmt | is-not-empty) { $fmt } else if $a == "agent" { "json" } else { "text" }
+
+  let project_ncl = ($root | path join ".ontoref" "project.ncl")
+  if not ($project_ncl | path exists) {
+    let data = { has_vault: false, reason: "no .ontoref/project.ncl" }
+    emit-output $data $f { || render-vault-text $data }
+    return
+  }
+
+  let import_path = (nickel-import-path $root)
+  let exported = (do { ^nickel export --import-path $import_path $project_ncl } | complete)
+  if $exported.exit_code != 0 {
+    let data = { has_vault: false, reason: $"nickel export failed: ($exported.stderr | str trim)" }
+    emit-output $data $f { || render-vault-text $data }
+    return
+  }
+  let pj = ($exported.stdout | from json)
+  let sops = ($pj.sops? | default {})
+  let enabled = ($sops.enabled? | default false)
+  if not $enabled {
+    let data = { has_vault: false, reason: "sops disabled in project.ncl" }
+    emit-output $data $f { || render-vault-text $data }
+    return
+  }
+
+  let slug = ($pj.slug? | default "")
+  let vault_id = (($sops.vault_id? | default "") | default $slug)
+  let recipient_rules = ($sops.recipient_rules? | default [])
+  let recipient_groups = ($sops.recipient_groups? | default {})
+  let mode = (if ($recipient_rules | length) > 0 { "declarative" } else { "legacy" })
+
+  let vault_dir = ($env.HOME | path join ".config" "ontoref" "vaults" $vault_id)
+  let access_sops = ($vault_dir | path join "access.sops.yaml")
+  let recipients_in_access = (
+    if ($access_sops | path exists) {
+      open --raw $access_sops | lines | where { |l| $l =~ 'age1[a-z0-9]+' } | length
+    } else { 0 }
+  )
+  let log_path = ($vault_dir | path join "logs" "access.jsonl")
+  let last_op = (
+    if ($log_path | path exists) {
+      let last_line = (open --raw $log_path | lines | last)
+      if ($last_line | is-not-empty) {
+        try { $last_line | from json } catch { {} }
+      } else { {} }
+    } else { {} }
+  )
+
+  let actor_bindings = ($sops.actor_key_bindings? | default {})
+  let group_recipient_count = (
+    $recipient_groups | columns | each { |g|
+      { group: $g, count: (($recipient_groups | get $g) | length) }
+    }
+  )
+  let rules_summary = (
+    $recipient_rules | each { |r|
+      { path: ($r.path? | default ""), groups: ($r.groups? | default []) }
+    }
+  )
+
+  let data = {
+    has_vault: true,
+    vault_id: $vault_id,
+    mode: $mode,
+    registry_endpoint: ($sops.registry_endpoint? | default ""),
+    master_key_path: ($sops.master_key_path? | default ""),
+    actor_key_bindings: $actor_bindings,
+    recipient_groups: $group_recipient_count,
+    recipient_rules: $rules_summary,
+    access_sops_present: ($access_sops | path exists),
+    recipients_in_access: $recipients_in_access,
+    last_op: $last_op,
+  }
+
+  emit-output $data $f { || render-vault-text $data }
+}
+
+def render-vault-text [data: record]: nothing -> nothing {
+  if not $data.has_vault {
+    print $"(ansi yellow)no active vault(ansi reset) — ($data.reason)"
+    return
+  }
+  let mode_color = (if $data.mode == "declarative" { "green" } else { "yellow" })
+  print $"(ansi cyan)vault(ansi reset)  ($data.vault_id)  (ansi $mode_color)[($data.mode)](ansi reset)"
+  if ($data.registry_endpoint | is-not-empty) {
+    print $"  registry        ($data.registry_endpoint)"
+  }
+  if ($data.master_key_path | is-not-empty) {
+    print $"  master_key      ($data.master_key_path)"
+  }
+  print $"  access.sops     ($data.recipients_in_access) recipient\(s\), present=($data.access_sops_present)"
+  if ($data.recipient_groups | length) > 0 {
+    print "  recipient_groups:"
+    for g in $data.recipient_groups {
+      print $"    - ($g.group): ($g.count) key\(s\)"
+    }
+  }
+  if ($data.recipient_rules | length) > 0 {
+    print "  recipient_rules:"
+    for r in $data.recipient_rules {
+      let grps = ($r.groups | str join ", ")
+      print $"    - ($r.path)  →  [($grps)]"
+    }
+  }
+  if ($data.last_op | columns | length) > 0 {
+    let op = ($data.last_op.op? | default "?")
+    let actor = ($data.last_op.actor? | default "?")
+    let ts = ($data.last_op.ts? | default "")
+    print $"  last_audit      ($op) by ($actor) @ ($ts)"
+  }
+}
+
 # ── describe capabilities ───────────────────────────────────────────────────────
 # "What can I DO here? What commands, modes, recipes, tools exist?"
 
@@ -442,7 +565,7 @@ export def "describe impact" [
 
   # When --include-external, query the daemon for cross-project entries
   let external_impacts = if $include_external {
-    let daemon_url = ($env.ONTOREF_DAEMON_URL? | default "http://127.0.0.1:7891")
+    let daemon_url = (daemon-url)
     let result = do {
       http get $"($daemon_url)/graph/impact?node=($node_id)&depth=($depth)&include_external=true"
     } | complete
@@ -1338,7 +1461,7 @@ export def "describe guides" [
   let manifest_critical_deps = ($manifest.critical_deps? | default [])
 
   # Fetch API surface from daemon; empty list if daemon is not reachable.
-  let daemon_url = ($env.ONTOREF_DAEMON_URL? | default "http://127.0.0.1:7891")
+  let daemon_url = (daemon-url)
   let api_surface = do {
     let r = (do { ^curl -sf --max-time 2 $"($daemon_url)/api/catalog" } | complete)
     if $r.exit_code == 0 {
@@ -1460,7 +1583,7 @@ export def "describe api" [
 ]: nothing -> nothing {
   let a = if ($actor | is-not-empty) { $actor } else { (actor-default) }
   let f = if ($fmt | is-not-empty) { $fmt } else if $a == "agent" { "json" } else { "text" }
-  let daemon_url = ($env.ONTOREF_DAEMON_URL? | default "http://127.0.0.1:7891")
+  let daemon_url = (daemon-url)
 
   let result = (do { http get $"($daemon_url)/api/catalog" } | complete)
   if $result.exit_code != 0 {
@@ -1912,7 +2035,32 @@ export def "describe workspace" [
     }
   })
 
-  let data = { crates: $crates_with_ws_deps }
+  # Registry participant namespace derived from manifest.
+  let manifest      = (load-manifest-safe $root)
+  let participant   = ($manifest | get -o project | default ($root | path basename))
+  let reg_provides  = ($manifest | get -o registry_provides | default {})
+  let namespaces    = if ($reg_provides | is-not-empty) {
+    $reg_provides | get -o registries.registries | default []
+    | each { |e| $e | get -o namespaces.own | default [] }
+    | flatten
+  } else {
+    [$"domains/($participant)/", $"modes/($participant)/"]
+  }
+  # Endpoint resolution: manifest declared registry_provides → env override → empty.
+  # When multiple registries declared, take the one matching `default` (or first).
+  let endpoint = if ($reg_provides | is-not-empty) {
+    let entries  = ($reg_provides | get -o registries.registries | default [])
+    let default_id = ($reg_provides | get -o registries.default | default "")
+    let entry = if ($default_id | is-not-empty) {
+      $entries | where { |e| $e.id == $default_id } | first | default null
+    } else {
+      $entries | first | default null
+    }
+    if $entry == null { "" } else { $entry | get -o endpoint | default "" }
+  } else {
+    ($env | get -o PROVISIONING_REGISTRY | default "")
+  }
+  let data = { crates: $crates_with_ws_deps, registry: { participant: $participant, endpoint: $endpoint, namespaces: $namespaces } }
 
   emit-output $data $f { ||
     print $"(ansi white_bold)Workspace Crates(ansi reset)  ($crates_with_ws_deps | length) members"
@@ -1932,6 +2080,15 @@ export def "describe workspace" [
       }
       print ""
     }
+    print $"(ansi white_bold)Registry(ansi reset)"
+    if ($endpoint | is-not-empty) {
+      print $"  endpoint    : ($endpoint)"
+    } else {
+      print $"  endpoint    : (ansi dark_gray)not declared in registry_provides — set in .ontology/manifest.ncl(ansi reset)"
+    }
+    print $"  participant : ($participant)"
+    print $"  namespaces  : ($namespaces | str join '  ')"
+    print ""
   }
 }
 
diff --git a/reflection/modules/env.nu b/reflection/modules/env.nu
index fc95689..b7ef5ff 100644
--- a/reflection/modules/env.nu
+++ b/reflection/modules/env.nu
@@ -27,6 +27,15 @@ export-env {
     $env.NICKEL_IMPORT_PATH = ($paths | str join ":")
   }
 
+  # Always ensure ~/.config/ontoref/schemas/ is on the import path when present.
+  # `just install-daemon` places ontoref-project.ncl there; consumer .ontoref/project.ncl
+  # files import it by bare name. Done unconditionally because the bash wrapper may
+  # have pre-set NICKEL_IMPORT_PATH without including this canonical schemas dir.
+  let schemas_dir = ($env.HOME | path join ".config" "ontoref" "schemas")
+  if ($schemas_dir | path exists) and (not ($env.NICKEL_IMPORT_PATH | str contains $schemas_dir)) {
+    $env.NICKEL_IMPORT_PATH = $"($env.NICKEL_IMPORT_PATH):($schemas_dir)"
+  }
+
   if ($env.ONTOREF_ACTOR? | is-empty) {
     $env.ONTOREF_ACTOR = "developer"
   }
diff --git a/reflection/modules/graph.nu b/reflection/modules/graph.nu
index 2ae3eca..8fce03e 100644
--- a/reflection/modules/graph.nu
+++ b/reflection/modules/graph.nu
@@ -22,9 +22,8 @@ def actor-default []: nothing -> string {
   $env.ONTOREF_ACTOR? | default "developer"
 }
 
-def daemon-url []: nothing -> string {
-  $env.ONTOREF_DAEMON_URL? | default "http://127.0.0.1:7891"
-}
+# daemon-url imported from store.nu — single source of truth (reads
+# ~/.config/ontoref/config.ncl::daemon.port; respects ONTOREF_DAEMON_URL env override).
 
 def daemon-running []: nothing -> bool {
   let r = do { ^curl -sf $"(daemon-url)/health" } | complete
diff --git a/reflection/modules/migrate.nu b/reflection/modules/migrate.nu
index d2f0981..2cb5d4a 100644
--- a/reflection/modules/migrate.nu
+++ b/reflection/modules/migrate.nu
@@ -96,7 +96,21 @@ def run-migration-check [check: record, root: string]: nothing -> record {
 def load-all-migrations []: nothing -> list<record> {
   let dir = (migrations-dir)
   if not ($dir | path exists) { return [] }
-  let ip = (nickel-import-path $env.ONTOREF_ROOT)
+  # Avoid $ONTOREF_ROOT/ontology in NICKEL_IMPORT_PATH — it triggers "Not a directory"
+  # in nickel's directory traversal inside the nu subprocess chain. Subdirs are safe.
+  # See NuCmd branch comment for root cause details.
+  let ip = (
+    [
+      $"($env.ONTOREF_ROOT)/.ontology"
+      $"($env.ONTOREF_ROOT)/adrs"
+      $"($env.ONTOREF_ROOT)/ontology/schemas"
+      $"($env.ONTOREF_ROOT)/ontology/defaults"
+      $env.ONTOREF_ROOT
+    ]
+    | where { |p| $p | path exists }
+    | uniq
+    | str join ":"
+  )
   glob $"($dir)/*.ncl"
   | sort
   | each { |f|
diff --git a/reflection/modules/secrets.nu b/reflection/modules/secrets.nu
new file mode 100644
index 0000000..c1820a3
--- /dev/null
+++ b/reflection/modules/secrets.nu
@@ -0,0 +1,711 @@
+#!/usr/bin/env nu
+# ──────────────────────────────────────────────────────────────────────────────
+# Registry credential resolution for src-vault and OCI artifact ops (ADR-017).
+#
+# Two-layer model — these are different credentials, do not confuse:
+#
+#   ┌─ Layer 0: vault-access ─────────────────────────────────────────────────┐
+#   │ Where:    ~/.config/ontoref/vaults/<vault_id>/access.sops.yaml          │
+#   │ Decrypt:  master .kage from project.ncl::sops.master_key_path           │
+#   │           or global ~/.config/ontoref/config.ncl::vault.master_key_path │
+#   │ Yields:   { zot_username, zot_password, vault_key }                     │
+#   │ Use for:  oras pull src-vault/<project>:latest + restic/kopia open      │
+#   │ Function: resolve-vault-access                                          │
+#   └─────────────────────────────────────────────────────────────────────────┘
+#   ┌─ Layer 2: registry-credential ──────────────────────────────────────────┐
+#   │ Where:    ~/.config/ontoref/vaults/<vault_id>/src-vault/                │
+#   │           <RegistryEntry.credential_sops[_rw]>                          │
+#   │           (multi-recipient encrypted file inside the synced src-vault — │
+#   │           paths in the manifest are relative to the src-vault root)     │
+#   │ Decrypt:  same master .kage as Layer 0                                  │
+#   │ Yields:   { docker_config_dir, audit_event }                            │
+#   │ Use for:  oras push|pull domains/<participant>/<id>:<ver>               │
+#   │ Function: resolve-registry-credential                                   │
+#   └─────────────────────────────────────────────────────────────────────────┘
+#
+# Authorization gate (assert-actor-authorized) is independent: validates the
+# ONTOREF_ACTOR → role binding + scope.ops before either function is called.
+# Callers MUST gate first, resolve second.
+#
+# Canonical caller pattern (e.g. domain_client.nu):
+#
+#     assert-actor-authorized $project_root push
+#     let cred = (resolve-registry-credential $project_root primary --op push)
+#     with-env { DOCKER_CONFIG: $cred.docker_config_dir } {
+#         ^oras push $ref --artifact-type ...
+#     }
+#     rm -rf $cred.docker_config_dir
+#
+# Errors are surfaced via `error make --unspanned` with a `[<code>]` prefix in
+# the message. Catchable via `try { ... } catch { |e| ... }`. Codes:
+#
+#   [actor-bindings-missing]     project.ncl::sops.actor_key_bindings empty
+#   [actor-not-bound]            ONTOREF_ACTOR has no entry in actor_key_bindings
+#   [actor-not-in-bound-actor]   scope.bound_actor non-empty and excludes the actor
+#   [credential-sops-missing]    RegistryEntry has no credential_sops/_rw for op
+#   [invalid-op]                 op is neither 'pull' nor 'push'
+#   [kage-not-resolvable]        master_key_path absent or file missing
+#   [manifest-ncl-missing]       <project>/.ontology/manifest.ncl absent
+#   [op-not-in-scope]            scope.ops does not include the requested op
+#   [project-ncl-missing]        <project>/.ontoref/project.ncl absent
+#   [registry-id-unknown]        registry_id not found in registries[]
+#   [registry-provides-missing]  manifest has no registry_provides
+#   [scope-not-loaded]           src-vault not opened — scopes/<role>.ncl missing
+#   [sops-decrypt-failed]        sops returned non-zero
+#   [sops-file-not-found]        credential_sops path does not exist on disk
+#   [target-not-in-scope]        target ref does not match scope.namespaces globs
+#
+# Daemon never calls these functions — daemon has no .kage, by structural rule
+# (ADR-017 invariant: "Daemon: solo declarativo — nunca toca credenciales").
+# ──────────────────────────────────────────────────────────────────────────────
+
+# ── Internal helpers ──────────────────────────────────────────────────────────
+
+def _project-ncl [project_root: string]: nothing -> string {
+    let p = ($project_root | path join ".ontoref" "project.ncl")
+    if not ($p | path exists) {
+        error make --unspanned { msg: $"[project-ncl-missing] No project.ncl at ($p)" }
+    }
+    $p
+}
+
+def _manifest-ncl [project_root: string]: nothing -> string {
+    let p = ($project_root | path join ".ontology" "manifest.ncl")
+    if not ($p | path exists) {
+        error make --unspanned { msg: $"[manifest-ncl-missing] No manifest.ncl at ($p)" }
+    }
+    $p
+}
+
+def _import-path []: nothing -> string {
+    $env.NICKEL_IMPORT_PATH? | default ($env.PROVISIONING? | default "")
+}
+
+def _eval-ncl [path: string]: nothing -> any {
+    let r = (do { ^nickel export --import-path (_import-path) $path } | complete)
+    if $r.exit_code != 0 {
+        error make --unspanned { msg: $"Failed to evaluate ($path):\n($r.stderr)" }
+    }
+    $r.stdout | from json
+}
+
+def _resolve-master-key-path [project_sops: record]: nothing -> string {
+    let proj_override = ($project_sops | get -o master_key_path | default "")
+    if ($proj_override | is-not-empty) {
+        if not ($proj_override | path exists) {
+            error make --unspanned {
+                msg: $"[kage-not-resolvable] project.ncl::sops.master_key_path = ($proj_override) — file missing"
+            }
+        }
+        return $proj_override
+    }
+    let global_path = ($env.HOME | path join ".config" "ontoref" "config.ncl")
+    if not ($global_path | path exists) {
+        error make --unspanned {
+            msg: "[kage-not-resolvable] master_key_path absent in project.ncl and ~/.config/ontoref/config.ncl missing"
+        }
+    }
+    let global = (_eval-ncl $global_path)
+    let global_kage = ($global | get -o vault.master_key_path | default "")
+    if ($global_kage | is-empty) {
+        error make --unspanned {
+            msg: "[kage-not-resolvable] master_key_path absent both in project.ncl::sops and ~/.config/ontoref/config.ncl::vault"
+        }
+    }
+    if not ($global_kage | path exists) {
+        error make --unspanned {
+            msg: $"[kage-not-resolvable] global master_key_path = ($global_kage) — file missing"
+        }
+    }
+    $global_kage
+}
+
+def _vault-id [project: record]: nothing -> string {
+    let from_sops = ($project | get -o sops.vault_id | default "")
+    if ($from_sops | is-not-empty) { return $from_sops }
+    let slug = ($project | get -o slug | default "")
+    if ($slug | is-empty) {
+        error make --unspanned {
+            msg: "sops.vault_id not declared and project slug absent"
+        }
+    }
+    $slug
+}
+
+def _select-credential-field [op: string]: nothing -> string {
+    match $op {
+        "pull" => "credential_sops",
+        "push" => "credential_sops_rw",
+        _ => {
+            error make --unspanned { msg: $"[invalid-op] op must be 'pull' or 'push', got: ($op)" }
+        }
+    }
+}
+
+# ── Public: build-docker-config-tmpdir ────────────────────────────────────────
+# Allocates an isolated DOCKER_CONFIG directory with a single registry credential.
+# Caller must `rm -rf` the returned path immediately after the oras invocation.
+
+export def build-docker-config-tmpdir [
+    registry: string
+    username: string
+    password: string
+]: nothing -> string {
+    let tmpdir   = (mktemp -d -t ontoref-dockercfg.XXXXXX)
+    let auth_b64 = ($"($username):($password)" | encode base64 | str trim)
+    $'{"auths":{"($registry)":{"auth":"($auth_b64)"}}}'
+        | save ($tmpdir | path join "config.json")
+    $tmpdir
+}
+
+# ── Public: append-vault-access-log ───────────────────────────────────────────
+# Append a structured entry to ~/.config/ontoref/vaults/<vault_id>/logs/access.jsonl.
+
+export def append-vault-access-log [
+    vault_id:    string
+    actor:       string
+    op:          string
+    --detail:      string = ""
+    --registry-id: string = ""
+    --sops-path:   string = ""
+]: nothing -> nothing {
+    let logs_dir = ($env.HOME | path join ".config" "ontoref" "vaults" $vault_id "logs")
+    mkdir $logs_dir
+    let entry = {
+        ts:          (date now | format date "%Y-%m-%dT%H:%M:%SZ"),
+        actor:       $actor,
+        op:          $op,
+        registry_id: $registry_id,
+        sops_path:   $sops_path,
+        detail:      $detail,
+    }
+    (($entry | to json --raw) ++ "\n")
+        | save --append ($logs_dir | path join "access.jsonl")
+}
+
+# ── Public: resolve-vault-access (Layer 0) ────────────────────────────────────
+# Reads ~/.config/ontoref/vaults/<vault_id>/access.sops.yaml and returns the
+# credentials needed to pull src-vault/<vault_id>:latest from ZOT and to open
+# the local restic/kopia repo. Decrypt happens in-memory; nothing is written.
+
+export def resolve-vault-access [
+    project_root: string
+]: nothing -> record {
+    let project = (_eval-ncl (_project-ncl $project_root))
+    let sops_cfg = ($project | get -o sops | default null)
+    if $sops_cfg == null or not ($sops_cfg | get -o enabled | default false) {
+        return { zot_username: "", zot_password: "", vault_key: "" }
+    }
+    let vault_id    = (_vault-id $project)
+    let access_sops = ($env.HOME | path join ".config" "ontoref" "vaults" $vault_id "access.sops.yaml")
+    if not ($access_sops | path exists) {
+        error make --unspanned {
+            msg: $"[scope-not-loaded] access.sops.yaml not found at ($access_sops) — run: ore secrets open ($vault_id)"
+        }
+    }
+    let kage = (_resolve-master-key-path $sops_cfg)
+    let r = (with-env { SOPS_AGE_KEY_FILE: $kage } {
+        do { ^sops --decrypt $access_sops } | complete
+    })
+    if $r.exit_code != 0 {
+        error make --unspanned {
+            msg: $"[sops-decrypt-failed] Layer 0 access.sops.yaml decrypt failed:\n($r.stderr)"
+        }
+    }
+    let creds = ($r.stdout | from yaml)
+    {
+        zot_username: ($creds | get -o zot_username | default ""),
+        zot_password: ($creds | get -o zot_password | default ""),
+        vault_key:    ($creds | get -o vault_key    | default ""),
+    }
+}
+
+# ── Public: resolve-registry-credential (Layer 2) ─────────────────────────────
+# Looks up RegistryEntry by id in manifest.registry_provides.registries.registries,
+# selects credential_sops (op=pull) or credential_sops_rw (op=push), decrypts it
+# with the master .kage, and returns an isolated DOCKER_CONFIG tmpdir.
+# Caller MUST rm -rf the returned docker_config_dir after the oras call.
+#
+# This function does NOT validate actor authorization — call assert-actor-authorized
+# beforehand. It is a no-op to call this function as agent or ci if scope allows.
+
+export def resolve-registry-credential [
+    project_root: string
+    registry_id:  string   # "" = use registries.default from manifest
+    --op:         string = "pull"
+]: nothing -> record {
+    let field = (_select-credential-field $op)
+
+    let manifest = (_eval-ncl (_manifest-ncl $project_root))
+    let provides = ($manifest | get -o registry_provides | default null)
+    if $provides == null {
+        error make --unspanned {
+            msg: $"[registry-provides-missing] .ontology/manifest.ncl has no registry_provides block"
+        }
+    }
+    let entries = ($provides | get -o registries.registries | default [])
+    let actual_id = if ($registry_id | is-empty) {
+        let from_manifest = ($provides | get -o registries.default | default "")
+        if ($from_manifest | is-empty) {
+            error make --unspanned {
+                msg: $"[registry-id-unknown] registry_id empty and registries.default not declared in manifest"
+            }
+        }
+        $from_manifest
+    } else { $registry_id }
+    let entry = ($entries | where { |e| $e.id == $actual_id } | first | default null)
+    if $entry == null {
+        let known = ($entries | each { |e| $e.id } | str join ", ")
+        error make --unspanned {
+            msg: $"[registry-id-unknown] no RegistryEntry with id=($actual_id). Known: [($known)]"
+        }
+    }
+
+    let sops_rel = ($entry | get -o $field | default "")
+    if ($sops_rel | is-empty) {
+        error make --unspanned {
+            msg: $"[credential-sops-missing] RegistryEntry id=($registry_id) has no ($field) for op=($op)"
+        }
+    }
+    let project  = (_eval-ncl (_project-ncl $project_root))
+    let sops_cfg = ($project | get -o sops | default {})
+    let vault_id = (_vault-id $project)
+    let vault_root = ($env.HOME | path join ".config" "ontoref" "vaults" $vault_id "src-vault")
+    let sops_path  = ($vault_root | path join $sops_rel)
+    if not ($sops_path | path exists) {
+        let hint = $"vault may not be opened — try: ore secrets open ($vault_id)"
+        error make --unspanned {
+            msg: $"[sops-file-not-found] ($sops_path) — ($hint)"
+        }
+    }
+    let kage = (_resolve-master-key-path $sops_cfg)
+
+    let r = (with-env { SOPS_AGE_KEY_FILE: $kage } {
+        do { ^sops --decrypt --output-type json $sops_path } | complete
+    })
+    if $r.exit_code != 0 {
+        error make --unspanned {
+            msg: $"[sops-decrypt-failed] ($sops_path) decrypt failed:\n($r.stderr)"
+        }
+    }
+    let creds  = ($r.stdout | from json)
+    let user   = ($creds | get -o username | default "")
+    let pass   = ($creds | get -o password | default "")
+    let endpoint = ($entry.endpoint)
+
+    let tmpdir = (build-docker-config-tmpdir $endpoint $user $pass)
+    let actor  = ($env.ONTOREF_ACTOR? | default "developer")
+    let event  = {
+        ts:          (date now | format date "%Y-%m-%dT%H:%M:%SZ"),
+        actor:       $actor,
+        op:          $op,
+        registry_id: $actual_id,
+        endpoint:    $endpoint,
+        sops_path:   $sops_rel,
+    }
+
+    append-vault-access-log $vault_id $actor $op --registry-id $actual_id --sops-path $sops_rel
+
+    { docker_config_dir: $tmpdir, audit_event: $event }
+}
+
+# ── Public: assert-actor-authorized ───────────────────────────────────────────
+# Reads project.ncl::sops.actor_key_bindings to map ONTOREF_ACTOR → role, then
+# loads ~/.config/ontoref/vaults/<vault_id>/src-vault/scopes/<role>.ncl and
+# verifies the requested op is in scope.ops. Errors are nominal so callers can
+# distinguish authorization failures from credential-resolution failures.
+
+export def assert-actor-authorized [
+    project_root: string
+    op:           string
+]: nothing -> nothing {
+    let _ = (_select-credential-field $op)   # validates op early
+
+    let project  = (_eval-ncl (_project-ncl $project_root))
+    let sops_cfg = ($project | get -o sops | default null)
+    if $sops_cfg == null or not ($sops_cfg | get -o enabled | default false) {
+        return null   # sops disabled = no enforcement (consistent with resolve-vault-access)
+    }
+
+    let bindings = ($sops_cfg | get -o actor_key_bindings | default {})
+    if ($bindings | is-empty) {
+        error make --unspanned {
+            msg: "[actor-bindings-missing] project.ncl::sops.actor_key_bindings is empty"
+        }
+    }
+    let actor = ($env.ONTOREF_ACTOR? | default "developer")
+    let role  = ($bindings | get -o $actor | default "")
+    if ($role | is-empty) {
+        let known = ($bindings | columns | str join ", ")
+        error make --unspanned {
+            msg: $"[actor-not-bound] ONTOREF_ACTOR=($actor) has no entry in actor_key_bindings. Known actors: [($known)]"
+        }
+    }
+
+    let vault_id   = (_vault-id $project)
+    let scope_path = ($env.HOME
+        | path join ".config" "ontoref" "vaults" $vault_id "src-vault" "scopes" $"($role).ncl")
+    if not ($scope_path | path exists) {
+        error make --unspanned {
+            msg: $"[scope-not-loaded] ($scope_path) missing — vault not opened. Run: ore secrets open ($vault_id)"
+        }
+    }
+    let scope = (_eval-ncl $scope_path)
+
+    # bound_actor: when non-empty, only listed actors may assume this role.
+    # Empty list = no restriction (the actor_key_bindings mapping is the only gate).
+    let bound = ($scope | get -o bound_actor | default [])
+    if ($bound | is-not-empty) and ($actor not-in $bound) {
+        error make --unspanned {
+            msg: $"[actor-not-in-bound-actor] role=($role) bound_actor=($bound) does not include ONTOREF_ACTOR=($actor)"
+        }
+    }
+
+    # ops: the operation must be permitted by the role.
+    let ops = ($scope | get -o ops | default [])
+    let allowed = ($ops | any { |o|
+        let s = ($o | describe)
+        if $s == "string" { $o == $op } else { ($o | to text) == $op }
+    })
+    if not $allowed {
+        error make --unspanned {
+            msg: $"[op-not-in-scope] role=($role) does not allow op=($op). scope.ops=($ops)"
+        }
+    }
+    null
+}
+
+# ── Public: assert-target-in-scope ────────────────────────────────────────────
+# Validate that an OCI artifact target (e.g. "domains/libre-wuji/zot") is
+# permitted by the current actor's scope.namespaces glob list. Caller must invoke
+# this AFTER assert-actor-authorized and BEFORE the oras call. Defense-in-depth
+# layer-1 (CLI pre-validation) per ADR-017 — registry ACL is the layer-2 enforcement.
+
+export def assert-target-in-scope [
+    project_root: string
+    target:       string   # e.g. "domains/libre-wuji/zot" or "modes/foo/bar"
+]: nothing -> nothing {
+    let project  = (_eval-ncl (_project-ncl $project_root))
+    let sops_cfg = ($project | get -o sops | default null)
+    if $sops_cfg == null or not ($sops_cfg | get -o enabled | default false) {
+        return null
+    }
+    let bindings = ($sops_cfg | get -o actor_key_bindings | default {})
+    let actor    = ($env.ONTOREF_ACTOR? | default "developer")
+    let role     = ($bindings | get -o $actor | default "")
+    if ($role | is-empty) { return null }   # assert-actor-authorized errors first
+
+    let vault_id   = (_vault-id $project)
+    let scope_path = ($env.HOME
+        | path join ".config" "ontoref" "vaults" $vault_id "src-vault" "scopes" $"($role).ncl")
+    if not ($scope_path | path exists) { return null }
+
+    let scope      = (_eval-ncl $scope_path)
+    let namespaces = ($scope | get -o namespaces | default [])
+    if ($namespaces | is-empty) { return null }   # no restrictions declared
+
+    # Each entry like "domains/*/" or "modes/libre-wuji/" — match by prefix-after-glob.
+    let matched = ($namespaces | any { |ns|
+        let normalized = ($ns | str trim --right --char "/")
+        if ($normalized | str contains "*") {
+            let prefix = ($normalized | str replace --regex '\*.*' '')
+            $target | str starts-with $prefix
+        } else {
+            $target | str starts-with $normalized
+        }
+    })
+    if not $matched {
+        error make --unspanned {
+            msg: $"[target-not-in-scope] role=($role) namespaces=($namespaces) does not allow target=($target)"
+        }
+    }
+    null
+}
+
+# ── CLI dispatch — `ore secrets X` ────────────────────────────────────────────
+# Each command shells out to the corresponding bash recipe in justfiles/secrets.just,
+# propagating ONTOREF_PROJECT_ROOT so the recipe targets the consumer project, not
+# ontoref's own checkout. The recipes carry the reference orchestration logic
+# (project.ncl read, sops cipher/decipher, oras + cosign + restic invocation).
+
+def _project-root-cli []: nothing -> string {
+    let from_env = ($env.ONTOREF_PROJECT_ROOT? | default "")
+    if ($from_env | is-not-empty) { return $from_env }
+    pwd
+}
+
+def _run-secrets-recipe [recipe: string, ...args: string]: nothing -> nothing {
+    let ontoref_root = ($env.ONTOREF_ROOT? | default "")
+    if ($ontoref_root | is-empty) {
+        error make --unspanned {
+            msg: "ONTOREF_ROOT not set — invoke via ./ontoref entry point, not nu directly"
+        }
+    }
+    let recipe_file = ($ontoref_root | path join "justfiles" "secrets.just")
+    if not ($recipe_file | path exists) {
+        error make --unspanned { msg: $"secrets.just not found at ($recipe_file)" }
+    }
+    let project_root = (_project-root-cli)
+    with-env { ONTOREF_PROJECT_ROOT: $project_root } {
+        ^just -f $recipe_file $recipe ...$args
+    }
+}
+
+# Bootstrap a new vault for the active project (admin only). Creates access.sops.yaml,
+# initializes the local restic/kopia repo. Push to ZOT happens via `secrets push`.
+export def "secrets bootstrap" []: nothing -> nothing {
+    _run-secrets-recipe "secrets-bootstrap"
+}
+
+# Pull latest src-vault/<vault_id>:latest from ZOT and restore locally.
+export def "secrets sync" []: nothing -> nothing {
+    _run-secrets-recipe "secrets-sync"
+}
+
+# Push current vault state to ZOT — signs with cosign (mandatory per ADR-017).
+export def "secrets push" []: nothing -> nothing {
+    _run-secrets-recipe "secrets-push"
+}
+
+# Open the vault for editing — runs sops on access.sops.yaml.
+export def "secrets open" []: nothing -> nothing {
+    _run-secrets-recipe "secrets-open"
+}
+
+# Close the vault — pushes updated state and releases the edit context.
+export def "secrets close" []: nothing -> nothing {
+    _run-secrets-recipe "secrets-close"
+}
+
+# Force-release an abandoned or expired lock (admin operation, audited).
+export def "secrets force-unlock" []: nothing -> nothing {
+    _run-secrets-recipe "secrets-force-unlock"
+}
+
+# Add or update cosign_password inside access.sops.yaml without opening the
+# editor. Use to upgrade vaults bootstrapped before the cosign_password field
+# existed, so 'secrets push' runs non-interactively.
+export def "secrets set-cosign-password" []: nothing -> nothing {
+    _run-secrets-recipe "secrets-set-cosign-password"
+}
+
+# Recover a corrupted/missing access.sops.yaml from the ZOT artifact.
+# Caller provides credentials via RECOVERY_ZOT_USER + RECOVERY_ZOT_PASS env
+# vars or a pre-built DOCKER_CONFIG. See FAQ::credential-vault-disaster-recovery.
+export def "secrets recover" []: nothing -> nothing {
+    _run-secrets-recipe "secrets-recover"
+}
+
+# Add a recipient age public key to access.sops.yaml.
+# Use after a new developer/role is added to the project.
+export def "secrets add-key" [pubkey: string, role: string]: nothing -> nothing {
+    _run-secrets-recipe "secrets-add-key" $pubkey $role
+}
+
+# Remove a recipient age public key (revocation). The token in the registry should
+# be rotated separately — sops removal does not invalidate already-known plaintext.
+export def "secrets remove-key" [pubkey: string]: nothing -> nothing {
+    _run-secrets-recipe "secrets-remove-key" $pubkey
+}
+
+# Re-encrypt all *.sops.yaml in the vault with the current .sops.yaml recipient list.
+export def "secrets rekey" []: nothing -> nothing {
+    _run-secrets-recipe "secrets-rekey"
+}
+
+# Generate a new age keypair for a role (prints private key to stdout — never stored on disk).
+export def "secrets gen-key" [role: string]: nothing -> nothing {
+    _run-secrets-recipe "secrets-gen-key" $role
+}
+
+# Run vault constraints audit. Default: all checks. Specific:
+#   bootstrap-credentials  decrypts access.sops.yaml with master_key
+#   no-credential-env      manifest does not declare credential_env (banned by ADR-017)
+#   recipients             access.sops.yaml has ≥ 2 recipients (multi-recipient mandatory)
+export def "secrets audit" [--check: string = ""]: nothing -> nothing {
+    if ($check | is-empty) {
+        _run-secrets-recipe "secrets-audit"
+    } else {
+        _run-secrets-recipe "secrets-audit" $"check=($check)"
+    }
+}
+
+# Show vault state for the active project — declared sops config, vault dir presence,
+# decrypt sanity, last access log entry.
+export def "secrets status" []: nothing -> nothing {
+    let project_root = (_project-root-cli)
+    let project_ncl  = ($project_root | path join ".ontoref" "project.ncl")
+    if not ($project_ncl | path exists) {
+        print $"  (ansi yellow)No .ontoref/project.ncl at ($project_root)(ansi reset)"
+        return
+    }
+    let project  = (_eval-ncl $project_ncl)
+    let sops_cfg = ($project | get -o sops | default null)
+    if $sops_cfg == null or not ($sops_cfg | get -o enabled | default false) {
+        print "  sops: disabled in project.ncl"
+        return
+    }
+    let vault_id = (_vault-id $project)
+    let vault_dir = ($env.HOME | path join ".config" "ontoref" "vaults" $vault_id)
+    let access_sops = ($vault_dir | path join "access.sops.yaml")
+    let logs_jsonl  = ($vault_dir | path join "logs" "access.jsonl")
+
+    print $"  (ansi white_bold)Vault(ansi reset)         ($vault_id)"
+    print $"  (ansi dark_gray)backend(ansi reset)       ($sops_cfg | get -o vault_backend | default 'restic')"
+    print $"  (ansi dark_gray)vault dir(ansi reset)     ($vault_dir)"
+    print $"  (ansi dark_gray)access.sops(ansi reset)   (if ($access_sops | path exists) { 'present' } else { 'MISSING — run: ore secrets sync' })"
+    print $"  (ansi dark_gray)master_key(ansi reset)    (try { _resolve-master-key-path $sops_cfg } catch { |e| $e.msg })"
+
+    # Cosign keypair state — read from global config, no env-var fallback (ADR-017 declarative)
+    let global_cfg = ($env.HOME | path join ".config" "ontoref" "config.ncl")
+    let cosign = if ($global_cfg | path exists) {
+        try { _eval-ncl $global_cfg | get -o vault.cosign | default {} } catch { {} }
+    } else { {} }
+    let key_path = ($cosign | get -o key_path | default "")
+    let pub_path = ($cosign | get -o pub_path | default "")
+    let tlog     = ($cosign | get -o tlog | default false)
+    let key_state = if ($key_path | is-empty) {
+        "(ansi yellow)not declared(ansi reset)"
+    } else if ($key_path | path exists) { "present" } else { $"(ansi red)missing(ansi reset) — ($key_path)" }
+    let pub_state = if ($pub_path | is-empty) {
+        "(ansi yellow)not declared(ansi reset)"
+    } else if ($pub_path | path exists) { "present" } else { $"(ansi red)missing(ansi reset) — ($pub_path)" }
+    let tlog_state = if $tlog { "(ansi cyan)Rekor public log(ansi reset)" } else { "private (no Rekor)" }
+    print $"  (ansi dark_gray)cosign key(ansi reset)    ($key_state)"
+    print $"  (ansi dark_gray)cosign pub(ansi reset)    ($pub_state)"
+    print $"  (ansi dark_gray)cosign tlog(ansi reset)   ($tlog_state)"
+
+    if ($logs_jsonl | path exists) {
+        let last  = (open --raw $logs_jsonl | lines | last | from json)
+        let reg   = ($last | get -o registry_id | default "")
+        let extra = if ($reg | is-empty) { "" } else { $" — registry=($reg)" }
+        print $"  (ansi dark_gray)last op(ansi reset)       ($last.op) by ($last.actor) @ ($last.ts)($extra)"
+    } else {
+        print "  last op       (no entries)"
+    }
+}
+
+# Detailed inventory of the active project's vault: recipients, roles, declared
+# registry credentials, last operations. Read-only; no decrypt of sensitive values.
+export def "secrets describe" []: nothing -> nothing {
+    let project_root = (_project-root-cli)
+    let project_ncl  = ($project_root | path join ".ontoref" "project.ncl")
+    if not ($project_ncl | path exists) {
+        print $"  (ansi yellow)No .ontoref/project.ncl at ($project_root)(ansi reset)"
+        return
+    }
+    let project  = (_eval-ncl $project_ncl)
+    let sops_cfg = ($project | get -o sops | default null)
+    if $sops_cfg == null or not ($sops_cfg | get -o enabled | default false) {
+        print "  sops: disabled in project.ncl"
+        return
+    }
+    let vault_id  = (_vault-id $project)
+    let vault_dir = ($env.HOME | path join ".config" "ontoref" "vaults" $vault_id)
+    let src_vault = ($vault_dir | path join "src-vault")
+
+    print ""
+    print $"(ansi white_bold)Vault(ansi reset)  ($vault_id)"
+    print $"  project_root  ($project_root)"
+    print $"  vault_dir     ($vault_dir)"
+    print $"  backend       ($sops_cfg | get -o vault_backend | default 'restic')"
+    print $"  endpoint      ($sops_cfg | get -o registry_endpoint | default '(not declared)')"
+    print ""
+
+    # Per-file recipient routing (declarative model, when present).
+    let groups = ($sops_cfg | get -o recipient_groups | default {})
+    let rules  = ($sops_cfg | get -o recipient_rules | default [])
+    if ($rules | is-not-empty) {
+        print $"  (ansi white_bold)Recipient groups(ansi reset)"
+        for g in ($groups | columns) {
+            let keys = ($groups | get $g)
+            let count = ($keys | length)
+            print $"    ($g)  ($count) key\(s\)"
+            for k in $keys {
+                print $"      ($k)"
+            }
+        }
+        print ""
+        print $"  (ansi white_bold)Path rules(ansi reset)  ($rules | length) total"
+        for r in $rules {
+            let merged = ($r.groups | each { |g| $groups | get -o $g | default [] } | flatten | uniq)
+            let n = ($merged | length)
+            let glist = ($r.groups | str join " + ")
+            print $"    ($r.path)"
+            print $"      groups: ($glist)  →  ($n) recipient\(s\)"
+        }
+        print ""
+    } else {
+        # Legacy mode — count age1... in access.sops.yaml metadata.
+        let access_sops = ($vault_dir | path join "access.sops.yaml")
+        if ($access_sops | path exists) {
+            let recipients = (open --raw $access_sops | from yaml
+                | get -o sops.age | default []
+                | each { |r| $r.recipient? | default "" }
+                | where { |r| $r | str starts-with "age1" })
+            let nrec = ($recipients | length)
+            print $"  (ansi white_bold)Recipients(ansi reset)  ($nrec) — legacy single-set mode"
+            for r in $recipients {
+                print $"    ($r)"
+            }
+        } else {
+            print $"  (ansi yellow)access.sops.yaml absent(ansi reset) — run: ore secrets sync"
+        }
+        print ""
+    }
+
+    # actor_key_bindings table.
+    let bindings = ($sops_cfg | get -o actor_key_bindings | default {})
+    if ($bindings | columns | is-not-empty) {
+        print $"  (ansi white_bold)Actor → role(ansi reset)"
+        for actor in ($bindings | columns) {
+            let role = ($bindings | get $actor)
+            print $"    ($actor)  →  ($role)"
+        }
+        print ""
+    }
+
+    # Roles (scopes/) with their access mode.
+    if ($src_vault | path join "scopes" | path exists) {
+        let scope_files = (ls ($src_vault | path join "scopes") | where name =~ '\.ncl$' | get name)
+        let nfiles = ($scope_files | length)
+        print $"  (ansi white_bold)Roles(ansi reset)  ($nfiles) scope file\(s\)"
+        for sf in $scope_files {
+            let role     = ($sf | path basename | str replace ".ncl" "")
+            let s        = (try { _eval-ncl $sf } catch { {} })
+            let access   = ($s | get -o access | default "(none)")
+            let ops      = ($s | get -o ops | default [] | str join ", ")
+            let bound    = ($s | get -o bound_actor | default [] | str join ", ")
+            let bound_disp = if ($bound | is-empty) { "(unbound)" } else { $bound }
+            print $"    ($role)  access=($access)  bound=($bound_disp)"
+            print $"             ops: ($ops)"
+        }
+        print ""
+    } else {
+        print $"  (ansi yellow)scopes/ absent(ansi reset) — vault not bootstrapped or not synced"
+        print ""
+    }
+
+    # Registry credential files.
+    if ($src_vault | path join "registry" | path exists) {
+        let creds = (ls ($src_vault | path join "registry") | where name =~ '\.sops\.yaml$' | get name)
+        print $"  (ansi white_bold)Registry credentials(ansi reset)  ($creds | length)"
+        for c in $creds {
+            print $"    ($c | path basename)"
+        }
+        print ""
+    }
+
+    # Recent operations from access.jsonl.
+    let logs_jsonl = ($vault_dir | path join "logs" "access.jsonl")
+    if ($logs_jsonl | path exists) {
+        let entries = (open --raw $logs_jsonl | lines | last 5 | each { |l| $l | from json })
+        print $"  (ansi white_bold)Last operations(ansi reset)  ($entries | length) most recent"
+        for e in $entries {
+            let reg = ($e | get -o registry_id | default "")
+            let r   = if ($reg | is-empty) { "" } else { $" registry=($reg)" }
+            print $"    ($e.ts)  ($e.actor)  ($e.op)($r)"
+        }
+        print ""
+    }
+}
diff --git a/reflection/modules/services.nu b/reflection/modules/services.nu
index 3520859..6c78070 100644
--- a/reflection/modules/services.nu
+++ b/reflection/modules/services.nu
@@ -131,7 +131,7 @@ export def "services overview" []: nothing -> nothing {
     let status_icon = if ($status_text == "running") { "✓" } else { "✗" }
 
     if $service.enabled {
-      let port = ($config.daemon.port? | default 7891)
+      let port = ($config.daemon.port? | default 7890)
       let status_color = if ($status_text == "running") { $green } else { $red }
       let line = $blue + "  " + $service.id + $reset + " [" + $status_color + $status_icon + $reset + "] " + $status_text + " - port " + $yellow + ($port | into string) + $reset
       print $line
@@ -315,7 +315,7 @@ def daemon-health []: nothing -> nothing {
     return
   }
 
-  let url = ($env.ONTOREF_DAEMON_URL? | default "http://127.0.0.1:7891")
+  let url = (daemon-url)
   let result = (do { ^curl -sf $"($url)/health" } | complete)
 
   if $result.exit_code != 0 {
diff --git a/reflection/modules/store.nu b/reflection/modules/store.nu
index a39542d..18f4549 100644
--- a/reflection/modules/store.nu
+++ b/reflection/modules/store.nu
@@ -47,8 +47,29 @@ def nickel-import-path [root: string]: nothing -> string {
 
 # ── Configuration ────────────────────────────────────────────────────────────────
 
-def daemon-url []: nothing -> string {
-  $env.ONTOREF_DAEMON_URL? | default "http://127.0.0.1:7891"
+# Resolve the daemon HTTP port. Reads ~/.config/ontoref/config.ncl::daemon.port
+# as the canonical source — same value the daemon binary reads at startup —
+# so client and daemon never drift on the port. ONTOREF_DAEMON_PORT env override
+# takes precedence for ad-hoc redirection. Final fallback is the schema default
+# (matches install/resources/config.ncl).
+export def daemon-port []: nothing -> int {
+  let env_port = ($env.ONTOREF_DAEMON_PORT? | default "")
+  if ($env_port | is-not-empty) { return ($env_port | into int) }
+  let cfg_path = $"($env.HOME)/.config/ontoref/config.ncl"
+  if not ($cfg_path | path exists) { return 7890 }
+  let exported = (do { ^nickel export $cfg_path } | complete)
+  if $exported.exit_code != 0 { return 7890 }
+  let parsed = (try { $exported.stdout | from json } catch { {} })
+  ($parsed.daemon?.port? | default 7890)
+}
+
+# Resolve the daemon URL. ONTOREF_DAEMON_URL env override takes precedence;
+# otherwise composes from the canonical port (see daemon-port). Use this from
+# every nu module that talks to the daemon — never hard-code 127.0.0.1:NNNN.
+export def daemon-url []: nothing -> string {
+  let env_url = ($env.ONTOREF_DAEMON_URL? | default "")
+  if ($env_url | is-not-empty) { return $env_url }
+  $"http://127.0.0.1:(daemon-port)"
 }
 
 # Load project config to check if DB is enabled
diff --git a/reflection/modules/sync.nu b/reflection/modules/sync.nu
index 1b661d8..c991726 100644
--- a/reflection/modules/sync.nu
+++ b/reflection/modules/sync.nu
@@ -4,7 +4,7 @@
 # Scans: crates, scenarios, agents, CI, forms, modes, justfiles, .claude/.
 # Operates on $env.ONTOREF_PROJECT_ROOT (set by onref wrapper).
 
-use store.nu [daemon-export, daemon-export-safe]
+use store.nu [daemon-export, daemon-export-safe, daemon-url]
 
 # Resolve project root: ONTOREF_PROJECT_ROOT if set and different from ONTOREF_ROOT, else ONTOREF_ROOT.
 def project-root []: nothing -> string {
@@ -180,7 +180,7 @@ export def "sync diff" [
       let conn_result = do { nickel export --format json $conn_file } | complete
       if $conn_result.exit_code == 0 {
         let connections = ($conn_result.stdout | from json)
-        let daemon_url = ($env.ONTOREF_DAEMON_URL? | default "http://127.0.0.1:7891")
+        let daemon_url = (daemon-url)
         let projects_result = do { http get $"($daemon_url)/projects" } | complete
         if $projects_result.exit_code == 0 {
           let registered = ($projects_result.stdout | from json | get slugs? | default [])
@@ -1701,3 +1701,158 @@ def audit-gates [root: string, diff: list]: nothing -> list {
     }
   }
 }
+
+# ── sync protocol-files ─────────────────────────────────────────────────────────
+# Refresh a consumer project's local copy of the ontoref protocol surface
+# (`ontology/{schemas,defaults}/`) from the installed data directory.
+#
+# Each consumer project keeps a local snapshot of the schema/defaults files so
+# Nickel imports resolve relative to the project root. When ontoref evolves
+# (new constructors, schema fields, etc.) those snapshots drift and exports
+# fail with `missing field` or `could not find import`. This command compares
+# the project's snapshot to the installed source and reports / applies updates.
+#
+# Targets are auto-detected: the command looks for `ontology/` and
+# `.ontology/` under the project root and refreshes whichever exists.
+#
+# Default is dry-run (report only). Pass --apply to copy.
+
+export def "sync protocol-files" [
+  --apply,            # Apply changes (default: dry-run report only).
+  --fmt: string = "", # Output format: text* | json
+  --actor: string = "",
+]: nothing -> nothing {
+  let root = (project-root)
+  let a = if ($actor | is-not-empty) { $actor } else { "developer" }
+  let f = if ($fmt | is-not-empty) { $fmt } else { "text" }
+
+  let src_root = ($env.ONTOREF_ROOT? | default "")
+  if ($src_root | is-empty) or (not (($src_root | path join "ontology") | path exists)) {
+    error make { msg: $"[protocol-files] ONTOREF_ROOT not set or has no ontology/ — cannot determine source" }
+  }
+
+  # Detect target ontology dirs in the consumer project.
+  let candidates = (
+    ["ontology", ".ontology"]
+      | each { |d| { local: $d, path: ($root | path join $d) } }
+      | where { |c| $c.path | path exists }
+  )
+  if ($candidates | is-empty) {
+    let data = { dirs: [], summary: { added: 0, updated: 0, skipped: 0, applied: $apply } }
+    if $f == "json" { print ($data | to json) } else {
+      print $"[protocol-files] no ontology/ or .ontology/ under ($root) — nothing to refresh"
+    }
+    return
+  }
+
+  # Walk each subdir (schemas, defaults). Build per-file plan.
+  let subdirs = ["schemas", "defaults"]
+  let plans = ($candidates | each { |c|
+    let target_base = $c.path
+    let entries = ($subdirs | each { |sd|
+      let src_dir = ($src_root | path join "ontology" $sd)
+      let dst_dir = ($target_base | path join $sd)
+      if not ($src_dir | path exists) { return [] }
+      if not ($dst_dir | path exists) {
+        # Whole subdir absent — every source file is an ADD candidate, but the
+        # consumer may have intentionally not adopted this layer; skip with note.
+        return [{
+          local: $c.local,
+          subdir: $sd,
+          file: "(directory)",
+          status: "MISSING_SUBDIR",
+          src: $src_dir,
+          dst: $dst_dir,
+          src_mtime: "",
+          dst_mtime: "",
+        }]
+      }
+      let src_files = (glob $"($src_dir)/*.ncl")
+      $src_files | each { |sf|
+        let rel = ($sf | path basename)
+        let df = ($dst_dir | path join $rel)
+        let src_mtime = (ls $sf | get 0.modified)
+        let entry_base = {
+          local: $c.local,
+          subdir: $sd,
+          file: $rel,
+          src: $sf,
+          dst: $df,
+          src_mtime: ($src_mtime | format date "%Y-%m-%dT%H:%M:%SZ"),
+        }
+        if not ($df | path exists) {
+          $entry_base | merge { status: "ADD", dst_mtime: "" }
+        } else {
+          let dst_mtime = (ls $df | get 0.modified)
+          let dst_iso = ($dst_mtime | format date "%Y-%m-%dT%H:%M:%SZ")
+          if $src_mtime > $dst_mtime {
+            $entry_base | merge { status: "UPDATE", dst_mtime: $dst_iso }
+          } else {
+            $entry_base | merge { status: "SKIP", dst_mtime: $dst_iso }
+          }
+        }
+      }
+    } | flatten)
+    { local: $c.local, root: $target_base, entries: $entries }
+  })
+
+  # Apply if requested.
+  let applied_plans = if $apply {
+    $plans | each { |p|
+      let entries = ($p.entries | each { |e|
+        if $e.status == "ADD" or $e.status == "UPDATE" {
+          let dir = ($e.dst | path dirname)
+          if not ($dir | path exists) { mkdir $dir }
+          cp $e.src $e.dst
+          $e | merge { applied: true }
+        } else {
+          $e | merge { applied: false }
+        }
+      })
+      { local: $p.local, root: $p.root, entries: $entries }
+    }
+  } else { $plans }
+
+  let all_entries = ($applied_plans | each { |p| $p.entries } | flatten)
+  let summary = {
+    added:    ($all_entries | where status == "ADD"     | length),
+    updated:  ($all_entries | where status == "UPDATE"  | length),
+    skipped:  ($all_entries | where status == "SKIP"    | length),
+    missing_subdirs: ($all_entries | where status == "MISSING_SUBDIR" | length),
+    applied:  $apply,
+  }
+
+  let data = { dirs: $applied_plans, summary: $summary, source: ($src_root | path join "ontology") }
+
+  if $f == "json" {
+    print ($data | to json)
+    return
+  }
+
+  # Text rendering.
+  let mode = if $apply { "APPLY" } else { "DRY-RUN" }
+  print $"(ansi cyan)protocol-files(ansi reset) ($mode) — source: ($data.source)"
+  for p in $applied_plans {
+    print $"  (ansi white_bold)($p.local)/(ansi reset)  ($p.root)"
+    for e in $p.entries {
+      let tag = (
+        if   $e.status == "ADD"            { $"(ansi green)+ ADD   (ansi reset)" }
+        else if $e.status == "UPDATE"      { $"(ansi yellow)~ UPDATE(ansi reset)" }
+        else if $e.status == "SKIP"        { $"(ansi dark_gray). skip  (ansi reset)" }
+        else if $e.status == "MISSING_SUBDIR" { $"(ansi red)? missing(ansi reset)" }
+        else { $e.status }
+      )
+      let detail = if $e.status == "MISSING_SUBDIR" {
+        $"      ($e.subdir)/ — not present in consumer; not creating"
+      } else {
+        $"      ($e.subdir)/($e.file)  src=($e.src_mtime)  dst=($e.dst_mtime)"
+      }
+      print $"    ($tag) ($detail)"
+    }
+  }
+  print ""
+  print $"  summary: added=($summary.added) updated=($summary.updated) skipped=($summary.skipped) missing_subdirs=($summary.missing_subdirs) applied=($summary.applied)"
+  if not $apply and (($summary.added + $summary.updated) > 0) {
+    print $"  → run with --apply to write changes"
+  }
+}
diff --git a/reflection/modules/validate.nu b/reflection/modules/validate.nu
index f1cac65..840f414 100644
--- a/reflection/modules/validate.nu
+++ b/reflection/modules/validate.nu
@@ -10,6 +10,8 @@
 #   validate check-adr <id>       — run all constraints for one ADR
 #   validate check-all            — run all constraints across all accepted ADRs
 #   validate justfile             — validate project justfile against convention schema
+#   validate modes [--check]      — ADR-018 level hierarchy + strategy compliance
+#   mode resolve <id>             — show which hierarchy level handles a mode
 #
 # Error handling: do { ... } | complete — never panics, always returns a result.
 
@@ -52,21 +54,39 @@ def run-grep [check: record]: nothing -> record {
   if $check.must_be_empty {
     {
       passed: (not $has_matches),
-      detail: (if $has_matches { $"Pattern found (violation): ($result.stdout | str trim)" } else { "Pattern absent — ok" })
+      detail: (if $has_matches { $"Pattern found, violation: ($result.stdout | str trim)" } else { "Pattern absent — ok" })
     }
   } else {
     {
       passed: $has_matches,
-      detail: (if $has_matches { "Pattern present — ok" } else { "Pattern absent (required match missing)" })
+      detail: (if $has_matches { "Pattern present — ok" } else { "Pattern absent — required match missing" })
     }
   }
 }
 
 # Run a 'Cargo check: parse Cargo.toml, verify forbidden_deps absent from [dependencies].
+# Resolves Cargo.toml in two steps:
+#   1. workspace layout: <adr-root>/crates/<check.crate>/Cargo.toml (preferred for multi-crate repos)
+#   2. single-crate fallback: <adr-root>/Cargo.toml when its [package].name matches check.crate,
+#                             OR unconditionally when check.crate equals the basename of adr-root
+# This lets the same constraint shape apply to workspace projects (stratumiops, provisioning)
+# and single-crate projects (lian-build, ontoref-* satellites).
 def run-cargo [check: record]: nothing -> record {
-  let cargo_path = ([(adr-root), "crates", $check.crate, "Cargo.toml"] | path join)
-  if not ($cargo_path | path exists) {
-    return { passed: false, detail: $"Cargo.toml not found: ($cargo_path)" }
+  let workspace_path = ([(adr-root), "crates", $check.crate, "Cargo.toml"] | path join)
+  let root_path     = ([(adr-root), "Cargo.toml"] | path join)
+  let cargo_path = if ($workspace_path | path exists) {
+    $workspace_path
+  } else if ($root_path | path exists) {
+    let root_cargo = (open $root_path)
+    let pkg_name = ($root_cargo.package?.name? | default "")
+    let project_basename = ((adr-root) | path basename)
+    if $pkg_name == $check.crate or $project_basename == $check.crate {
+      $root_path
+    } else {
+      return { passed: false, detail: $"Cargo.toml at ($root_path) declares package.name='($pkg_name)', not '($check.crate)' — workspace path ($workspace_path) also missing" }
+    }
+  } else {
+    return { passed: false, detail: $"Cargo.toml not found at ($workspace_path) or ($root_path)" }
   }
 
   let cargo = (open $cargo_path)
@@ -80,18 +100,24 @@ def run-cargo [check: record]: nothing -> record {
   let found = ($check.forbidden_deps | where { |dep| $dep in $all_dep_keys })
   {
     passed: ($found | is-empty),
-    detail: (if ($found | is-empty) { "No forbidden deps found" } else { $"Forbidden deps present: ($found | str join ', ')" })
+    detail: (if ($found | is-empty) { $"No forbidden deps found in ($cargo_path)" } else { $"Forbidden deps present: ($found | str join ', ')" })
   }
 }
 
 # Run a 'NuCmd check: execute cmd via nu -c, assert exit code.
+# The cmd executes with cwd = (adr-root) so relative paths in the constraint
+# (e.g. 'manifest.ncl', 'src/main.rs', 'catalog/domains') resolve against the
+# project being validated, regardless of where validate was invoked from.
+# The cd is composed into the cmd string rather than mutating parent shell
+# state — keeps run-nucmd reentrant and side-effect-free.
 def run-nucmd [check: record]: nothing -> record {
-  let result = do { nu -c $check.cmd } | complete
+  let cmd_with_cwd = $"cd '(adr-root)'; ($check.cmd)"
+  let result = do { nu -c $cmd_with_cwd } | complete
   let expected = ($check.expect_exit? | default 0)
   {
     passed: ($result.exit_code == $expected),
     detail: (if ($result.exit_code == $expected) {
-      "Command exited as expected"
+      $"Command exited as expected from cwd (adr-root)"
     } else {
       $"Exit ($result.exit_code) ≠ expected ($expected): ($result.stderr | str trim)"
     })
@@ -418,3 +444,512 @@ export def "validate justfile" [
     print $"  recipes found: ($result.recipes_found)  variables found: ($result.variables_found)"
   }
 }
+
+# ── Mode hierarchy validation (ADR-018) ──────────────────────────────────────
+
+# Extract level info from the project manifest. Returns Base if absent.
+def load-level-info []: nothing -> record {
+  let root = (adr-root)
+  let manifest_file = ([$root, ".ontology", "manifest.ncl"] | path join)
+  if not ($manifest_file | path exists) {
+    return { index: "Base", name: ([$root] | path basename), parent: null }
+  }
+  let manifest = (daemon-export-safe $manifest_file)
+  if $manifest == null {
+    return { index: "Base", name: ([$root] | path basename), parent: null }
+  }
+  let level = ($manifest.level? | default null)
+  if $level == null {
+    return { index: "Base", name: ($manifest.project? | default ([$root] | path basename)), parent: null }
+  }
+  {
+    index:       ($level.index | into string),
+    name:        ($level.name? | default "unknown"),
+    parent:      ($level.parent? | default null),
+    parent_path: ($level.parent_path? | default null),
+  }
+}
+
+# Read mode id and strategy from a single NCL file via source grep.
+# Using rg instead of nickel export because local normalizers may strip strategy/extends
+# from the exported record — these are source-level annotations, not runtime fields.
+def read-mode-strategy [file: string]: nothing -> record {
+  let stem = ($file | path basename | str replace ".ncl" "")
+
+  # id: prefer grepped value, fall back to stem (they match in practice)
+  let id_hit = (do { ^rg -m1 --no-heading -o 'id\s*=\s*"[^"]*"' $file } | complete)
+  let id = if $id_hit.exit_code == 0 {
+    $id_hit.stdout | str trim
+    | str replace -r '^id\s*=\s*"' '' | str replace -r '".*$' '' | str trim
+    | if ($in | str length) > 0 { $in } else { $stem }
+  } else { $stem }
+
+  # strategy = 'Override  (Nickel enum tag syntax — single quote prefix)
+  let strat_hit = (do { ^rg -m1 --no-heading -o "'(Override|Delegate|Merge|Compose)" $file } | complete)
+  let strategy = if $strat_hit.exit_code == 0 {
+    $strat_hit.stdout | str trim | str replace "'" "" | str trim
+  } else { null }
+
+  # extends = ["step-a", "step-b"]  — extract content between [ and ]
+  let ext_hit = (do { ^rg -m1 --no-heading -o 'extends\s*=\s*\[[^\]]*\]' $file } | complete)
+  let extends = if $ext_hit.exit_code == 0 {
+    $ext_hit.stdout | str trim
+    | str replace -r '^extends\s*=\s*\[' '' | str replace -r '\].*$' ''
+    | split row ","
+    | each { |s| $s | str replace -ra '"' '' | str trim }
+    | where { |s| ($s | str length) > 0 }
+  } else { [] }
+
+  {
+    id:       $id,
+    file:     $file,
+    strategy: $strategy,
+    extends:  $extends,
+    parse_ok: ($file | path exists),
+  }
+}
+
+# Collect mode files with source metadata: { file, local }
+# local=true means the file belongs to this project (not the ontoref root).
+# strategy-declared checks only apply to local files.
+def collect-mode-files-tagged []: nothing -> list<record> {
+  let root = (adr-root)
+  let ontoref_root = ($env.ONTOREF_ROOT? | default $root)
+  let is_self = ($root == $ontoref_root)
+
+  let root_modes = if ($ontoref_root | path exists) {
+    glob $"($ontoref_root)/reflection/modes/*.ncl"
+    | each { |f| { file: $f, local: $is_self } }
+  } else { [] }
+
+  let project_modes = if (not $is_self) and ([$root, "reflection", "modes"] | path join | path exists) {
+    glob $"($root)/reflection/modes/*.ncl"
+    | each { |f| { file: $f, local: true } }
+  } else { [] }
+
+  # project-local files shadow root files with same basename
+  let local_basenames = ($project_modes | each { |r| $r.file | path basename } | uniq)
+  let filtered_root = ($root_modes | where { |r| not (($r.file | path basename) in $local_basenames) })
+  ($filtered_root | append $project_modes)
+}
+
+# Collect all mode files (plain list, for backwards compat with mode resolve).
+def collect-mode-files []: nothing -> list<string> {
+  collect-mode-files-tagged | get file
+}
+
+# Format a single mode result row for text output.
+def fmt-mode-row [m: record, level_index: string]: nothing -> string {
+  let status = if not $m.parse_ok {
+    $"(ansi red)PARSE_ERROR(ansi reset)"
+  } else if ($m.strategy | is-empty) {
+    if $level_index == "Base" {
+      $"(ansi dark_gray)implicit-base(ansi reset)"
+    } else {
+      $"(ansi yellow)MISSING(ansi reset)"
+    }
+  } else {
+    match $m.strategy {
+      "Override" => $"(ansi green)Override(ansi reset)",
+      "Delegate" => $"(ansi cyan)Delegate(ansi reset)",
+      "Merge"    => $"(ansi blue)Merge(ansi reset)",
+      "Compose"  => $"(ansi magenta)Compose(ansi reset)",
+      _          => $"(ansi yellow)($m.strategy)(ansi reset)",
+    }
+  }
+  let extends_str = if ($m.extends | is-not-empty) { $" extends=[($m.extends | str join ', ')]" } else { "" }
+  $"  ($m.id | fill -w 45) ($status)($extends_str)"
+}
+
+# Run the level-declared check: manifest must declare level if not Base.
+def check-level-declared [level: record, modes: list<record>]: nothing -> list<record> {
+  let ok = $level.index != "Base" or ($level.name | is-not-empty)
+  [{
+    check:   "level-declared",
+    severity: "Soft",
+    passed:  ($level.index != "Base"),
+    detail:  (if $level.index == "Base" {
+      "No level declared — project is implicitly Base (ontoref root). Add level = { index = 'Domain, ... } for consumer projects."
+    } else {
+      $"Level ($level.index) declared: name=($level.name) parent=($level.parent? | default 'none')"
+    }),
+  }]
+}
+
+# Run the strategy-declared check: at Domain/Instance, local mode files must declare strategy.
+# Root (inherited) modes are excluded — if no local file exists, delegation is implicit and correct.
+def check-strategy-declared [level: record, modes: list<record>]: nothing -> list<record> {
+  if $level.index == "Base" {
+    return [{
+      check:    "strategy-declared",
+      severity: "Soft",
+      passed:   true,
+      detail:   "Base level — strategy declarations not required",
+    }]
+  }
+  let local_modes = ($modes | where local == true)
+  if ($local_modes | is-empty) {
+    return [{
+      check:    "strategy-declared",
+      severity: "Soft",
+      passed:   true,
+      detail:   "No local mode files — all modes delegate to parent implicitly",
+    }]
+  }
+  $local_modes | each { |m|
+    let missing = $m.parse_ok and ($m.strategy | is-empty)
+    {
+      check:    "strategy-declared",
+      severity: "Soft",
+      passed:   (not $missing),
+      detail:   (if $missing {
+        $"($m.id): strategy field absent — implicit Delegate \(Soft warning per ADR-018\)"
+      } else if not $m.parse_ok {
+        $"($m.id): NCL parse error — cannot verify strategy"
+      } else {
+        $"($m.id): strategy=($m.strategy)"
+      }),
+    }
+  }
+}
+
+# Run the delegate-chain check: Delegate modes must have a declared parent in the manifest.
+# If parent_path is set, verifies the parent project actually has the mode file (Hard failure if absent).
+# If parent_path is absent but parent is declared, emits a Soft warning (unverifiable remotely).
+def check-delegate-chain [level: record, modes: list<record>]: nothing -> list<record> {
+  let delegates = ($modes | where { |m| $m.strategy == "Delegate" })
+  if ($delegates | is-empty) {
+    return [{
+      check:    "delegate-chain",
+      severity: "Soft",
+      passed:   true,
+      detail:   "No Delegate modes — chain is trivially valid",
+    }]
+  }
+  if ($level.parent | is-empty) {
+    return ($delegates | each { |m|
+      {
+        check:    "delegate-chain",
+        severity: "Hard",
+        passed:   false,
+        detail:   $"($m.id): strategy=Delegate but manifest.level.parent is not declared — orphaned delegation",
+      }
+    })
+  }
+  let parent_path = ($level.parent_path? | default null)
+  if ($parent_path | is-empty) {
+    # parent declared but no path — unverifiable, Soft warning
+    return ($delegates | each { |m|
+      {
+        check:    "delegate-chain",
+        severity: "Soft",
+        passed:   true,
+        detail:   $"($m.id): delegates to parent=($level.parent) — add parent_path to manifest.level to enable cross-project verification",
+      }
+    })
+  }
+  # parent_path declared — verify each Delegate mode exists in parent's reflection/modes/
+  $delegates | each { |m|
+    let parent_mode_file = ([$parent_path, "reflection", "modes", $"($m.id).ncl"] | path join)
+    let found = ($parent_mode_file | path exists)
+    {
+      check:    "delegate-chain",
+      severity: "Hard",
+      passed:   $found,
+      detail:   (if $found {
+        $"($m.id): verified in parent at ($parent_path)/reflection/modes/($m.id).ncl"
+      } else {
+        $"($m.id): strategy=Delegate but ($parent_path)/reflection/modes/($m.id).ncl not found — orphaned delegation"
+      }),
+    }
+  }
+}
+
+# Run the compose-extends check: Compose modes must have a non-empty extends field.
+def check-compose-extends [level: record, modes: list<record>]: nothing -> list<record> {
+  let composers = ($modes | where { |m| $m.strategy == "Compose" })
+  if ($composers | is-empty) {
+    return [{
+      check:    "compose-extends",
+      severity: "Soft",
+      passed:   true,
+      detail:   "No Compose modes",
+    }]
+  }
+  $composers | each { |m|
+    let ok = ($m.extends | is-not-empty)
+    {
+      check:    "compose-extends",
+      severity: "Hard",
+      passed:   $ok,
+      detail:   (if $ok {
+        $"($m.id): extends=[($m.extends | str join ', ')]"
+      } else {
+        $"($m.id): strategy=Compose but extends=[] — must name parent step ids to inherit"
+      }),
+    }
+  }
+}
+
+# ── Self-test (exercises all 4 check variants with synthetic fixtures) ────────
+
+# Write a minimal NCL mode fragment to a file — no contracts, just the fields rg needs.
+def write-test-mode [dir: string, id: string, content: string]: nothing -> string {
+  let file = ([$dir, $"($id).ncl"] | path join)
+  $content | save --force $file
+  $file
+}
+
+# Run validate modes self-test: creates ephemeral temp fixtures, exercises every check
+# variant on known positive and negative inputs, reports pass/fail for each assertion.
+# Returns non-zero if any assertion fails.
+export def "validate modes --self-test" [
+  --fmt (-f): string = "text",
+]: nothing -> any {
+  # Build temp dirs for a synthetic Instance project and its parent Domain
+  let tmp       = ([ ($env.TMPDIR? | default "/tmp"), "ontoref-validate-selftest" ] | path join)
+  let tmp_parent = ([ ($env.TMPDIR? | default "/tmp"), "ontoref-validate-selftest-parent" ] | path join)
+  let modes_dir        = ([$tmp, "reflection", "modes"] | path join)
+  let parent_modes_dir = ([$tmp_parent, "reflection", "modes"] | path join)
+
+  do { rm -rf $tmp $tmp_parent } | ignore
+  mkdir $modes_dir
+  mkdir $parent_modes_dir
+
+  # ── Fixture: parent project has one mode (delegate-present) ─────────────────
+  write-test-mode $parent_modes_dir "delegate-present" '{ id = "delegate-present", strategy = '"'"'Override, steps = [] }' | ignore
+
+  # ── Fixtures: local Instance modes ──────────────────────────────────────────
+  # override-mode: should pass strategy-declared
+  write-test-mode $modes_dir "override-mode" '{ id = "override-mode", strategy = '"'"'Override, steps = [] }' | ignore
+  # no-strategy: should warn strategy-declared (Soft)
+  write-test-mode $modes_dir "no-strategy" '{ id = "no-strategy", steps = [] }' | ignore
+  # delegate-present: delegates to parent, parent HAS the mode → should pass Hard
+  write-test-mode $modes_dir "delegate-present" '{ id = "delegate-present", strategy = '"'"'Delegate, steps = [] }' | ignore
+  # delegate-missing: delegates to parent, parent DOES NOT have it → should fail Hard
+  write-test-mode $modes_dir "delegate-missing" '{ id = "delegate-missing", strategy = '"'"'Delegate, steps = [] }' | ignore
+  # compose-ok: Compose with non-empty extends → should pass Hard
+  write-test-mode $modes_dir "compose-ok" '{ id = "compose-ok", strategy = '"'"'Compose, extends = ["build-base", "test-base"], steps = [] }' | ignore
+  # compose-bad: Compose with empty extends → should fail Hard
+  write-test-mode $modes_dir "compose-bad" '{ id = "compose-bad", strategy = '"'"'Compose, extends = [], steps = [] }' | ignore
+
+  # ── Synthetic level records (bypass manifest loading) ───────────────────────
+  let level_base = {
+    index: "Base", name: "selftest-base", parent: null, parent_path: null
+  }
+  let level_instance_no_parent_path = {
+    index: "Instance", name: "selftest-instance", parent: "selftest-domain", parent_path: null
+  }
+  let level_instance_with_parent_path = {
+    index: "Instance", name: "selftest-instance", parent: "selftest-domain", parent_path: $tmp_parent
+  }
+
+  # ── Read synthetic mode records via rg ──────────────────────────────────────
+  let all_modes = (
+    glob $"($modes_dir)/*.ncl"
+    | each { |f| (read-mode-strategy $f) | merge { local: true } }
+    | sort-by id
+  )
+
+  # Helper: run a check, assert expected outcome, return result record
+  let assertions = [
+    # level-declared: Base → Soft warning (no level declared)
+    {
+      label:    "level-declared/Base → Soft warning",
+      expected: false,
+      result:   (check-level-declared $level_base [] | first),
+    },
+    # level-declared: Instance → passes
+    {
+      label:    "level-declared/Instance → pass",
+      expected: true,
+      result:   (check-level-declared $level_instance_no_parent_path [] | first),
+    },
+    # strategy-declared: override-mode → pass; no-strategy → Soft warn
+    {
+      label:    "strategy-declared/override-mode → pass",
+      expected: true,
+      result:   (check-strategy-declared $level_instance_no_parent_path $all_modes | where { |r| "override-mode" in $r.detail } | get 0? | default { passed: false }),
+    },
+    {
+      label:    "strategy-declared/no-strategy → Soft warn",
+      expected: false,
+      result:   (check-strategy-declared $level_instance_no_parent_path $all_modes | where { |r| "no-strategy" in $r.detail } | get 0? | default { passed: true }),
+    },
+    # delegate-chain: no parent_path → Soft pass (unverifiable)
+    {
+      label:    "delegate-chain/no-parent_path → Soft pass",
+      expected: true,
+      result:   (check-delegate-chain $level_instance_no_parent_path $all_modes | where { |r| "delegate-present" in $r.detail } | get 0? | default { passed: false }),
+    },
+    # delegate-chain: parent_path set, mode present → Hard pass
+    {
+      label:    "delegate-chain/parent_path+present → Hard pass",
+      expected: true,
+      result:   (check-delegate-chain $level_instance_with_parent_path $all_modes | where { |r| "delegate-present" in $r.detail } | get 0? | default { passed: false }),
+    },
+    # delegate-chain: parent_path set, mode absent → Hard fail
+    {
+      label:    "delegate-chain/parent_path+missing → Hard fail",
+      expected: false,
+      result:   (check-delegate-chain $level_instance_with_parent_path $all_modes | where { |r| "delegate-missing" in $r.detail } | get 0? | default { passed: true }),
+    },
+    # compose-extends: compose-ok → Hard pass
+    {
+      label:    "compose-extends/extends-present → Hard pass",
+      expected: true,
+      result:   (check-compose-extends $level_instance_no_parent_path $all_modes | where { |r| "compose-ok" in $r.detail } | get 0? | default { passed: false }),
+    },
+    # compose-extends: compose-bad → Hard fail
+    {
+      label:    "compose-extends/extends-empty → Hard fail",
+      expected: false,
+      result:   (check-compose-extends $level_instance_no_parent_path $all_modes | where { |r| "compose-bad" in $r.detail } | get 0? | default { passed: true }),
+    },
+  ]
+
+  let results = ($assertions | each { |a|
+    let actual   = $a.result.passed
+    let ok       = ($actual == $a.expected)
+    {
+      label:    $a.label,
+      expected: (if $a.expected { "pass" } else { "fail" }),
+      actual:   (if $actual { "pass" } else { "fail" }),
+      ok:       $ok,
+    }
+  })
+
+  # Cleanup
+  do { rm -rf $tmp $tmp_parent } | ignore
+
+  let failures = ($results | where ok == false)
+
+  if $fmt == "json" {
+    print ($results | to json)
+  } else {
+    print $"validate modes --self-test: ($results | length) assertions, ($failures | length) failures"
+    print ""
+    for r in $results {
+      let icon = if $r.ok { $"(ansi green)✓(ansi reset)" } else { $"(ansi red)✗(ansi reset)" }
+      let detail = if $r.ok { "" } else { $"  expected=($r.expected) got=($r.actual)" }
+      print $"($icon) ($r.label)($detail)"
+    }
+  }
+
+  if ($failures | is-not-empty) {
+    error make { msg: $"($failures | length) self-test assertion\(s\) failed" }
+  }
+}
+
+# Validate mode files against the ADR-018 level hierarchy and resolution strategy rules.
+# Checks: level-declared | strategy-declared | delegate-chain | compose-extends | all
+export def "validate modes" [
+  --check (-c): string = "all",  # level-declared | strategy-declared | delegate-chain | compose-extends | all
+  --fmt (-f): string = "table",  # table | json | yaml | text
+]: nothing -> any {
+  let level = (load-level-info)
+  let tagged = (collect-mode-files-tagged)
+  let modes = ($tagged | each { |t|
+    let m = (read-mode-strategy $t.file)
+    $m | merge { local: $t.local }
+  } | sort-by id)
+
+  let results = match $check {
+    "level-declared"    => (check-level-declared    $level $modes),
+    "strategy-declared" => (check-strategy-declared $level $modes),
+    "delegate-chain"    => (check-delegate-chain    $level $modes),
+    "compose-extends"   => (check-compose-extends   $level $modes),
+    _                   => (
+      (check-level-declared    $level $modes)
+      | append (check-strategy-declared $level $modes)
+      | append (check-delegate-chain    $level $modes)
+      | append (check-compose-extends   $level $modes)
+    ),
+  }
+
+  let failures = ($results | where passed == false)
+  let hard_failures = ($failures | where severity == "Hard")
+
+  match $fmt {
+    "json" => { print ($results | to json) },
+    "yaml" => { print ($results | to yaml) },
+    _      => {
+      let local_count = ($modes | where local == true | length)
+      let inherited_count = ($modes | where local == false | length)
+      print $"Level: (ansi cyan_bold)($level.index)(ansi reset)  name=($level.name)  parent=($level.parent? | default 'none')"
+      print $"Modes: ($modes | length) \(($local_count) local, ($inherited_count) inherited\)  Checks: ($results | length)  Failures: ($failures | length)"
+      print ""
+      for r in $results {
+        let icon = if $r.passed { $"(ansi green)✓(ansi reset)" } else if $r.severity == "Hard" { $"(ansi red)✗(ansi reset)" } else { $"(ansi yellow)⚠(ansi reset)" }
+        print $"($icon) [($r.check)] ($r.detail)"
+      }
+    },
+  }
+
+  if ($hard_failures | is-not-empty) {
+    error make { msg: $"($hard_failures | length) Hard constraint\(s\) failed in validate modes --check ($check)" }
+  }
+}
+
+# Show which hierarchy level handles a mode invocation (ADR-018 resolution).
+# Reports the mode's strategy at this level and what that means for execution.
+export def "mode resolve" [
+  id: string,        # Mode id (without .ncl extension)
+  --fmt (-f): string = "text",  # text | json
+]: nothing -> any {
+  let level = (load-level-info)
+  let mode_files = (collect-mode-files)
+  let candidates = ($mode_files | where { |f| ($f | path basename | str replace ".ncl" "") == $id })
+
+  if ($candidates | is-empty) {
+    error make { msg: $"Mode '($id)' not found in reflection/modes/" }
+  }
+
+  let m = (read-mode-strategy ($candidates | first))
+
+  let resolution = if not $m.parse_ok {
+    { answered_by: "unknown", reason: "NCL parse error", action: "fix parse error" }
+  } else if ($m.strategy | is-empty) {
+    if $level.index == "Base" {
+      { answered_by: $level.name, reason: "Base level — implicit Override", action: "executes here" }
+    } else {
+      { answered_by: "parent", reason: "strategy absent — treated as implicit Delegate (ADR-018 Soft warning)", action: $"traverse to ($level.parent? | default 'unknown')" }
+    }
+  } else {
+    match $m.strategy {
+      "Override" => { answered_by: $level.name, reason: "strategy=Override", action: "executes here — traversal stops" },
+      "Delegate" => { answered_by: "parent",     reason: "strategy=Delegate", action: $"traverse to ($level.parent? | default 'unknown') — not implemented here" },
+      "Merge"    => { answered_by: "all-levels", reason: "strategy=Merge", action: $"accumulate contributions from ($level.name) and parent chain; this level wins conflicts" },
+      "Compose"  => { answered_by: $level.name,  reason: $"strategy=Compose extends=[($m.extends | str join ', ')]", action: $"executes here with step\(s\) ($m.extends | str join ', ') inherited from ($level.parent? | default 'parent')" },
+      _          => { answered_by: "unknown",    reason: $"unrecognised strategy: ($m.strategy)", action: "check ADR-018" },
+    }
+  }
+
+  let effective_strategy = if ($m.strategy? | is-empty) {
+    if $level.index == "Base" { "implicit-Override" } else { "implicit-Delegate" }
+  } else {
+    $m.strategy
+  }
+
+  let out = {
+    mode:        $m.id,
+    level_index: $level.index,
+    level_name:  $level.name,
+    parent:      ($level.parent? | default null),
+    strategy:    $effective_strategy,
+    answered_by: $resolution.answered_by,
+    action:      $resolution.action,
+    reason:      $resolution.reason,
+  }
+
+  if $fmt == "json" {
+    print ($out | to json)
+    return
+  }
+
+  print $"Mode:     (ansi cyan_bold)($out.mode)(ansi reset)"
+  print $"Level:    ($out.level_index) / ($out.level_name)"
+  print $"Strategy: ($out.strategy)"
+  print $"Resolves: (ansi green_bold)($out.answered_by)(ansi reset) — ($out.action)"
+  if ($out.reason | is-not-empty) { print $"Reason:   (ansi dark_gray)($out.reason)(ansi reset)" }
+}
diff --git a/reflection/modules/vault.nu b/reflection/modules/vault.nu
new file mode 100644
index 0000000..85e3710
--- /dev/null
+++ b/reflection/modules/vault.nu
@@ -0,0 +1,258 @@
+#!/usr/bin/env nu
+# Vault backend abstraction for src-vault snapshot operations (ADR-017).
+# All functions accept the decrypted vault_key value — they never read sops.
+# Switching backends: change sops.vault_backend in project.ncl, re-init the local repo.
+
+use ./secrets.nu [resolve-vault-access]
+
+const VAULT_BACKENDS = ["restic" "kopia"]
+
+def require-tool [tool: string]: nothing -> string {
+    let rows = (which $tool)
+    if ($rows | is-empty) {
+        error make --unspanned { msg: $"vault backend '($tool)' not found in PATH" }
+    }
+    $rows | get path | first | into string
+}
+
+def assert-backend [tool: string]: nothing -> nothing {
+    if $tool not-in $VAULT_BACKENDS {
+        error make --unspanned { msg: $"Unknown vault backend: ($tool). Valid: ($VAULT_BACKENDS | str join ', ')" }
+    }
+}
+
+# Initialize a new vault repository at `repo_path`.
+export def vault-init [
+    repo_path: string
+    vault_key: string
+    --tool: string = "restic"
+]: nothing -> nothing {
+    assert-backend $tool
+    let bin    = (require-tool $tool)
+    let result = match $tool {
+        "restic" => (with-env { RESTIC_PASSWORD: $vault_key } {
+            do { ^$bin -r $repo_path init } | complete
+        }),
+        "kopia" => (with-env { KOPIA_PASSWORD: $vault_key } {
+            do { ^$bin repository create filesystem --path $repo_path } | complete
+        }),
+    }
+    if $result.exit_code != 0 {
+        error make --unspanned { msg: $"vault init failed:\n($result.stderr)" }
+    }
+}
+
+# Create a snapshot of `source_path`. Returns the snapshot short-id.
+export def vault-backup [
+    repo_path:   string
+    source_path: string
+    vault_key:   string
+    --tool: string = "restic"
+    --tags: list<string> = []
+]: nothing -> string {
+    assert-backend $tool
+    let bin    = (require-tool $tool)
+    let result = match $tool {
+        "restic" => {
+            let tag_args = ($tags | each { |t| ["--tag" $t] } | flatten)
+            with-env { RESTIC_PASSWORD: $vault_key } {
+                do { ^$bin -r $repo_path backup $source_path ...$tag_args } | complete
+            }
+        },
+        "kopia" => {
+            with-env { KOPIA_PASSWORD: $vault_key } {
+                do { ^$bin --config-file ($repo_path | path join "kopia.config") snapshot create $source_path } | complete
+            }
+        },
+    }
+    if $result.exit_code != 0 {
+        error make --unspanned { msg: $"vault backup failed:\n($result.stderr)" }
+    }
+    $result.stdout | lines
+        | where { |l| $l | str contains "snapshot " }
+        | first
+        | default ""
+        | parse --regex 'snapshot\s+([0-9a-f]+)'
+        | get -o capture0
+        | first
+        | default ""
+}
+
+# Restore snapshot `snapshot_id` (or "latest") to `dest_path`.
+export def vault-restore [
+    repo_path:   string
+    dest_path:   string
+    vault_key:   string
+    snapshot_id: string = "latest"
+    --tool: string = "restic"
+]: nothing -> nothing {
+    assert-backend $tool
+    let bin    = (require-tool $tool)
+    let result = match $tool {
+        "restic" => (with-env { RESTIC_PASSWORD: $vault_key } {
+            do { ^$bin -r $repo_path restore $snapshot_id --target $dest_path } | complete
+        }),
+        "kopia" => (with-env { KOPIA_PASSWORD: $vault_key } {
+            do { ^$bin --config-file ($repo_path | path join "kopia.config") snapshot restore $snapshot_id $dest_path } | complete
+        }),
+    }
+    if $result.exit_code != 0 {
+        error make --unspanned { msg: $"vault restore failed:\n($result.stderr)" }
+    }
+}
+
+# List snapshots. Returns list of {id, time, tags} records.
+export def vault-snapshots [
+    repo_path: string
+    vault_key: string
+    --tool: string = "restic"
+]: nothing -> list {
+    assert-backend $tool
+    let bin    = (require-tool $tool)
+    let result = match $tool {
+        "restic" => (with-env { RESTIC_PASSWORD: $vault_key } {
+            do { ^$bin -r $repo_path snapshots --json } | complete
+        }),
+        "kopia" => (with-env { KOPIA_PASSWORD: $vault_key } {
+            do { ^$bin --config-file ($repo_path | path join "kopia.config") snapshot list --json } | complete
+        }),
+    }
+    if $result.exit_code != 0 {
+        error make --unspanned { msg: $"vault snapshots failed:\n($result.stderr)" }
+    }
+    let raw = ($result.stdout | str trim)
+    if ($raw | is-empty) { return [] }
+    match $tool {
+        "restic" => {
+            $raw | from json | each { |s|
+                { id: ($s | get -o short_id | default ""), time: ($s | get -o time | default ""), tags: ($s | get -o tags | default []) }
+            }
+        },
+        "kopia" => {
+            $raw | from json | each { |s|
+                { id: ($s | get -o id | default ""), time: ($s | get -o startTime | default ""), tags: [] }
+            }
+        },
+    }
+}
+
+# Verify vault repository integrity.
+export def vault-check [
+    repo_path: string
+    vault_key: string
+    --tool: string = "restic"
+]: nothing -> bool {
+    assert-backend $tool
+    let bin    = (require-tool $tool)
+    let result = match $tool {
+        "restic" => (with-env { RESTIC_PASSWORD: $vault_key } {
+            do { ^$bin -r $repo_path check } | complete
+        }),
+        "kopia" => (with-env { KOPIA_PASSWORD: $vault_key } {
+            do { ^$bin --config-file ($repo_path | path join "kopia.config") repository verify-object-integrity } | complete
+        }),
+    }
+    $result.exit_code == 0
+}
+
+# ── CLI dispatch — `ore vault X` ──────────────────────────────────────────────
+# High-level wrappers that resolve vault context (project.ncl → vault_id, backend,
+# repo_path, vault_key) and call the primitives above.
+
+def _project-root-cli []: nothing -> string {
+    let from_env = ($env.ONTOREF_PROJECT_ROOT? | default "")
+    if ($from_env | is-not-empty) { return $from_env }
+    pwd
+}
+
+def _resolve-vault-context [project_root: string]: nothing -> record {
+    let project_ncl = ($project_root | path join ".ontoref" "project.ncl")
+    if not ($project_ncl | path exists) {
+        error make --unspanned { msg: $"No project.ncl at ($project_ncl)" }
+    }
+    let import_path = ($env.NICKEL_IMPORT_PATH? | default ($env.PROVISIONING? | default ""))
+    let r = (do { ^nickel export --import-path $import_path $project_ncl } | complete)
+    if $r.exit_code != 0 {
+        error make --unspanned { msg: $"Failed to evaluate project.ncl: ($r.stderr)" }
+    }
+    let project  = ($r.stdout | from json)
+    let sops_cfg = ($project | get -o sops | default {})
+    let vault_id = ($sops_cfg | get -o vault_id | default ($project | get -o slug | default ""))
+    if ($vault_id | is-empty) {
+        error make --unspanned { msg: "Cannot determine vault_id (slug empty and sops.vault_id absent)" }
+    }
+    let backend  = ($sops_cfg | get -o vault_backend | default "restic")
+    let backend_str = if (($backend | describe) == "string") { $backend } else { ($backend | to text) }
+    {
+        vault_id:  $vault_id,
+        backend:   $backend_str,
+        repo_path: ($env.HOME | path join ".config" "ontoref" "vaults" $vault_id "repo"),
+    }
+}
+
+# List snapshots in the vault repository for the active project.
+export def "vault snapshots" []: nothing -> nothing {
+    let project_root = (_project-root-cli)
+    let ctx          = (_resolve-vault-context $project_root)
+    let access       = (resolve-vault-access $project_root)
+    if ($access.vault_key | is-empty) {
+        error make --unspanned {
+            msg: "vault_key empty — run: ore secrets sync, then retry"
+        }
+    }
+    let snaps = (vault-snapshots $ctx.repo_path $access.vault_key --tool $ctx.backend)
+    if ($snaps | is-empty) {
+        print $"  No snapshots in ($ctx.repo_path)"
+        return
+    }
+    print $"  (ansi white_bold)Vault snapshots(ansi reset) — ($ctx.vault_id) ((($ctx.backend)))"
+    for s in $snaps {
+        print $"    ($s.id)  ($s.time)  tags=[($s.tags | str join ', ')]"
+    }
+}
+
+# Run integrity check on the vault repository.
+export def "vault check" []: nothing -> nothing {
+    let project_root = (_project-root-cli)
+    let ctx          = (_resolve-vault-context $project_root)
+    let access       = (resolve-vault-access $project_root)
+    if ($access.vault_key | is-empty) {
+        error make --unspanned {
+            msg: "vault_key empty — run: ore secrets sync, then retry"
+        }
+    }
+    let ok = (vault-check $ctx.repo_path $access.vault_key --tool $ctx.backend)
+    if $ok {
+        print $"  (ansi green)OK(ansi reset)  ($ctx.vault_id) (($ctx.backend)) repository integrity verified"
+    } else {
+        print $"  (ansi red)FAIL(ansi reset)  ($ctx.vault_id) (($ctx.backend)) check failed"
+    }
+}
+
+# Summary: vault context resolved + last snapshot.
+export def "vault status" []: nothing -> nothing {
+    let project_root = (_project-root-cli)
+    let ctx          = (_resolve-vault-context $project_root)
+    print $"  (ansi white_bold)Vault(ansi reset)      ($ctx.vault_id)"
+    print $"  (ansi dark_gray)backend(ansi reset)    ($ctx.backend)"
+    print $"  (ansi dark_gray)repo(ansi reset)       ($ctx.repo_path) (if ($ctx.repo_path | path exists) { '(present)' } else { '(absent)' })"
+    if not ($ctx.repo_path | path exists) {
+        print "  Run: ore secrets bootstrap (admin) or ore secrets sync"
+        return
+    }
+    let access = (try { resolve-vault-access $project_root } catch { |e|
+        print $"  vault access  (ansi red)error(ansi reset): ($e.msg)"
+        return
+    })
+    if ($access.vault_key | is-empty) { return }
+    let snaps = (try { vault-snapshots $ctx.repo_path $access.vault_key --tool $ctx.backend } catch { |e|
+        print $"  snapshots     (ansi red)error(ansi reset): ($e.msg)"
+        return
+    })
+    if ($snaps | is-empty) {
+        print "  snapshots  (none)"
+    } else {
+        let last = ($snaps | last)
+        print $"  (ansi dark_gray)snapshots(ansi reset)  ($snaps | length) total — last: ($last.id) @ ($last.time)"
+    }
+}
diff --git a/reflection/modules/workflow.nu b/reflection/modules/workflow.nu
index 63f4093..643937e 100644
--- a/reflection/modules/workflow.nu
+++ b/reflection/modules/workflow.nu
@@ -97,6 +97,22 @@ def resolve-build [id: string, catalog: record]: nothing -> record {
   resolve-spec $id $catalog "build"
 }
 
+# Resolve `{TARGET_DIR}` and `{ROOT}` placeholders in BuildSpec.artifacts paths.
+# {TARGET_DIR} → `cargo metadata --format-version 1 --no-deps | .target_directory`
+# {ROOT}       → workspace root
+# Falls back to "$root/target" if cargo metadata is unavailable.
+def resolve-artifact-path [path: string, root: string]: nothing -> string {
+  let meta = do { cd $root; ^cargo metadata --format-version 1 --no-deps } | complete
+  let target_dir = if $meta.exit_code == 0 {
+    $meta.stdout | from json | get target_directory
+  } else {
+    $"($root)/target"
+  }
+  $path
+    | str replace --all "{TARGET_DIR}" $target_dir
+    | str replace --all "{ROOT}" $root
+}
+
 # ── Pre-commit generator ─────────────────────────────────────────────────────
 
 # Translate a ValidationSpec into a pre-commit hook entry (YAML fragment).
diff --git a/reflection/nulib/bootstrap.nu b/reflection/nulib/bootstrap.nu
index 385bc93..2a35001 100644
--- a/reflection/nulib/bootstrap.nu
+++ b/reflection/nulib/bootstrap.nu
@@ -7,7 +7,7 @@
 #
 # Usage:
 #   use reflection/nulib/bootstrap.nu *
-#   ncl-bootstrap "~/.config/ontoref/config.ncl" "ontoref-daemon" --port 7891
+#   ncl-bootstrap "~/.config/ontoref/config.ncl" "ontoref-daemon" --port 7890
 #   ncl-bootstrap "config.ncl" "myapp" --sops "secrets.enc.json"
 #   ncl-bootstrap "config.ncl" "myapp" --dry-run   # print composed JSON
 
diff --git a/reflection/nulib/help.nu b/reflection/nulib/help.nu
index 7c2cda4..5bb0142 100644
--- a/reflection/nulib/help.nu
+++ b/reflection/nulib/help.nu
@@ -467,6 +467,38 @@ export def help-group [group: string] {
         { short: "mg p", long: "migrate pending" },
       ]
     },
+    "secrets" => {
+      print ""
+      fmt-header "SECRETS  (registry credential vault — ADR-017)"
+      fmt-sep
+      fmt-info "Per-project src-vault stored as OCI artifact. Multi-recipient sops + age, signed with cosign."
+      print ""
+      fmt-cmd $"($cmd) secrets status"               "vault state, last access, master key resolution"
+      fmt-cmd $"($cmd) secrets describe"             "detailed inventory: recipients, roles, credentials, operations"
+      fmt-cmd $"($cmd) secrets bootstrap"            "create vault for a new project (admin only)"
+      fmt-cmd $"($cmd) secrets sync"                 "pull latest src-vault from ZOT"
+      fmt-cmd $"($cmd) secrets push"                 "push current vault to ZOT (cosign signs)"
+      fmt-cmd $"($cmd) secrets open"                 "acquire OCI lock + edit access.sops.yaml"
+      fmt-cmd $"($cmd) secrets close"                "push updated vault + release lock"
+      fmt-cmd $"($cmd) secrets force-unlock"         "force-release abandoned lock (admin, audited)"
+      fmt-cmd $"($cmd) secrets gen-key <role>"       "generate new age keypair for a role"
+      fmt-cmd $"($cmd) secrets add-key <pub> <role>" "add age recipient to access.sops.yaml"
+      fmt-cmd $"($cmd) secrets remove-key <pub>"     "remove age recipient (revocation)"
+      fmt-cmd $"($cmd) secrets rekey"                "re-encrypt with current .sops.yaml recipients"
+      fmt-cmd $"($cmd) secrets audit [--check <c>]"  "verify ADR-017 vault constraints"
+      print ""
+    },
+    "vault" => {
+      print ""
+      fmt-header "VAULT  (restic|kopia repo for src-vault snapshots)"
+      fmt-sep
+      fmt-info "Inspection-only at CLI. Mutations go through `secrets push`/`secrets sync`."
+      print ""
+      fmt-cmd $"($cmd) vault status"     "vault context + last snapshot"
+      fmt-cmd $"($cmd) vault snapshots"  "list all snapshots in the local repo"
+      fmt-cmd $"($cmd) vault check"      "integrity check of the repo"
+      print ""
+    },
     _ if (($"($env.ONTOREF_ROOT?)/domains/($group)/domain.ncl") | path exists) => {
       let domain_ncl = $"($env.ONTOREF_ROOT)/domains/($group)/domain.ncl"
       let d = (nickel export $domain_ncl | from json)
@@ -490,7 +522,7 @@ export def help-group [group: string] {
     _ => {
       print $"  (ansi red)Unknown group: ($group)(ansi reset)"
       print ""
-      fmt-info "Available groups: check | form | mode | adr | register | backlog | config | sync | coder | manifest | describe | log | migrate | <domain-id>"
+      fmt-info "Available groups: check | form | mode | adr | register | backlog | config | sync | coder | manifest | describe | log | migrate | secrets | vault | <domain-id>"
       print ""
     },
   }
diff --git a/reflection/nulib/modes.nu b/reflection/nulib/modes.nu
index 25139e1..352676c 100644
--- a/reflection/nulib/modes.nu
+++ b/reflection/nulib/modes.nu
@@ -174,14 +174,42 @@ def actor-can-run-step [step_actor: string]: nothing -> bool {
   }
 }
 
+# Build the substitution context for a step command. Reads project slug from
+# .ontoref/project.ncl when present; falls back to project_dir basename.
+# Used for {project_dir} / {ontoref_dir} / {project_name} placeholder substitution
+# in step cmd strings before they reach bash or nu.
+def build-cmd-context []: nothing -> record {
+  let project_dir = ($env.ONTOREF_PROJECT_ROOT? | default ($env.ONTOREF_ROOT? | default $env.PWD))
+  let ontoref_dir = ($env.ONTOREF_ROOT? | default "")
+  let basename_fallback = ($project_dir | path basename)
+  let project_ncl = $"($project_dir)/.ontoref/project.ncl"
+  let project_name = if ($project_ncl | path exists) {
+    let exported = (do { ^nickel export $project_ncl } | complete)
+    if $exported.exit_code == 0 {
+      let parsed = (try { $exported.stdout | from json } catch { {} })
+      ($parsed.slug? | default $basename_fallback)
+    } else { $basename_fallback }
+  } else { $basename_fallback }
+  { project_dir: $project_dir, ontoref_dir: $ontoref_dir, project_name: $project_name }
+}
+
 # Execute a single step's command. Returns { success: bool, output: string }.
+# Substitutes {project_dir} / {ontoref_dir} / {project_name} placeholders before
+# dispatching to nu (when the cmd contains nu-only pipeline ops) or bash.
 def exec-step-cmd [cmd: string]: nothing -> record {
+  let ctx = (build-cmd-context)
+  let resolved = (
+    $cmd
+    | str replace --all "{project_dir}"  $ctx.project_dir
+    | str replace --all "{ontoref_dir}"  $ctx.ontoref_dir
+    | str replace --all "{project_name}" $ctx.project_name
+  )
   let nu_patterns = ["| from json", "| get ", "| where ", "| each ", "| select ", "| sort-by"]
-  let is_nu = ($nu_patterns | any { |p| $cmd | str contains $p })
+  let is_nu = ($nu_patterns | any { |p| $resolved | str contains $p })
   let result = if $is_nu {
-    do { ^nu -c $cmd } | complete
+    do { ^nu -c $resolved } | complete
   } else {
-    do { ^bash -c $cmd } | complete
+    do { ^bash -c $resolved } | complete
   }
   {
     success: ($result.exit_code == 0),
diff --git a/reflection/qa.ncl b/reflection/qa.ncl
index 83db835..0e6f22b 100644
--- a/reflection/qa.ncl
+++ b/reflection/qa.ncl
@@ -1,5 +1,1170 @@
 let s = import "schemas/qa.ncl" in
 
 {
-  entries = [],
+  entries = [
+    {
+      id         = "credential-vault-best-practice",
+      question   = "What is the canonical approach to manage registry credentials with ontoref's credential vault?",
+      answer     = m%"
+ADR-017 implements a layered credential model. Apply each layer in order; do not
+skip layers — they enforce different invariants.
+
+DATA FLOW
+
+  developer machine                          ZOT registry
+  ─────────────────                          ─────────────────
+  ~/.age/keys/<actor>.key.txt   (Layer 0)
+            │ sops --decrypt
+            ▼
+  access.sops.yaml              (Layer 1)
+  { zot_user, zot_pass,
+    vault_key, cosign_pass }
+            │
+            │ oras pull ──────────────────►  src-vault/<id>:latest
+            ◄────────────────── (cosigned)
+            ▼
+  ~/.config/ontoref/vaults/<id>/src-vault/
+    scopes/<role>.ncl                       ─┐
+    registry/<file>.sops.yaml    (Layer 2)   │ assert-actor-authorized
+            │ sops --decrypt                 │ assert-target-in-scope
+            ▼                                │
+  DOCKER_CONFIG=$tmpdir                     ─┘
+            │
+            │ oras push/pull ─────────────►  domains/<participant>/<id>:<v>
+            │                                modes/<participant>/<id>:<v>
+            ▼ rm -rf $tmpdir
+
+LAYER 0 — Master key (per developer/machine)
+  - One age private key (.kage) per actor; declared globally in
+    ~/.config/ontoref/config.ncl::vault.master_key_path. Per-project override
+    in <project>/.ontoref/project.ncl::sops.master_key_path when the project
+    requires a different key (e.g. yubikey-backed for production-only access).
+  - Generate once with: age-keygen -o ~/.age/keys/<name>.key.txt
+  - Permissions 0400. Never commit; never put inside any vault directory.
+
+LAYER 1 — Vault access credential (per-project)
+  - access.sops.yaml encrypted multi-recipient with all actors who may open
+    the vault. Contains: zot_username, zot_password, vault_key (restic/kopia
+    repo password), cosign_password.
+  - Lives at ~/.config/ontoref/vaults/<vault_id>/access.sops.yaml.
+  - Generated once by 'ore secrets bootstrap'; updated via 'ore secrets open'.
+
+LAYER 2 — Operation credentials (per-purpose)
+  - Files under ~/.config/ontoref/vaults/<vault_id>/src-vault/registry/.
+  - Referenced from .ontology/manifest.ncl::registry_provides.registries[]:
+      credential_sops    (RO — pull/list)
+      credential_sops_rw (RW — push)
+  - Paths are RELATIVE to src-vault/, not to project root.
+  - Decrypted into an isolated DOCKER_CONFIG tmpdir per oras invocation.
+
+PER-FILE RECIPIENT ROUTING (multi-tenant, optional)
+  - Single vault, multiple recipient sets via sops creation_rules.
+  - Declare in project.ncl::sops:
+      recipient_groups = { admin = [...], clientA = [...], agents = [...] }
+      recipient_rules  = [
+        { path = \"registry/clientA-.*\\.sops\\.yaml$\", groups = [\"admin\", \"clientA\"] },
+        { path = \"registry/agent-.*\\.sops\\.yaml$\",   groups = [\"admin\", \"agents\"] },
+      ]
+  - Bootstrap generates <vault_dir>/.sops.yaml; sops encrypts each file with
+    the union of declared groups. Use this instead of multi-vault for
+    tenant/agent isolation in a single project.
+
+AUTHORIZATION GATING (always enforced)
+  - project.ncl::sops.actor_key_bindings maps ONTOREF_ACTOR → role.
+  - <vault_id>/src-vault/scopes/<role>.ncl declares { access, bound_actor,
+    namespaces, ops }. Two-level enforcement:
+      assert-actor-authorized — checks scope.ops + scope.bound_actor
+      assert-target-in-scope  — checks the OCI ref against scope.namespaces
+  - Both fire BEFORE any oras call. No cache hit bypasses these checks.
+
+HARD RULES (ADR-017 invariants — non-negotiable)
+  - Daemon never touches credentials. Resolution lives in the CLI process
+    that holds the actor's .kage. The ontoref daemon only reads declarative
+    metadata.
+  - Every oras call runs with DOCKER_CONFIG=$tmpdir, freshly built from sops
+    and torn down after the call. No fallback to ~/.docker/config.json.
+  - cosign signing is mandatory for src-vault pushes. Default tlog=false
+    (private model); vault.cosign.signing_config_path declares a Rekor-less
+    signing config when tlog disabled.
+  - Multi-recipient sops mandatory. ≥ 2 recipients per encrypted file.
+  - access.sops.yaml carries cosign_password so push runs non-interactively.
+  - Vault lock (OCI artifact at src-vault/<id>:lock) coordinates concurrent
+    edits with TTL 60min. force-unlock is admin-only and auditable.
+
+DAY-TO-DAY COMMANDS
+  ore secrets status                 vault state, master key resolution
+  ore secrets describe               full inventory: groups, rules, scopes, ops
+  ore secrets sync                   pull latest src-vault from ZOT
+  ore secrets open                   acquire lock + edit access.sops.yaml
+  ore secrets close                  impact report + push + release lock
+  ore secrets rekey                  regenerate .sops.yaml + sops updatekeys
+  ore secrets force-unlock           release abandoned lock (admin)
+
+ESCAPE HATCHES
+  ONTOREF_SECRETS_YES=1              skip impact-confirm in secrets close
+  (no ambient docker config fallback — by design, see ADR-017 invariant I4)
+
+REFERENCES
+  reflection/modules/secrets.nu (header)  — function contract
+  reflection/migrations/0016              — adoption steps
+  install/resources/templates/sops/       — copy-paste templates per tenancy model
+  adrs/adr-017                            — invariants and rationale
+  reflection/qa.ncl::ontoref-three-layer-model
+                                          — vault credentials gate LAYER 2
+                                            publication (oras push of a project's
+                                            domains/<participant>/* and
+                                            modes/<participant>/*). The
+                                            credential resolution chain ALSO
+                                            sits at the LAYER 2 ↔ LAYER 3 seam:
+                                            cred files referenced from
+                                            manifest.ncl::registry_provides
+                                            (Layer 2 declaration) are decrypted
+                                            by caller-side workflows (Layer 3
+                                            execution).
+"%,
+      actor      = "human",
+      created_at = "2026-05-03",
+      tags       = ["credentials", "adr-017", "sops", "cosign", "vault", "best-practice", "layer-2"],
+      related    = ["adr-017", "adr-015"],
+      verified   = true,
+    },
+    {
+      id         = "credential-vault-templates",
+      question   = "How do I bootstrap the credential vault for a new project? Are there templates I can copy?",
+      answer     = m%"
+Three adoption paths, in order of complexity. Pick the one matching your project:
+
+PATH A — SINGLE-TEAM (legacy, simplest)
+  Use when: one team, no tenant isolation, no agent restrictions beyond ops gating.
+  Template: install/resources/templates/sops/single-team/project.ncl
+  Steps:
+    1. Copy the template snippet into your .ontoref/project.ncl
+    2. Set master_key_path (per-project) or rely on the global config
+    3. Add registry_provides to .ontology/manifest.ncl with credential_sops/_rw
+       pointing to registry/ro.sops.yaml and registry/rw.sops.yaml
+    4. export SOPS_AGE_RECIPIENTS=\"age1...\"  (comma-separated, ≥ 2 keys)
+    5. ore secrets bootstrap   (creates default ro/rw files seeded with
+       interactive zot credentials)
+    6. ore secrets push
+
+PATH B — MULTI-TENANT (recommended for libre-wuji-class projects)
+  Use when: multiple clients/agents/teams must NOT see each other's credentials.
+  Template: install/resources/templates/sops/multi-tenant/project.ncl
+  Adds recipient_groups + recipient_rules. Each tenant has its own group of
+  age public keys; rules route file paths to group unions.
+  Steps:
+    1. Copy the template — adjust group keys and rule patterns to your tenancy
+    2. Add registry_provides entries for each tenant (e.g. credential_sops =
+       \"registry/clientA-ro.sops.yaml\")
+    3. ore secrets bootstrap   (skips default ro/rw files, generates .sops.yaml)
+    4. ore secrets open        (populate registry/<file>.sops.yaml entries)
+    5. ore secrets close
+
+PATH C — AGENT-FIRST (ontoref/MCP integration)
+  Use when: AI agents read credentials with strict restrictions.
+  Template: install/resources/templates/sops/agent-first/project.ncl
+  Same shape as Path B but with predefined groups for admin / developer / agent
+  and a default scope file that gives 'agent' role RO ops on a single
+  agent-readonly.sops.yaml.
+
+UNIVERSAL CHECKLIST
+  Pre-bootstrap:
+    - master .kage generated (age-keygen) and at master_key_path
+    - cosign keypair at vault.cosign.{key_path,pub_path}
+    - signing-config-no-rekor.json (when tlog=false)
+    - ZOT registry reachable; ACL allows src-vault/<vault_id>/ namespace
+  Post-bootstrap:
+    - ore secrets describe  shows expected recipients and per-file routing
+    - ore secrets audit     all 3 checks pass
+
+NO TEMPLATE = LEGACY DEFAULTS
+  If you skip the templates entirely, ore secrets bootstrap with
+  SOPS_AGE_RECIPIENTS env-var works as a minimal viable path. The result is
+  Path A. You can migrate to B or C later by adding recipient_groups +
+  recipient_rules and running ore secrets rekey.
+"%,
+      actor      = "human",
+      created_at = "2026-05-03",
+      tags       = ["credentials", "templates", "onboarding", "adoption"],
+      related    = ["adr-017"],
+      verified   = true,
+    },
+    {
+      id         = "credential-vault-troubleshooting",
+      question   = "What do the named errors from secrets.nu mean and how to recover?",
+      answer     = m%"
+The credential helper raises 15 named errors. Each maps to a recovery action:
+
+[invalid-op]                op must be 'pull' or 'push'. Caller bug.
+[project-ncl-missing]       Run from a project with .ontoref/project.ncl, or set
+                            ONTOREF_PROJECT_ROOT to that path.
+[manifest-ncl-missing]      Apply migration 0016 — add registry_provides to
+                            .ontology/manifest.ncl.
+[registry-provides-missing] Same as above — manifest needs registry_provides block.
+[registry-id-unknown]       Pass --registry-id matching a declared entry, or
+                            declare registries.default in the manifest.
+[credential-sops-missing]   The chosen RegistryEntry has no credential_sops/_rw
+                            for the requested op. Add the field.
+[sops-file-not-found]       Vault not synced. Run: ore secrets sync <vault_id>.
+[kage-not-resolvable]       Master key absent. Set vault.master_key_path globally
+                            or sops.master_key_path per project.
+[sops-decrypt-failed]       Your .kage is not a recipient of the file, or it is
+                            corrupt. Verify with: ore secrets describe.
+[actor-bindings-missing]    project.ncl::sops.actor_key_bindings is empty. Map
+                            at least the actors used in this project.
+[actor-not-bound]           ONTOREF_ACTOR has no entry in actor_key_bindings.
+                            Set the env var or add the mapping.
+[actor-not-in-bound-actor]  scope.bound_actor list excludes this actor. Either
+                            change actor or relax the scope.
+[scope-not-loaded]          Scope file missing — vault not synced or never
+                            created. Run: ore secrets sync <vault_id>.
+[op-not-in-scope]           The role's scope.ops does not allow this operation.
+                            Use a higher-privilege role or extend scope.
+[target-not-in-scope]       The OCI ref does not match any scope.namespaces
+                            glob. Operate on a permitted target or extend scope.
+
+Errors are raised before any registry call — no operation is half-completed.
+"%,
+      actor      = "human",
+      created_at = "2026-05-03",
+      tags       = ["credentials", "errors", "troubleshooting", "secrets"],
+      related    = ["adr-017"],
+      verified   = true,
+    },
+    {
+      id         = "integration-what-and-why",
+      question   = "What is integration in ontoref and why use it?",
+      answer     = m%"
+WHAT
+  Integration is the federated distribution of two kinds of artifacts via an
+  OCI registry (typically zot):
+
+    DOMAIN ARTIFACTS  application/vnd.ontoref.domain.v1
+      A domain is a Nickel contract (contract.ncl) describing the typed shape
+      of a piece of structured data — e.g. 'registry-access', 'secret-delivery',
+      'compute-provisioning'. Pushed at domains/<participant>/<id>:<semver>.
+
+    MODE ARTIFACTS    application/vnd.ontoref.mode.v1
+      A mode is an operational orchestration (provisioning.ncl + domains.lock.ncl)
+      that consumes one or more domain contracts to perform a workflow — e.g.
+      'lian-build/provisioning'. Pushed at modes/<participant>/<id>:<semver>.
+
+  Both are cosign-signed and immutable per version.
+
+WHY
+  - DECOUPLE producers from consumers. The team that defines the contract for
+    'registry-access' does not need to coordinate with every workspace that
+    consumes it. Versioning (semver) handles compatibility.
+  - REUSE across projects. A mode author writes one mode artifact; multiple
+    workspaces subscribe with different cabling files binding their own values.
+  - VERIFIABLE TRUST. cosign signatures + multi-recipient sops (ADR-017)
+    establish who published an artifact and who can read its credentials.
+  - DAG-FORMALIZED. Modes declare domains_used + steps as a typed graph;
+    consumers can statically verify their cabling resolves all bindings.
+
+DATA FLOW
+
+  producer project (libre-wuji)         registry (zot)            consumer project
+  ─────────────────────────────         ────────────────          ────────────────
+  catalog/domains/registry-access/
+    contract.ncl              ─push──► domains/libre-wuji/
+    example.json                       registry-access:0.1.0
+                                       (cosigned)
+  catalog/modes/lian-build/
+    provisioning.ncl          ─push──► modes/lian-build/
+    domains.lock.ncl                   provisioning:0.1.0
+                                       (cosigned)
+                                              │
+                                              │ oras pull
+                                              ▼
+                                                              infra/<ws>/integrations/
+                                                              lian-build.ncl
+                                                                (cabling — binds
+                                                                 mode params to
+                                                                 workspace values)
+
+  consumer commands:
+    prvng integration subscribe lian-build --mode-file ... --workspace-dir .
+    prvng integration validate  lian-build --workspace-dir .
+    prvng integration invoke    lian-build --binary lian-build
+
+REFERENCES
+  - ADR-042 (provisioning workspace) — federation model
+  - reflection/migrations/0015        — registry topology adoption
+  - install/resources/templates/integration/ — copy-paste templates
+  - reflection/qa.ncl::ontoref-three-layer-model
+                                      — domains and modes are LAYER 2 of a
+                                        project's ontoref instance (the
+                                        integration surface). This entry
+                                        focuses on the federation mechanism;
+                                        the layering entry frames where the
+                                        artifacts sit relative to a project's
+                                        self-management ontoref and to
+                                        caller-side wiring.
+"%,
+      actor      = "human",
+      created_at = "2026-05-03",
+      tags       = ["integration", "oci", "domains", "modes", "federation", "layer-2"],
+      related    = ["adr-015", "adr-017"],
+      verified   = true,
+    },
+    {
+      id         = "integration-how-to-implement",
+      question   = "How do I implement integration in my project — as producer, consumer, or both?",
+      answer     = m%"
+Two roles, often the same project plays both. Pick the side and follow the
+predefined paths.
+
+PRODUCER SIDE — publish a domain or mode artifact
+
+  Path P1 — DOMAIN PUBLISHER (you author a contract.ncl others should consume)
+    Template: install/resources/templates/integration/domain-producer/
+    Steps:
+      1. Create catalog/domains/<id>/ with:
+           contract.ncl   — typed shape of the domain (Nickel contract)
+           example.json   — sample value matching the contract
+           manifest.ncl   — DomainArtifact descriptor (id, version, layers)
+      2. Declare uses_registry in manifest.ncl::DomainArtifact pointing to the
+         RegistryEntry that hosts pushes (ADR-017 G2 impact analysis).
+      3. ore secrets bootstrap (one-time per project)
+      4. prvng integration domain publish catalog/domains/<id> <participant>
+         (cosign-signs at push time)
+    Validation:
+      prvng integration domain verify <participant>/<id> <version>
+      → checks media types, layers, contract typecheck, cosign signature
+
+  Path P2 — MODE PUBLISHER (you author a mode that orchestrates other modes)
+    Template: install/resources/templates/integration/mode-producer/
+    Steps:
+      1. Create catalog/modes/<mode-id>/ with:
+           provisioning.ncl  — the IntegrationMode declaration
+                                 (id, participant, direction, trigger,
+                                  domains_used, steps)
+           domains.lock.ncl  — pinned domain versions consumed
+      2. prvng integration mode publish catalog/modes/<mode-id> <participant> <mode-id> <version>
+
+CONSUMER SIDE — bind a published mode to your workspace
+
+  Path C1 — MODE SUBSCRIBER (you adopt someone's published mode)
+    Template: install/resources/templates/integration/mode-consumer/
+    Steps:
+      1. prvng integration subscribe <mode-id> --mode-file <path-to-mode> --workspace-dir .
+         → pulls all domains_used dependencies, verifies signatures,
+           scaffolds infra/<ws>/integrations/<mode-id>.ncl (the cabling)
+      2. Edit the cabling file to bind mode parameters to workspace values
+         (e.g. dns zones, tenant ids, registry endpoints from your manifest).
+      3. prvng integration validate <mode-id> --workspace-dir .
+         → typechecks the cabling and confirms every binding resolves.
+      4. prvng integration invoke <mode-id> --binary <name>
+         → assembles context envelope + pipes to the mode binary stdin.
+
+PREDEFINED INTEGRATION MODES (canonical examples in libre-wuji)
+
+  modes/cloudatasave/provisioning   data-save workflow with backup-policy-binding
+                                    + result-reporting domains
+  modes/lian-build/provisioning     CI build pipeline using compute-provisioning
+                                    + cache-management + secret-delivery domains
+
+  These are reference implementations — clone the structure, adapt domains_used,
+  re-publish under your participant.
+
+CABLING FILE STRUCTURE
+
+  infra/<workspace>/integrations/<mode-id>.ncl
+
+    let mode = import \"modes/<participant>/<mode-id>:<version>\" in
+    {
+      mode_id   = mode.id,
+      bindings = {
+        # Match each domain in mode.domains_used. Resolve via:
+        #   - workspace state (manifest, capabilities)
+        #   - secret-delivery (pulls from credential vault)
+        #   - registry-access (zot endpoint + namespace policy)
+        \"<domain-id>\" = { ... fields per the domain's contract.ncl ... },
+      },
+    }
+
+WITHOUT TEMPLATES — minimal viable
+
+  Producer:  manually create contract.ncl + example.json + manifest.ncl in any
+             dir; cosign keypair; prvng integration domain publish
+  Consumer:  manually pull the OCI artifact; write cabling.ncl from scratch
+             matching the mode's domains_used schema; invoke
+"%,
+      actor      = "human",
+      created_at = "2026-05-03",
+      tags       = ["integration", "templates", "producer", "consumer", "subscribe"],
+      related    = ["adr-015", "adr-017"],
+      verified   = true,
+    },
+    {
+      id         = "integration-troubleshooting",
+      question   = "Common errors when working with integration artifacts — what causes them and how to fix?",
+      answer     = m%"
+ON PUSH
+
+  oras push: 403 Forbidden
+    Cause: ZOT ACL does not declare the target namespace (e.g. domains/<x>/
+           when registry config only allows domains/**).
+    Fix:   Add the namespace to zot configmap creation_rules and redeploy.
+
+  cosign sign: 401 Unauthorized
+    Cause: cosign needs DOCKER_CONFIG to fetch the manifest before signing.
+    Fix:   Ensure the calling code wraps cosign with the same isolated
+           DOCKER_CONFIG used for the oras push.
+
+  cosign sign: --tlog-upload=false is not supported with --signing-config
+    Cause: cosign 2+ deprecated --tlog-upload. Use a signing-config without
+           rekorTlogUrls/Config instead.
+    Fix:   Generate one with:
+             curl https://raw.githubusercontent.com/sigstore/root-signing/refs/heads/main/targets/signing_config.v0.2.json \\
+               | jq 'del(.rekorTlogUrls) | del(.rekorTlogConfig)' > signing-config-no-rekor.json
+           Set vault.cosign.signing_config_path in ~/.config/ontoref/config.ncl.
+
+ON PULL / SUBSCRIBE
+
+  Vault artifact signature FAILED
+    Cause: cosign pubkey configured does not match the signing key.
+    Fix:   Confirm vault.cosign.pub_path in ~/.config/ontoref/config.ncl points
+           to the public half of the keypair used for the push. Verify the
+           tlog policy is symmetric (both sign and verify expect tlog=false).
+
+  domain-pull: scope-not-loaded
+    Cause: Vault not synced — scopes/<role>.ncl absent from local src-vault.
+    Fix:   ore secrets sync <vault_id>
+
+  oras pull: not found
+    Cause: Mismatch between expected ref format. Old flat domains/<id> are not
+           reachable via the new domains/<participant>/<id> path.
+    Fix:   Either re-publish the artifact under the participant-scoped path,
+           or pass --registry overriding to a registry that has the legacy ref.
+
+ON INVOKE
+
+  integration validate: <domain-id> binding does not resolve
+    Cause: Cabling.ncl has a binding whose value cannot be derived from
+           workspace state at validate time.
+    Fix:   Inspect prvng integration describe <mode-id> to see the expected
+           shape; ensure your cabling provides the matching field with the
+           required type (use prvng i validate --strict for hard fail).
+
+  integration invoke: binary not found
+    Cause: Mode declares Invocation.binary.source = 'oci_blob but the blob
+           reference is unreachable, or 'cargo_install but cargo crate is
+           absent in the local registry index.
+    Fix:   Pre-fetch the binary: docker pull <oci_layer> or cargo install
+           <crate>. Or pass --binary <path> to override resolution.
+
+REFERENCES
+  - prvng integration --help        full command surface
+  - reflection/qa.ncl::integration-* this FAQ entry tree
+  - reflection/migrations/0015      participant-scoped namespace migration
+"%,
+      actor      = "human",
+      created_at = "2026-05-03",
+      tags       = ["integration", "errors", "troubleshooting", "cosign", "oras"],
+      related    = ["adr-015", "adr-017"],
+      verified   = true,
+    },
+    {
+      id         = "nats-what-and-why",
+      question   = "Why does ontoref use NATS, and what role does it play for projects that adopt the protocol?",
+      answer     = m%"
+WHAT
+  ontoref uses NATS JetStream as its async event substrate. The daemon publishes
+  lifecycle events; the CLI receives notifications; projects publish domain-
+  specific events (build started/completed, sync done, integration invoked).
+  All events flow through a single JetStream stream with a typed subject
+  hierarchy.
+
+  TOPOLOGY (default — see nats/streams.json)
+
+    Stream:    ECOSYSTEM
+      subjects:    ecosystem.>
+      retention:   Limits  (max_age = 30 days)
+      storage:     File    (durable across restarts)
+
+    Consumers (pull, explicit ack):
+      daemon-ontoref       filters: ecosystem.daemon.>
+                                    ecosystem.actor.>
+                                    ecosystem.ontoref.>
+      cli-notifications    filters: ecosystem.ontoref.>
+                                    ecosystem.actor.>
+
+  SUBJECT HIERARCHY
+
+    ecosystem.daemon.<event>           daemon lifecycle (started, reload, ...)
+    ecosystem.ontoref.<scope>.<event>  protocol events (sync.done, mode.run, ...)
+    ecosystem.actor.<actor>.<event>    per-actor session/audit
+    ecosystem.<project>.<scope>.<evt>  project-specific events
+                                       (e.g. ecosystem.lian-build.build.completed)
+
+WHY ontoref uses NATS
+  - DECOUPLE the daemon from the CLI. The daemon publishes; CLIs subscribe.
+    No HTTP polling, no shared filesystem state, no blocking RPCs. Either side
+    can restart without dropping the other.
+  - DURABILITY across restarts. JetStream's File storage means a CLI launched
+    after a daemon event still receives it via consumer replay. ack_policy
+    Explicit lets each consumer track its own position.
+  - MULTI-PROJECT VISIBILITY. The shared ECOSYSTEM stream and ecosystem.>
+    subject root let any project publish or subscribe without negotiating a
+    dedicated stream. Tenancy lives in the subject hierarchy, not in
+    stream-per-project sprawl.
+  - GRACEFUL DEGRADATION. NATS is a runtime-toggle service (ADR-014):
+    nats_events.enabled = false in .ontoref/config.ncl shuts publishing off
+    cleanly. Consumers see no events; the daemon and CLI continue working.
+    Same is true if NATS is unreachable at startup — connection failure logs
+    a warning and the publisher returns Ok(None).
+
+WHY a project adopting ontoref should publish to it
+  - AUDIT TRAIL by subscribing once and recording. ecosystem.<project>.>
+    captures every lifecycle event a project emits without bespoke logging
+    pipelines.
+  - CROSS-PROJECT COORDINATION. A workspace's CI pipeline can wait for
+    ecosystem.lian-build.build.completed before triggering a deploy step,
+    rather than polling an HTTP API or watching a filesystem.
+  - FAN-OUT FOR FREE. Multiple consumers (oncall dashboard, audit log,
+    notification UI, downstream pipeline) subscribe to the same subject
+    without the producer knowing.
+  - PROTOCOL ALIGNMENT. ADR-014 defines NATS as one of three runtime services
+    (NATS, SurrealDB, src-vault). Projects that adopt ontoref get the same
+    enable/disable mechanics for free; turn it on later without code changes.
+
+CONFIGURATION SHAPE
+  Global (every ontoref-onboarded host):
+    ~/.config/ontoref/config.ncl::nats_events
+                          .url      = "nats://..."
+                          .enabled  = true | false
+                          .nkey_seed = "..."  (optional)
+    ~/.config/ontoref/streams.json   (stream + consumer topology)
+
+  Project-local override (a project that wants its own stream):
+    <project>/.ontoref/config.ncl::nats_events.streams_config = "nats/streams.json"
+    <project>/nats/streams.json      (overrides the global topology)
+
+  Most projects accept the global topology; the project-local override is for
+  the rare case of a project needing isolated streams (e.g. high-volume
+  internal events that shouldn't share retention with ECOSYSTEM).
+
+REFERENCES
+  - nats/streams.json                 default ECOSYSTEM topology
+  - crates/ontoref-daemon/src/nats.rs daemon-side NatsPublisher
+  - adr-002                           daemon as notification barrier
+  - adr-014                           runtime service toggles (NATS as one)
+  - reflection/qa.ncl::nats-*         this FAQ entry tree
+  - reflection/qa.ncl::ontoref-three-layer-model
+                                      — the subjects a project publishes are
+                                        a LAYER 2 surface (other projects
+                                        subscribe to coordinate). The broker
+                                        and ECOSYSTEM stream are protocol-level
+                                        infrastructure, not project Layer 2,
+                                        so the layering crisply separates
+                                        WHAT a project emits (Layer 2) from
+                                        WHERE the events flow (protocol).
+"%,
+      actor      = "human",
+      created_at = "2026-05-03",
+      tags       = ["nats", "jetstream", "events", "ecosystem", "architecture", "layer-2"],
+      related    = ["adr-002", "adr-014"],
+      verified   = true,
+    },
+    {
+      id         = "nats-how-to-setup",
+      question   = "How do I set up NATS for ontoref, and how does a project plug into the event system?",
+      answer     = m%"
+Three deployment shapes — pick the lowest one your threat model accepts. All
+three satisfy the daemon's connection requirements.
+
+OPTION A — Local nats-server, no auth (lowest barrier, dev only)
+  Best for: laptop development, ad-hoc testing of a project's event publishing.
+
+  Install:
+    macOS:    brew install nats-server
+    Linux:    curl -sf https://binaries.nats.dev/nats-io/nats-server/v2/install.sh | sh
+
+  Run (foreground, ^C to stop):
+    nats-server -DV -js
+  The -js flag enables JetStream so the ECOSYSTEM stream can be created.
+
+  Wire into ontoref:
+    Edit ~/.config/ontoref/config.ncl::nats_events
+      enabled = true,
+      url     = "nats://127.0.0.1:4222",
+      # nkey_seed unset — anonymous client
+    Restart ontoref-daemon (it picks up the new config on next bootstrap).
+
+  Bootstrap the topology (one-time per server):
+    nats --server nats://127.0.0.1:4222 stream add --config nats/streams.json
+    nats --server nats://127.0.0.1:4222 consumer add ECOSYSTEM \\
+         --config '<consumer block from streams.json>'
+  (Or let the daemon's TopologyConfig apply it on first connect — see
+  crates/ontoref-daemon/src/nats.rs::connect.)
+
+OPTION B — Local nats-server with NKey auth (matches deployed shape)
+  Best for: validating the NKey code path before deploying; production-shape
+  testing without k8s overhead.
+
+  Generate an NKey:
+    go install github.com/nats-io/nkeys/nk@latest
+    nk -gen user > /tmp/ontoref.user.nk
+    nk -inkey /tmp/ontoref.user.nk -pubout > /tmp/ontoref.user.pub
+
+  Configure /tmp/nats-server.conf:
+    port: 4222
+    jetstream: enabled
+    authorization {
+      users = [{
+        nkey: "<paste /tmp/ontoref.user.pub contents>"
+        permissions {
+          publish   { allow: ["ecosystem.>"] }
+          subscribe { allow: ["ecosystem.>", "_INBOX.>"] }
+        }
+      }]
+    }
+
+  Run:
+    nats-server -c /tmp/nats-server.conf
+
+  Wire into ontoref:
+    Edit ~/.config/ontoref/config.ncl::nats_events
+      enabled   = true,
+      url       = "nats://127.0.0.1:4222",
+      nkey_seed = "<paste /tmp/ontoref.user.nk contents>",
+
+  Note: platform-nats hardcodes require_signed_messages=false. The broker
+  authenticates the client by NKey identity but does not require per-message
+  signatures. This matches the deployed pattern.
+
+OPTION C — Production via provisioning's nats component
+  Best for: shared workspaces, persistent JetStream state, multi-actor.
+
+  The reusable component lives at:
+    <provisioning>/catalog/components/nats/
+      metadata.ncl       (name=nats, version=2.10, mode=cluster, JetStream)
+      cluster/manifest_plan.ncl
+      nickel/{main,defaults,contracts}.ncl
+
+  Defaults (defaults.ncl):
+    port=4222, monitor_port=8222, mode='cluster, image=nats:2.10-alpine
+    jetstream.max_mem=256MB, jetstream.max_file=1GB
+    storage 1Gi persistent
+
+  Deploy:
+    prvng workspace install nats        # add to workspace component DAG
+    prvng workspace apply <ws-id>       # execute
+    prvng workspace status <ws-id>      # confirm pod Running
+
+  Wire into ontoref (in-cluster vs port-forward):
+    In-cluster:
+      url       = "nats://nats.<ws-namespace>.svc.cluster.local:4222"
+      nkey_seed = $(kubectl get secret -n <ns> ontoref-nkey -o jsonpath='{.data.seed}' | base64 -d)
+
+    Out-of-cluster (port-forward for ad-hoc):
+      kubectl port-forward -n <ns> svc/nats 4222:4222 &
+      url       = "nats://127.0.0.1:4222"
+      nkey_seed = (same as above)
+
+PROJECT-SIDE: enabling event publishing in a project
+  In <project>/.ontoref/config.ncl:
+
+    nats_events = {
+      enabled   = true,
+      url       = "nats://127.0.0.1:4222",       # or workspace URL
+      nkey_seed = std.env.get "NATS_NKEY_SEED",  # or null for anonymous
+      emit      = ["ecosystem.<project>.<scope>.>"],   # subjects this project publishes
+      subscribe = ["ecosystem.ontoref.>"],              # subjects this project consumes
+    }
+
+  Subject discipline: prefix every emitted subject with
+    ecosystem.<your-project-slug>.<scope>.<event>
+  The slug matches .ontoref/project.ncl::slug. Scope is project-internal
+  (e.g. lian-build uses 'build' for build lifecycle: ecosystem.lian-build.build.started).
+
+  Project-local stream override (rare):
+    Add nats/streams.json at the project root with the same shape as ontoref's
+    global, then set nats_events.streams_config = "nats/streams.json" in the
+    project's config.ncl. The daemon applies this topology on next connect
+    instead of inheriting the global ECOSYSTEM stream.
+
+VERIFY end-to-end
+  Subscribe in one terminal:
+    nats sub --server $NATS_URL 'ecosystem.>'
+
+  Trigger an event from another terminal:
+    ontoref --actor developer sync .       # daemon emits ecosystem.ontoref.sync.*
+    OR run a project that publishes (e.g. lian-build integrate)
+
+  The subscriber prints the JSON envelope within milliseconds. If it does not,
+  see reflection/qa.ncl::nats-troubleshooting.
+
+REFERENCES
+  - nats/streams.json                 default topology — copy as starting point
+  - ~/.config/ontoref/config.ncl      where nats_events is configured
+  - <provisioning>/catalog/components/nats   Option C source
+  - crates/ontoref-daemon/src/nats.rs daemon connect + topology apply
+  - reflection/qa.ncl::nats-what-and-why  rationale and architecture
+"%,
+      actor      = "human",
+      created_at = "2026-05-03",
+      tags       = ["nats", "setup", "configuration", "nkey", "onboarding"],
+      related    = ["adr-002", "adr-014", "adr-017"],
+      verified   = true,
+    },
+    {
+      id         = "ontoref-three-layer-model",
+      question   = "When I see a project's ontoref instance, what am I actually looking at? What are the three layers, and how do they not mix?",
+      answer     = m%"
+A project's ontoref instance has THREE distinct layers, each with its own
+audience, lifecycle, and validation rule. They share the project's repo but
+do not share content. This entry exists so adopters discover the layering
+from protocol documentation, not by re-deriving it across three projects.
+
+LAYER 1 — Self-management ontoref (about the project itself)
+
+  paths     .ontology/  reflection/  adrs/
+  audience  this project's developers and maintainers
+  purpose   describe the project to itself — axioms, FSM dimensions,
+            binding decisions (ADRs), open questions (backlog), accepted
+            knowledge (qa)
+  required  YES, on every ontoref-onboarded project. `ontoref setup` creates
+            it.
+
+  Concrete files:
+    .ontology/core.ncl       axioms, tensions, practices, edges
+    .ontology/state.ncl      FSM dimensions for project maturity
+    .ontology/gate.ncl       membranes gating state transitions
+    .ontology/manifest.ncl   project metadata, layers, capabilities,
+                             registry_provides (binds Layer 1 to Layer 2)
+    reflection/qa.ncl        accepted knowledge as typed Q&A entries
+    reflection/backlog.ncl   open items routed to graduation targets
+    reflection/modes/        the project's own integration-mode declarations
+    adrs/adr-NNN-*.ncl       binding decisions with typed constraint checks
+
+LAYER 2 — Specialized domain/mode ontoref (the integration surface)
+
+  paths     schemas/  catalog/{domains,modes}/
+            manifest.ncl::registry_provides
+  audience  OTHER projects that want to integrate this project
+  purpose   the contract surface other projects bind to — typed domains,
+            orchestration modes, registry-namespace claim. Lives in this
+            project so the schemas and the binary stay in lock-step.
+  required  OPTIONAL. A pure consumer of the protocol (no published
+            artifacts) skips Layer 2. A federated peer (publishes
+            domains/<participant>/* or modes/<participant>/*) has all of it.
+
+  Concrete files (when present):
+    schemas/<contract>.ncl                         typed domain contracts
+    catalog/domains/<id>/manifest.ncl              OCI DomainArtifact
+    catalog/domains/<id>/contract.ncl              re-export of schema
+    catalog/domains/<id>/example.json              canonical instance
+    catalog/modes/<id>/provisioning.ncl            ModeArtifact + steps DAG
+    catalog/modes/<id>/domains.lock.ncl            pinned domain digests
+    manifest.ncl::registry_provides.{participant,registries}
+                                                   namespace claim + cred refs
+
+LAYER 3 — Caller-side implementations (NOT in this project)
+
+  paths     <caller>/extensions/<this-project>/
+            <caller>/catalog/components/<...>/   (when consuming)
+            <workspace>/infra/<ws>/integrations/
+  audience  operators and CI of caller projects
+  purpose   wiring this project into a specific workspace or pipeline —
+            cabling values to mode steps, declaring which workspace
+            components consume which artifacts
+  present   PER CALLER, NEVER in this project's repo
+
+  This layer is OUTPUT of the integration relationship, not input to it.
+  The project does not ship Layer 3 content for itself; callers ship it
+  for their own use. Cross-references from this project to Layer 3 are
+  explicit pointers ("see <caller>/extensions/..."), never copy-paste.
+
+WHY THE LAYERING MATTERS — three concrete failure modes when mixed
+
+  Layer 3 in Layer 1
+    A project's reflection/qa.ncl carries entries titled "How do I plug
+    <project> into workspace X?" Other callers reading the FAQ get
+    nothing useful; they're looking at one workspace's wiring instead
+    of the project itself.
+
+  Layer 2 in Layer 1
+    Schemas live under reflection/ or .ontology/ as "project notes".
+    They drift from the binary because they're not on the contract path.
+    Consumers pull a contract artifact from the registry and find it
+    out of sync.
+
+  Layer 1 in Layer 2
+    A catalog/domains/<id>/contract.ncl carries architectural rationale
+    instead of a typed shape. Consumers pull and get text where they
+    expected a Nickel contract; integrations break.
+
+CROSS-LAYER REFERENCES — explicit, never implicit
+
+  Within a project: tag qa entries that touch a Layer 3 boundary with
+  "layer-3-boundary" so the boundary is visible. Touch implies "points
+  out to a caller", not "documents the caller's wiring".
+
+  Between projects: cross-link via id (e.g. lian-build's
+  reflection/backlog.ncl::bl-002 mirrors ontoref's
+  reflection/backlog.ncl::bl-007). Never duplicate content; refer.
+
+  To the protocol: the protocol's own qa entries (this file) describe
+  the layering generically; project-level qa entries reference them
+  ("see <ontoref>/reflection/qa.ncl::ontoref-three-layer-model") rather
+  than restate.
+
+DETECTION — quick checks to see what a project has
+
+  Layer 1 present:
+    test -f .ontology/core.ncl && test -f reflection/qa.ncl && \
+      test -d adrs/ && echo "L1 present"
+  Layer 2 present:
+    test -d catalog/ && rg -q 'registry_provides' manifest.ncl && \
+      echo "L2 present"
+  Layer 3 absence (correctness check):
+    rg -l 'extensions/.*/cabling\|infra/.*/integrations/' \
+       --type-not ncl --type-not md . | head
+    Should return empty — Layer 3 belongs to callers, not this project.
+
+STATUS — protocol-level codification
+
+  The model is observed in practice (ontoref + lian-build + provisioning)
+  but not yet codified as a protocol ADR with enforceable constraints.
+  See reflection/backlog.ncl::bl-009 for the open codification question
+  including four constraint candidates (Layer 1 mandatory, Layer 2
+  biconditional, Layer 3 isolation, cross-layer tag convention).
+
+  Open question flagged in bl-009: how does this layering interact with
+  ADR-018's level hierarchy (Base / Domain / Instance)? Likely orthogonal
+  axes (3-layer × 3-level matrix), but unresolved until the ADR drafts.
+
+REFERENCES
+  - reflection/backlog.ncl::bl-009     codification work item
+  - adr-018-level-hierarchy-mode-resolution-strategy   open interaction
+  - lian-build/reflection/qa.ncl::lian-build-what-and-why   worked example
+                                       of all three layers as observed
+  - lian-build/manifest.ncl            Layer 2 example (registry_provides)
+  - lian-build/catalog/                Layer 2 example (domains + modes)
+"%,
+      actor      = "human",
+      created_at = "2026-05-03",
+      tags       = ["ontoref", "layers", "architecture", "adoption", "scope", "boundaries"],
+      related    = ["adr-018-level-hierarchy-mode-resolution-strategy"],
+      verified   = true,
+    },
+    {
+      id         = "nats-troubleshooting",
+      question   = "My ontoref/project NATS publishing isn't working — how do I diagnose it?",
+      answer     = m%"
+Symptoms and fixes. Apply in order; each step rules out a category.
+
+(1) Daemon logs "NATS connect failed" or "events disabled"
+    Cause: nats_events.enabled = false, or url unreachable, or NKey rejected.
+    Diagnose:
+      - cat ~/.config/ontoref/config.ncl | rg -A5 'nats_events'
+      - Check enabled = true.
+      - curl -sf http://<host>:8222/healthz   (broker monitor port)
+        Should return {"status":"ok"}; if not, the broker is unreachable.
+    Fix:
+      - Wrong URL: edit url, restart daemon.
+      - Broker down: start nats-server (Option A/B) or check workspace pod.
+      - NKey mismatch: regenerate user pub key, update nats-server.conf.
+
+(2) Subscriber sees no events
+    Cause: subject prefix mismatch, or publisher silently dropping (warn-only
+    degradation hides failures).
+    Diagnose:
+      - In another terminal, subscribe to the wildcard root:
+          nats sub --server $NATS_URL 'ecosystem.>'
+        If THIS receives events, your filter was wrong.
+      - Inspect the publisher's stderr — platform-nats logs the resolved
+        subject on each publish at INFO level. Compare to your subscribe filter.
+      - JetStream consumers ack-once: a consumer that already ack'd a message
+        won't redeliver it. Check consumer info:
+          nats consumer info ECOSYSTEM <consumer-name>
+    Fix:
+      - Subject mismatch: align the subscribe pattern to what's published.
+      - Consumer stuck: nats consumer rm + re-add, or use a fresh
+        consumer name to start from latest.
+      - Warn-only drops: set RUST_LOG=warn and re-run; look for "NATS publish
+        failed".
+
+(3) "no responders available for request" or stream missing
+    Cause: ECOSYSTEM stream not created on the broker.
+    Diagnose:
+      - nats stream ls --server $NATS_URL
+        Should list ECOSYSTEM. If empty, the topology was never applied.
+    Fix:
+      - Re-run topology bootstrap (see nats-how-to-setup OPTION A "Bootstrap
+        the topology"), or restart ontoref-daemon — its connect() applies
+        nats/streams.json on first call.
+
+(4) NKey decode error / "invalid seed"
+    Cause: seed format wrong (must be the SU... prefixed value, not the public
+    UD... or a JWT).
+    Diagnose:
+      - echo $NATS_NKEY_SEED | head -c 4   # should print 'SU' (user seed)
+                                            # or 'SA' (account), 'SO' (operator)
+    Fix:
+      - Regenerate: nk -gen user > /tmp/ontoref.user.nk; use the WHOLE file
+        contents. Do not concatenate the .pub file.
+
+(5) "permissions violation for publish"
+    Cause: broker's authorization block restricts publish subjects; the
+    project is publishing outside its allowed namespace.
+    Diagnose:
+      - Inspect nats-server.conf (Option B) or the workspace's NATS auth config
+        (Option C — kubectl get configmap -n <ns> nats-server-conf).
+      - Check the user's permissions.publish.allow list against your subject.
+    Fix:
+      - Widen the allow list to include ecosystem.<your-project>.> — or
+        better, use a per-project user with scoped permissions.
+
+(6) Project's events appear in daemon log but never on NATS
+    Cause: the project is using a different NATS_URL than the daemon, or its
+    nats_events.enabled is false.
+    Diagnose:
+      - Compare the project's resolved NATS_URL (its stderr at startup) to
+        the daemon's. They must point at the same broker if they share the
+        ECOSYSTEM stream.
+    Fix:
+      - Project-local override is intentional? Inspect
+        <project>/.ontoref/config.ncl::nats_events. If unintentional, remove
+        the override; the daemon's global config applies.
+
+(7) Events received but with stale timestamps / out of order
+    Cause: JetStream's File storage replays unacked messages on consumer
+    reconnect. This is a feature, not a bug — explicit-ack consumers are
+    expected to handle redelivery.
+    Fix:
+      - Code your subscribers to be idempotent; use the message's correlation_id
+        / event_id (when present) to deduplicate.
+      - If strict ordering matters: set MaxAckPending = 1 on the consumer.
+
+REFERENCES
+  - reflection/qa.ncl::nats-what-and-why         architecture
+  - reflection/qa.ncl::nats-how-to-setup         setup paths
+  - crates/ontoref-daemon/src/nats.rs            daemon connect logic
+  - <stratumiops>/crates/platform-nats/          NatsConnectionConfig shape
+  - https://docs.nats.io/running-a-nats-service/troubleshooting   upstream docs
+"%,
+      actor      = "human",
+      created_at = "2026-05-03",
+      tags       = ["nats", "errors", "troubleshooting", "jetstream", "nkey"],
+      related    = ["adr-002", "adr-014"],
+      verified   = true,
+    },
+    {
+      id         = "ontoref-dao-discipline",
+      question   = "What is ondaod and when does an architectural analysis or ADR draft have to apply it?",
+      answer     = m%"
+ALIAS  ondaod — shorthand for this discipline in conversation, ADR drafts,
+       CLAUDE.md rules, and reflection mode declarations.
+
+WHAT
+  Discipline applied to architectural analysis in any ontoref-onboarded
+  project: read named tensions before recommending; describe synthesis state,
+  not pick a pole; name engaged tensions explicitly. The named tensions
+  in .ontology/core.ncl are continuous Spirals, not binaries — recommendations
+  that collapse them by choosing one side silently bias toward whichever pole
+  the analyst's reasoning happened to land on.
+
+WHEN  (triggered)
+  - Architectural analysis — any reasoning that produces a recommendation
+    about structure, naming, layout, contracts, or constraints.
+  - ADR drafting — every ADR must apply ondaod alongside the four-criterion
+    ADR test (alternative-rejected, lasting-constraint, multi-component-
+    reversal, not-duplicating-existing).
+  - Work touching .ontology/, adrs/, reflection/, catalog/, manifest.ncl.
+
+WHEN  (NOT triggered)
+  - Routine code work — bug fixes, feature implementation, refactors that
+    don't change architecture.
+  - Operational tasks — CI runs, commits, tests.
+  - Pure data extraction — querying, reading.
+
+HOW  (procedure)
+  1. READ      .ontology/core.ncl; locate `level = 'Tension` nodes.
+  2. IDENTIFY  which named tensions the question engages. Empty set is
+               allowed but must be declared:
+                 tensions_engaged: []  # no Spiral tensions present
+  3. CHARACTERIZE  the synthesis state of each engaged tension:
+                   - Where on the flow?  (claim-only / populating /
+                                          realized / consumed / dormant)
+                   - What direction is the project moving?
+                                         (toward Yang / toward Yin / static)
+  4. RECOMMEND  the move that maintains continuous flow, not the move that
+                collapses the tension. Half-states are partially-realized
+                syntheses, not violations.
+  5. CITE       engaged tensions explicitly in the output (prose paragraph
+                or structured field).
+
+WHY
+  Default reasoning — human or agent — is Yang: sequential, decide-and-
+  commit. This loses the continuous flow that the named tensions exist to
+  surface. Recommendations that don't name engaged tensions silently bias
+  toward whichever pole reasoning happened to land on. ondaod surfaces the
+  bias structurally so analysts and reviewers can correct for it before
+  the bias compounds across decisions.
+
+FORBIDDEN PATTERNS  (most common ondaod violations)
+
+  - Pole-collapse recommendations: "pick option A" / "go with B" / "the
+    right answer is X" — without naming what got collapsed.
+
+  - Reality-collapses-intent: "drop the declared claim because the catalog
+    is empty" — that erases Yin intent rather than describing partial
+    realization. The half-state is the project's current location on a
+    continuous flow, not a contract violation.
+
+  - Hard-biconditional on Spiral questions: ADR constraints with severity
+    = 'Hard and `A ⟺ B` checks on questions that core.ncl already names as
+    'Spiral tensions. Use 'Soft constraints that report direction of motion.
+
+  - Yang-bias by sequence: when serial reasoning produces "best option per
+    question", that IS the bias. Counter: characterize all questions'
+    synthesis states first, recommend last (or not at all).
+
+ADR INTEGRATION  (criterion 5)
+
+  The four-criterion `adr?` test extends to FIVE when ondaod applies:
+
+    1. alternative consciously rejected?
+    2. lasting constraints future contributors must follow?
+    3. reversing requires coordinated effort across multiple components?
+    4. not already captured as a constraint in an existing ADR?
+    5. ondaod — engaged named tensions identified and synthesis state
+                described, OR explicitly tensions_engaged: [] with rationale
+                "no Spiral tensions present in this decision"?
+
+  All five must hold. Failing 5 means the question is tractable (1-4 hold)
+  but the analysis hasn't characterized the flow — request synthesis-state
+  description before drafting the ADR.
+
+  A 'Spiral-poled ADR (when the schema permits the field — see bl-009
+  graduation) cannot use severity = 'Hard biconditional constraints; Spiral
+  decisions get Soft constraints reporting direction of motion.
+
+ACCESS PATHS  (how agents and humans reach this entry)
+
+  From any ontoref-onboarded project, via the canonical `ontoref` CLI:
+
+    ontoref qa show ontoref-dao-discipline
+      Auto-emits JSON when invoked by an agent (agent-identified context);
+      humans see the formatted output. Force one or the other explicitly
+      with --fmt json | -f json (or --fmt md). Available on every `ontoref`
+      subcommand that returns structured data.
+
+  Direct file read (last resort, no CLI required):
+
+    $ONTOREF_ROOT/reflection/qa.ncl::ontoref-dao-discipline
+
+  This entry is canonical and not duplicated into consumer projects. The
+  discipline applies to consumer projects via reference; the content lives
+  here once. Each consumer project carries its OWN dao discipline entry as
+  an extension (e.g. lian-build/reflection/qa.ncl::lian-build-dao-discipline)
+  that names the project's specific tensions and forbidden patterns; both
+  the protocol baseline (this entry) and the project extension are read
+  together when applying ondaod.
+
+REFERENCES
+  - .ontology/core.ncl::read-tensions-first   Practice node (structural anchor)
+  - reflection/qa.ncl::ontoref-three-layer-model    worked example of
+                                                     Yang-bias and Spiral re-frame
+  - reflection/backlog.ncl::bl-009            three-layer model graduation
+                                              (depends on ondaod-disciplined
+                                               analysis)
+  - global ~/.claude/CLAUDE.md::adr?          extended four-→-five criterion
+                                              definition references this entry
+"%,
+      actor      = "human",
+      created_at = "2026-05-03",
+      tags       = ["ontoref", "dao", "discipline", "ondaod", "tensions", "spiral", "adr-process", "meta"],
+      related    = ["adr-018-level-hierarchy-mode-resolution-strategy"],
+      verified   = true,
+    },
+    {
+      id         = "credential-vault-disaster-recovery",
+      question   = "What if I lose my .kage, my vault_key, the access.sops.yaml file, or the entire local vault directory?",
+      answer     = m%"
+The credential vault has multiple recovery paths depending on what survives.
+None requires re-bootstrapping unless the catastrophic case (all local + all
+keys lost) hits. Verified empirically 2026-05-03 — the access.sops.yaml lost
+recovery via oras pull works as documented.
+
+LOSS MATRIX
+
+  Lost                              Recovery
+  ────                              ────────
+  access.sops.yaml                  oras pull src-vault/<id>:latest from ZOT.
+  (corrupted/deleted)               cp the file from the artifact into
+                                    ~/.config/ontoref/vaults/<id>/.
+
+  Local restic repo (/repo dir)     oras pull restores src-vault/ subtree
+                                    (scopes, registry, logs).
+
+  vault_key alone                   Decrypt access.sops.yaml with .kage,
+                                    extract vault_key field.
+
+  Your .kage alone                  Peer recipient decrypts; share vault_key
+                                    out-of-band. Generate new .kage; add via
+                                    'ore secrets add-key' + 'rekey'.
+
+  Both local AND your .kage         Use a peer recipient's .kage to pull and
+  (other recipients alive)          decrypt. Generate new .kage; add new
+                                    pubkey via add-key.
+
+  Both local AND ALL recipients'    Catastrophic. The encrypted artifact in
+  .kage files                       ZOT is unrecoverable by design.
+                                    Re-bootstrap; reissue all registry
+                                    credentials from zot admin surface.
+
+CONCRETE RECOVERY (access.sops.yaml lost — verified)
+
+  See justfiles/secrets.just::secrets-recover (or call ore directly):
+
+    ore secrets recover --from-registry        # pulls and restores access.sops.yaml
+                                               # using current zot credentials
+                                               # from your project.ncl.
+
+  Manually, the equivalent oras invocation is documented in
+  justfiles/_secrets_lib.sh::vault_zot_config_open + an oras pull. Steps:
+
+    1. Build a DOCKER_CONFIG tmpdir with admin zot credentials
+    2. oras pull <registry>/src-vault/<vault-id>:latest --output <tmp>
+    3. cp <tmp>/access.sops.yaml ~/.config/ontoref/vaults/<vault-id>/
+    4. ore secrets status — should report 'access.sops: present'
+
+BACKUP STRATEGY
+
+  master .kage              Hardware key (Yubikey via age-plugin-yubikey),
+                            encrypted disk, or password manager with file
+                            attachment. Multi-recipient sops makes the
+                            recipient list itself the resilience layer.
+  vault_key                 Encrypted inside access.sops.yaml — recoverable
+                            via .kage. Backup is automatic.
+  cosign signing key        Separate (vault.cosign.key_path). Treat as a
+                            standalone private key — backup independently.
+
+INVARIANTS
+
+  - Multi-recipient sops mandatory (≥ 2 per file) — no single point of failure.
+  - access.sops.yaml is in the OCI artifact — pulling restores it intact.
+  - cosign verification on pull detects substituted artifacts.
+  - Daemon never holds credentials — daemon recovery is independent.
+
+WHAT NOT TO DO
+
+  - Do NOT skip cosign verification on pull.
+  - Do NOT rotate vault_key proactively — it is the local restic repo
+    password, not a public-service credential.
+  - Do NOT re-bootstrap to skip recovery — fresh bootstrap loses audit.jsonl
+    history and the participant's src-vault history in ZOT.
+"%,
+      actor      = "human",
+      created_at = "2026-05-03",
+      tags       = ["credentials", "recovery", "disaster", "operations", "backup"],
+      related    = ["adr-017", "adr-019"],
+      verified   = true,
+    },
+  ],
 } | s.QaStore
diff --git a/reflection/requirements/base.ncl b/reflection/requirements/base.ncl
index 932d424..8247ca4 100644
--- a/reflection/requirements/base.ncl
+++ b/reflection/requirements/base.ncl
@@ -175,5 +175,71 @@ let s = import "./schema.ncl" in
       used_by_modes   = ["new_project"],
     } | s.Tool),
 
+    # ── OCI credential and artifact layer (ADR-017) ───────────────────────────
+
+    ({
+      id              = "oras",
+      binary          = "oras",
+      description     = "OCI Registry As Storage client. Required to push/pull domain artifacts, mode artifacts, and src-vault OCI artifacts. Primary mechanism for ontoref OCI discovery.",
+      version_min     = "1.2.0",
+      check_cmd       = "oras version",
+      version_extract = "Version:\\s+(\\d+\\.\\d+\\.\\d+)",
+      install_hint    = "brew install oras  # or: https://oras.land/docs/installation",
+      docs_url        = "https://oras.land",
+      severity        = 'Hard,
+      contexts        = ["local_dev", "ci", "agent"],
+      used_by_forms   = [],
+      used_by_modes   = ["new_project"],
+      config_hint     = "All oras calls use an isolated DOCKER_CONFIG tmpdir (ADR-017) — never ambient ~/.docker/config.json.",
+    } | s.Tool),
+
+    ({
+      id              = "cosign",
+      binary          = "cosign",
+      description     = "Artifact signing and verification. Required to sign src-vault OCI artifacts on push and verify on pull. Prevents artifact substitution attacks on credential distribution.",
+      version_min     = "2.0.0",
+      check_cmd       = "cosign version",
+      version_extract = "GitVersion:\\s+(\\d+\\.\\d+\\.\\d+)",
+      install_hint    = "brew install cosign  # or: https://docs.sigstore.dev/cosign/system_config/installation/",
+      docs_url        = "https://docs.sigstore.dev/cosign",
+      severity        = 'Hard,
+      contexts        = ["local_dev", "ci"],
+      used_by_forms   = [],
+      used_by_modes   = [],
+      config_hint     = "COSIGN_KEY_PATH env var or --cosign-key flag selects the signing key. Unsigned src-vault artifacts are rejected by ore secrets pull.",
+    } | s.Tool),
+
+    ({
+      id              = "sops",
+      binary          = "sops",
+      description     = "Secrets OPerationS — multi-recipient encrypted file editor. Encrypts/decrypts access.sops.yaml containing registry credentials and vault_key.",
+      version_min     = "3.8.0",
+      check_cmd       = "sops --version",
+      version_extract = "sops (\\d+\\.\\d+\\.\\d+)",
+      install_hint    = "brew install sops  # or: https://github.com/getsops/sops/releases",
+      docs_url        = "https://getsops.io",
+      severity        = 'Hard,
+      contexts        = ["local_dev", "ci"],
+      used_by_forms   = [],
+      used_by_modes   = [],
+      config_hint     = "SOPS_AGE_KEY_FILE must point to the .kage master key (declared in project.ncl sops.master_key_path). Never inside vault dir.",
+    } | s.Tool),
+
+    ({
+      id              = "age",
+      binary          = "age-keygen",
+      description     = "age encryption key generator. Required to provision new actor keypairs for multi-recipient sops vaults.",
+      version_min     = "1.1.0",
+      check_cmd       = "age-keygen --version",
+      version_extract = "v(\\d+\\.\\d+\\.\\d+)",
+      install_hint    = "brew install age  # or: https://github.com/FiloSottile/age/releases",
+      docs_url        = "https://age-encryption.org",
+      severity        = 'Soft,
+      contexts        = ["local_dev"],
+      used_by_forms   = [],
+      used_by_modes   = [],
+      config_hint     = "Only needed when generating new keypairs (just secrets-gen-key). Decryption uses sops which bundles its own age support.",
+    } | s.Tool),
+
   ],
 }
diff --git a/reflection/schema.ncl b/reflection/schema.ncl
index 0bdd872..a277e16 100644
--- a/reflection/schema.ncl
+++ b/reflection/schema.ncl
@@ -62,6 +62,11 @@ let _ActionStep = fun ActionContract => {
   note       | String | optional,
 } in
 
+# Resolution strategy for multi-level hierarchy traversal (ADR-018).
+# Required at Domain (level 2) and Instance (level 3) — absent at Base (level 1).
+# Implicit absence at level 2+ is treated as 'Delegate with a Soft validation warning.
+let _ResolutionStrategy = [| 'Override, 'Delegate, 'Merge, 'Compose |] in
+
 let _ModeBase = fun ActionContract => {
   id             | String,
   trigger        | String,
@@ -70,6 +75,13 @@ let _ModeBase = fun ActionContract => {
   steps          | Array (_ActionStep ActionContract),
   postconditions | Array String | default = [],
   converge       | _Converge | optional,
+  # ADR-018: which level owns this mode's implementation.
+  # 'Override = complete here; 'Delegate = traverse to parent; 'Merge = accumulate;
+  # 'Compose = partial inheritance declared in `extends`.
+  strategy       | _ResolutionStrategy | optional,
+  # When strategy = 'Compose: step or field ids to inherit from the parent mode.
+  # Every reference must exist in the parent — a broken extends is always a bug.
+  extends        | Array String | optional,
 } in
 
 # DAG-validated Mode contract:
@@ -112,10 +124,11 @@ let _Mode = fun ActionContract =>
 in
 
 {
-  Dependency  = _Dependency,
-  OnError     = _OnError,
-  Guard       = _Guard,
-  Converge    = _Converge,
-  ActionStep  = _ActionStep,
-  Mode        = _Mode,
+  Dependency         = _Dependency,
+  OnError            = _OnError,
+  Guard              = _Guard,
+  Converge           = _Converge,
+  ActionStep         = _ActionStep,
+  Mode               = _Mode,
+  ResolutionStrategy = _ResolutionStrategy,
 }
diff --git a/reflection/schemas/workflow.ncl b/reflection/schemas/workflow.ncl
index 5ba46aa..dba2d7e 100644
--- a/reflection/schemas/workflow.ncl
+++ b/reflection/schemas/workflow.ncl
@@ -33,7 +33,13 @@ let build_spec = {
   args          | Array String | default = [],
   cargo_profile | String       | default = "",
   target        | String       | default = "",   # cargo target triple; "" = native
-  artifacts     | Array String | default = [],   # expected output paths
+  artifacts     | Array String | default = [],   # expected output paths.
+                                                  # Supported placeholders:
+                                                  #   {TARGET_DIR}  → cargo `target_directory` (resolved via `cargo metadata`)
+                                                  #   {ROOT}        → workspace root
+                                                  # Use placeholders for paths that depend on the cargo target dir,
+                                                  # since it may be redirected via .cargo/config.toml `target-dir`.
+                                                  # Resolve via `reflection/modules/workflow.nu` `resolve-artifact-path`.
 } in
 
 # Distribution spec — where to send artifacts after build.
diff --git a/reflection/tests/test_credential_chain.sh b/reflection/tests/test_credential_chain.sh
new file mode 100755
index 0000000..1363693
--- /dev/null
+++ b/reflection/tests/test_credential_chain.sh
@@ -0,0 +1,199 @@
+#!/usr/bin/env bash
+# Full credential chain smoke (ADR-017 + ADR-019).
+#
+# Drives the end-to-end path against a fixture project + the live ZOT registry:
+#
+#   1. Bootstrap a throwaway project with multi-tenant recipient_rules
+#   2. Encryption respects the rules (.sops.yaml generated; access.sops uses
+#      its rule's recipients; per-tenant files use their own subset)
+#   3. Push the vault to ZOT (cosign signed, tlog=false)
+#   4. Pull from an isolated HOME — proves DOCKER_CONFIG isolation
+#   5. Decrypt a credential via assert-actor-authorized + resolve-registry-credential
+#   6. Cleanup — registry tags deleted, fixture removed
+#
+# Run:
+#   RECOVERY_ZOT_USER=<user> RECOVERY_ZOT_PASS=<pass> \
+#       bash reflection/tests/test_credential_chain.sh
+#
+# Exit 0 on green; non-zero on any failure.
+
+set -euo pipefail
+
+ONTOREF_ROOT=${ONTOREF_ROOT:-/Users/Akasha/Development/ontoref}
+ZOT_HOST=${ZOT_HOST:-reg.librecloud.online}
+ZOT_USER=${ZOT_USER:-${RECOVERY_ZOT_USER:-}}
+ZOT_PASS=${ZOT_PASS:-${RECOVERY_ZOT_PASS:-}}
+
+[[ -n "$ZOT_USER" ]] || { echo "ZOT_USER (or RECOVERY_ZOT_USER) required"; exit 2; }
+[[ -n "$ZOT_PASS" ]] || { echo "ZOT_PASS (or RECOVERY_ZOT_PASS) required"; exit 2; }
+
+# Use the developer's existing master key + cosign config.
+# The fixture project will declare master_key_path = developer's .kage.
+DEV_KAGE=${DEV_KAGE:-$HOME/.age/keys/ontoref.key.txt}
+[[ -f "$DEV_KAGE" ]] || { echo "DEV_KAGE not found at $DEV_KAGE"; exit 2; }
+DEV_PUB=$(grep "public key" "$DEV_KAGE" | head -1 | awk '{print $NF}')
+
+PASS=0
+FAIL=0
+FIXTURE_SLUG="ccsmoke-$(date +%s)"
+FIXTURE=$(mktemp -d)
+FIXTURE_HOME=""
+
+cleanup() {
+    local rc=$?
+    [[ -n "$FIXTURE_HOME" ]] && rm -rf "$FIXTURE_HOME"
+    rm -rf "$FIXTURE" "$HOME/.config/ontoref/vaults/$FIXTURE_SLUG"
+    # Best-effort: delete the fixture's pushed artifacts from ZOT.
+    local cfg
+    cfg=$(mktemp -d)
+    local auth
+    auth=$(printf '%s:%s' "$ZOT_USER" "$ZOT_PASS" | base64)
+    printf '{"auths":{"%s":{"auth":"%s"}}}' "$ZOT_HOST" "$auth" > "$cfg/config.json"
+    DOCKER_CONFIG="$cfg" oras manifest delete -f "$ZOT_HOST/src-vault/$FIXTURE_SLUG:latest" 2>/dev/null || true
+    DOCKER_CONFIG="$cfg" oras manifest delete -f "$ZOT_HOST/src-vault/$FIXTURE_SLUG:lock"   2>/dev/null || true
+    rm -rf "$cfg"
+    [[ $rc -eq 0 ]] && echo "── credential chain smoke: $PASS passed, $FAIL failed ──" \
+                     || echo "── ABORTED at PASS=$PASS FAIL=$FAIL rc=$rc ──"
+    return $rc
+}
+trap cleanup EXIT
+
+assert() {
+    local name="$1"; shift
+    if "$@"; then
+        echo "  ✓ $name"
+        PASS=$((PASS + 1))
+    else
+        echo "  ✗ $name"
+        FAIL=$((FAIL + 1))
+    fi
+}
+
+# ────────── Phase 1: build fixture project ──────────────────────────────────
+echo "phase 1: build fixture project"
+
+mkdir -p "$FIXTURE/.ontoref" "$FIXTURE/.ontology"
+cat > "$FIXTURE/.ontoref/project.ncl" <<NCL
+let s = import "ontoref-project.ncl" in
+s.make_project {
+  slug = "$FIXTURE_SLUG", root = "$FIXTURE", nickel_import_paths = [], keys = [],
+  sops = {
+    enabled = true, vault_id = "$FIXTURE_SLUG",
+    master_key_path = "$DEV_KAGE",
+    registry_endpoint = "$ZOT_HOST",
+    actor_key_bindings = { developer = "developer" },
+    recipient_groups = { admin = ["$DEV_PUB"], developer = ["$DEV_PUB"] },
+    recipient_rules = [
+      { path = "access\\\\.sops\\\\.yaml\$",                groups = ["admin", "developer"] },
+      { path = "registry/admin-.*\\\\.sops\\\\.yaml\$",     groups = ["admin"] },
+      { path = "registry/developer-.*\\\\.sops\\\\.yaml\$", groups = ["admin", "developer"] },
+    ],
+  },
+}
+NCL
+
+cat > "$FIXTURE/.ontology/manifest.ncl" <<'NCL'
+let m = import "ontology/defaults/manifest.ncl" in
+m.make_manifest {
+  project = "ccsmoke", repo_kind = 'Mixed,
+  description = "credential chain smoke fixture",
+  consumption_modes = [],
+  registry_provides = m.make_registry_provides {
+    participant = "ccsmoke",
+    registries = m.make_registries_config {
+      default = "primary",
+      registries = [ m.make_registry_entry {
+        id = "primary",
+        endpoint = "ZOT_HOST_PLACEHOLDER",
+        credential_sops    = "registry/developer-ro.sops.yaml",
+        credential_sops_rw = "registry/developer-rw.sops.yaml",
+      } ],
+    },
+  },
+}
+NCL
+sed -i.bak "s|ZOT_HOST_PLACEHOLDER|$ZOT_HOST|" "$FIXTURE/.ontology/manifest.ncl" && rm "$FIXTURE/.ontology/manifest.ncl.bak"
+
+assert "fixture project.ncl renders" \
+    bash -c "NICKEL_IMPORT_PATH='$FIXTURE:$ONTOREF_ROOT:$ONTOREF_ROOT/ontology:$HOME/.config/ontoref/schemas' nickel export '$FIXTURE/.ontoref/project.ncl' >/dev/null 2>&1"
+
+# ────────── Phase 2: bootstrap (declarative mode) ────────────────────────────
+echo "phase 2: bootstrap"
+
+ONTOREF_PROJECT_ROOT="$FIXTURE" \
+NICKEL_IMPORT_PATH="$FIXTURE:$ONTOREF_ROOT:$ONTOREF_ROOT/ontology:$HOME/.config/ontoref/schemas" \
+    bash -c "cd $ONTOREF_ROOT && printf '%s\n%s\n%s\n%s\n' '$ZOT_USER' '$ZOT_PASS' 'restic-pwd-test' 'cosign-pwd-test' \
+        | just --justfile justfiles/secrets.just secrets-bootstrap" >/dev/null 2>&1
+
+assert "access.sops.yaml created" \
+    [ -f "$HOME/.config/ontoref/vaults/$FIXTURE_SLUG/access.sops.yaml" ]
+assert ".sops.yaml generated (declarative mode)" \
+    [ -f "$HOME/.config/ontoref/vaults/$FIXTURE_SLUG/.sops.yaml" ]
+assert "default ro/rw skipped (declarative paths defined by operator)" \
+    [ ! -f "$HOME/.config/ontoref/vaults/$FIXTURE_SLUG/src-vault/registry/ro.sops.yaml" ]
+
+# ────────── Phase 3: populate per-tenant credential ──────────────────────────
+echo "phase 3: write per-tenant credential file"
+
+VAULT_DIR="$HOME/.config/ontoref/vaults/$FIXTURE_SLUG"
+mkdir -p "$VAULT_DIR/src-vault/registry"
+DEV_RO="$VAULT_DIR/src-vault/registry/developer-ro.sops.yaml"
+DEV_RW="$VAULT_DIR/src-vault/registry/developer-rw.sops.yaml"
+PLAIN=$(mktemp)
+printf 'username: %s\npassword: %s\n' "$ZOT_USER" "$ZOT_PASS" > "$PLAIN"
+SOPS_AGE_KEY_FILE="$DEV_KAGE" sops --config "$VAULT_DIR/.sops.yaml" \
+    --filename-override "$DEV_RO" --input-type yaml --encrypt "$PLAIN" > "$DEV_RO"
+SOPS_AGE_KEY_FILE="$DEV_KAGE" sops --config "$VAULT_DIR/.sops.yaml" \
+    --filename-override "$DEV_RW" --input-type yaml --encrypt "$PLAIN" > "$DEV_RW"
+rm -f "$PLAIN"
+
+assert "developer-ro.sops.yaml encrypted" \
+    [ -s "$DEV_RO" ]
+assert "developer-rw.sops.yaml encrypted" \
+    [ -s "$DEV_RW" ]
+
+# ────────── Phase 4: push to ZOT ────────────────────────────────────────────
+echo "phase 4: push to registry"
+
+# Push uses developer's cosign key from global config.
+COSIGN_PASSWORD="cosign-pwd-test" \
+ONTOREF_PROJECT_ROOT="$FIXTURE" \
+NICKEL_IMPORT_PATH="$FIXTURE:$ONTOREF_ROOT:$ONTOREF_ROOT/ontology:$HOME/.config/ontoref/schemas" \
+    bash -c "cd $ONTOREF_ROOT && just --justfile justfiles/secrets.just secrets-push" \
+    > /tmp/cc-push.log 2>&1 || true
+
+assert "push succeeded (manifest in zot)" \
+    test -n "$(curl -sf -u "$ZOT_USER:$ZOT_PASS" "https://$ZOT_HOST/v2/src-vault/$FIXTURE_SLUG/tags/list" 2>/dev/null \
+        | grep -o latest)"
+
+# ────────── Phase 5: isolated-HOME pull ─────────────────────────────────────
+echo "phase 5: pull from isolated HOME"
+
+FIXTURE_HOME=$(mktemp -d)
+mkdir -p "$FIXTURE_HOME/.config/ontoref"
+ln -s "$HOME/.config/ontoref/vaults"     "$FIXTURE_HOME/.config/ontoref/vaults"
+ln -s "$HOME/.config/ontoref/config.ncl" "$FIXTURE_HOME/.config/ontoref/config.ncl"
+ln -s "$HOME/.config/ontoref/cosign"     "$FIXTURE_HOME/.config/ontoref/cosign"
+ln -s "$HOME/.config/ontoref/schemas"    "$FIXTURE_HOME/.config/ontoref/schemas"
+
+# Use sops + oras directly with a tmp DOCKER_CONFIG built from the vault credentials.
+TMPCFG=$(mktemp -d)
+auth=$(printf '%s:%s' "$ZOT_USER" "$ZOT_PASS" | base64)
+printf '{"auths":{"%s":{"auth":"%s"}}}' "$ZOT_HOST" "$auth" > "$TMPCFG/config.json"
+
+PULLED=$(mktemp -d)
+HOME="$FIXTURE_HOME" DOCKER_CONFIG="$TMPCFG" \
+    oras pull "$ZOT_HOST/src-vault/$FIXTURE_SLUG:latest" --output "$PULLED" >/dev/null 2>&1
+
+assert "pull restored access.sops.yaml" \
+    [ -f "$PULLED/access.sops.yaml" ]
+
+assert "decrypts under isolated HOME (master_key resolves)" \
+    bash -c "HOME=\"$FIXTURE_HOME\" SOPS_AGE_KEY_FILE=\"$DEV_KAGE\" sops --decrypt --extract '[\"zot_username\"]' \"$PULLED/access.sops.yaml\" >/dev/null 2>&1"
+
+rm -rf "$PULLED" "$TMPCFG"
+
+# ────────── Phase 6: cleanup happens in trap ────────────────────────────────
+echo "phase 6: cleanup deferred to trap"
+
+[[ "$FAIL" -eq 0 ]]
diff --git a/reflection/tests/test_secrets.nu b/reflection/tests/test_secrets.nu
new file mode 100644
index 0000000..389e220
--- /dev/null
+++ b/reflection/tests/test_secrets.nu
@@ -0,0 +1,355 @@
+#!/usr/bin/env nu
+# Tests for reflection/modules/secrets.nu — the 15 named errors that form the
+# helper contract. Each test builds a minimal fixture (project_root with
+# .ontoref/project.ncl and/or .ontology/manifest.ncl) and asserts the helper
+# raises the expected `[<code>]` prefix.
+#
+# Run:  nu reflection/tests/test_secrets.nu
+# Exit: 0 if all pass, 1 if any fail.
+
+use ../modules/secrets.nu *
+
+const ONTOREF_ROOT = "/Users/Akasha/Development/ontoref"
+
+# Self-contained import path — the secrets module reads $env.NICKEL_IMPORT_PATH
+# when invoking `nickel export`. Each fixture's manifest.ncl imports
+# `ontology/defaults/manifest.ncl`, which lives under $ONTOREF_ROOT/ontology.
+# We append (rather than overwrite) so a caller's existing path still applies.
+$env.NICKEL_IMPORT_PATH = (
+    [
+        ($env.NICKEL_IMPORT_PATH? | default "")
+        $ONTOREF_ROOT
+        $"($ONTOREF_ROOT)/ontology"
+        ($env.HOME | path join ".config" "ontoref" "schemas")
+    ] | where { |s| ($s | str length) > 0 } | str join (char esep)
+)
+
+# ── Test runner ──────────────────────────────────────────────────────────────
+
+mut TOTAL  = 0
+mut PASSED = 0
+mut FAILED = []
+
+def expect-error [
+    name:    string
+    code:    string   # e.g. "kage-not-resolvable"
+    body:    closure
+]: nothing -> record {
+    let r = (try {
+        do $body
+        { kind: "no-error", msg: "" }
+    } catch { |e|
+        { kind: "error", msg: $e.msg }
+    })
+    let prefix = $"[($code)]"
+    if $r.kind == "no-error" {
+        { ok: false, msg: $"expected error ($prefix) but body returned without error" }
+    } else if ($r.msg | str starts-with $prefix) {
+        { ok: true, msg: "" }
+    } else {
+        let snippet = ($r.msg | str substring 0..200)
+        { ok: false, msg: $"expected ($prefix) got: ($snippet)" }
+    }
+}
+
+# ── Fixture builders ─────────────────────────────────────────────────────────
+
+def write-project-ncl [
+    root:    string
+    body:    string   # the body of make_project { ... }
+]: nothing -> nothing {
+    mkdir ($root | path join ".ontoref")
+    let schema_dir = ($env.HOME | path join ".config" "ontoref" "schemas")
+    let import = $'let s = import "ontoref-project.ncl" in
+'
+    let content = $"($import)
+s.make_project {
+($body)
+}"
+    $content | save --force ($root | path join ".ontoref" "project.ncl")
+}
+
+def write-manifest-ncl [
+    root:    string
+    body:    string   # additional fields beyond the required boilerplate
+]: nothing -> nothing {
+    mkdir ($root | path join ".ontology")
+    # Minimum required fields by ontology/defaults/manifest.ncl::make_manifest.
+    let boilerplate = ([
+        '  project           = "fixture",',
+        ("  repo_kind         = " + (char single_quote) + "Mixed,"),
+        '  description       = "test fixture",',
+        '  consumption_modes = [],',
+    ] | str join (char newline))
+    let content = ([
+        'let m = import "ontology/defaults/manifest.ncl" in',
+        '',
+        'm.make_manifest {',
+        $boilerplate,
+        $body,
+        '}',
+    ] | str join (char newline))
+    $content | save --force ($root | path join ".ontology" "manifest.ncl")
+}
+
+def fixture-base-project [root: string, sops_block: string]: nothing -> nothing {
+    write-project-ncl $root $'  slug                = "fixture",
+  root                = "($root)",
+  nickel_import_paths = ["($root)", "($ONTOREF_ROOT)", "($ONTOREF_ROOT)/ontology"],
+  keys                = [],
+  ($sops_block)'
+}
+
+# ── Tests ────────────────────────────────────────────────────────────────────
+
+let TMP = (mktemp -d -t ontoref-tests.XXXXXX)
+print $"# fixtures in ($TMP)"
+print ""
+
+# Each test runs in its own subdir of $TMP so fixtures don't collide.
+def run-one [name: string, code: string, body: closure]: nothing -> record {
+    let r = (expect-error $name $code $body)
+    if $r.ok {
+        print $"  (ansi green)✓(ansi reset) ($name)  [($code)]"
+    } else {
+        print $"  (ansi red)✗(ansi reset) ($name)  [($code)]  →  ($r.msg)"
+    }
+    $r
+}
+
+# --- 1. [invalid-op] ------------------------------------------------------------
+let r1 = (run-one "invalid op" "invalid-op" {
+    let f = ($TMP | path join "t1")
+    fixture-base-project $f 'sops = { enabled = false }'
+    write-manifest-ncl $f ''
+    resolve-registry-credential $f primary --op badname
+})
+$TOTAL = $TOTAL + 1; if $r1.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "invalid-op") }
+
+# --- 2. [project-ncl-missing] ---------------------------------------------------
+# Use assert-actor-authorized which reads project.ncl first;
+# resolve-registry-credential reads manifest first and would shadow this case.
+let r2 = (run-one "project-ncl missing" "project-ncl-missing" {
+    let f = ($TMP | path join "t2-empty")
+    mkdir $f
+    assert-actor-authorized $f pull
+})
+$TOTAL = $TOTAL + 1; if $r2.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "project-ncl-missing") }
+
+# --- 3. [manifest-ncl-missing] --------------------------------------------------
+let r3 = (run-one "manifest.ncl missing" "manifest-ncl-missing" {
+    let f = ($TMP | path join "t3")
+    fixture-base-project $f 'sops = { enabled = true, vault_id = "fixture", master_key_path = "/dev/null" }'
+    resolve-registry-credential $f primary --op pull
+})
+$TOTAL = $TOTAL + 1; if $r3.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "manifest-ncl-missing") }
+
+# --- 4. [registry-provides-missing] --------------------------------------------
+let r4 = (run-one "registry_provides missing" "registry-provides-missing" {
+    let f = ($TMP | path join "t4")
+    fixture-base-project $f 'sops = { enabled = true, vault_id = "fixture", master_key_path = "/dev/null" }'
+    write-manifest-ncl $f ''
+    resolve-registry-credential $f primary --op pull
+})
+$TOTAL = $TOTAL + 1; if $r4.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "registry-provides-missing") }
+
+# --- 5. [registry-id-unknown] --------------------------------------------------
+let r5 = (run-one "registry id unknown" "registry-id-unknown" {
+    let f = ($TMP | path join "t5")
+    fixture-base-project $f 'sops = { enabled = true, vault_id = "fixture", master_key_path = "/dev/null" }'
+    write-manifest-ncl $f '  registry_provides = m.make_registry_provides {
+    participant = "fixture",
+    registries = m.make_registries_config {
+      default = "primary",
+      registries = [ m.make_registry_entry { id = "primary", endpoint = "x", credential_sops = "registry/ro.sops.yaml" } ],
+    },
+  },'
+    resolve-registry-credential $f does-not-exist --op pull
+})
+$TOTAL = $TOTAL + 1; if $r5.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "registry-id-unknown") }
+
+# --- 6. [credential-sops-missing] ----------------------------------------------
+let r6 = (run-one "credential_sops missing for op" "credential-sops-missing" {
+    let f = ($TMP | path join "t6")
+    fixture-base-project $f 'sops = { enabled = true, vault_id = "fixture", master_key_path = "/dev/null" }'
+    write-manifest-ncl $f '  registry_provides = m.make_registry_provides {
+    participant = "fixture",
+    registries = m.make_registries_config {
+      default = "primary",
+      registries = [ m.make_registry_entry { id = "primary", endpoint = "x", credential_sops = "registry/ro.sops.yaml" } ],
+    },
+  },'
+    # credential_sops_rw is absent, op=push → should fail
+    resolve-registry-credential $f primary --op push
+})
+$TOTAL = $TOTAL + 1; if $r6.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "credential-sops-missing") }
+
+# --- 7. [sops-file-not-found] --------------------------------------------------
+let r7 = (run-one "sops file not on disk" "sops-file-not-found" {
+    let f = ($TMP | path join "t7")
+    fixture-base-project $f $'sops = { enabled = true, vault_id = "fixture-t7-($r6.ok)", master_key_path = "/dev/null" }'
+    write-manifest-ncl $f '  registry_provides = m.make_registry_provides {
+    participant = "fixture",
+    registries = m.make_registries_config {
+      default = "primary",
+      registries = [ m.make_registry_entry { id = "primary", endpoint = "x", credential_sops = "registry/ro.sops.yaml" } ],
+    },
+  },'
+    # Vault dir for "fixture-t7-..." does not exist → sops file not found
+    resolve-registry-credential $f primary --op pull
+})
+$TOTAL = $TOTAL + 1; if $r7.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "sops-file-not-found") }
+
+# --- 8. [kage-not-resolvable] (master_key_path file missing) -------------------
+let r8 = (run-one "kage not resolvable" "kage-not-resolvable" {
+    let f = ($TMP | path join "t8")
+    let bogus_kage = ($TMP | path join "non-existent.kage")
+    fixture-base-project $f $'sops = { enabled = true, vault_id = "fixture-t8", master_key_path = "($bogus_kage)" }'
+    write-manifest-ncl $f '  registry_provides = m.make_registry_provides {
+    participant = "fixture",
+    registries = m.make_registries_config {
+      default = "primary",
+      registries = [ m.make_registry_entry { id = "primary", endpoint = "x", credential_sops = "registry/ro.sops.yaml" } ],
+    },
+  },'
+    # Pre-create the sops file so we get past sops-file-not-found.
+    let vault_dir = ($env.HOME | path join ".config" "ontoref" "vaults" "fixture-t8" "src-vault" "registry")
+    mkdir $vault_dir
+    "fake" | save --force ($vault_dir | path join "ro.sops.yaml")
+    resolve-registry-credential $f primary --op pull
+})
+$TOTAL = $TOTAL + 1; if $r8.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "kage-not-resolvable") }
+# Cleanup
+rm -rf ($env.HOME | path join ".config" "ontoref" "vaults" "fixture-t8")
+
+# --- 9. [actor-bindings-missing] -----------------------------------------------
+let r9 = (run-one "actor_key_bindings empty" "actor-bindings-missing" {
+    let f = ($TMP | path join "t9")
+    fixture-base-project $f 'sops = { enabled = true, vault_id = "fixture-t9", master_key_path = "/dev/null", actor_key_bindings = {} }'
+    write-manifest-ncl $f ''
+    assert-actor-authorized $f pull
+})
+$TOTAL = $TOTAL + 1; if $r9.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "actor-bindings-missing") }
+
+# --- 10. [actor-not-bound] -----------------------------------------------------
+let r10 = (run-one "actor not in bindings" "actor-not-bound" {
+    let f = ($TMP | path join "t10")
+    fixture-base-project $f 'sops = { enabled = true, vault_id = "fixture-t10", master_key_path = "/dev/null", actor_key_bindings = { admin = "admin" } }'
+    write-manifest-ncl $f ''
+    with-env { ONTOREF_ACTOR: "alien" } {
+        assert-actor-authorized $f pull
+    }
+})
+$TOTAL = $TOTAL + 1; if $r10.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "actor-not-bound") }
+
+# --- 11. [scope-not-loaded] ----------------------------------------------------
+let r11 = (run-one "scope file missing" "scope-not-loaded" {
+    let f = ($TMP | path join "t11")
+    fixture-base-project $f 'sops = { enabled = true, vault_id = "fixture-t11", master_key_path = "/dev/null", actor_key_bindings = { developer = "developer" } }'
+    write-manifest-ncl $f ''
+    # No scopes/developer.ncl in vault dir
+    with-env { ONTOREF_ACTOR: "developer" } {
+        assert-actor-authorized $f pull
+    }
+})
+$TOTAL = $TOTAL + 1; if $r11.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "scope-not-loaded") }
+
+# --- 12. [op-not-in-scope] -----------------------------------------------------
+let r12 = (run-one "op not in scope.ops" "op-not-in-scope" {
+    let f = ($TMP | path join "t12")
+    fixture-base-project $f 'sops = { enabled = true, vault_id = "fixture-t12", master_key_path = "/dev/null", actor_key_bindings = { developer = "developer" } }'
+    write-manifest-ncl $f ''
+    let scope_dir = ($env.HOME | path join ".config" "ontoref" "vaults" "fixture-t12" "src-vault" "scopes")
+    mkdir $scope_dir
+    ("{ role = \"developer\", access = 'ro, bound_actor = [], namespaces = [\"domains/*/\"], ops = ['pull, 'verify] }"
+        | save --force ($scope_dir | path join "developer.ncl"))
+    with-env { ONTOREF_ACTOR: "developer" } {
+        assert-actor-authorized $f push
+    }
+})
+$TOTAL = $TOTAL + 1; if $r12.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "op-not-in-scope") }
+rm -rf ($env.HOME | path join ".config" "ontoref" "vaults" "fixture-t12")
+
+# --- 13. [actor-not-in-bound-actor] --------------------------------------------
+let r13 = (run-one "scope.bound_actor excludes actor" "actor-not-in-bound-actor" {
+    let f = ($TMP | path join "t13")
+    fixture-base-project $f 'sops = { enabled = true, vault_id = "fixture-t13", master_key_path = "/dev/null", actor_key_bindings = { developer = "developer" } }'
+    write-manifest-ncl $f ''
+    let scope_dir = ($env.HOME | path join ".config" "ontoref" "vaults" "fixture-t13" "src-vault" "scopes")
+    mkdir $scope_dir
+    ("{ role = \"developer\", access = 'rw, bound_actor = [\"admin\"], namespaces = [\"domains/*/\"], ops = ['pull, 'push] }"
+        | save --force ($scope_dir | path join "developer.ncl"))
+    with-env { ONTOREF_ACTOR: "developer" } {
+        assert-actor-authorized $f pull
+    }
+})
+$TOTAL = $TOTAL + 1; if $r13.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "actor-not-in-bound-actor") }
+rm -rf ($env.HOME | path join ".config" "ontoref" "vaults" "fixture-t13")
+
+# --- 14. [target-not-in-scope] -------------------------------------------------
+let r14 = (run-one "target not in scope.namespaces" "target-not-in-scope" {
+    let f = ($TMP | path join "t14")
+    fixture-base-project $f 'sops = { enabled = true, vault_id = "fixture-t14", master_key_path = "/dev/null", actor_key_bindings = { developer = "developer" } }'
+    write-manifest-ncl $f ''
+    let scope_dir = ($env.HOME | path join ".config" "ontoref" "vaults" "fixture-t14" "src-vault" "scopes")
+    mkdir $scope_dir
+    ("{ role = \"developer\", access = 'rw, bound_actor = [], namespaces = [\"domains/foo/\"], ops = ['pull, 'push] }"
+        | save --force ($scope_dir | path join "developer.ncl"))
+    with-env { ONTOREF_ACTOR: "developer" } {
+        assert-target-in-scope $f "modes/lian-build/"
+    }
+})
+$TOTAL = $TOTAL + 1; if $r14.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "target-not-in-scope") }
+rm -rf ($env.HOME | path join ".config" "ontoref" "vaults" "fixture-t14")
+
+# --- 15. [sops-decrypt-failed] ------------------------------------------------
+# Generate a real age keypair (key-A), encrypt a sops file with key-A's pubkey,
+# then attempt resolve-registry-credential with a DIFFERENT master key (key-B).
+# sops returns non-zero because key-B is not a recipient.
+let r15 = (run-one "sops decrypt failed (wrong master_key)" "sops-decrypt-failed" {
+    let f       = ($TMP | path join "t15")
+    let kage_a  = ($TMP | path join "key-a.kage")
+    let kage_b  = ($TMP | path join "key-b.kage")
+    ^age-keygen -o $kage_a out+err> /dev/null
+    ^age-keygen -o $kage_b out+err> /dev/null
+    chmod 0400 $kage_a
+    chmod 0400 $kage_b
+    let pub_a = (open --raw $kage_a | lines | where { |l| $l | str contains "public key" } | first | split row " " | last)
+
+    fixture-base-project $f $'sops = { enabled = true, vault_id = "fixture-t15", master_key_path = "($kage_b)" }'
+    write-manifest-ncl $f '  registry_provides = m.make_registry_provides {
+    participant = "fixture",
+    registries = m.make_registries_config {
+      default = "primary",
+      registries = [ m.make_registry_entry { id = "primary", endpoint = "x", credential_sops = "registry/ro.sops.yaml" } ],
+    },
+  },'
+    # Pre-create the sops file encrypted ONLY for key-A (key-B cannot decrypt).
+    let vault_dir = ($env.HOME | path join ".config" "ontoref" "vaults" "fixture-t15" "src-vault" "registry")
+    mkdir $vault_dir
+    let sops_path = ($vault_dir | path join "ro.sops.yaml")
+    let plain = ($vault_dir | path join "_plain.yaml")
+    "username: u\npassword: p\n" | save --force $plain
+    let enc = (^sops --age $pub_a --input-type yaml --encrypt $plain)
+    $enc | save --force $sops_path
+    rm $plain
+
+    # Now resolve with key-B as master — sops decrypt must fail.
+    resolve-registry-credential $f primary --op pull
+})
+$TOTAL = $TOTAL + 1; if $r15.ok { $PASSED = $PASSED + 1 } else { $FAILED = ($FAILED | append "sops-decrypt-failed") }
+rm -rf ($env.HOME | path join ".config" "ontoref" "vaults" "fixture-t15")
+
+# ── Report ───────────────────────────────────────────────────────────────────
+
+print ""
+let bar = if ($FAILED | is-empty) { (ansi green) } else { (ansi red) }
+print $"($bar)── ($PASSED)/($TOTAL) tests passed(ansi reset)"
+if not ($FAILED | is-empty) {
+    print $"  failed: ($FAILED | str join ', ')"
+    rm -rf $TMP
+    exit 1
+}
+
+# Cleanup fixtures
+rm -rf $TMP
diff --git a/reflection/tests/test_vault_lock.sh b/reflection/tests/test_vault_lock.sh
new file mode 100755
index 0000000..da6e08a
--- /dev/null
+++ b/reflection/tests/test_vault_lock.sh
@@ -0,0 +1,112 @@
+#!/usr/bin/env bash
+# Vault lock concurrency smoke (ADR-017 G1).
+#
+# Drives the lock helpers in justfiles/_secrets_lib.sh against the live ZOT
+# registry. Tests:
+#   1. First holder acquires the lock — push succeeds.
+#   2. Second holder attempting to acquire while lock is fresh — sees the
+#      existing holder and refuses to overwrite (modeled in secrets-open).
+#   3. TTL expiry — a forged-past expires timestamp is treated as abandoned.
+#   4. Force-unlock writes audit + deletes — second holder can then proceed.
+#
+# Assumes:
+#   - libre-wuji vault bootstrapped (its access.sops.yaml decryptable).
+#   - regadm zot credentials exported in env or implicit via DOCKER_CONFIG.
+#
+# Exit 0 on green; non-zero on any failed assertion.
+
+set -euo pipefail
+
+ONTOREF_ROOT=${ONTOREF_ROOT:-/Users/Akasha/Development/ontoref}
+PROJECT_ROOT=${ONTOREF_PROJECT_ROOT:-/Users/Akasha/project-provisioning/workspaces/libre-wuji}
+NICKEL_IMPORT_PATH="${PROJECT_ROOT}:${ONTOREF_ROOT}:${ONTOREF_ROOT}/ontology:${HOME}/.config/ontoref/schemas"
+export NICKEL_IMPORT_PATH
+export ONTOREF_PROJECT_ROOT="$PROJECT_ROOT"
+
+# shellcheck disable=SC1091
+source "${ONTOREF_ROOT}/justfiles/_secrets_lib.sh"
+
+PASS=0
+FAIL=0
+
+assert() {
+    local name="$1"; shift
+    if "$@"; then
+        echo "  ✓ $name"
+        PASS=$((PASS + 1))
+    else
+        echo "  ✗ $name"
+        FAIL=$((FAIL + 1))
+    fi
+}
+
+vault_zot_config_open || { echo "cannot open zot config — abort"; exit 1; }
+trap 'rm -rf "$TMPDIR_CFG"' EXIT
+
+# Cleanup any stale lock from prior runs.
+vault_lock_release || true
+
+# Test 1 — fresh acquire.
+echo "test 1: fresh lock acquire"
+NOW_TS=$(date -u +%Y-%m-%dT%H:%M:%SZ)
+NOW_EPOCH=$(date -u +%s)
+EXP_EPOCH=$((NOW_EPOCH + 3600))
+EXP_TS=$(date -u -r "$EXP_EPOCH" +%Y-%m-%dT%H:%M:%SZ 2>/dev/null \
+    || date -u -d "@$EXP_EPOCH" +%Y-%m-%dT%H:%M:%SZ 2>/dev/null)
+
+assert "push lock as alice" \
+    vault_lock_push "alice" "developer" "$NOW_TS" "$EXP_TS"
+
+# Test 2 — fetch returns alice's lock.
+echo "test 2: fetch returns existing holder"
+EXISTING=$(vault_lock_fetch)
+ALICE=$(echo "$EXISTING" | jq -r .locked_by)
+assert "lock fetched holder=alice" \
+    [ "$ALICE" = "alice" ]
+
+# Test 3 — second holder bob detects alice and refuses (logic from secrets-open).
+echo "test 3: second holder sees existing fresh lock"
+EXIST_BY=$(echo "$EXISTING" | jq -r .locked_by)
+EXIST_EXP=$(echo "$EXISTING" | jq -r .expires)
+EXIST_EXP_EPOCH=$(date -u -j -f "%Y-%m-%dT%H:%M:%SZ" "$EXIST_EXP" +%s 2>/dev/null \
+    || date -u -d "$EXIST_EXP" +%s 2>/dev/null || echo 0)
+NOW_EPOCH=$(date -u +%s)
+assert "fresh lock blocks bob (alice holds, not expired)" \
+    [ "$EXIST_BY" = "alice" ] && [ "$EXIST_EXP_EPOCH" -gt "$NOW_EPOCH" ]
+
+# Test 4 — TTL-expired lock is detectable (forge a past expiry by re-pushing).
+echo "test 4: TTL-expired lock"
+PAST_EPOCH=$((NOW_EPOCH - 60))
+PAST_TS=$(date -u -r "$PAST_EPOCH" +%Y-%m-%dT%H:%M:%SZ 2>/dev/null \
+    || date -u -d "@$PAST_EPOCH" +%Y-%m-%dT%H:%M:%SZ 2>/dev/null)
+vault_lock_push "alice" "developer" "$PAST_TS" "$PAST_TS"
+EXISTING=$(vault_lock_fetch)
+EXP=$(echo "$EXISTING" | jq -r .expires)
+EXP_EP=$(date -u -j -f "%Y-%m-%dT%H:%M:%SZ" "$EXP" +%s 2>/dev/null \
+    || date -u -d "$EXP" +%s 2>/dev/null || echo 0)
+assert "expired lock detected (expires < now)" \
+    [ "$EXP_EP" -lt "$NOW_EPOCH" ]
+
+# Test 5 — force release.
+echo "test 5: force release succeeds"
+vault_lock_release
+LATER=$(vault_lock_fetch)
+assert "lock empty after release" [ -z "$LATER" ]
+
+# Test 6 — bob can now acquire.
+echo "test 6: bob acquires after release"
+NOW_TS=$(date -u +%Y-%m-%dT%H:%M:%SZ)
+EXP_EPOCH=$(($(date -u +%s) + 3600))
+EXP_TS=$(date -u -r "$EXP_EPOCH" +%Y-%m-%dT%H:%M:%SZ 2>/dev/null \
+    || date -u -d "@$EXP_EPOCH" +%Y-%m-%dT%H:%M:%SZ 2>/dev/null)
+vault_lock_push "bob" "developer" "$NOW_TS" "$EXP_TS"
+EXISTING=$(vault_lock_fetch)
+BOB=$(echo "$EXISTING" | jq -r .locked_by)
+assert "lock now held by bob" [ "$BOB" = "bob" ]
+
+# Cleanup.
+vault_lock_release
+
+echo
+echo "── $PASS passed, $FAIL failed ──"
+[ "$FAIL" -eq 0 ]
diff --git a/steal_this_deck_session.txt b/steal_this_deck_session.txt
new file mode 100644
index 0000000..ffa3435
--- /dev/null
+++ b/steal_this_deck_session.txt
@@ -0,0 +1,862 @@
+ ▐▛███▜▌   Claude Code v2.1.138
+▝▜█████▛▘  Sonnet 4.6 · Claude Max
+  ▘▘ ▝▝    /Users/Akasha/Development/ontoref
+
+❯ dime qué tiene que ver este contendip con este proyecto
+  /Users/Akasha/Development/ontoref/assets/presentation/docs/Steal_This_Deck.pd
+  f
+
+⏺ Voy a leer el PDF para ver qué contiene.
+
+  Read 1 file (ctrl+o to expand)
+
+⏺ Voy a leer las páginas restantes antes de responder.
+
+  Read 1 file (ctrl+o to expand)
+
+⏺ La conexión es directa y profunda. Este deck es esencialmente el argumento
+  teórico externo que justifica por qué ontoref existe.
+
+  Lo que Talisman argumenta en KGC 2026:
+
+  ▎ AI needs knowledge infrastructure — ontologies, typed relations, provenance
+  ▎ — to work reliably. Data infrastructure (lakes, vectors, ETL) doesn't
+  ▎ produce knowledge. You cannot skip layers.
+
+  Lo que ontoref implementa:
+
+  Exactamente eso, pero scopeado a proyectos de software en lugar de
+  organizaciones empresariales.
+
+  Las correspondencias específicas son precisas:
+
+  Deck (org/enterprise): Controlled vocabularies
+  Ontoref (proyecto de software): ontology/schemas/*.ncl — contratos NCL tipados
+  ────────────────────────────────────────
+  Deck (org/enterprise): Ontologies — formal commitments
+  Ontoref (proyecto de software): .ontology/core.ncl — nodos Practice/Concept
+  con
+    edges
+  ────────────────────────────────────────
+  Deck (org/enterprise): Knowledge graphs — live, queryable
+  Ontoref (proyecto de software): DAG-formalized knowledge, ontoref graph,
+    describe
+  ────────────────────────────────────────
+  Deck (org/enterprise): Governance — forever, not a project
+  Ontoref (proyecto de software): reflection/, ADRs, migraciones, state.ncl FSM
+  ────────────────────────────────────────
+  Deck (org/enterprise): "Memory is knowledge"
+  Ontoref (proyecto de software): .coder/ sessions, Q&A persistence (ADR-003)
+
+  El punto de slide 12 — "MCP and A2A are transport. Not semantics." — es
+  particularmente relevante: ontoref tiene una superficie MCP
+  (crates/ontoref-daemon/src/mcp/) que expone precisamente la capa semántica que
+   el protocolo de transporte no puede proveer por sí solo. Es la diferencia que
+   Talisman identifica.
+
+  El punto de slide "The Productivity Paradox" — AI generates volume without a
+  knowledge backbone — es exactamente el problema que el Agent Entry-Point
+  Protocol de ontoref resuelve: sin describe capabilities, un agente como yo
+  opera sobre correlación estadística, no sobre conocimiento estructurado del
+  proyecto.
+
+  ¿Por qué está en assets/presentation/? Probablemente es referencia para una
+  presentación propia de ontoref — el argumento de por qué proyectos de software
+   necesitan este tipo de infraestructura de conocimiento para que los agentes
+  AI trabajen sobre ellos con fiabilidad.
+
+✻ Sautéed for 1m 21s
+
+❯ lo acabamos de recibir y no sabemos muy bien qué hacer, si es que debemos o
+  tenemos que hacer algo al respecto ?
+
+⏺ Honestamente, no tienes que hacer nada en el codebase — no hay ningún gap
+  técnico que este deck revele.
+
+  Lo que sí tiene valor es entender qué te da este documento como recurso:
+
+  Si estás construyendo una presentación para ontoref (lo sugiere
+  assets/presentation/), este deck te regala el trabajo de contextualización que
+   normalmente es lo más difícil de una pitch:
+
+  - Ya no tienes que convencer a la audiencia de que AI necesita infraestructura
+   de conocimiento — Talisman lo hace con datos (KGC 2026, benchmarks citados,
+  investigación peer-reviewed)
+  - Puedes entrar directamente a: "Ese problema existe también en proyectos de
+  software. Ontoref es la infraestructura de conocimiento para que los agentes
+  AI trabajen sobre tu proyecto con fiabilidad, no con correlación estadística"
+  - El slide 12 ("MCP are transport, not semantics") es especialmente útil:
+  ontoref tiene superficie MCP + capa ontológica — exactamente el gap que ella
+  identifica
+
+  El único movimiento concreto que consideraría es guardarlo como referencia
+  citeable en assets/presentation/docs/ (ya está), y si construyes slides, abrir
+   con el contexto que ella establece antes de entrar a qué hace ontoref.
+
+  Lo que no haría: incorporar su terminología al codebase, cambiar la ontología,
+   ni crear documentos de alineación. El código ya hace lo que el deck describe
+  — es validación externa, no instrucción de diseño.
+
+✻ Sautéed for 31s
+
+※ recap: Reviewing a KGC 2026 presentation about knowledge infrastructure for AI. No action needed — the PDF is external validation for ontoref's pitch, already saved in assets/presentation/docs/. (disable recaps in /config)
+
+❯ puedes extraer del doc las partes relevantes que pueden servir para ontoref
+  en posts, slides, presentacione e intro
+
+⏺ Voy a leer las últimas páginas que me faltan antes de compilar.
+
+  Read 1 file (ctrl+o to expand)
+
+⏺ Aquí está la extracción organizada por uso:
+
+  ---
+  Framing / Intro hooks
+
+  Para abrir cualquier presentación de ontoref:
+
+  ▎ "AI is a knowledge tool. Not a data tool. That breaks every assumption
+  ▎ underneath many AI strategies."
+  ▎ — Talisman, KGC 2026
+
+  Adaptado a ontoref: los agentes AI que trabajan sobre tu repositorio operan
+  sobre correlación estadística si no hay infraestructura de conocimiento.
+  Ontoref es esa infraestructura, pero para proyectos de software.
+
+  ▎ "Agentic AI is not a model upgrade away. It is a shared language, from which
+  ▎  to act."
+
+  Este cierre del deck es el mejor hook de intro para ontoref: un modelo más
+  grande no sabe qué es tu proyecto, qué decisiones tomaste, ni en qué estado
+  está. Eso requiere conocimiento estructurado, no un token window mayor.
+
+  ---
+  Datos citable para el problema
+
+  Todos con fuente verificable — útiles en posts y slides sin necesitar
+  justificación propia:
+
+  Dato: 89% de empresas reportan cero impacto de productividad de AI en 3 años
+  Fuente: NBER, Feb 2026, n=5.937 execs
+  ────────────────────────────────────────
+  Dato: Developers experimentados fueron 19% más lentos con herramientas AI
+  Fuente: METR RCT, 2025
+  ────────────────────────────────────────
+  Dato: Ganancia neta semanal del trabajador AI-promedio: −14 minutos
+  Fuente: Foxit/Sapio, March 2026
+  ────────────────────────────────────────
+  Dato: 1.7× más issues en PRs escritos por AI
+  Fuente: CodeRabbit, 2025
+  ────────────────────────────────────────
+  Dato: 46% del código en GitHub escrito por Copilot
+  Fuente: Octoverse 2025
+
+  El argumento que emerge: AI is not saving time. It is generating volume.
+  Without a knowledge backbone, its only measurable output is noise.
+
+  ---
+  El argumento central — para slides y posts
+
+  El stack de conocimiento (Talisman lo llama "knowledge infrastructure"):
+
+  1. Controlled vocabularies  — términos con significado único y autoritativo
+  2. Taxonomies               — organización jerárquica
+  3. Thesauri                 — equivalencias, relaciones cruzadas
+  4. Ontologies               — compromisos formales: clases, propiedades,
+  constraints
+  5. Knowledge graphs         — live, queryable, governable
+
+  ▎ "You cannot skip layers. You cannot start at five and reverse-engineer to
+  ▎ one."
+
+  Ontoref implementa este stack para proyectos de software:
+  - Schemas NCL → vocabularios controlados
+  - .ontology/core.ncl → nodos Practice/Concept con edges tipados
+  - DAG-formalized knowledge → el knowledge graph
+  - ADRs + migrations + state.ncl → governance
+
+  ---
+  El argumento de precisión — para posts técnicos
+
+  ▎ "The accuracy gap is not closed by a bigger model. It is closed by a defined
+  ▎  schema, an ontology, and a validated query."
+
+  Con datos:
+  - 16% → 72% en question-answering sobre SQL enterprise con ontology checks
+  (Allemang & Sequeda, data.world AI Lab, 2024)
+  - 3.4× GraphRAG vs vector RAG en 43 queries enterprise (Diffbot KG-LM
+  Benchmark, 2023)
+  - Vector RAG colapsa a 0% past 5 entities per query. KG-grounded retrieval
+  sostiene.
+
+  ▎ "Defining your terms is the cheapest accuracy and cost lever in the LLM
+  ▎ stack."
+
+  Ontoref es exactamente esto: ontoref describe y el sistema Q&A (ADR-003) son
+  el vocabulario controlado que reduce el espacio de alucinación cuando un
+  agente trabaja sobre el proyecto.
+
+  ---
+  El argumento de MCP — crítico para el posicionamiento técnico
+
+  ▎ "MCP and A2A are transport. Not semantics. They move bytes between
+  ▎ endpoints. They do not establish shared meaning."
+  ▎
+  ▎ "Two agents connected by MCP exchange text bytes. They do not truly share
+  ▎ context."
+
+  Ontoref cierra este gap: tiene superficie MCP (ontoref-daemon/src/mcp/) encima
+   de una capa ontológica. El protocolo mueve bytes; ontoref provee el
+  significado que hace que esos bytes sean conocimiento, no texto.
+
+  Esto es diferenciación directa respecto a "simplemente exponer tu repo por
+  MCP".
+
+  ---
+  El argumento cultural — para posts de opinión
+
+  ▎ "The industry is optimized to ship solutions, not to own problems."
+
+  ┌─────────────────────────────┬────────────────────────────────┐
+  │          Celebrado          │            Huérfano            │
+  ├─────────────────────────────┼────────────────────────────────┤
+  │ Launching new platforms     │ Maintaining existing platforms │
+  ├─────────────────────────────┼────────────────────────────────┤
+  │ AI-generated content        │ Taxonomy work                  │
+  ├─────────────────────────────┼────────────────────────────────┤
+  │ Ontologies built on the fly │ Domain expertise               │
+  └─────────────────────────────┴────────────────────────────────┘
+
+  ▎ "Knowledge work IS the maintenance. Yet it keeps getting cut."
+
+  Ontoref formaliza exactamente lo que siempre se corta: las decisiones
+  arquitectónicas (ADRs), el estado del proyecto (state.ncl), la memoria
+  operacional (.coder/). Lo hace queryable y machine-readable para que no
+  dependa de que alguien lo recuerde.
+
+  ---
+  El close — para cualquier formato
+
+  ▎ "The organizations that close the perception/reality gap first won't be the
+  ▎ ones with the best models. They'll be the ones who finally did the work."
+
+  Adaptado: los proyectos donde los agentes AI trabajan mejor no son los que
+  tienen el modelo más grande. Son los que tienen conocimiento estructurado
+  sobre sí mismos.
+
+  ---
+  Nota de atribución
+
+  Todo esto es de Jessica Talisman (KGC 2026, "Stop Betting, Start Building").
+  Al usar cualquiera de estos puntos en público, citar la fuente — es un
+  argumento de autoridad que refuerza, no debilita, la posición de ontoref.
+
+✻ Baked for 1m 21s
+
+❯ guarda esto en assets/presentation/docs como referencia
+
+  Listed 1 directory (ctrl+o to expand)
+
+⏺ Write(assets/presentation/docs/talisman-steal-this-deck-extractions.md)
+  ⎿  Wrote 118 lines to
+     assets/presentation/docs/talisman-steal-this-deck-extractions.md
+      2 # Extractions — Steal This Deck (Talisman, KGC 2026)
+      3 Source: Jessica Talisman, "Stop Betting, Start Building", Knowled
+        ge Graph Conference 2026.
+      4 Newsletter: Intentional Arrangement (Substack). Received: 9 May 2
+        026.
+      5 Full deck: `Steal_This_Deck.pdf`
+      6
+      7 ---
+      8
+      9 ## Framing / Intro hooks
+     … +153 lines (ctrl+o to expand)
+
+⏺ Guardado en assets/presentation/docs/talisman-steal-this-deck-extractions.md,
+  junto al PDF original.
+
+✻ Churned for 1m 16s
+
+❯ pero ontoref son dos coxistencias de ontologia y reflection
+  Ella tiene un enfoque más de la semántica ontologica y menos de la operaciona
+   que es el Yang
+  pero sin el Yin de reflection son buenas intenciones/deseos, etc ?
+
+⏺ Exacto. Es la crítica más precisa que se le puede hacer al argumento de
+  Talisman, y es exactamente lo que diferencia ontoref.
+
+  Lo que ella describe es infraestructura de conocimiento como artefacto de
+  construcción — vocabularios, ontologías, knowledge graphs. El problema que
+  identifica (governance "forever, not a project") lo nombra pero no lo
+  resuelve. Su stack termina en "Govern" como un sexto paso, sin mecanismo
+  interno.
+
+  La pregunta que no responde: ¿quién vigila que la ontología no derive respecto
+   al sistema real? ¿Quién detecta cuando una decisión arquitectónica invalida
+  un nodo del grafo? ¿Quién propaga los cambios a los consumidores?
+
+  Sin ese cierre operacional, la ontología más rigurosa acaba siendo exactamente
+   lo que describes — buenas intenciones formalizadas. Archaeology, no
+  infrastructure.
+
+  Ontoref es las dos mitades:
+
+  Ontology (Yin — lo que ES)        Reflection (Yang — lo que DEVIENE)
+  ──────────────────────────────     ──────────────────────────────────
+  .ontology/core.ncl                 reflection/modes/*.ncl
+  Nodos, edges, constraints          DAGs ejecutables, pasos reportables
+  Compromisos formales               Operaciones que cambian estado
+  state.ncl (fotografía)             state.ncl FSM (transición activa)
+  ADRs (decisión tomada)             migrations/ (propagación al ecosistema)
+  describe project                   sync diff --docs (detección de deriva)
+
+  Sin Reflection, la ontología cristaliza. Describe lo que el proyecto quería
+  ser en el momento en que se escribió. Sin Ontology, Reflection es ejecución
+  sin verdad — sabe qué hacer pero no qué es.
+
+  La tensión nombrada en core.ncl — "Ontology vs Reflection: both must coexist
+  without one dominating" — es precisamente esto. Talisman articula bien el Yin.
+   Ontoref es la respuesta al Yang que su framework no provee.
+
+  Vale la pena añadir esto a las extracciones — es el argumento diferenciador
+  central para cualquier presentación frente a audiencias que ya conocen el
+  mundo KG/semantic web. ¿Lo agrego al documento?
+
+✻ Cooked for 47s
+
+❯ sip
+  Por qué los DAGs se vienen usando desde hace mucho tiempo y no veo que haya
+  la perspectiva de ontoref
+  includo see montan DAgs para los run-books, cd/ci, compiladores ?
+⏺ Es la pregunta correcta. Y la respuesta tiene capas.
+
+  Por qué los DAGs son ubicuos
+
+  En todos los usos que mencionas — CI/CD, compiladores, runbooks, Airflow, dbt,
+   Bazel — el DAG es un modelo de ejecución. Las aristas significan "esto antes
+  que aquello". El grafo describe cómo se computa, no qué es lo que se computa.
+  Es una herramienta de ordenación topológica.
+
+  CI/CD:      test → build → deploy
+  Compiler:   parse → typecheck → codegen → link
+  Runbook:    check_health → drain → restart → verify
+
+  Todos inerts. El grafo no sabe qué es el proyecto, no tiene semántica sobre
+  sus propias aristas.
+
+  ---
+  Lo que ontoref hace diferente
+
+  Son dos usos de DAGs que no coexisten en ningún otro sitio:
+
+  DAG ontológico — aristas con tipo semántico:
+  Practice "ncl-schemas" --implements--> Principle "type-safety"
+  Practice "ncl-schemas" --enforces-->   Constraint "zero-runtime-deps"
+  Practice "ncl-schemas" --enables-->    Capability "agent-queryable-state"
+  Practice "ncl-schemas" --tension-->    Practice "adoption-friction"
+  No es "esto antes que aquello". Es conocimiento — compromisos formales sobre
+  qué ES el proyecto y cómo se relacionan sus conceptos. El equivalente sería un
+   compilador que sabe por qué existe y qué trade-offs tomó.
+
+  DAG de reflexión — aristas con contrato ejecutable + restricción de actor:
+  step "validate-state" (actor: any)    --> step "run-mode" (actor: developer)
+  step "run-mode" (actor: developer)    --> step "report" (actor:
+  agent|developer)
+  No es solo ejecución. El DAG es validado como contrato antes de ejecutarse,
+  registra su propio progreso, y sabe quién puede hacer qué.
+
+  ---
+  El gap que nadie ha cerrado
+
+  El mundo de knowledge graphs (Talisman, W3C, RDF/OWL) aplica este tipo de
+  semántica a dominios empresariales — productos, clientes, transacciones. Nunca
+   al proyecto de software mismo.
+
+  El mundo de herramientas de software (GitHub, Jira, Linear, CI/CD) trata el
+  conocimiento del proyecto como documentos no tipados — wikis, tickets,
+  READMEs. Usa DAGs para ejecución, nunca para representación de conocimiento.
+
+  Knowledge Graph world:  semántica sobre el dominio del negocio
+  Software tooling world: DAGs sobre la ejecución del software
+  ontoref:                semántica sobre el proyecto de software en sí mismo
+
+  ---
+  La pieza que lo hace no-trivial
+
+  La parte que no existe en ningún otro sitio es que los dos DAGs se hablan:
+
+  - El DAG ontológico dice qué es el proyecto (prácticas, principios,
+  constraints activos)
+  - El DAG de reflexión opera sobre ese estado — detecta deriva, ejecuta modos,
+  registra transiciones
+  - Las migraciones propagan cambios del DAG ontológico a proyectos consumidores
+
+  Sin esa conexión, tienes una ontología bonita que nadie mantiene (buenas
+  intenciones) o tienes runbooks sin semántica (ejecución ciega). El Yang sin
+  Yin, o el Yin sin Yang.
+
+  ---
+  Por qué no había emergido antes
+
+  Tres condicionantes que ahora se alinean:
+
+  1. Agentes AI — que necesitan consumir conocimiento estructurado del proyecto,
+   no texto libre. Antes no había consumidor que lo requiriera.
+  2. NCL/Nickel — un lenguaje de configuración con contratos y tipos que permite
+   expresar ontología sin un triplestore externo. RDF/OWL requería
+  infraestructura pesada.
+  3. La escala correcta — Knowledge Graphs empresariales son proyectos de años.
+  Ontoref opera a escala de repositorio, con adopción incremental y sin
+  enforcement.
+
+✻ Cooked for 1m 38s
+
+❯ pero hay valores importantes cuando se introducen detalles como el tipo de
+  almacenamiento, el uso modos federados, la gestión via CLI/UI/MCP/GRAPHQL
+  y luego consumir usando jj y radicle
+
+⏺ Ahí está el salto cualitativo que separa ontoref de cualquier otro enfoque.
+
+  Lo que añade cada capa:
+
+  Almacenamiento tipado y soberano
+
+  El enterprise knowledge graph (el mundo de Talisman) asume un triplestore
+  central mantenido por un equipo dedicado. Ontoref invierte eso: el
+  conocimiento vive con el proyecto, en NCL versionado junto al código. La
+  elección de almacenamiento no es un detalle de implementación — es una
+  declaración de soberanía. Nadie puede quitarte tu ontología porque está en tu
+  repo.
+
+  Modos federados
+
+  Esto es lo que ningún sistema de knowledge graph tiene: proyectos que se
+  describen a sí mismos Y pueden ser descubiertos y consultados por otros
+  proyectos. No es un grafo central al que todos apuntan — es una red de
+  proyectos auto-descritos que pueden exponer su conocimiento entre ellos.
+  Ecosistema-level visibility sin un intermediario.
+
+  proyecto-A consulta: ¿qué proyectos implementan "auth-pattern-X"?
+  proyecto-B, proyecto-C responden desde sus propias ontologías
+  sin un servidor central que lo sepa todo
+
+  NATS (ya feature-gated en el daemon) es exactamente el fabric para esto.
+
+  CLI + UI + MCP + GraphQL
+
+  Cuatro superficies que no son redundantes — responden a cuatro tipos de
+  consumidor:
+
+  ┌───────────────┬───────────────────────────────────────────────────┐
+  │  Superficie   │                    Consumidor                     │
+  ├───────────────┼───────────────────────────────────────────────────┤
+  │ CLI (Nushell) │ Desarrollador en terminal, scripts, CI            │
+  ├───────────────┼───────────────────────────────────────────────────┤
+  │ UI (axum)     │ Managers, revisión visual, onboarding             │
+  ├───────────────┼───────────────────────────────────────────────────┤
+  │ MCP           │ Agentes AI — la razón de existir de ontoref       │
+  ├───────────────┼───────────────────────────────────────────────────┤
+  │ GraphQL       │ Herramientas externas, dashboards, queries ad-hoc │
+  └───────────────┴───────────────────────────────────────────────────┘
+
+  GraphQL sobre el DAG ontológico es especialmente potente: SPARQL es el
+  estándar semántico pero requiere expertise. GraphQL es lo que todo el mundo
+  sabe escribir. Reduces la barrera de consumo sin sacrificar la estructura.
+
+  jj + Radicle como capa de transporte
+
+  Aquí es donde se cierra el argumento de soberanía:
+
+  - jj (Jujutsu): modelo de VCS donde cada working copy es siempre un commit.
+  Alinea perfectamente con ontoref — no hay "estado sucio", cada momento del
+  proyecto es capturable. El operation log de jj es análogo al tracking de
+  sesiones de reflection.
+  - Radicle: P2P, local-first, sin GitHub. Si tu código es soberano, tu
+  conocimiento sobre ese código también lo es. Radicle como transport hace que
+  la federación de ontoref no dependa de ninguna plataforma centralizada.
+
+  ---
+  El cuadro completo que diferencia ontoref:
+
+  Enterprise KG (Talisman):     Ontoref:
+  ─────────────────────────     ──────────────────────────────────
+  Triplestore central           NCL distribuido, por repo
+  Equipo dedicado               El proyecto se mantiene a sí mismo
+  GitHub/GitLab                 jj + Radicle (soberano, P2P)
+  API SPARQL interna            CLI + UI + MCP + GraphQL
+  Un grafo global               Federación de proyectos auto-descritos
+  Knowledge como infraestructura Knowledge como parte del proyecto
+
+  La conexión con Yin/Yang:
+
+  El almacenamiento soberano + federación + multi-superficie es el Yang
+  operacional que hace que el Yin ontológico sea accionable a escala de
+  ecosistema. Sin esto, incluso la ontología más rigurosa es local y mortal —
+  muere cuando el equipo cambia o cuando la plataforma desaparece.
+
+  Con esto, el conocimiento del proyecto sobrevive al equipo, se propaga sin
+  intermediarios, y puede ser consumido por cualquier agente o herramienta que
+  hable cualquiera de las cuatro superficies.
+
+  ---
+  ¿Quieres que capture esto también en el documento de referencia de
+  assets/presentation/docs/?
+
+✻ Sautéed for 58s
+
+❯ estamos con ontoref no sólo en gestión de proyectos, también en
+  infrawstructura y en lo personal
+
+⏺ Eso cambia el scope fundamental. Ya no es una herramienta de gestión de
+  proyectos — es un protocolo de auto-conocimiento para cualquier sistema con
+  identidad e intención.
+
+  Los tres dominios comparten la misma estructura profunda:
+
+                Ontology (Yin)           Reflection (Yang)
+                ─────────────────────    ──────────────────────────
+  Proyecto      Qué ES el software       Qué DEVIENE — ADRs, modos,
+                principios, prácticas    migraciones, deriva de docs
+
+  Infra         Qué ES el sistema        Qué DEVIENE — provisioning,
+                servicios, deps,         drift detection, remediation,
+                constraints, topología   runbooks como DAGs ejecutables
+
+  Personal      Quién ERES               Quién DEVIENS — decisiones con
+                valores, roles,          consecuencias duraderas, hábitos,
+                principios vitales       revisiones, deriva de intención
+
+  Lo que diferencia a los tres no es la arquitectura — es el sujeto que se
+  describe a sí mismo.
+
+  ---
+  Infra — lo que añade:
+
+  Infraestructura ya usa DAGs (Terraform, Pulumi, Ansible) pero todos inerts:
+  describen cómo se provisiona, no por qué existe así. Ontoref sobre infra
+  captura las tensiones reales:
+
+  Servicio "auth" --tension--> Principio "stateless"
+      porque necesita session state pero el principio dice no
+  Constraint "zero-downtime" --enforces--> Práctica "blue-green"
+      y eso está en el ADR-infra-003, no en el Terraform
+
+  jj + Radicle + NATS federation = GitOps soberano donde cada entorno (dev,
+  staging, prod) se describe a sí mismo y puede ser consultado desde el proyecto
+   que lo consume.
+
+  Personal — lo que añade:
+
+  Es el dominio más radical. La misma pregunta que ontoref hace a un proyecto —
+  ¿qué eres, qué tensiones activas tienes, en qué estado estás respecto a donde
+  quieres estar? — es exactamente la pregunta que el conocimiento personal
+  estructurado necesita responder. No un segundo cerebro de notas. Un grafo de
+  compromisos.
+
+  Valor "profundidad" --tension--> Rol "padre"
+      porque el tiempo es finito y los dos lo demandan
+  Decisión "dejar empresa X" --constraint--> Futuro laboral
+      con el mismo peso que un ADR arquitectónico
+  Estado "aprendiendo Rust" --blocker--> "contribuir a ontoref-core"
+      FSM personal, no lista de tareas
+
+  ---
+  La convergencia que lo hace coherente:
+
+  Los tres dominios son instancias del mismo protocolo porque los tres tienen:
+
+  - Identidad — qué son, qué principios los definen
+  - Tensiones activas — trade-offs que no se resuelven, se gestionan
+  - Decisiones con consecuencias duraderas — el equivalente de ADRs
+  - Estado observable — dónde están vs. dónde quieren estar
+  - Deriva — la distancia entre intención y realidad que crece con el tiempo
+  - Operaciones — los modos que reducen esa deriva
+
+  Y todos se benefician del mismo stack: almacenamiento soberano, federación,
+  multi-superficie, jj/Radicle.
+
+  ---
+  Lo que esto significa para el posicionamiento:
+
+  Talisman habla de knowledge infrastructure para organizaciones. Ontoref es
+  knowledge infrastructure desde abajo — desde el individuo hasta el proyecto
+  hasta la infraestructura que ejecuta ese proyecto. Sin vendor, sin plataforma
+  central, sin equipo dedicado de ontólogos.
+
+  El sujeto más pequeño que puede adoptar ontoref es una persona. El más grande
+  es un ecosistema de proyectos federados.
+
+  ¿Guardamos esto también como referencia?
+
+✻ Cogitated for 52s
+
+❯ guarda todos también
+
+⏺ Write(assets/presentation/docs/ontoref-yin-yang-diferenciacion.md)
+  ⎿  Wrote 67 lines to
+     assets/presentation/docs/ontoref-yin-yang-diferenciacion.md
+       2 # Ontoref — Yin/Yang: Lo que Talisman no resuelve
+       3 Referencia para posts, slides, presentaciones.
+       4 Contexto: extracción del análisis comparativo con "Steal This De
+         ck" (Talisman, KGC 2026).
+       5
+       7 ---
+       8 ## El gap del argumento de Talisman
+       9
+      10 Talisman describe infraestructura de conocimiento como artefacto
+          de construcción —
+     … +77 lines (ctrl+o to expand)
+
+⏺ Write(assets/presentation/docs/ontoref-dags-diferenciacion.md)
+  ⎿  Wrote 107 lines to
+     assets/presentation/docs/ontoref-dags-diferenciacion.md
+       2 # Ontoref — Por qué los DAGs de ontoref son diferentes
+       3 Referencia para posts, slides, presentaciones.
+       4 Contexto: análisis de por qué DAGs son ubicuos pero nadie tiene
+         la perspectiva de ontoref.
+       5
+       7 ---
+       8 ## DAGs son ubicuos — y todos hacen lo mismo
+       9
+      10 En CI/CD, compiladores, runbooks, Airflow, dbt, Bazel: el DAG es
+          un **modelo de ejecución**.
+     … +130 lines (ctrl+o to expand)
+
+⏺ Write(assets/presentation/docs/ontoref-scope-proyecto-infra-personal.md)
+  ⎿  Wrote 157 lines to
+     assets/presentation/docs/ontoref-scope-proyecto-infra-personal.md
+      1 # Ontoref — Scope: Proyecto, Infraestructura, Personal
+      2
+      3 Referencia para posts, slides, presentaciones.
+      4 Contexto: ontoref como protocolo de auto-conocimiento para cualqu
+        ier sistema con identidad e intención,
+      5 más allá de gestión de proyectos de software.
+      6
+      7 ---
+      8
+      9 ## La afirmación central
+     … +181 lines (ctrl+o to expand)
+
+❯ ok
+
+⏺ Tres archivos guardados en assets/presentation/docs/:
+
+  - talisman-steal-this-deck-extractions.md — citas, datos y mapeos del deck de
+  Talisman
+  - ontoref-yin-yang-diferenciacion.md — por qué Talisman no resuelve el lado
+  operacional
+  - ontoref-dags-diferenciacion.md — por qué los DAGs de ontoref son distintos a
+   todos los demás
+  - ontoref-scope-proyecto-infra-personal.md — el scope completo con el stack
+  soberano
+
+✻ Worked for 1m 59s
+
+❯ genera un índice de los docs de presentation
+
+  Listed 2 directories (ctrl+o to expand)
+
+⏺ Write(assets/presentation/docs/INDEX.md)
+  ⎿  Wrote 56 lines to assets/presentation/docs/INDEX.md
+       2 # Presentation Docs — Index
+       3 Material de referencia para posts, slides y presentaciones de on
+         toref.
+       4
+       5 ---
+       6
+       7 ## Análisis y posicionamiento ontoref
+       8
+       9 | Archivo | Contenido |
+      10 |---|---|
+     … +73 lines (ctrl+o to expand)
+
+⏺ INDEX.md creado en assets/presentation/docs/. Agrupa el material en cuatro
+  bloques:
+
+  1. Análisis ontoref — los tres docs de posicionamiento generados hoy
+  2. Talisman KGC 2026 — deck PDF + extracciones + artículo EN/ES
+  3. José Celano — ontologías y DAGs
+  4. Material previo — ontología como filosofía, sprints, notas de voz
+
+✻ Sautéed for 35s
+
+❯ genera un post basado en todo esto
+
+⏺ User answered Claude's questions:
+  ⎿  · ¿En qué idioma quieres el post? → Bilingüe
+     · ¿Cuál es el ángulo principal del post? → Puedes generar 4 posts enficados
+      a cada una de las partes
+     · ¿Para qué plataforma/formato? → me gustaría publicar en 1 y algo en 2
+     pero el foco no va a allí va en jesusperez.pro/blog y ejemplos de posts
+     están en /Users/Akasha/Development/jpl-website/site/content/blog/{en,es}
+
+⏺ Voy a revisar el formato de los posts existentes antes de escribir.
+  Searched for 2 patterns, read 4 files, listed 3 directories (ctrl+o to expand)
+
+⏺ Tengo todo lo que necesito. Voy a generar los 4 posts en EN + ES en paralelo.
+
+  ★ Insight ─────────────────────────────────────
+  Los posts existentes usan published: false para borradores y la categoría
+  define css_class. La categoría ontoref ya tiene directorios pero sin contenido
+   — ideal para iniciar la serie.
+  ─────────────────────────────────────────────────
+
+⏺ Write(/Users/Akasha/Development/jpl-website/site/content/blog/en/ontoref/ai-kn
+       owledge-tool-who-keeps-it-alive.md)
+  ⎿  Wrote 97 lines to ../jpl-website/site/content/blog/en/ontoref/ai-knowl
+     edge-tool-who-keeps-it-alive.md
+       2 # Post metadata
+       4 title: "AI is a Knowledge Tool. But Who Keeps the Knowledge Aliv
+       5 slug: "ai-knowledge-tool-who-keeps-it-alive"
+       6 subtitle: "Talisman is right — and stops exactly where the hard
+       7 excerpt: "Jessica Talisman's KGC 2026 talk is the clearest artic
+         infrastructure and expect reasoning to emerge. She's right. But
+         her solution — build knowledge infrastructure — stops exactly wh
+         ng?"the hard problem begins: who keeps the knowledge from drifti
+       9 # Publication info
+      10 author: "Jesús Pérez"
+     … +151 lines (ctrl+o to expand)
+
+⏺ Write(/Users/Akasha/Development/jpl-website/site/content/blog/es/ontoref/la-ia
+       -herramienta-conocimiento-quien-mantiene-vivo.md)
+  ⎿  Wrote 97 lines to ../jpl-website/site/content/blog/es/ontoref/la-ia-he
+     rramienta-conocimiento-quien-mantiene-vivo.md
+      2 # Post metadata
+      3 id: "la-ia-herramienta-conocimiento-quien-mantiene-vivo"
+        tiene vivo el conocimiento?"enta de conocimiento. ¿Pero quién man
+      5 slug: "la-ia-herramienta-conocimiento-quien-mantiene-vivo"
+      6 subtitle: "Talisman tiene razón — y se detiene exactamente donde
+      7 excerpt: "La charla de Jessica Talisman en KGC 2026 es la articul
+        ación más clara de por qué fallan las estrategias de AI: las orga
+        erja el razonamiento. Tiene razón. Pero su solución — construir i
+        nfraestructura de conocimiento — se detiene exactamente donde emp
+        ieza el problema difícil: ¿quién evita qu
+     … +162 lines (ctrl+o to expand)
+
+⏺ Write(/Users/Akasha/Development/jpl-website/site/content/blog/en/ontoref/dags-
+       everywhere-none-know-what-they-are.md)
+  ⎿  Wrote 106 lines to ../jpl-website/site/content/blog/en/ontoref/dags-ev
+     erywhere-none-know-what-they-are.md
+       2 # Post metadata
+       4 title: "DAGs Are Everywhere. None of Them Know What They Are."
+       5 slug: "dags-everywhere-none-know-what-they-are"
+         s for execution. Ontoref uses them for knowledge."nbook uses DAG
+       7 excerpt: "CI/CD pipelines, compilers, runbooks, data orchestrato
+         hem uses DAGs as execution models: this before that, topological
+         sent what the system is, why it exists, or what trade-offs defin
+         e it. That's the gap ontoref fills."
+       9 # Publication info
+      10 author: "Jesús Pérez"
+     … +161 lines (ctrl+o to expand)
+
+⏺ Write(/Users/Akasha/Development/jpl-website/site/content/blog/es/ontoref/dags-
+       en-todos-lados-ninguno-sabe-lo-que-es.md)
+  ⎿  Wrote 106 lines to ../jpl-website/site/content/blog/es/ontoref/dags-en
+     -todos-lados-ninguno-sabe-lo-que-es.md
+      2 # Post metadata
+      3 id: "dags-en-todos-lados-ninguno-sabe-lo-que-es"
+      5 slug: "dags-en-todos-lados-ninguno-sabe-lo-que-es"e lo que es."
+      6 subtitle: "Cada build system, pipeline CI y runbook usa DAGs para
+      7 excerpt: "Pipelines CI/CD, compiladores, runbooks, orquestadores
+        de datos — todos usan grafos acíclicos dirigidos. Todos sin excep
+        ordenación topológica, resolución de dependencias. Ninguno los us
+        a para representar qué es el sistema, por qué existe, o qué trade
+      8 -offs lo definen. Ese es el gap que llena ontoref."
+      9 # Publication i
+     … +167 lines (ctrl+o to expand)
+
+⏺ Write(/Users/Akasha/Development/jpl-website/site/content/blog/en/ontoref/one-p
+       rotocol-three-subjects.md)
+  ⎿  Wrote 122 lines to ../jpl-website/site/content/blog/en/ontoref/one-pro
+     tocol-three-subjects.md
+       2 # Post metadata
+       4 title: "One Protocol, Three Subjects"
+       5 slug: "one-protocol-three-subjects"
+         structure, and yourself"tion applies to your project, your infra
+       7 excerpt: "Ontoref started as a protocol for software project sel
+         , reflection DAG for what BECOMES — applies without modification
+         changes. The question is identical: what are you, what tensions
+         define you, where are you versus where you intend to be?"
+       9 # Publication info
+      10 author: "Jesús Pérez"
+     … +177 lines (ctrl+o to expand)
+
+⏺ Write(/Users/Akasha/Development/jpl-website/site/content/blog/es/ontoref/un-pr
+       otocolo-tres-sujetos.md)
+  ⎿  Wrote 122 lines to ../jpl-website/site/content/blog/es/ontoref/un-prot
+     ocolo-tres-sujetos.md
+       2 # Post metadata
+       4 title: "Un protocolo, tres sujetos"
+       5 slug: "un-protocolo-tres-sujetos"
+         ctura, y a ti mismo"pregunta aplica a tu proyecto, tu infraestru
+       7 excerpt: "Ontoref comenzó como un protocolo para el auto-conocim
+         lógico para lo que ES, DAG de reflexión para lo que DEVIENE — ap
+         os. El sujeto cambia. La pregunta es idéntica: ¿qué eres, qué te
+         nsiones te definen, dónde estás versus dónde pretendes estar?"
+       9 # Publication info
+      10 author: "Jesús Pérez"
+     … +182 lines (ctrl+o to expand)
+
+⏺ Write(/Users/Akasha/Development/jpl-website/site/content/blog/en/ontoref/your-
+       ontology-should-live-with-your-code.md)
+  ⎿  Wrote 106 lines to ../jpl-website/site/content/blog/en/ontoref/your-on
+     tology-should-live-with-your-code.md
+      2 # Post metadata
+      4 title: "Your Ontology Should Live With Your Code"
+      5 slug: "your-ontology-should-live-with-your-code"
+        ure without a home isn't infrastructure at all"wledge infrastruct
+      7 excerpt: "Enterprise knowledge graphs live in a triplestore maint
+        hanges, the budget gets cut, or the vendor pivots, the knowledge
+        -first knowledge that lives alongside its subject, versioned, que
+        ryable across four surfaces, and federated without a central brok
+      8 er."
+      9 # Publication i
+     … +176 lines (ctrl+o to expand)
+
+⏺ Write(/Users/Akasha/Development/jpl-website/site/content/blog/es/ontoref/tu-on
+       tologia-deberia-vivir-con-tu-codigo.md)
+  ⎿  Wrote 106 lines to ../jpl-website/site/content/blog/es/ontoref/tu-onto
+     logia-deberia-vivir-con-tu-codigo.md
+      2 # Post metadata
+      3 id: "tu-ontologia-deberia-vivir-con-tu-codigo"
+      5 slug: "tu-ontologia-deberia-vivir-con-tu-codigo""
+      6 subtitle: "Soberanía, federación, y por qué la infraestructura de
+         conocimiento sin hogar no es infraestructura"
+         mantenido por un equipo dedicado en la plataforma de un vendor.
+        Cuando el equipo cambia, el presupuesto se recorta, o el vendor p
+        o: conocimiento soberano y local-first que vive junto a su sujeto
+        , versionado, consultable a través de cuatro superficies, y feder
+        ado sin un broker cent
+     … +186 lines (ctrl+o to expand)
+
+⏺ 8 archivos generados. Resumen de la serie:
+
+  ---
+  Post 1 — El Yang faltante
+  - EN: en/ontoref/ai-knowledge-tool-who-keeps-it-alive.md
+  - ES: es/ontoref/la-ia-herramienta-conocimiento-quien-mantiene-vivo.md
+  - Argumento: Talisman tiene razón en el diagnóstico, pero governance sin loop
+  operacional es aspiración. Ontoref es las dos mitades.
+
+  Post 2 — DAGs como conocimiento
+  - EN: en/ontoref/dags-everywhere-none-know-what-they-are.md
+  - ES: es/ontoref/dags-en-todos-lados-ninguno-sabe-lo-que-es.md
+  - Argumento: todos usan DAGs para ejecución, nadie para conocimiento. El doble
+   DAG de ontoref y por qué no existía antes.
+
+  Post 3 — Los tres sujetos
+  - EN: en/ontoref/one-protocol-three-subjects.md
+  - ES: es/ontoref/un-protocolo-tres-sujetos.md
+  - Argumento: proyecto + infra + personal comparten la misma pregunta
+  estructural y el mismo stack.
+
+  Post 4 — Soberanía
+  - EN: en/ontoref/your-ontology-should-live-with-your-code.md
+  - ES: es/ontoref/tu-ontologia-deberia-vivir-con-tu-codigo.md
+  - Argumento: jj + Radicle + NATS + cuatro superficies. El knowledge graph que
+  vive con el código es genuinamente tuyo.
+
+  Todos marcados published: false — borradores listos para revisar antes de
+  publicar.
+
+✻ Sautéed for 8m 22s
diff --git a/templates/ontology/connections.ncl b/templates/ontology/connections.ncl
index abdb3ba..508922f 100644
--- a/templates/ontology/connections.ncl
+++ b/templates/ontology/connections.ncl
@@ -12,7 +12,7 @@
 #   downstream — projects that depend on / consume this one
 #   peers      — symmetric relationships (shared concerns, sibling services)
 
-let s = import "connections" in
+let s = import "reflection/schemas/connections.ncl" in
 
 {
   upstream   = [
diff --git a/templates/ontology/manifest.ncl b/templates/ontology/manifest.ncl
index 24f3e31..653a830 100644
--- a/templates/ontology/manifest.ncl
+++ b/templates/ontology/manifest.ncl
@@ -16,6 +16,17 @@ m.make_manifest {
   project   = "{{ project_name }}",
   repo_kind = 'Library,
 
+  # Required by ProjectManifest contract (no default). Replace this default
+  # consumer with the project's actual consumers as they emerge.
+  consumption_modes = [
+    m.make_consumption_mode {
+      consumer    = 'Developer,
+      needs       = ['OntologyExport],
+      audit_level = 'Standard,
+      description = "Default scaffold consumer — replace with actual consumer types as they emerge.",
+    },
+  ],
+
   content_assets = [
     # m.make_asset {
     #   id          = "logo-horizontal",