chore: update crates, names and clippy fixes

.gitignore

@@ -1,5 +1,7 @@
 .p
 .claude
+.opencode
+AGENTS.md
 .vscode
 .shellcheckrc
 .coder

.gitmodules

@@ -0,0 +1,9 @@
+[submodule "secretumvault"]
+	path = secretumvault
+	url = ssh://git@repo.jesusperez.pro:32225/jesus/secretumvault.git
+[submodule "stratumiops"]
+	path = stratumiops
+	url = ssh://git@repo.jesusperez.pro:32225/jesus/stratumiops.git
+[submodule "syntaxis"]
+	path = syntaxis
+	url = ssh://git@repo.jesusperez.pro:32225/jesus/syntaxis.git

Cargo.toml

@@ -9,11 +9,26 @@ members = [
   "crates/control-center",
   "crates/control-center-ui",
   "crates/vault-service",
-  "crates/rag",
   "crates/detector",
   "crates/mcp-server",
-  "crates/provisioning-daemon",
+  "crates/daemon",
+  "prov-ecosystem/crates/daemon-cli",
+  "prov-ecosystem/crates/machines",
+  "prov-ecosystem/crates/encrypt",
+  "prov-ecosystem/crates/backup",
+  "prov-ecosystem/crates/observability",
 ]
+
+exclude = [
+  "syntaxis",
+  "syntaxis/core",
+  "prov-ecosystem/crates/syntaxis-integration",
+  "prov-ecosystem/crates/audit",
+  "prov-ecosystem/crates/valida",
+  "prov-ecosystem/crates/runtime",
+  "prov-ecosystem/crates/gitops",
+]
+
 resolver = "2"
 
   [workspace.package]
@@ -39,7 +54,7 @@ resolver = "2"
   serde = { version = "1.0", features = ["derive"] }
   serde_json = "1.0"
   toml = "0.9"
-  uuid = { version = "1.19", features = ["v4", "serde"] }
+  uuid = { version = "1.20", features = ["v4", "serde"] }
 
   # ============================================================================
   # ERROR HANDLING
@@ -80,7 +95,7 @@ resolver = "2"
   # DATABASE AND STORAGE
   # ============================================================================
   sqlx = { version = "0.8", features = ["runtime-tokio-rustls", "sqlite", "chrono", "uuid"] }
-  surrealdb = { version = "2.4", features = ["kv-mem", "protocol-ws", "protocol-http"] }
+  surrealdb = { version = "2.6", features = ["kv-mem", "protocol-ws", "protocol-http"] }
 
   # ============================================================================
   # SECURITY AND CRYPTOGRAPHY
@@ -89,7 +104,7 @@ resolver = "2"
   argon2 = "0.5"
   base64 = "0.22"
   hmac = "0.12"
-  jsonwebtoken = { version = "10.2", features = ["rust_crypto"] }
+  jsonwebtoken = { version = "10.3", features = ["rust_crypto"] }
   rand = { version = "0.9", features = ["std_rng", "os_rng"] }
   ring = "0.17"
   sha2 = "0.10"
@@ -127,7 +142,7 @@ resolver = "2"
 
   # Additional cryptography
   hkdf = "0.12"
-  rsa = "0.9.9"
+  rsa = "0.9.10"
   zeroize = { version = "1.8", features = ["derive"] }
 
   # Additional security
@@ -186,7 +201,7 @@ resolver = "2"
   wasm-bindgen-futures = "0.4"
 
   # Random number generation
-  getrandom = { version = "0.3" }
+  getrandom = { version = "0.4" }
 
   # ============================================================================
   # TUI (Terminal User Interface)
@@ -216,28 +231,164 @@ resolver = "2"
   parking_lot = "0.12"
   which = "8"
   yaml-rust = "0.4"
+  humantime-serde = "1.1"
+
+  # Metrics
+  prometheus = "0.14"
+
+  approx = "0.5"
+
+  # Utilities
+  xxhash-rust = { version = "0.8", features = ["xxh3"] }
 
   # ============================================================================
   # RAG FRAMEWORK DEPENDENCIES (Rig)
   # ============================================================================
-  rig-core = "0.27"
+  rig-core = "0.30"
   rig-surrealdb = "0.1"
   tokenizers = "0.22"
 
   # ============================================================================
-  # PROV-ECOSYSTEM DAEMON (replaces cli-daemon)
+  # STRATUM ECOSYSTEM DEPENDENCIES (for RAG embeddings & LLM)
   # ============================================================================
-  daemon-cli = { path = "../../submodules/prov-ecosystem/crates/daemon-cli" }
+  moka = { version = "0.12", features = ["future"] }
+  sled = "0.34"
+  fastembed = "5.8"
+  lancedb = "0.23"
+  arrow = "=56"
 
   # ============================================================================
-  # SECRETUMVAULT (Enterprise Secrets Management)
+  # INTERNAL WORKSPACE CRATES (Local path dependencies)
   # ============================================================================
-  secretumvault = { path = "../../submodules/secretumvault" }
+  platform-config = { path = "./crates/platform-config" }
+  service-clients = { path = "./crates/service-clients" }
+  rag = { path = "./crates/rag" }
+  mcp-server = { path = "./crates/mcp-server" }
+  ai-service = { path = "./crates/ai-service" }
+
+  # ============================================================================
+  # PROV-ECOSYSTEM (Now members of workspace)
+  # ============================================================================
+  daemon-cli = { path = "./prov-ecosystem/crates/daemon-cli" }
+  machines = { path = "./prov-ecosystem/crates/machines" }
+  encrypt = { path = "./prov-ecosystem/crates/encrypt" }
+  backup = { path = "./prov-ecosystem/crates/backup" }
+  observability = { path = "./prov-ecosystem/crates/observability" }
+  init-servs = { path = "./prov-ecosystem/crates/init-servs" }
+
+  # stratum-embeddings and stratum-llm are built in isolated Docker context for RAG
+  # See: crates/rag/docker/Dockerfile
+  stratum-embeddings = { path = "./stratumiops/crates/stratum-embeddings", features = ["openai-provider", "ollama-provider", "fastembed-provider", "memory-cache"] }
+  stratum-llm = { path = "./stratumiops/crates/stratum-llm", features = ["anthropic", "openai", "ollama"] }
+
+  # ============================================================================
+  # SECRETUMVAULT (Enterprise Secrets Management - optional)
+  # ============================================================================
+  secretumvault = { path = "./secretumvault" }
+
+  # ============================================================================
+  # WASM/WEB-SPECIFIC DEPENDENCIES
+  # ============================================================================
+  web-sys = { version = "0.3", features = [
+    "console",
+    "Window",
+    "Document",
+    "Element",
+    "HtmlElement",
+    "HtmlCanvasElement",
+    "CanvasRenderingContext2d",
+    "EventTarget",
+    "Event",
+    "DragEvent",
+    "DataTransfer",
+    "HtmlInputElement",
+    "HtmlSelectElement",
+    "HtmlTextAreaElement",
+    "HtmlButtonElement",
+    "HtmlDivElement",
+    "Storage",
+    "Location",
+    "History",
+    "Navigator",
+    "ServiceWorkerRegistration",
+    "ServiceWorker",
+    "NotificationPermission",
+    "Notification",
+    "Headers",
+    "Request",
+    "RequestInit",
+    "RequestMode",
+    "Response",
+    "AbortController",
+    "AbortSignal",
+    "WebSocket",
+    "MessageEvent",
+    "CloseEvent",
+    "ErrorEvent",
+    "Blob",
+    "Url",
+    "FileReader",
+    "File",
+    "HtmlAnchorElement",
+    "MouseEvent",
+    "TouchEvent",
+    "KeyboardEvent",
+    "ResizeObserver",
+    "ResizeObserverEntry",
+    "IntersectionObserver",
+    "IntersectionObserverEntry",
+    "MediaQueryList",
+    "MediaQueryListEvent",
+    "CredentialsContainer",
+    "PublicKeyCredential",
+    "PublicKeyCredentialCreationOptions",
+    "PublicKeyCredentialRequestOptions",
+    "AuthenticatorResponse",
+    "AuthenticatorAttestationResponse",
+    "AuthenticatorAssertionResponse",
+    "Crypto",
+    "SubtleCrypto",
+    "CryptoKey",
+  ] }
+
+  # ============================================================================
+  # ADDITIONAL MISSING DEPENDENCIES (Not in original workspace)
+  # ============================================================================
+  ed25519-dalek = "2.2"
+  http-body-util = "0.1"
 
   # ============================================================================
   # BYTES MANIPULATION
   # ============================================================================
-  bytes = "1.5"
+  bytes = "1.11"
+
+  # ============================================================================
+  # HTTP AND PROTOCOL UTILITIES
+  # ============================================================================
+  http = "1"
+
+  # ============================================================================
+  # CONTAINER MANAGEMENT AND SSH
+  # ============================================================================
+  bollard = "0.20"
+  russh = "0.57"
+  russh-keys = "0.49"
+
+  # ============================================================================
+  # SECRETS MANAGEMENT
+  # ============================================================================
+  age = "0.11"
+  rusty_vault = "0.2.1"
+
+  # ============================================================================
+  # ADDITIONAL DATA FORMAT SERIALIZATION
+  # ============================================================================
+  serde_yaml = "0.9"
+
+  # ============================================================================
+  # PATH AND SHELL UTILITIES
+  # ============================================================================
+  shellexpand = "3.1"
 
   [workspace.metadata]
   description = "Provisioning Platform - Rust workspace for cloud infrastructure automation tools"

crates/ai-service/Cargo.toml

@@ -5,6 +5,10 @@ edition.workspace = true
 name = "ai-service"
 version.workspace = true
 
+[[bin]]
+name = "provisioning-ai-service"
+path = "src/main.rs"
+
 [dependencies]
 # Workspace dependencies
 async-trait = { workspace = true }
@@ -22,7 +26,7 @@ serde_json = { workspace = true }
 toml = { workspace = true }
 
 # Platform configuration
-platform-config = { path = "../platform-config" }
+platform-config = { workspace = true }
 
 # Error handling
 anyhow = { workspace = true }
@@ -40,14 +44,18 @@ uuid = { workspace = true, features = ["v4", "serde"] }
 clap = { workspace = true, features = ["derive"] }
 
 # RAG crate for AI capabilities
-provisioning-rag = { path = "../rag" }
+rag = { workspace = true }
 
 # MCP server tools for real implementations
-provisioning-mcp-server = { path = "../mcp-server" }
+mcp-server = { workspace = true }
 
 # Graph operations for DAG
 petgraph = { workspace = true }
 
+# Stratum ecosystem - embeddings and LLM abstraction (optional - requires external setup)
+stratum-embeddings = { workspace = true }
+stratum-llm = { workspace = true }
+
 [dev-dependencies]
 tempfile = { workspace = true }
 tokio-test = { workspace = true }
@@ -56,8 +64,3 @@ tokio-test = { workspace = true }
 [lib]
 name = "ai_service"
 path = "src/lib.rs"
-
-# Binary target
-[[bin]]
-name = "ai-service"
-path = "src/main.rs"

crates/ai-service/src/main.rs

@@ -13,12 +13,24 @@ use tracing_subscriber::{layer::SubscriberExt, util::SubscriberInitExt};
 #[command(name = "ai-service")]
 #[command(about = "HTTP service for AI capabilities including RAG, MCP tool invocation, DAG operations, and knowledge graphs", long_about = None)]
 struct Args {
+    /// Configuration file path (highest priority)
+    #[arg(short = 'c', long, env = "AI_SERVICE_CONFIG")]
+    config: Option<std::path::PathBuf>,
+
+    /// Configuration directory (searches for ai-service.ncl|toml|json)
+    #[arg(long, env = "PROVISIONING_CONFIG_DIR")]
+    config_dir: Option<std::path::PathBuf>,
+
+    /// Deployment mode (solo, multiuser, cicd, enterprise)
+    #[arg(short = 'm', long, env = "AI_SERVICE_MODE")]
+    mode: Option<String>,
+
     /// Service bind address
-    #[arg(short, long, default_value = "127.0.0.1")]
+    #[arg(short = 'H', long, default_value = "127.0.0.1")]
     host: String,
 
     /// Service bind port
-    #[arg(short, long, default_value_t = DEFAULT_PORT)]
+    #[arg(short = 'p', long, default_value_t = DEFAULT_PORT)]
     port: u16,
 }
 

crates/control-center-ui/Cargo.toml

@@ -1,9 +1,11 @@
 [package]
-authors = ["Control Center Team"]
+authors.workspace = true
-autobins = false                                                                       # Disable auto-detection of binary targets
+autobins = false
 description = "Control Center UI - Leptos CSR App for Cloud Infrastructure Management"
 edition.workspace = true
+license.workspace = true
 name = "control-center-ui"
+repository.workspace = true
 version.workspace = true
 
 [lib]
@@ -87,89 +89,21 @@ js-sys = { workspace = true }
 wasm-bindgen-futures = { workspace = true }
 
 # Random number generation (WASM-specific override with js feature)
-getrandom = { version = "0.3.4", features = ["wasm_js"] }
+getrandom = { workspace = true, features = ["wasm_js"] }
 
-# ============================================================================
+# HTTP client
-# PROJECT-SPECIFIC DEPENDENCIES (not in workspace)
+reqwest = { workspace = true, features = ["json"] }
-# ============================================================================
 
-# Web APIs
+# Tokio with time features
-web-sys = { version = "0.3", features = [
+tokio = { workspace = true, features = ["time"] }
-  "console",
-  "Window",
-  "Document",
-  "Element",
-  "HtmlElement",
-  "HtmlCanvasElement",
-  "CanvasRenderingContext2d",
-  "EventTarget",
-  "Event",
-  "DragEvent",
-  "DataTransfer",
-  "HtmlInputElement",
-  "HtmlSelectElement",
-  "HtmlTextAreaElement",
-  "HtmlButtonElement",
-  "HtmlDivElement",
-  "Storage",
-  "Location",
-  "History",
-  "Navigator",
-  "ServiceWorkerRegistration",
-  "ServiceWorker",
-  "NotificationPermission",
-  "Notification",
-  "Headers",
-  "Request",
-  "RequestInit",
-  "RequestMode",
-  "Response",
-  "AbortController",
-  "AbortSignal",
-  "WebSocket",
-  "MessageEvent",
-  "CloseEvent",
-  "ErrorEvent",
-  "Blob",
-  "Url",
-  "FileReader",
-  "File",
-  "HtmlAnchorElement",
-  "MouseEvent",
-  "TouchEvent",
-  "KeyboardEvent",
-  "ResizeObserver",
-  "ResizeObserverEntry",
-  "IntersectionObserver",
-  "IntersectionObserverEntry",
-  # Media Query APIs
-  "MediaQueryList",
-  "MediaQueryListEvent",
-  # WebAuthn APIs
-  "CredentialsContainer",
-  "PublicKeyCredential",
-  "PublicKeyCredentialCreationOptions",
-  "PublicKeyCredentialRequestOptions",
-  "AuthenticatorResponse",
-  "AuthenticatorAttestationResponse",
-  "AuthenticatorAssertionResponse",
-  # Crypto APIs
-  "Crypto",
-  "SubtleCrypto",
-  "CryptoKey",
-] }
 
-# HTTP client (project-specific for WASM features)
+# Web APIs (WASM browser APIs)
-reqwest = { version = "0.13", features = ["json"] }
+web-sys = { workspace = true }
-
-# Tokio with time features for WASM (project-specific version)
-tokio = { version = "1.49", features = ["time"] }
 
 # Profile configurations moved to workspace root
 
-  # WASM pack settings
+[package.metadata.wasm-pack.profile.release]
-  [package.metadata.wasm-pack.profile.release]
+wasm-opt = ['-Oz', '--enable-mutable-globals']
-  wasm-opt = ['-Oz', '--enable-mutable-globals']
 
-  [package.metadata.wasm-pack.profile.dev]
+[package.metadata.wasm-pack.profile.dev]
-  wasm-opt = false
+wasm-opt = false

crates/control-center/Cargo.toml

@@ -52,10 +52,10 @@ validator = { workspace = true }
 reqwest = { workspace = true }
 
 # HTTP service clients (machines, init, AI) - enables remote service calls
-service-clients = { path = "../service-clients" }
+service-clients = { workspace = true }
 
 # Platform configuration management
-platform-config = { path = "../platform-config" }
+platform-config = { workspace = true }
 
 # Security and cryptography
 aes-gcm = { workspace = true }
@@ -153,9 +153,8 @@ compliance = ["core"]
 # Modules: anomaly (detection)
 experimental = ["core"]
 
-# Default: Recommended for standard deployments
+# Default: All features enabled
-# Includes auth, KMS, audit - the essentials
+default = ["core", "kms", "audit", "mfa", "compliance", "experimental"]
-default = ["core", "kms", "audit"]
 
 # Full: All features enabled (development and testing)
 all = ["core", "kms", "audit", "mfa", "compliance", "experimental"]
@@ -165,8 +164,7 @@ all = ["core", "kms", "audit", "mfa", "compliance", "experimental"]
 name = "control_center"
 path = "src/lib.rs"
 
-# Binary target (uses all features)
+# Binary target (uses all features by default)
 [[bin]]
-name = "control-center"
+name = "provisioning-control-center"
 path = "src/main.rs"
-required-features = ["all"]

crates/control-center/Dockerfile

@@ -1,65 +0,0 @@
-# Multi-stage build for Control-Center
-# Builds from platform workspace root
-
-# Build stage - Using nightly for edition2024 support (required by async-graphql 7.x)
-FROM rustlang/rust:nightly-bookworm AS builder
-
-WORKDIR /workspace
-
-# Install build dependencies
-RUN apt-get update && apt-get install -y \
-    pkg-config \
-    libssl-dev \
-    && rm -rf /var/lib/apt/lists/*
-
-# Copy entire platform workspace (required for workspace dependencies)
-COPY Cargo.toml Cargo.lock ./
-COPY orchestrator ./orchestrator
-COPY control-center ./control-center
-COPY control-center-ui ./control-center-ui
-COPY mcp-server ./mcp-server
-COPY installer ./installer
-
-# Build control-center (workspace-aware)
-WORKDIR /workspace
-RUN cargo build --release --package control-center
-
-# Runtime stage
-FROM debian:bookworm-slim
-
-# Install runtime dependencies
-RUN apt-get update && apt-get install -y \
-    ca-certificates \
-    curl \
-    && rm -rf /var/lib/apt/lists/*
-
-# Create non-root user
-RUN useradd -m -u 1000 provisioning && \
-    mkdir -p /data /var/log/control-center && \
-    chown -R provisioning:provisioning /data /var/log/control-center
-
-# Copy binary from builder
-COPY --from=builder /workspace/target/release/control-center /usr/local/bin/control-center
-RUN chmod +x /usr/local/bin/control-center
-
-# Copy default configuration
-COPY control-center/config.defaults.toml /etc/provisioning/config.defaults.toml
-
-# Switch to non-root user
-USER provisioning
-WORKDIR /app
-
-# Expose port
-EXPOSE 8081
-
-# Set environment variables
-ENV RUST_LOG=info
-ENV DATA_DIR=/data
-ENV CONTROL_CENTER_DATABASE_URL=rocksdb:///data/control-center.db
-
-# Health check
-HEALTHCHECK --interval=30s --timeout=10s --start-period=30s --retries=3 \
-    CMD curl -f http://localhost:8081/health || exit 1
-
-# Run the binary with config path
-CMD ["control-center", "--config", "/etc/provisioning/config.defaults.toml"]

crates/control-center/src/main.rs

@@ -47,13 +47,30 @@ use tracing_subscriber::EnvFilter;
 #[command(name = "control-center")]
 #[command(about = "Control Center - JWT Authentication & User Management Service")]
 #[command(version = env!("CARGO_PKG_VERSION"))]
+#[command(after_help = "CONFIGURATION HIERARCHY (highest to lowest priority):\n  1. CLI: -c/--config <path> (explicit file)\n  2. CLI: --config-dir <dir> --mode <mode> (directory + mode)\n  3. CLI: --config-dir <dir> (searches for control-center.ncl|toml|json)\n  4. CLI: --mode <mode> (searches in provisioning/platform/config/)\n  5. ENV: CONTROL_CENTER_CONFIG (explicit file)\n  6. ENV: PROVISIONING_CONFIG_DIR (searches for control-center.ncl|toml|json)\n  7. ENV: CONTROL_CENTER_MODE (mode-based in default path)\n  8. Built-in defaults")]
 struct Cli {
-    /// Configuration file path
+    /// Configuration file path (highest priority)
-    #[arg(short, long, default_value = "config.toml")]
+    ///
+    /// Accepts absolute or relative path. Supports .ncl, .toml, and .json formats.
+    #[arg(short = 'c', long, env = "CONTROL_CENTER_CONFIG")]
     config: Option<PathBuf>,
 
+    /// Configuration directory (searches for control-center.ncl|toml|json)
+    ///
+    /// Searches for configuration files in order of preference: .ncl > .toml > .json
+    /// Can also search for mode-specific files: control-center.{mode}.{ncl|toml|json}
+    #[arg(long, env = "PROVISIONING_CONFIG_DIR")]
+    config_dir: Option<PathBuf>,
+
+    /// Deployment mode (solo, multiuser, cicd, enterprise)
+    ///
+    /// Determines which configuration profile to use. Searches in:
+    /// provisioning/platform/config/control-center.{mode}.{ncl|toml}
+    #[arg(short = 'm', long, env = "CONTROL_CENTER_MODE")]
+    mode: Option<String>,
+
     /// Server port (overrides config file)
-    #[arg(short, long)]
+    #[arg(short = 'p', long)]
     port: Option<u16>,
 
     /// Server host (overrides config file)
@@ -90,9 +107,15 @@ async fn main() -> Result<()> {
         .with_target(false)
         .init();
 
+    // Resolve config file path using new resolver
+    let resolver = platform_config::ConfigResolver::new()
+        .with_cli_config(cli.config.clone())
+        .with_cli_config_dir(cli.config_dir.clone())
+        .with_cli_mode(cli.mode.clone());
+
     // Load configuration
-    let mut config = if let Some(config_path) = cli.config {
+    let mut config = if let Some(path) = resolver.resolve("control-center") {
-        Config::load_from_file(config_path)?
+        Config::load_from_file(path)?
     } else {
         Config::load()?
     };

crates/daemon/Cargo.toml

@@ -2,10 +2,14 @@
 authors.workspace = true
 edition.workspace = true
 license.workspace = true
-name = "provisioning-daemon"
+name = "daemon"
 repository.workspace = true
 version.workspace = true
 
+[[bin]]
+name = "provisioning-daemon"
+path = "src/main.rs"
+
 [dependencies]
 # Core daemon library from prov-ecosystem
 daemon-cli = { workspace = true }
@@ -22,7 +26,7 @@ serde_json = { workspace = true }
 toml = { workspace = true }
 
 # Platform configuration
-platform-config = { path = "../platform-config" }
+platform-config = { workspace = true }
 
 # Error handling
 anyhow = { workspace = true }

crates/daemon/Dockerfile.runtime

@@ -0,0 +1,38 @@
+FROM debian:bookworm-slim
+
+# Install runtime dependencies
+RUN apt-get update && apt-get install -y \
+    ca-certificates \
+    curl \
+    && rm -rf /var/lib/apt/lists/*
+
+# Create user
+RUN useradd -m -u 1000 provisioning && \
+    mkdir -p /data /var/log/provisioning-daemon /etc/provisioning && \
+    chown -R provisioning:provisioning /data /var/log/provisioning-daemon /etc/provisioning
+
+# Copy pre-built binary
+COPY target/release/provisioning-daemon /usr/local/bin/provisioning-daemon
+RUN chmod +x /usr/local/bin/provisioning-daemon
+
+# Copy default configuration files (assumes they're available at build time)
+COPY provisioning/platform/config/runtime/generated/provisioning-daemon.*.toml /etc/provisioning/
+
+USER provisioning
+WORKDIR /app
+
+EXPOSE 8079
+
+ENV RUST_LOG=info
+ENV DATA_DIR=/data
+ENV PROVISIONING_DAEMON_MODE=solo
+ENV PROVISIONING_CONFIG_DIR=/etc/provisioning
+
+HEALTHCHECK --interval=30s --timeout=10s --start-period=30s --retries=3 \
+    CMD curl -f http://localhost:8079/api/v1/health || exit 1
+
+# Configuration precedence:
+# 1. PROVISIONING_DAEMON_CONFIG (explicit path)
+# 2. PROVISIONING_DAEMON_MODE (mode-specific file)
+# 3. Default fallback
+CMD ["provisioning-daemon"]

crates/daemon/src/main.rs

@@ -30,12 +30,20 @@ use tracing_subscriber::EnvFilter;
 #[command(about = "Provisioning platform daemon with Nushell execution and config rendering")]
 #[command(version = env!("CARGO_PKG_VERSION"))]
 struct Args {
-    /// Configuration file path
+    /// Configuration file path (highest priority)
-    #[arg(short, long)]
+    #[arg(short = 'c', long, env = "PROVISIONING_DAEMON_CONFIG")]
     config: Option<PathBuf>,
 
+    /// Configuration directory (searches for provisioning-daemon.ncl|toml|json)
+    #[arg(long, env = "PROVISIONING_CONFIG_DIR")]
+    config_dir: Option<PathBuf>,
+
+    /// Deployment mode (solo, multiuser, cicd, enterprise)
+    #[arg(short = 'm', long, env = "PROVISIONING_DAEMON_MODE")]
+    mode: Option<String>,
+
     /// Enable verbose logging
-    #[arg(short, long)]
+    #[arg(short = 'v', long)]
     verbose: bool,
 
     /// Validate configuration and exit

crates/extension-registry/Cargo.toml

@@ -5,6 +5,10 @@ edition.workspace = true
 name = "extension-registry"
 version.workspace = true
 
+[[bin]]
+name = "provisioning-extension-registry"
+path = "src/main.rs"
+
 [dependencies]
 # Workspace dependencies
 async-trait = { workspace = true }
@@ -21,7 +25,7 @@ serde = { workspace = true, features = ["derive"] }
 serde_json = { workspace = true }
 
 # Platform configuration
-platform-config = { path = "../platform-config" }
+platform-config = { workspace = true }
 
 # Error handling
 anyhow = { workspace = true }
@@ -61,7 +65,7 @@ parking_lot = { workspace = true }
 toml = { workspace = true }
 
 [dev-dependencies]
-http-body-util = "0.1"
+http-body-util = { workspace = true }
 hyper = { workspace = true }
 tempfile = { workspace = true }
 tokio-test = { workspace = true }
@@ -70,8 +74,3 @@ tokio-test = { workspace = true }
 [lib]
 name = "extension_registry"
 path = "src/lib.rs"
-
-# Binary target
-[[bin]]
-name = "extension-registry"
-path = "src/main.rs"

crates/extension-registry/Dockerfile

@@ -1,7 +1,34 @@
-# Build stage
+# Multi-stage build for extension-registry
-FROM rust:1.75-slim as builder
+# Generated from Nickel template - DO NOT EDIT DIRECTLY
+# Source: provisioning/schemas/platform/templates/docker/Dockerfile.chef.ncl
 
-WORKDIR /app
+# ============================================================================
+# Stage 1: PLANNER - Generate dependency recipe
+# ============================================================================
+FROM rust:1.82-trixie AS planner
+
+WORKDIR /workspace
+
+# Install cargo-chef
+RUN cargo install cargo-chef --version 0.1.67
+
+# Copy workspace manifests
+COPY Cargo.toml Cargo.lock ./
+COPY crates ./crates
+COPY daemon-cli ./daemon-cli
+COPY secretumvault ./secretumvault
+COPY prov-ecosystem ./prov-ecosystem
+COPY stratumiops ./stratumiops
+
+# Generate recipe.json (dependency graph)
+RUN cargo chef prepare --recipe-path recipe.json --bin extension-registry
+
+# ============================================================================
+# Stage 2: CACHER - Build dependencies only
+# ============================================================================
+FROM rust:1.82-trixie AS cacher
+
+WORKDIR /workspace
 
 # Install build dependencies
 RUN apt-get update && apt-get install -y \
@@ -9,40 +36,84 @@ RUN apt-get update && apt-get install -y \
     libssl-dev \
     && rm -rf /var/lib/apt/lists/*
 
-# Copy manifests
+# Install cargo-chef
-COPY Cargo.toml Cargo.lock ./
+RUN cargo install cargo-chef --version 0.1.67
+
+# sccache disabled
+
+# Copy recipe from planner
+COPY --from=planner /workspace/recipe.json recipe.json
+
+# Build dependencies - This layer will be cached
+RUN cargo chef cook --release --recipe-path recipe.json 
+
+# ============================================================================
+# Stage 3: BUILDER - Build source code
+# ============================================================================
+FROM rust:1.82-trixie AS builder
+
+WORKDIR /workspace
+
+# Install build dependencies
+RUN apt-get update && apt-get install -y \
+    pkg-config \
+    libssl-dev \
+    && rm -rf /var/lib/apt/lists/*
+
+# sccache disabled
+
+# Copy cached dependencies from cacher stage
+COPY --from=cacher /workspace/target target
+COPY --from=cacher /usr/local/cargo /usr/local/cargo
 
 # Copy source code
-COPY src ./src
+COPY Cargo.toml Cargo.lock ./
+COPY crates ./crates
+COPY daemon-cli ./daemon-cli
+COPY secretumvault ./secretumvault
+COPY prov-ecosystem ./prov-ecosystem
+COPY stratumiops ./stratumiops
 
-# Build release binary
+# Build release binary with parallelism
-RUN cargo build --release
+ENV CARGO_BUILD_JOBS=4
+RUN cargo build --release --package extension-registry 
 
-# Runtime stage
+# ============================================================================
-FROM debian:bookworm-slim
+# Stage 4: RUNTIME - Minimal runtime image
+# ============================================================================
+FROM debian:trixie-slim
 
 # Install runtime dependencies
 RUN apt-get update && apt-get install -y \
     ca-certificates \
+    curl \
     && rm -rf /var/lib/apt/lists/*
 
 # Create non-root user
-RUN useradd -m -u 1000 registry && \
+RUN useradd -m -u 1000 provisioning && \
-    mkdir -p /app/data && \
+    mkdir -p /data /var/log/extension-registry && \
-    chown -R registry:registry /app
+    chown -R provisioning:provisioning /data /var/log/extension-registry
-
-USER registry
-WORKDIR /app
 
 # Copy binary from builder
-COPY --from=builder /app/target/release/extension-registry /usr/local/bin/
+COPY --from=builder /workspace/target/release/extension-registry /usr/local/bin/extension-registry
+RUN chmod +x /usr/local/bin/extension-registry
 
-# Expose port
+# No config file to copy
-EXPOSE 8082
+
+# Switch to non-root user
+USER provisioning
+WORKDIR /app
+
+# Expose service port
+EXPOSE 9093
+
+# Environment variables
+ENV RUST_LOG=info
+ENV DATA_DIR=/data
 
 # Health check
-HEALTHCHECK --interval=30s --timeout=3s --start-period=5s --retries=3 \
+HEALTHCHECK --interval=30s --timeout=10s --start-period=30s --retries=3 \
-  CMD curl -f http://localhost:8082/api/v1/health || exit 1
+    CMD curl -f http://localhost:9093/health || exit 1
 
-# Run service
+# Run the binary
 CMD ["extension-registry"]

crates/extension-registry/src/handlers.rs

@@ -287,5 +287,6 @@ pub fn routes(state: AppState) -> Router {
         .route("/extensions/:name", get(get_extension))
         // Health
         .route("/health", get(health))
+        .route("/api/v1/health", get(health))
         .with_state(state)
 }

crates/extension-registry/src/main.rs

@@ -13,6 +13,18 @@ use handlers::{routes, AppState};
 #[command(name = "extension-registry")]
 #[command(about = "OCI-compliant extension registry proxy", long_about = None)]
 struct Cli {
+    /// Configuration file path (highest priority)
+    #[arg(short = 'c', long, env = "EXTENSION_REGISTRY_CONFIG")]
+    config: Option<std::path::PathBuf>,
+
+    /// Configuration directory (searches for extension-registry.ncl|toml|json)
+    #[arg(long, env = "PROVISIONING_CONFIG_DIR")]
+    config_dir: Option<std::path::PathBuf>,
+
+    /// Deployment mode (solo, multiuser, cicd, enterprise)
+    #[arg(short = 'm', long, env = "EXTENSION_REGISTRY_MODE")]
+    mode: Option<String>,
+
     /// Host to bind to
     #[arg(long, default_value = "127.0.0.1")]
     host: String,

crates/mcp-server/Cargo.toml

@@ -1,14 +1,18 @@
 [package]
-authors = ["Jesús Pérez Lorenzo <jpl@jesusperez.pro>"]
+authors.workspace = true
 categories = ["command-line-utilities", "development-tools"]
 description = "Rust-native MCP server for Infrastructure Automation system"
 edition.workspace = true
 keywords = ["mcp", "rust", "infrastructure", "provisioning", "ai"]
 license.workspace = true
-name = "provisioning-mcp-server"
+name = "mcp-server"
 repository.workspace = true
 version.workspace = true
 
+[[bin]]
+name = "provisioning-mcp-server"
+path = "src/simple_main.rs"
+
 [dependencies]
 # ============================================================================
 # WORKSPACE DEPENDENCIES
@@ -23,7 +27,7 @@ serde_json = { workspace = true }
 toml = { workspace = true }
 
 # Platform configuration
-platform-config = { path = "../platform-config" }
+platform-config = { workspace = true }
 
 # Error handling
 anyhow = { workspace = true }
@@ -63,13 +67,13 @@ walkdir = { workspace = true }
 # rust-mcp-sdk = "0.7.0"
 
 # RAG System (from provisioning-rag crate)
-provisioning-rag = { path = "../rag", features = [] }
+rag = { path = "../rag", features = [] }
 
 # Date/time utilities
 chrono = { workspace = true }
 
 # YAML parsing
-serde_yaml = "0.9"
+serde_yaml = { workspace = true }
 
 # Directory utilities
 dirs = { workspace = true }
@@ -83,14 +87,8 @@ tokio-test = { workspace = true }
 debug = ["tracing-subscriber/json"]
 default = []
 
-[[bin]]
+# Note: simple_main.rs is the active entry point
-name = "provisioning-mcp-server"
+# main.rs uses incompatible rust_mcp_sdk v0.7.0 API and is disabled
-path = "src/simple_main.rs"
-
-# Disabled: main.rs uses incompatible rust_mcp_sdk v0.7.0 API
-# [[bin]]
-# name = "provisioning-mcp-server-full"
-# path = "src/main.rs"
 
 [lib]
 name = "provisioning_mcp_server"

crates/mcp-server/Dockerfile

@@ -1,63 +0,0 @@
-# Build stage
-FROM rust:1.75-slim as builder
-
-WORKDIR /app
-
-# Install build dependencies
-RUN apt-get update && apt-get install -y \
-    pkg-config \
-    libssl-dev \
-    && rm -rf /var/lib/apt/lists/*
-
-# Copy manifests
-COPY Cargo.toml Cargo.lock ./
-
-# Create dummy source to cache dependencies
-RUN mkdir -p src && \
-    echo "fn main() {}" > src/main.rs && \
-    cargo build --release && \
-    rm -rf src
-
-# Copy actual source code
-COPY src ./src
-
-# Build release binary
-RUN cargo build --release --bin mcp-server
-
-# Runtime stage
-FROM debian:bookworm-slim
-
-# Install runtime dependencies
-RUN apt-get update && apt-get install -y \
-    ca-certificates \
-    curl \
-    && rm -rf /var/lib/apt/lists/*
-
-# Create non-root user
-RUN useradd -m -u 1000 provisioning && \
-    mkdir -p /data /var/log/mcp-server && \
-    chown -R provisioning:provisioning /data /var/log/mcp-server
-
-# Copy binary from builder
-COPY --from=builder /app/target/release/mcp-server /usr/local/bin/
-
-# Copy default configuration
-COPY config.defaults.toml /etc/provisioning/mcp-config.defaults.toml
-
-# Switch to non-root user
-USER provisioning
-WORKDIR /app
-
-# Expose port
-EXPOSE 8084
-
-# Set environment variables
-ENV RUST_LOG=info
-ENV MCP_PROTOCOL=http
-
-# Health check
-HEALTHCHECK --interval=30s --timeout=10s --start-period=30s --retries=3 \
-    CMD curl -f http://localhost:8084/health || exit 1
-
-# Run the binary
-CMD ["mcp-server"]

crates/orchestrator/Cargo.toml

@@ -2,7 +2,7 @@
 authors.workspace = true
 description = "Cloud-native infrastructure orchestrator with Nushell integration"
 edition.workspace = true
-name = "provisioning-orchestrator"
+name = "orchestrator"
 version.workspace = true
 
 [dependencies]
@@ -45,28 +45,28 @@ clap = { workspace = true }
 tracing-subscriber = { workspace = true }
 
 # Docker/Container management
-bollard = "0.17"
+bollard = { workspace = true }
 
 # HTTP client for DNS/OCI/services
 reqwest = { workspace = true }
 
 # HTTP service clients (machines, init, AI) - enables remote service calls
-service-clients = { path = "../service-clients" }
+service-clients = { workspace = true }
 
 # Platform configuration management
-platform-config = { path = "../platform-config" }
+platform-config = { workspace = true }
 
 # LRU cache for OCI manifests
-lru = "0.12"
+lru = { workspace = true }
 
 # Authorization policy engine
-cedar-policy = "4.2"
+cedar-policy = { workspace = true }
 
 # File system watcher for hot reload
-notify = "6.1"
+notify = { workspace = true }
 
 # Base64 encoding/decoding
-base64 = "0.22"
+base64 = { workspace = true }
 
 # JWT token validation
 jsonwebtoken = { workspace = true }
@@ -78,14 +78,14 @@ rsa = { workspace = true }
 sha2 = { workspace = true }
 
 # SSH key management
-ed25519-dalek = "2.1"
+ed25519-dalek = { workspace = true }
 
 # SSH client library (pure Rust, async-first)
-russh = "0.44"
+russh = { workspace = true }
-russh-keys = "0.44"
+russh-keys = { workspace = true }
 
 # Path expansion for tilde (~) handling
-shellexpand = "3.1"
+shellexpand = { workspace = true }
 
 # ============================================================================
 # FEATURE-GATED OPTIONAL DEPENDENCIES
@@ -141,9 +141,18 @@ http-api = ["core"]
 # SurrealDB: Optional storage backend
 surrealdb = ["dep:surrealdb"]
 
-# Default: Recommended for standard deployments
+# Default: All features enabled
-# Includes core, audit, compliance, platform, ssh, workflow
+default = [
-default = ["core", "audit", "compliance", "platform", "ssh", "workflow", "http-api"]
+  "core",
+  "audit",
+  "compliance",
+  "platform",
+  "ssh",
+  "workflow",
+  "testing",
+  "http-api",
+  "surrealdb",
+]
 
 # Full: All features enabled (development and testing)
 all = [
@@ -170,11 +179,10 @@ tower = { workspace = true, features = ["util"] }
 name = "provisioning_orchestrator"
 path = "src/lib.rs"
 
-# Binary target (requires testing feature for test environment API)
+# Binary target (uses all features by default)
 [[bin]]
 name = "provisioning-orchestrator"
 path = "src/main.rs"
-required-features = ["all"]
 
 [[bench]]
 harness = false

crates/orchestrator/Dockerfile

@@ -1,8 +1,32 @@
-# Multi-stage build for Orchestrator
+# Multi-stage build for provisioning-orchestrator
-# Builds from platform workspace root
+# Generated from Nickel template - DO NOT EDIT DIRECTLY
+# Source: provisioning/schemas/platform/templates/docker/Dockerfile.chef.ncl
 
-# Build stage - Using nightly for consistency with control-center
+# ============================================================================
-FROM rustlang/rust:nightly-bookworm AS builder
+# Stage 1: PLANNER - Generate dependency recipe
+# ============================================================================
+FROM rust:1.82-trixie AS planner
+
+WORKDIR /workspace
+
+# Install cargo-chef
+RUN cargo install cargo-chef --version 0.1.67
+
+# Copy workspace manifests
+COPY Cargo.toml Cargo.lock ./
+COPY crates ./crates
+COPY daemon-cli ./daemon-cli
+COPY secretumvault ./secretumvault
+COPY prov-ecosystem ./prov-ecosystem
+COPY stratumiops ./stratumiops
+
+# Generate recipe.json (dependency graph)
+RUN cargo chef prepare --recipe-path recipe.json --bin provisioning-orchestrator
+
+# ============================================================================
+# Stage 2: CACHER - Build dependencies only
+# ============================================================================
+FROM rust:1.82-trixie AS cacher
 
 WORKDIR /workspace
 
@@ -12,20 +36,52 @@ RUN apt-get update && apt-get install -y \
     libssl-dev \
     && rm -rf /var/lib/apt/lists/*
 
-# Copy entire platform workspace (required for workspace dependencies)
+# Install cargo-chef
-COPY Cargo.toml Cargo.lock ./
+RUN cargo install cargo-chef --version 0.1.67
-COPY orchestrator ./orchestrator
+
-COPY control-center ./control-center
+# sccache disabled
-COPY control-center-ui ./control-center-ui
+
-COPY mcp-server ./mcp-server
+# Copy recipe from planner
-COPY installer ./installer
+COPY --from=planner /workspace/recipe.json recipe.json
+
+# Build dependencies - This layer will be cached
+RUN cargo chef cook --release --recipe-path recipe.json 
+
+# ============================================================================
+# Stage 3: BUILDER - Build source code
+# ============================================================================
+FROM rust:1.82-trixie AS builder
 
-# Build orchestrator (workspace-aware)
 WORKDIR /workspace
-RUN cargo build --release --package provisioning-orchestrator
 
-# Runtime stage
+# Install build dependencies
-FROM debian:bookworm-slim
+RUN apt-get update && apt-get install -y \
+    pkg-config \
+    libssl-dev \
+    && rm -rf /var/lib/apt/lists/*
+
+# sccache disabled
+
+# Copy cached dependencies from cacher stage
+COPY --from=cacher /workspace/target target
+COPY --from=cacher /usr/local/cargo /usr/local/cargo
+
+# Copy source code
+COPY Cargo.toml Cargo.lock ./
+COPY crates ./crates
+COPY daemon-cli ./daemon-cli
+COPY secretumvault ./secretumvault
+COPY prov-ecosystem ./prov-ecosystem
+COPY stratumiops ./stratumiops
+
+# Build release binary with parallelism
+ENV CARGO_BUILD_JOBS=4
+RUN cargo build --release --package provisioning-orchestrator 
+
+# ============================================================================
+# Stage 4: RUNTIME - Minimal runtime image
+# ============================================================================
+FROM debian:trixie-slim
 
 # Install runtime dependencies
 RUN apt-get update && apt-get install -y \
@@ -35,30 +91,29 @@ RUN apt-get update && apt-get install -y \
 
 # Create non-root user
 RUN useradd -m -u 1000 provisioning && \
-    mkdir -p /data /var/log/orchestrator && \
+    mkdir -p /data /var/log/provisioning-orchestrator && \
-    chown -R provisioning:provisioning /data /var/log/orchestrator
+    chown -R provisioning:provisioning /data /var/log/provisioning-orchestrator
 
 # Copy binary from builder
 COPY --from=builder /workspace/target/release/provisioning-orchestrator /usr/local/bin/provisioning-orchestrator
 RUN chmod +x /usr/local/bin/provisioning-orchestrator
 
-# Copy default configuration
+COPY crates/provisioning-orchestrator/config.defaults.toml /etc/provisioning/config.defaults.toml
-COPY orchestrator/config.defaults.toml /etc/provisioning/config.defaults.toml
 
 # Switch to non-root user
 USER provisioning
 WORKDIR /app
 
-# Expose port
+# Expose service port
-EXPOSE 8080
+EXPOSE 9090
 
-# Set environment variables
+# Environment variables
 ENV RUST_LOG=info
 ENV DATA_DIR=/data
 
 # Health check
 HEALTHCHECK --interval=30s --timeout=10s --start-period=30s --retries=3 \
-    CMD curl -f http://localhost:8080/health || exit 1
+    CMD curl -f http://localhost:9090/health || exit 1
 
 # Run the binary
 CMD ["provisioning-orchestrator"]

crates/orchestrator/rollback_instructions_14043518-e459-4316-aadd-6ee6d221e644.txt

@@ -1,20 +0,0 @@
-Manual Rollback Instructions for Migration migration-123
-=============================================
-
-Migration ID: migration-123
-Rollback ID: 14043518-e459-4316-aadd-6ee6d221e644
-Timestamp: 2026-01-06 12:47:28 UTC
-
-Steps to rollback:
-1. Stop the orchestrator service
-2. If backup exists at '/var/folders/my/jvl5hd5x6rgbhks6yszk213c0000gn/T/.tmpdtjZLt/backup.json', restore it to target storage
-3. Verify target storage contains original data
-4. Remove any partially migrated data
-5. Restart orchestrator service
-
-Storage-specific cleanup commands:
-- For filesystem: Remove or rename target directory
-- For SurrealDB embedded: Delete database files in target directory
-- For SurrealDB server: Connect and drop/recreate database
-
-IMPORTANT: Test data integrity before resuming operations!

crates/orchestrator/rollback_instructions_1e9b4914-f290-4bec-80f2-35128250f9fd.txt

@@ -1,20 +0,0 @@
-Manual Rollback Instructions for Migration migration-123
-=============================================
-
-Migration ID: migration-123
-Rollback ID: 1e9b4914-f290-4bec-80f2-35128250f9fd
-Timestamp: 2026-01-06 12:50:41 UTC
-
-Steps to rollback:
-1. Stop the orchestrator service
-2. If backup exists at '/var/folders/my/jvl5hd5x6rgbhks6yszk213c0000gn/T/.tmp9wM3YA/backup.json', restore it to target storage
-3. Verify target storage contains original data
-4. Remove any partially migrated data
-5. Restart orchestrator service
-
-Storage-specific cleanup commands:
-- For filesystem: Remove or rename target directory
-- For SurrealDB embedded: Delete database files in target directory
-- For SurrealDB server: Connect and drop/recreate database
-
-IMPORTANT: Test data integrity before resuming operations!

crates/orchestrator/rollback_instructions_21c8a4af-2562-4304-b5ec-90fb1b5fd0ab.txt

@@ -1,20 +0,0 @@
-Manual Rollback Instructions for Migration migration-123
-=============================================
-
-Migration ID: migration-123
-Rollback ID: 21c8a4af-2562-4304-b5ec-90fb1b5fd0ab
-Timestamp: 2026-01-06 13:10:16 UTC
-
-Steps to rollback:
-1. Stop the orchestrator service
-2. If backup exists at '/var/folders/my/jvl5hd5x6rgbhks6yszk213c0000gn/T/.tmpnoxnXR/backup.json', restore it to target storage
-3. Verify target storage contains original data
-4. Remove any partially migrated data
-5. Restart orchestrator service
-
-Storage-specific cleanup commands:
-- For filesystem: Remove or rename target directory
-- For SurrealDB embedded: Delete database files in target directory
-- For SurrealDB server: Connect and drop/recreate database
-
-IMPORTANT: Test data integrity before resuming operations!

crates/orchestrator/rollback_instructions_317e31fa-b549-49c9-a212-1f13445d913f.txt

@@ -1,20 +0,0 @@
-Manual Rollback Instructions for Migration migration-123
-=============================================
-
-Migration ID: migration-123
-Rollback ID: 317e31fa-b549-49c9-a212-1f13445d913f
-Timestamp: 2026-01-06 12:49:14 UTC
-
-Steps to rollback:
-1. Stop the orchestrator service
-2. If backup exists at '/var/folders/my/jvl5hd5x6rgbhks6yszk213c0000gn/T/.tmpe5B6TH/backup.json', restore it to target storage
-3. Verify target storage contains original data
-4. Remove any partially migrated data
-5. Restart orchestrator service
-
-Storage-specific cleanup commands:
-- For filesystem: Remove or rename target directory
-- For SurrealDB embedded: Delete database files in target directory
-- For SurrealDB server: Connect and drop/recreate database
-
-IMPORTANT: Test data integrity before resuming operations!

crates/orchestrator/rollback_instructions_5da5d888-527e-4aac-ab53-93e9a30014cc.txt

@@ -1,20 +0,0 @@
-Manual Rollback Instructions for Migration migration-123
-=============================================
-
-Migration ID: migration-123
-Rollback ID: 5da5d888-527e-4aac-ab53-93e9a30014cc
-Timestamp: 2026-01-06 12:53:52 UTC
-
-Steps to rollback:
-1. Stop the orchestrator service
-2. If backup exists at '/var/folders/my/jvl5hd5x6rgbhks6yszk213c0000gn/T/.tmpOI6ga7/backup.json', restore it to target storage
-3. Verify target storage contains original data
-4. Remove any partially migrated data
-5. Restart orchestrator service
-
-Storage-specific cleanup commands:
-- For filesystem: Remove or rename target directory
-- For SurrealDB embedded: Delete database files in target directory
-- For SurrealDB server: Connect and drop/recreate database
-
-IMPORTANT: Test data integrity before resuming operations!

crates/orchestrator/rollback_instructions_7c16746f-24b0-4bcc-8a49-b5dc6bc1f0c7.txt

@@ -1,20 +0,0 @@
-Manual Rollback Instructions for Migration migration-123
-=============================================
-
-Migration ID: migration-123
-Rollback ID: 7c16746f-24b0-4bcc-8a49-b5dc6bc1f0c7
-Timestamp: 2026-01-06 12:59:15 UTC
-
-Steps to rollback:
-1. Stop the orchestrator service
-2. If backup exists at '/var/folders/my/jvl5hd5x6rgbhks6yszk213c0000gn/T/.tmp8JuU01/backup.json', restore it to target storage
-3. Verify target storage contains original data
-4. Remove any partially migrated data
-5. Restart orchestrator service
-
-Storage-specific cleanup commands:
-- For filesystem: Remove or rename target directory
-- For SurrealDB embedded: Delete database files in target directory
-- For SurrealDB server: Connect and drop/recreate database
-
-IMPORTANT: Test data integrity before resuming operations!

crates/orchestrator/rollback_instructions_cb3ced5a-ab49-4754-ba90-c815ab0948ba.txt

@@ -1,20 +0,0 @@
-Manual Rollback Instructions for Migration migration-123
-=============================================
-
-Migration ID: migration-123
-Rollback ID: cb3ced5a-ab49-4754-ba90-c815ab0948ba
-Timestamp: 2026-01-06 12:53:37 UTC
-
-Steps to rollback:
-1. Stop the orchestrator service
-2. If backup exists at '/var/folders/my/jvl5hd5x6rgbhks6yszk213c0000gn/T/.tmpI1vYTj/backup.json', restore it to target storage
-3. Verify target storage contains original data
-4. Remove any partially migrated data
-5. Restart orchestrator service
-
-Storage-specific cleanup commands:
-- For filesystem: Remove or rename target directory
-- For SurrealDB embedded: Delete database files in target directory
-- For SurrealDB server: Connect and drop/recreate database
-
-IMPORTANT: Test data integrity before resuming operations!

crates/orchestrator/src/container_manager.rs

@@ -8,16 +8,14 @@ use std::default::Default;
 
 use anyhow::{anyhow, Context, Result};
 use bollard::{
-    container::{
-        Config, CreateContainerOptions, ListContainersOptions, LogsOptions, NetworkingConfig,
-        RemoveContainerOptions, StartContainerOptions, StopContainerOptions,
-    },
     exec::{CreateExecOptions, StartExecResults},
-    image::CreateImageOptions,
+    models::{
-    network::CreateNetworkOptions,
+        ContainerCreateBody, ContainerSummary, EndpointSettings, HostConfig, Ipam, IpamConfig,
-    service::{
+        NetworkingConfig, RestartPolicy, RestartPolicyNameEnum,
-        ContainerSummary, EndpointSettings, HostConfig, Ipam, IpamConfig, RestartPolicy,
+    },
-        RestartPolicyNameEnum,
+    query_parameters::{
+        CreateContainerOptions, CreateImageOptions, ListContainersOptions, LogsOptions,
+        RemoveContainerOptions, StartContainerOptions, StopContainerOptions,
     },
     Docker,
 };
@@ -58,6 +56,8 @@ impl ContainerManager {
 
     /// Create network for test environment
     pub async fn create_network(&self, config: &NetworkConfig) -> Result<String> {
+        use bollard::models::NetworkCreateRequest;
+
         let ipam_config = IpamConfig {
             subnet: Some(config.subnet.clone()),
             gateway: None,
@@ -71,36 +71,32 @@ impl ContainerManager {
             options: None,
         };
 
-        let create_options = CreateNetworkOptions {
+        let mut labels = HashMap::new();
+        labels.insert(
+            "managed_by".to_string(),
+            "provisioning_orchestrator".to_string(),
+        );
+        labels.insert("type".to_string(), "test_environment".to_string());
+
+        let create_request = NetworkCreateRequest {
             name: config.name.clone(),
-            check_duplicate: true,
+            driver: Some("bridge".to_string()),
-            driver: "bridge".to_string(),
+            internal: Some(false),
-            internal: false,
+            attachable: Some(true),
-            attachable: true,
+            ipam: Some(ipam),
-            ingress: false,
+            enable_ipv4: Some(true),
-            ipam,
+            options: Some(HashMap::new()),
-            enable_ipv6: false,
+            labels: Some(labels),
-            options: HashMap::new(),
+            ..Default::default()
-            labels: {
-                let mut labels = HashMap::new();
-                labels.insert(
-                    "managed_by".to_string(),
-                    "provisioning_orchestrator".to_string(),
-                );
-                labels.insert("type".to_string(), "test_environment".to_string());
-                labels
-            },
         };
 
         let response = self
             .docker
-            .create_network(create_options)
+            .create_network(create_request)
             .await
             .context("Failed to create Docker network")?;
 
-        let network_id = response
+        let network_id = response.id;
-            .id
-            .ok_or_else(|| anyhow!("Network ID not returned"))?;
 
         info!("Created network {} ({})", config.name, network_id);
         Ok(network_id)
@@ -119,12 +115,12 @@ impl ContainerManager {
 
     /// Pull image if not exists
     pub async fn ensure_image(&self, image: &str) -> Result<()> {
-        let options = Some(CreateImageOptions {
+        let options = CreateImageOptions {
-            from_image: image,
+            from_image: Some(image.to_string()),
             ..Default::default()
-        });
+        };
 
-        let mut stream = self.docker.create_image(options, None, None);
+        let mut stream = self.docker.create_image(Some(options), None, None);
 
         while let Some(result) = stream.next().await {
             match result {
@@ -176,7 +172,6 @@ impl ContainerManager {
             endpoint_config.insert(
                 net_id.to_string(),
                 EndpointSettings {
-                    network_id: Some(net_id.to_string()),
                     ..Default::default()
                 },
             );
@@ -190,31 +185,30 @@ impl ContainerManager {
             .collect();
 
         // Container configuration
-        let config = Config {
+        let mut labels = HashMap::new();
+        labels.insert(
+            "managed_by".to_string(),
+            "provisioning_orchestrator".to_string(),
+        );
+        labels.insert("type".to_string(), "test_container".to_string());
+        labels.insert("test_name".to_string(), name.to_string());
+
+        let config = ContainerCreateBody {
             image: Some(image.to_string()),
             hostname: Some(name.to_string()),
-            env: Some(env),
+            env: if env.is_empty() { None } else { Some(env) },
             cmd: command,
             host_config: Some(host_config),
             networking_config: Some(NetworkingConfig {
-                endpoints_config: endpoint_config,
+                endpoints_config: Some(endpoint_config),
-            }),
-            labels: Some({
-                let mut labels = HashMap::new();
-                labels.insert(
-                    "managed_by".to_string(),
-                    "provisioning_orchestrator".to_string(),
-                );
-                labels.insert("type".to_string(), "test_container".to_string());
-                labels.insert("test_name".to_string(), name.to_string());
-                labels
             }),
+            labels: Some(labels),
             ..Default::default()
         };
 
         let options = CreateContainerOptions {
-            name: name.to_string(),
+            name: Some(name.to_string()),
-            platform: None,
+            platform: String::new(),
         };
 
         let response = self
@@ -240,7 +234,7 @@ impl ContainerManager {
     /// Start container
     pub async fn start_container(&self, container_id: &str) -> Result<()> {
         self.docker
-            .start_container(container_id, None::<StartContainerOptions<String>>)
+            .start_container(container_id, None::<StartContainerOptions>)
             .await
             .context(format!("Failed to start container {}", container_id))?;
 
@@ -251,7 +245,8 @@ impl ContainerManager {
     /// Stop container
     pub async fn stop_container(&self, container_id: &str, timeout: Option<i64>) -> Result<()> {
         let options = StopContainerOptions {
-            t: timeout.unwrap_or(10),
+            t: Some(timeout.unwrap_or(10) as i32),
+            signal: None,
         };
 
         self.docker
@@ -329,7 +324,7 @@ impl ContainerManager {
         container_id: &str,
         tail: Option<&str>,
     ) -> Result<String> {
-        let options = LogsOptions::<String> {
+        let options = LogsOptions {
             stdout: true,
             stderr: true,
             tail: tail.unwrap_or("100").to_string(),
@@ -386,15 +381,15 @@ impl ContainerManager {
         let mut filters = HashMap::new();
         filters.insert("label".to_string(), vec![format!("{}={}", label, value)]);
 
-        let options = Some(ListContainersOptions {
+        let options = ListContainersOptions {
             all: true,
-            filters,
+            filters: Some(filters),
             ..Default::default()
-        });
+        };
 
         let containers = self
             .docker
-            .list_containers(options)
+            .list_containers(Some(options))
             .await
             .context("Failed to list containers")?;
 

crates/orchestrator/src/lib.rs

@@ -85,14 +85,36 @@ pub fn validate_storage_type(s: &str) -> Result<String, String> {
 
 // CLI arguments structure
 #[derive(clap::Parser, Clone)]
-#[command(author, version, about, long_about = None)]
+#[command(author, version, about = "Multi-service task orchestration and batch workflow engine")]
+#[command(long_about = "Orchestrator - Manages distributed task execution, batch workflows, and cluster provisioning with state management and rollback recovery")]
+#[command(after_help = "CONFIGURATION HIERARCHY (highest to lowest priority):\n  1. CLI: -c/--config <path> (explicit file)\n  2. CLI: --config-dir <dir> --mode <mode> (directory + mode)\n  3. CLI: --config-dir <dir> (searches for orchestrator.ncl|toml|json)\n  4. CLI: --mode <mode> (searches in provisioning/platform/config/)\n  5. ENV: ORCHESTRATOR_CONFIG (explicit file)\n  6. ENV: PROVISIONING_CONFIG_DIR (searches for orchestrator.ncl|toml|json)\n  7. ENV: ORCHESTRATOR_MODE (mode-based in default path)\n  8. Built-in defaults\n\nEXAMPLES:\n  # Explicit config file\n  orchestrator -c ~/my-config.toml\n\n  # Config directory with mode\n  orchestrator --config-dir ~/configs --mode enterprise\n\n  # Config directory (auto-discover file)\n  orchestrator --config-dir ~/.config/provisioning\n\n  # Via environment variables\n  export ORCHESTRATOR_CONFIG=~/.config/orchestrator.toml\n  orchestrator\n\n  # Mode-based configuration\n  orchestrator --mode solo")]
 pub struct Args {
+    /// Configuration file path (highest priority)
+    ///
+    /// Accepts absolute or relative path. Supports .ncl, .toml, and .json formats.
+    #[arg(short = 'c', long, env = "ORCHESTRATOR_CONFIG")]
+    pub config: Option<std::path::PathBuf>,
+
+    /// Configuration directory (searches for orchestrator.ncl|toml|json)
+    ///
+    /// Searches for configuration files in order of preference: .ncl > .toml > .json
+    /// Can also search for mode-specific files: orchestrator.{mode}.{ncl|toml|json}
+    #[arg(long, env = "PROVISIONING_CONFIG_DIR")]
+    pub config_dir: Option<std::path::PathBuf>,
+
+    /// Deployment mode (solo, multiuser, cicd, enterprise)
+    ///
+    /// Determines which configuration profile to use. Searches in:
+    /// provisioning/platform/config/orchestrator.{mode}.{ncl|toml}
+    #[arg(short = 'm', long, env = "ORCHESTRATOR_MODE")]
+    pub mode: Option<String>,
+
     /// Port to listen on
-    #[arg(short, long, default_value = "9090")]
+    #[arg(short = 'p', long, default_value = "9090")]
     pub port: u16,
 
     /// Data directory for storage
-    #[arg(short, long, default_value = "./data")]
+    #[arg(short = 'd', long, default_value = "./data")]
     pub data_dir: String,
 
     /// Storage backend type

crates/orchestrator/src/main.rs

@@ -1007,6 +1007,7 @@ async fn main() -> Result<()> {
 
     let app = Router::new()
         .route("/health", get(health_check))
+        .route("/api/v1/health", get(health_check))
         .route("/tasks", get(list_tasks))
         .route("/tasks/{id}", get(get_task_status))
         .route("/workflows/servers/create", post(create_server_workflow))

crates/platform-config/src/hierarchy.rs

@@ -8,7 +8,7 @@ const CONFIG_BASE_PATH: &str = "provisioning/platform/config";
 /// 2. Variable de entorno {SERVICE}_MODE + búsqueda de archivo
 /// 3. Fallback a defaults
 pub fn resolve_config_path(service_name: &str) -> Option<PathBuf> {
-    // Paso 1: Check {SERVICE}_CONFIG env var (explicit path)
+    // Priority 1: Check {SERVICE}_CONFIG env var (explicit path)
     let env_var = format!("{}_CONFIG", service_name.to_uppercase().replace('-', "_"));
     if let Ok(path) = env::var(&env_var) {
         let config_path = PathBuf::from(path);
@@ -22,7 +22,18 @@ pub fn resolve_config_path(service_name: &str) -> Option<PathBuf> {
         }
     }
 
-    // Paso 2: Check {SERVICE}_MODE env var + find config file
+    // Priority 2: Check PROVISIONING_CONFIG_DIR env var
+    if let Ok(dir) = env::var("PROVISIONING_CONFIG_DIR") {
+        if let Some(config) = super::resolver::find_config_in_dir(std::path::Path::new(&dir), service_name) {
+            tracing::debug!(
+                "Using config from PROVISIONING_CONFIG_DIR: {:?}",
+                config
+            );
+            return Some(config);
+        }
+    }
+
+    // Priority 3: Check {SERVICE}_MODE env var + find config file
     let mode_var = format!("{}_MODE", service_name.to_uppercase().replace('-', "_"));
     let mode = env::var(&mode_var).unwrap_or_else(|_| "solo".to_string());
 
@@ -36,7 +47,7 @@ pub fn resolve_config_path(service_name: &str) -> Option<PathBuf> {
         return Some(path);
     }
 
-    // Paso 3: Fallback - no config file found
+    // Fallback - no config file found
     tracing::debug!(
         "No config file found for {}.{} - using defaults",
         service_name,

crates/platform-config/src/lib.rs

@@ -53,6 +53,7 @@ pub mod format;
 pub mod hierarchy;
 pub mod loader;
 pub mod nickel;
+pub mod resolver;
 
 // Re-export main types
 pub use error::{ConfigError, Result};
@@ -60,3 +61,4 @@ pub use format::ConfigLoader;
 pub use hierarchy::{config_base_path, find_config_file, resolve_config_path};
 pub use loader::{ConfigLoaderExt, ConfigValidator};
 pub use nickel::is_nickel_available;
+pub use resolver::{ConfigResolver, find_config_in_dir, find_config_in_dir_with_mode};

crates/platform-config/src/resolver.rs

@@ -0,0 +1,212 @@
+use std::path::{Path, PathBuf};
+
+/// Resolves configuration file paths with CLI flags priority
+#[derive(Debug, Clone, Default)]
+pub struct ConfigResolver {
+    cli_config: Option<PathBuf>,
+    cli_config_dir: Option<PathBuf>,
+    cli_mode: Option<String>,
+}
+
+impl ConfigResolver {
+    /// Create a new ConfigResolver with no CLI overrides
+    pub fn new() -> Self {
+        Self {
+            cli_config: None,
+            cli_config_dir: None,
+            cli_mode: None,
+        }
+    }
+
+    /// Set explicit config file path (highest priority)
+    pub fn with_cli_config(mut self, path: Option<PathBuf>) -> Self {
+        self.cli_config = path;
+        self
+    }
+
+    /// Set config directory (searches for {service}.{ncl|toml|json})
+    pub fn with_cli_config_dir(mut self, dir: Option<PathBuf>) -> Self {
+        self.cli_config_dir = dir;
+        self
+    }
+
+    /// Set deployment mode for config resolution
+    pub fn with_cli_mode(mut self, mode: Option<String>) -> Self {
+        self.cli_mode = mode;
+        self
+    }
+
+    /// Resolve config file path with priority order:
+    /// 1. CLI flag: -config <path> (explicit file)
+    /// 2. CLI flag: --config-dir <dir> --mode <mode> (directory + mode)
+    /// 3. CLI flag: --config-dir <dir> (directory, default naming)
+    /// 4. CLI flag: --mode <mode> (mode in default path)
+    /// 5. ENV var: {SERVICE}_CONFIG (explicit file)
+    /// 6. ENV var: PROVISIONING_CONFIG_DIR (directory, default naming)
+    /// 7. ENV var: {SERVICE}_MODE (mode in default path)
+    /// 8. None (fallback to defaults)
+    pub fn resolve(&self, service_name: &str) -> Option<PathBuf> {
+        // Priority 1: CLI flag explicit path
+        if let Some(ref path) = self.cli_config {
+            tracing::debug!("Using CLI-provided config file: {:?}", path);
+            return Some(path.clone());
+        }
+
+        // Priority 2: CLI config-dir + mode
+        if let Some(ref dir) = self.cli_config_dir {
+            if let Some(ref mode) = self.cli_mode {
+                if let Some(config) = find_config_in_dir_with_mode(dir, service_name, mode) {
+                    tracing::debug!(
+                        "Using config file from CLI config-dir with mode: {:?}",
+                        config
+                    );
+                    return Some(config);
+                }
+            }
+        }
+
+        // Priority 3: CLI config-dir only
+        if let Some(ref dir) = self.cli_config_dir {
+            if let Some(config) = find_config_in_dir(dir, service_name) {
+                tracing::debug!("Using config file from CLI config-dir: {:?}", config);
+                return Some(config);
+            }
+        }
+
+        // Priority 4: CLI mode only (searches in default path)
+        if let Some(ref mode) = self.cli_mode {
+            if let Some(config) = super::hierarchy::find_config_file(service_name, mode) {
+                tracing::debug!("Using config file with CLI mode: {:?}", config);
+                return Some(config);
+            }
+        }
+
+        // Priority 5-7: Fall back to environment variable resolution
+        super::hierarchy::resolve_config_path(service_name)
+    }
+}
+
+/// Search for config file in directory with specific mode
+/// Searches in order: {service}.{mode}.ncl, {service}.{mode}.toml, {service}.{mode}.json
+pub fn find_config_in_dir_with_mode(dir: &Path, service_name: &str, mode: &str) -> Option<PathBuf> {
+    for ext in &["ncl", "toml", "json"] {
+        let path = dir.join(format!("{}.{}.{}", service_name, mode, ext));
+        if path.exists() {
+            tracing::trace!("Found config with mode: {:?}", path);
+            return Some(path);
+        }
+    }
+    None
+}
+
+/// Search for config file in directory with default naming
+/// Searches in order: {service}.ncl, {service}.toml, {service}.json
+pub fn find_config_in_dir(dir: &Path, service_name: &str) -> Option<PathBuf> {
+    for ext in &["ncl", "toml", "json"] {
+        let path = dir.join(format!("{}.{}", service_name, ext));
+        if path.exists() {
+            tracing::trace!("Found config in dir: {:?}", path);
+            return Some(path);
+        }
+    }
+    None
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use std::fs;
+    use tempfile::TempDir;
+
+    #[test]
+    fn test_cli_config_highest_priority() {
+        let resolver = ConfigResolver::new()
+            .with_cli_config(Some(PathBuf::from("/explicit/path.toml")));
+
+        let resolved = resolver.resolve("orchestrator");
+        assert_eq!(resolved, Some(PathBuf::from("/explicit/path.toml")));
+    }
+
+    #[test]
+    fn test_config_dir_searches_extensions_in_order() {
+        let temp_dir = TempDir::new().unwrap();
+        let ncl_path = temp_dir.path().join("orchestrator.ncl");
+        let toml_path = temp_dir.path().join("orchestrator.toml");
+
+        // Create both files
+        fs::write(&ncl_path, "{}").unwrap();
+        fs::write(&toml_path, "[orchestrator]").unwrap();
+
+        let resolver = ConfigResolver::new()
+            .with_cli_config_dir(Some(temp_dir.path().to_path_buf()));
+
+        let resolved = resolver.resolve("orchestrator").unwrap();
+        // Should prefer .ncl over .toml
+        assert_eq!(resolved, ncl_path);
+    }
+
+    #[test]
+    fn test_config_dir_with_mode() {
+        let temp_dir = TempDir::new().unwrap();
+        let enterprise_path = temp_dir.path().join("orchestrator.enterprise.toml");
+        fs::write(&enterprise_path, "[orchestrator]").unwrap();
+
+        let resolver = ConfigResolver::new()
+            .with_cli_config_dir(Some(temp_dir.path().to_path_buf()))
+            .with_cli_mode(Some("enterprise".to_string()));
+
+        let resolved = resolver.resolve("orchestrator").unwrap();
+        assert_eq!(resolved, enterprise_path);
+    }
+
+    #[test]
+    fn test_find_config_in_dir_prefers_ncl() {
+        let temp_dir = TempDir::new().unwrap();
+        let ncl_path = temp_dir.path().join("test-service.ncl");
+        let toml_path = temp_dir.path().join("test-service.toml");
+        let json_path = temp_dir.path().join("test-service.json");
+
+        fs::write(&ncl_path, "{}").unwrap();
+        fs::write(&toml_path, "[test]").unwrap();
+        fs::write(&json_path, "{}").unwrap();
+
+        let result = find_config_in_dir(temp_dir.path(), "test-service").unwrap();
+        assert_eq!(result, ncl_path);
+    }
+
+    #[test]
+    fn test_find_config_in_dir_with_mode_json_fallback() {
+        let temp_dir = TempDir::new().unwrap();
+        let json_path = temp_dir.path().join("test-service.solo.json");
+        fs::write(&json_path, "{}").unwrap();
+
+        let result = find_config_in_dir_with_mode(temp_dir.path(), "test-service", "solo").unwrap();
+        assert_eq!(result, json_path);
+    }
+
+    #[test]
+    fn test_no_config_found_returns_none() {
+        let temp_dir = TempDir::new().unwrap();
+
+        let resolver = ConfigResolver::new()
+            .with_cli_config_dir(Some(temp_dir.path().to_path_buf()));
+
+        let resolved = resolver.resolve("nonexistent-service");
+        assert!(resolved.is_none());
+    }
+
+    #[test]
+    fn test_cli_config_dir_overrides_env_var() {
+        let temp_dir = TempDir::new().unwrap();
+        let config_path = temp_dir.path().join("orchestrator.toml");
+        fs::write(&config_path, "[orchestrator]").unwrap();
+
+        // This test just verifies the resolver logic works correctly
+        // (actual env var override would need temp_env crate)
+        let resolver = ConfigResolver::new()
+            .with_cli_config_dir(Some(temp_dir.path().to_path_buf()));
+
+        let resolved = resolver.resolve("orchestrator").unwrap();
+        assert_eq!(resolved, config_path);
+    }
+}

crates/rag/Cargo.toml

@@ -2,9 +2,13 @@
 authors.workspace = true
 description = "RAG system for provisioning platform with Rig framework and SurrealDB"
 edition.workspace = true
-name = "provisioning-rag"
+name = "rag"
 version.workspace = true
 
+[[bin]]
+name = "provisioning-rag"
+path = "src/main.rs"
+
 [dependencies]
 # ============================================================================
 # WORKSPACE DEPENDENCIES - Core async runtime and traits
@@ -41,7 +45,7 @@ reqwest = { workspace = true }
 # REST API Framework (Phase 8)
 # ============================================================================
 axum = { workspace = true }
-http = "1"
+http = { workspace = true }
 hyper = { workspace = true, features = ["full"] }
 tower = { workspace = true }
 tower-http = { workspace = true, features = ["cors", "trace"] }
@@ -61,7 +65,11 @@ walkdir = { workspace = true }
 config = { workspace = true }
 
 # Platform configuration management
-platform-config = { path = "../platform-config" }
+platform-config = { workspace = true }
+
+# Stratum ecosystem - embeddings and LLM abstraction
+stratum-embeddings = { workspace = true, features = ["openai-provider", "ollama-provider", "fastembed-provider"] }
+stratum-llm = { workspace = true, features = ["anthropic", "openai", "ollama"] }
 
 # Regex for document parsing
 regex = { workspace = true }
@@ -97,13 +105,7 @@ name = "phase8_benchmarks"
 name = "provisioning_rag"
 path = "src/lib.rs"
 
-# Binary target (optional CLI tool)
-[[bin]]
-name = "provisioning-rag"
-path = "src/main.rs"
-required-features = ["cli"]
-
 # Features
 [features]
 cli = []
-default = []
+default = ["cli"]

crates/rag/docker/Dockerfile

@@ -1,59 +0,0 @@
-# Multi-stage build for Provisioning RAG Service
-# Stage 1: Builder
-FROM rust:1.80.1 as builder
-
-WORKDIR /app
-
-# Install dependencies
-RUN apt-get update && apt-get install -y \
-    pkg-config \
-    libssl-dev \
-    && rm -rf /var/lib/apt/lists/*
-
-# Copy workspace and source
-COPY Cargo.toml Cargo.lock ./
-COPY provisioning/platform/rag ./rag
-COPY provisioning/platform/rag/src ./src
-COPY provisioning/platform/rag/benches ./benches
-
-# Build the orchestrator binary in release mode
-RUN cd rag && cargo build --release \
-    && cp target/release/provisioning-rag /app/provisioning-rag
-
-# Stage 2: Runtime
-FROM debian:bookworm-slim
-
-WORKDIR /app
-
-# Install runtime dependencies
-RUN apt-get update && apt-get install -y \
-    ca-certificates \
-    curl \
-    openssl \
-    && rm -rf /var/lib/apt/lists/*
-
-# Copy binary from builder
-COPY --from=builder /app/provisioning-rag /app/
-
-# Create non-root user for security
-RUN useradd -m -u 1000 provisioning && \
-    chown -R provisioning:provisioning /app
-
-USER provisioning
-
-# Environment variables
-ENV PROVISIONING_LOG_LEVEL=info
-ENV PROVISIONING_API_HOST=0.0.0.0
-ENV PROVISIONING_API_PORT=9090
-ENV PROVISIONING_CACHE_SIZE=1000
-ENV PROVISIONING_CACHE_TTL_SECS=3600
-
-# Expose API port
-EXPOSE 9090
-
-# Health check
-HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
-    CMD curl -f http://localhost:9090/health || exit 1
-
-# Start the service
-CMD ["/app/provisioning-rag"]

crates/rag/src/embeddings.rs

@@ -1,197 +1,157 @@
-//! Embeddings module using Rig framework
+//! Embeddings module using stratum-embeddings
 
+use std::collections::HashMap;
 use std::sync::Arc;
+use std::time::Duration;
 
 use serde::{Deserialize, Serialize};
-use tokio::sync::Mutex;
+use stratum_embeddings::{
+    EmbeddingOptions, EmbeddingService, FastEmbedProvider, MemoryCache, OllamaModel,
+    OllamaProvider, OpenAiModel, OpenAiProvider,
+};
 
 use crate::config::EmbeddingConfig;
 use crate::error::Result;
 
-/// Document chunk to be embedded
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct DocumentChunk {
-    /// Unique chunk ID
     pub id: String,
-
-    /// Source document path
     pub source_path: String,
-
-    /// Document type (markdown, kcl, nushell, rust)
     pub doc_type: String,
-
-    /// Chunk content (text to embed)
     pub content: String,
-
-    /// Document category for filtering
     pub category: Option<String>,
-
+    pub metadata: HashMap<String, String>,
-    /// Metadata (headings, function names, etc.)
-    pub metadata: std::collections::HashMap<String, String>,
 }
 
-/// Embedded document with vector
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct EmbeddedDocument {
-    /// Chunk ID
     pub id: String,
-
-    /// Source document path
     pub source_path: String,
-
-    /// Document type
     pub doc_type: String,
-
-    /// Original content
     pub content: String,
-
-    /// Embedding vector
     pub embedding: Vec<f32>,
-
+    pub metadata: HashMap<String, String>,
-    /// Metadata
-    pub metadata: std::collections::HashMap<String, String>,
 }
 
-/// OpenAI HTTP client wrapper for embeddings
+enum Service {
-#[derive(Clone)]
+    OpenAi(EmbeddingService<OpenAiProvider, MemoryCache>),
-struct OpenAiClient {
+    Ollama(EmbeddingService<OllamaProvider, MemoryCache>),
-    api_key: String,
+    FastEmbed(EmbeddingService<FastEmbedProvider, MemoryCache>),
-    model: String,
-    dimension: usize,
 }
 
-impl OpenAiClient {
-    /// Create new OpenAI client
-    fn new(api_key: String, model: String, dimension: usize) -> Self {
-        Self {
-            api_key,
-            model,
-            dimension,
-        }
-    }
-
-    /// Generate embedding via OpenAI API
-    async fn embed(&self, text: &str) -> Result<Vec<f32>> {
-        let client = reqwest::Client::new();
-
-        #[derive(serde::Serialize)]
-        struct EmbeddingRequest {
-            input: String,
-            model: String,
-            dimensions: Option<usize>,
-        }
-
-        #[derive(serde::Deserialize)]
-        struct EmbeddingResponse {
-            data: Vec<EmbeddingData>,
-        }
-
-        #[derive(serde::Deserialize)]
-        struct EmbeddingData {
-            embedding: Vec<f32>,
-        }
-
-        let request = EmbeddingRequest {
-            input: text.to_string(),
-            model: self.model.clone(),
-            dimensions: if self.dimension == 1536 {
-                None
-            } else {
-                Some(self.dimension)
-            },
-        };
-
-        let response = client
-            .post("https://api.openai.com/v1/embeddings")
-            .header("Authorization", format!("Bearer {}", self.api_key))
-            .header("Content-Type", "application/json")
-            .json(&request)
-            .send()
-            .await
-            .map_err(|e| {
-                crate::error::RagError::embedding(format!("Failed to call OpenAI API: {}", e))
-            })?;
-
-        if !response.status().is_success() {
-            let error_text = response
-                .text()
-                .await
-                .unwrap_or_else(|_| "Unknown error".to_string());
-            return Err(crate::error::RagError::embedding(format!(
-                "OpenAI API error: {}",
-                error_text
-            )));
-        }
-
-        let embedding_response: EmbeddingResponse = response.json().await.map_err(|e| {
-            crate::error::RagError::embedding(format!("Failed to parse OpenAI response: {}", e))
-        })?;
-
-        embedding_response
-            .data
-            .first()
-            .map(|d| d.embedding.clone())
-            .ok_or_else(|| crate::error::RagError::embedding("No embedding data in response"))
-    }
-}
-
-/// Embedding engine using OpenAI via HTTP
 pub struct EmbeddingEngine {
     config: EmbeddingConfig,
-    client: Arc<Mutex<Option<OpenAiClient>>>,
+    service: Service,
 }
 
 impl EmbeddingEngine {
-    /// Create a new embedding engine
     pub fn new(config: EmbeddingConfig) -> Result<Self> {
-        // Validate configuration
+        let cache = MemoryCache::new(1000, Duration::from_secs(3600));
-        match config.provider.as_str() {
+
+        let service = match config.provider.as_str() {
             "openai" => {
-                if config.openai_api_key.is_none() && !config.fallback_local {
+                let api_key = config
+                    .openai_api_key
+                    .clone()
+                    .or_else(|| std::env::var("OPENAI_API_KEY").ok());
+
+                let api_key_str = if let Some(key) = api_key {
+                    key
+                } else if config.fallback_local {
+                    "dummy".to_string() // Will fail, but fallback will take over
+                } else {
                     return Err(crate::error::RagError::config(
-                        "OpenAI API key required for OpenAI provider (or enable fallback_local). \
+                        "OpenAI API key required. Set OPENAI_API_KEY or enable fallback_local",
-                         Set OPENAI_API_KEY environment variable.",
                     ));
+                };
+
+                let model = match config.model.as_str() {
+                    "text-embedding-3-small" => OpenAiModel::TextEmbedding3Small,
+                    "text-embedding-3-large" => OpenAiModel::TextEmbedding3Large,
+                    "text-embedding-ada-002" => OpenAiModel::TextEmbeddingAda002,
+                    _ => {
+                        tracing::warn!(
+                            "Unknown model '{}', using text-embedding-3-small",
+                            config.model
+                        );
+                        OpenAiModel::TextEmbedding3Small
+                    }
+                };
+
+                let provider = OpenAiProvider::new(api_key_str, model).map_err(|e| {
+                    crate::error::RagError::embedding(format!("OpenAI provider failed: {}", e))
+                })?;
+
+                let mut svc = EmbeddingService::new(provider).with_cache(cache);
+
+                if config.fallback_local {
+                    tracing::info!("Fallback to local embeddings enabled");
+                    let fallback = Arc::new(FastEmbedProvider::small().map_err(|e| {
+                        crate::error::RagError::embedding(format!("Fallback failed: {}", e))
+                    })?);
+                    svc = svc.with_fallback(fallback);
                 }
+
+                Service::OpenAi(svc)
             }
             "ollama" => {
-                tracing::info!("Using Ollama for local embeddings (no API costs)");
+                use stratum_embeddings::OllamaModel;
+
+                let model = if config.model == "nomic-embed-text" {
+                    OllamaModel::NomicEmbed
+                } else if config.model == "mxbai-embed-large" {
+                    OllamaModel::MxbaiEmbed
+                } else if config.model == "all-minilm" {
+                    OllamaModel::AllMiniLm
+                } else {
+                    OllamaModel::Custom(config.model.clone(), config.dimension)
+                };
+
+                let provider = OllamaProvider::new(model).map_err(|e| {
+                    crate::error::RagError::embedding(format!("Ollama provider failed: {}", e))
+                })?;
+
+                Service::Ollama(EmbeddingService::new(provider).with_cache(cache))
             }
             "local" => {
-                tracing::info!("Using local embedding model (no API costs)");
+                tracing::info!("Using FastEmbed local embeddings (no API costs)");
+                let provider = FastEmbedProvider::small().map_err(|e| {
+                    crate::error::RagError::embedding(format!("FastEmbed provider failed: {}", e))
+                })?;
+
+                Service::FastEmbed(EmbeddingService::new(provider).with_cache(cache))
             }
             _ => {
                 return Err(crate::error::RagError::config(format!(
-                    "Unknown embedding provider: {}. Supported: openai, ollama, local",
+                    "Unknown provider: {}. Supported: openai, ollama, local",
                     config.provider
                 )));
             }
-        }
+        };
 
         tracing::info!(
-            "Initialized embedding engine: {} (provider: {}, dimension: {})",
+            "Initialized stratum-embeddings: {} (provider: {}, dim: {})",
             config.model,
             config.provider,
             config.dimension
         );
 
-        Ok(Self {
+        Ok(Self { config, service })
-            config,
-            client: Arc::new(Mutex::new(None)),
-        })
     }
 
-    /// Get embedding configuration
     pub fn config(&self) -> &EmbeddingConfig {
         &self.config
     }
 
-    /// Embed a single chunk using Rig framework
-    ///
-    /// This is a stub implementation that creates zero vectors.
-    /// In production, this would call the Rig embeddings API.
     pub async fn embed_chunk(&self, chunk: &DocumentChunk) -> Result<EmbeddedDocument> {
-        let embedding = self.generate_embedding(&chunk.content).await?;
+        let options = EmbeddingOptions::default_with_cache();
+
+        let embedding = match &self.service {
+            Service::OpenAi(svc) => svc.embed(&chunk.content, &options).await,
+            Service::Ollama(svc) => svc.embed(&chunk.content, &options).await,
+            Service::FastEmbed(svc) => svc.embed(&chunk.content, &options).await,
+        }
+        .map_err(|e| crate::error::RagError::embedding(format!("Embedding failed: {}", e)))?;
 
         Ok(EmbeddedDocument {
             id: chunk.id.clone(),
@@ -203,133 +163,33 @@ impl EmbeddingEngine {
         })
     }
 
-    /// Embed multiple chunks in batch
     pub async fn embed_batch(&self, chunks: &[DocumentChunk]) -> Result<Vec<EmbeddedDocument>> {
-        let mut results = Vec::new();
+        let options = EmbeddingOptions::default_with_cache();
+        let texts: Vec<String> = chunks.iter().map(|c| c.content.clone()).collect();
 
-        for (idx, chunk) in chunks.iter().enumerate() {
+        let result = match &self.service {
-            let embedded = self.embed_chunk(chunk).await?;
+            Service::OpenAi(svc) => svc.embed_batch(texts, &options).await,
-            results.push(embedded);
+            Service::Ollama(svc) => svc.embed_batch(texts, &options).await,
-
+            Service::FastEmbed(svc) => svc.embed_batch(texts, &options).await,
-            // Log progress
-            if (idx + 1) % 10 == 0 {
-                tracing::debug!("Embedded {}/{} chunks", idx + 1, chunks.len());
-            }
         }
+        .map_err(|e| crate::error::RagError::embedding(format!("Batch embed failed: {}", e)))?;
 
-        tracing::info!("Embedded {} chunks total", chunks.len());
+        let results = chunks
+            .iter()
+            .zip(result.embeddings)
+            .map(|(chunk, embedding)| EmbeddedDocument {
+                id: chunk.id.clone(),
+                source_path: chunk.source_path.clone(),
+                doc_type: chunk.doc_type.clone(),
+                content: chunk.content.clone(),
+                embedding,
+                metadata: chunk.metadata.clone(),
+            })
+            .collect();
 
+        tracing::info!("Embedded {} chunks (stratum-embeddings)", chunks.len());
         Ok(results)
     }
-
-    /// Generate embedding for text using configured provider
-    async fn generate_embedding(&self, text: &str) -> Result<Vec<f32>> {
-        if text.is_empty() {
-            return Err(crate::error::RagError::embedding(
-                "Empty text cannot be embedded",
-            ));
-        }
-
-        tracing::trace!("Generating embedding for text: {} chars", text.len());
-
-        match self.config.provider.as_str() {
-            "openai" => self.embed_openai(text).await,
-            "ollama" => self.embed_ollama(text).await,
-            "local" => self.embed_local(text).await,
-            _ => Err(crate::error::RagError::embedding(format!(
-                "Unknown embedding provider: {}",
-                self.config.provider
-            ))),
-        }
-    }
-
-    /// Generate embedding via OpenAI API
-    async fn embed_openai(&self, text: &str) -> Result<Vec<f32>> {
-        let api_key = self
-            .config
-            .openai_api_key
-            .clone()
-            .or_else(|| std::env::var("OPENAI_API_KEY").ok())
-            .ok_or_else(|| {
-                crate::error::RagError::embedding(
-                    "OpenAI API key not found. Set OPENAI_API_KEY env var or config.",
-                )
-            })?;
-
-        let mut client_lock = self.client.lock().await;
-        if client_lock.is_none() {
-            *client_lock = Some(OpenAiClient::new(
-                api_key.clone(),
-                self.config.model.clone(),
-                self.config.dimension,
-            ));
-        }
-
-        let client = client_lock.as_ref().unwrap();
-        client.embed(text).await
-    }
-
-    /// Generate embedding via Ollama (local)
-    async fn embed_ollama(&self, text: &str) -> Result<Vec<f32>> {
-        let client = reqwest::Client::new();
-
-        #[derive(serde::Serialize)]
-        struct OllamaRequest {
-            model: String,
-            prompt: String,
-        }
-
-        #[derive(serde::Deserialize)]
-        struct OllamaResponse {
-            embedding: Vec<f32>,
-        }
-
-        let request = OllamaRequest {
-            model: self.config.model.clone(),
-            prompt: text.to_string(),
-        };
-
-        let response = client
-            .post("http://localhost:11434/api/embeddings")
-            .json(&request)
-            .send()
-            .await
-            .map_err(|e| {
-                crate::error::RagError::embedding(format!(
-                    "Failed to call Ollama API (ensure Ollama is running on localhost:11434): {}",
-                    e
-                ))
-            })?;
-
-        if !response.status().is_success() {
-            let error_text = response
-                .text()
-                .await
-                .unwrap_or_else(|_| "Unknown error".to_string());
-            return Err(crate::error::RagError::embedding(format!(
-                "Ollama API error: {}",
-                error_text
-            )));
-        }
-
-        let embedding_response: OllamaResponse = response.json().await.map_err(|e| {
-            crate::error::RagError::embedding(format!("Failed to parse Ollama response: {}", e))
-        })?;
-
-        Ok(embedding_response.embedding)
-    }
-
-    /// Generate embedding using local model (stub for future implementation)
-    async fn embed_local(&self, text: &str) -> Result<Vec<f32>> {
-        tracing::warn!(
-            "Local embeddings not fully implemented. Returning zero vector for now. For \
-             production local embeddings, use Ollama or integrate huggingface transformers."
-        );
-
-        // Return zero vector of correct dimension
-        // Future: integrate sentence-transformers or ONNX models
-        Ok(vec![0.0; self.config.dimension])
-    }
 }
 
 #[cfg(test)]
@@ -337,57 +197,54 @@ mod tests {
     use super::*;
 
     #[test]
-    fn test_embedding_engine_creation_openai() {
+    fn test_engine_openai() {
         let config = EmbeddingConfig {
             provider: "openai".to_string(),
             openai_api_key: Some("test-key".to_string()),
             ..Default::default()
         };
-
+        assert!(EmbeddingEngine::new(config).is_ok());
-        let result = EmbeddingEngine::new(config);
-        assert!(result.is_ok());
     }
 
     #[test]
-    fn test_embedding_engine_creation_openai_no_key() {
+    fn test_engine_ollama() {
         let config = EmbeddingConfig {
-            provider: "openai".to_string(),
+            provider: "ollama".to_string(),
-            openai_api_key: None,
+            model: "nomic-embed-text".to_string(),
-            fallback_local: false,
             ..Default::default()
         };
-
+        assert!(EmbeddingEngine::new(config).is_ok());
-        let result = EmbeddingEngine::new(config);
-        assert!(result.is_err());
     }
 
     #[test]
-    fn test_local_embedding_engine() {
+    fn test_engine_local() {
         let config = EmbeddingConfig {
             provider: "local".to_string(),
+            dimension: 384,
             ..Default::default()
         };
-
+        assert!(EmbeddingEngine::new(config).is_ok());
-        let engine = EmbeddingEngine::new(config);
-        assert!(engine.is_ok());
     }
 
     #[tokio::test]
     async fn test_embed_chunk() {
         let config = EmbeddingConfig {
             provider: "local".to_string(),
+            dimension: 384,
             ..Default::default()
         };
-
         let engine = EmbeddingEngine::new(config).unwrap();
 
+        let mut metadata = HashMap::new();
+        metadata.insert("section".to_string(), "test".to_string());
+
         let chunk = DocumentChunk {
             id: "test-1".to_string(),
-            source_path: "test.md".to_string(),
+            source_path: "/test/doc.md".to_string(),
             doc_type: "markdown".to_string(),
-            content: "This is test content".to_string(),
+            content: "Test document".to_string(),
-            category: None,
+            category: Some("test".to_string()),
-            metadata: std::collections::HashMap::new(),
+            metadata,
         };
 
         let result = engine.embed_chunk(&chunk).await;
@@ -395,6 +252,6 @@ mod tests {
 
         let embedded = result.unwrap();
         assert_eq!(embedded.id, "test-1");
-        assert_eq!(embedded.embedding.len(), 1536); // Default dimension
+        assert_eq!(embedded.embedding.len(), 384);
     }
 }

crates/rag/src/llm.rs

@@ -1,64 +1,52 @@
-//! LLM (Large Language Model) integration module
+//! LLM integration using stratum-llm
-//! Provides Claude API integration for RAG-based answer generation
 
-use serde::{Deserialize, Serialize};
+use stratum_llm::{
+    AnthropicProvider, ConfiguredProvider, CredentialSource, GenerationOptions, Message,
+    ProviderChain, Role, UnifiedClient,
+};
 use tracing::info;
 
 use crate::error::Result;
 
-/// Claude API request message
-#[derive(Debug, Clone, Serialize)]
-pub struct ClaudeMessage {
-    pub role: String,
-    pub content: String,
-}
-
-/// Claude API response
-#[derive(Debug, Clone, Deserialize)]
-pub struct ClaudeResponse {
-    pub content: Vec<ClaudeContent>,
-    pub stop_reason: String,
-}
-
-/// Claude response content
-#[derive(Debug, Clone, Deserialize)]
-pub struct ClaudeContent {
-    #[serde(rename = "type")]
-    pub content_type: String,
-    pub text: Option<String>,
-}
-
-/// LLM Client for Claude API
 pub struct LlmClient {
-    api_key: String,
+    client: UnifiedClient,
     pub model: String,
-    base_url: String,
 }
 
 impl LlmClient {
-    /// Create a new Claude LLM client
     pub fn new(model: String) -> Result<Self> {
-        // Get API key from environment
+        let api_key = std::env::var("ANTHROPIC_API_KEY").ok();
-        let api_key = std::env::var("ANTHROPIC_API_KEY").unwrap_or_else(|_| {
-            tracing::warn!("ANTHROPIC_API_KEY not set - Claude API calls will fail");
-            String::new()
-        });
 
-        Ok(Self {
+        if api_key.is_none() {
-            api_key,
+            tracing::warn!("ANTHROPIC_API_KEY not set - LLM calls will fail");
-            model,
-            base_url: "https://api.anthropic.com/v1".to_string(),
-        })
-    }
-
-    /// Generate an answer using Claude
-    pub async fn generate_answer(&self, query: &str, context: &str) -> Result<String> {
-        // If no API key, return placeholder
-        if self.api_key.is_empty() {
-            return Ok(self.generate_placeholder(query, context));
         }
 
-        // Build the system prompt
+        let provider =
+            AnthropicProvider::new(api_key.unwrap_or_default(), model.clone());
+
+        let configured = ConfiguredProvider {
+            provider: Box::new(provider),
+            credential_source: CredentialSource::EnvVar {
+                name: "ANTHROPIC_API_KEY".to_string(),
+            },
+            priority: 0,
+        };
+
+        let chain = ProviderChain::with_providers(vec![configured]);
+
+        let client = UnifiedClient::builder()
+            .with_chain(chain)
+            .build()
+            .map_err(|e| {
+                crate::error::RagError::LlmError(format!("Failed to build LLM client: {}", e))
+            })?;
+
+        info!("Initialized stratum-llm client: {}", model);
+
+        Ok(Self { client, model })
+    }
+
+    pub async fn generate_answer(&self, query: &str, context: &str) -> Result<String> {
         let system_prompt = format!(
             r#"You are a helpful assistant answering questions about a provisioning platform.
 You have been provided with relevant documentation context below.
@@ -71,118 +59,32 @@ Be concise and accurate.
             context
         );
 
-        // Build the user message
+        let messages = vec![
-        let user_message = query.to_string();
+            Message {
+                role: Role::System,
+                content: system_prompt,
+            },
+            Message {
+                role: Role::User,
+                content: query.to_string(),
+            },
+        ];
 
-        // Call Claude API
+        let options = GenerationOptions {
-        self.call_claude_api(&system_prompt, &user_message).await
+            max_tokens: Some(1024),
-    }
+            ..Default::default()
+        };
 
-    /// Call Claude API with messages
+        let response = self
-    async fn call_claude_api(&self, system: &str, user_message: &str) -> Result<String> {
+            .client
-        let client = reqwest::Client::new();
+            .generate(&messages, Some(&options))
-
-        // Build request payload
-        let payload = serde_json::json!({
-            "model": self.model,
-            "max_tokens": 1024,
-            "system": system,
-            "messages": [
-                {
-                    "role": "user",
-                    "content": user_message
-                }
-            ]
-        });
-
-        // Make the API request
-        let response = client
-            .post(format!("{}/messages", self.base_url))
-            .header("anthropic-version", "2023-06-01")
-            .header("x-api-key", &self.api_key)
-            .json(&payload)
-            .send()
             .await
             .map_err(|e| {
-                crate::error::RagError::LlmError(format!("Claude API request failed: {}", e))
+                crate::error::RagError::LlmError(format!("LLM generation failed: {}", e))
             })?;
 
-        // Check status
+        info!("Generated answer: {} characters", response.content.len());
-        if !response.status().is_success() {
+        Ok(response.content)
-            let status = response.status();
-            let error_text = response
-                .text()
-                .await
-                .unwrap_or_else(|_| "Unknown error".to_string());
-            return Err(crate::error::RagError::LlmError(format!(
-                "Claude API error {}: {}",
-                status, error_text
-            )));
-        }
-
-        // Parse response
-        let claude_response: ClaudeResponse = response.json().await.map_err(|e| {
-            crate::error::RagError::LlmError(format!("Failed to parse Claude response: {}", e))
-        })?;
-
-        // Extract text from response
-        let answer = claude_response
-            .content
-            .first()
-            .and_then(|c| c.text.clone())
-            .ok_or_else(|| {
-                crate::error::RagError::LlmError("No text in Claude response".to_string())
-            })?;
-
-        info!(
-            "Claude API call successful, generated {} characters",
-            answer.len()
-        );
-        Ok(answer)
-    }
-
-    /// Generate placeholder answer when API key is missing
-    fn generate_placeholder(&self, query: &str, context: &str) -> String {
-        format!(
-            "Based on the provided context about the provisioning platform:\n\n{}\n\n(Note: This \
-             is a placeholder response. Set ANTHROPIC_API_KEY environment variable for full \
-             Claude integration.)",
-            self.format_context_summary(query, context)
-        )
-    }
-
-    /// Format a summary of the context
-    fn format_context_summary(&self, query: &str, context: &str) -> String {
-        let context_lines = context.lines().count();
-        let query_lower = query.to_lowercase();
-
-        if query_lower.contains("deploy") || query_lower.contains("create") {
-            format!(
-                "Your question about deployment is addressed in {} lines of documentation. The \
-                 system supports multi-cloud deployment across AWS, UpCloud, and local \
-                 environments.",
-                context_lines
-            )
-        } else if query_lower.contains("architecture") || query_lower.contains("design") {
-            format!(
-                "The provisioning platform uses a modular architecture as described in {} lines \
-                 of documentation. Core components include Orchestrator, Control Center, and MCP \
-                 Server integration.",
-                context_lines
-            )
-        } else if query_lower.contains("security") || query_lower.contains("auth") {
-            format!(
-                "Security features are documented in {} lines. The system implements JWT-based \
-                 authentication, Cedar-based authorization, and dynamic secrets management.",
-                context_lines
-            )
-        } else {
-            format!(
-                "Your question is addressed in the provided {} lines of documentation. Please \
-                 review the context above for details.",
-                context_lines
-            )
-        }
     }
 }
 
@@ -192,38 +94,14 @@ mod tests {
 
     #[test]
     fn test_llm_client_creation() {
-        let client = LlmClient::new("claude-opus-4-1".to_string());
+        let client = LlmClient::new("claude-opus-4".to_string());
         assert!(client.is_ok());
     }
 
     #[test]
-    fn test_placeholder_generation() {
+    fn test_llm_client_model() {
-        let client = LlmClient {
+        let client = LlmClient::new("claude-sonnet-4".to_string());
-            api_key: String::new(),
+        assert!(client.is_ok());
-            model: "claude-opus-4-1".to_string(),
+        assert_eq!(client.unwrap().model, "claude-sonnet-4");
-            base_url: "https://api.anthropic.com/v1".to_string(),
-        };
-
-        let query = "How do I deploy the platform?";
-        let context = "Deployment is done using provisioning commands";
-
-        let answer = client.generate_placeholder(query, context);
-        assert!(answer.contains("deployment"));
-        assert!(answer.contains("placeholder"));
-    }
-
-    #[test]
-    fn test_context_summary_formatting() {
-        let client = LlmClient {
-            api_key: String::new(),
-            model: "claude-opus-4-1".to_string(),
-            base_url: "https://api.anthropic.com/v1".to_string(),
-        };
-
-        let deployment_query = "How do I deploy?";
-        let context = "Line 1\nLine 2\nLine 3";
-        let summary = client.format_context_summary(deployment_query, context);
-        assert!(summary.contains("deployment"));
-        assert!(summary.contains("3"));
     }
 }

crates/rag/src/main.rs

@@ -1,12 +1,34 @@
 //! RAG system command-line tool
 
+use clap::Parser;
 use provisioning_rag::config::RagConfig;
+use std::path::PathBuf;
+
+#[derive(Parser, Debug)]
+#[command(name = "rag")]
+#[command(about = "Retrieval-Augmented Generation system")]
+struct Args {
+    /// Configuration file path (highest priority)
+    #[arg(short = 'c', long, env = "RAG_CONFIG")]
+    config: Option<PathBuf>,
+
+    /// Configuration directory (searches for rag.ncl|toml|json)
+    #[arg(long, env = "PROVISIONING_CONFIG_DIR")]
+    config_dir: Option<PathBuf>,
+
+    /// Deployment mode (solo, multiuser, cicd, enterprise)
+    #[arg(short = 'm', long, env = "RAG_MODE")]
+    mode: Option<String>,
+}
 
 #[tokio::main]
 async fn main() -> anyhow::Result<()> {
     // Initialize logging
     tracing_subscriber::fmt::init();
 
+    // Parse CLI arguments
+    let _args = Args::parse();
+
     // Load configuration
     let _config = RagConfig::default();
 

crates/service-clients/Cargo.toml

@@ -1,11 +1,11 @@
 [package]
-authors = { workspace = true }
+authors.workspace = true
 description = "HTTP service client wrappers for provisioning platform services"
-edition = { workspace = true }
+edition.workspace = true
-license = { workspace = true }
+license.workspace = true
 name = "service-clients"
-repository = { workspace = true }
+repository.workspace = true
-version = { workspace = true }
+version.workspace = true
 
 [dependencies]
 async-trait = { workspace = true }
@@ -15,9 +15,7 @@ serde = { workspace = true, features = ["derive"] }
 serde_json = { workspace = true }
 thiserror = { workspace = true }
 tokio = { workspace = true, features = ["full"] }
-
+machines = { workspace = true }
-# Service types (optional - only if not using generic types)
-machines = { path = "../../../../submodules/prov-ecosystem/crates/machines" }
 
 [dev-dependencies]
 tempfile = { workspace = true }

crates/vault-service/Cargo.toml

@@ -1,9 +1,15 @@
 [package]
-authors = ["Provisioning Team"]
+authors.workspace = true
 description = "Vault Service for Provisioning Platform with secrets and key management (Age dev, Cosmian KMS prod, RustyVault self-hosted)"
-edition = "2021"
+edition.workspace = true
+license.workspace = true
 name = "vault-service"
-version = "0.2.0"
+repository.workspace = true
+version.workspace = true
+
+[[bin]]
+name = "provisioning-vault-service"
+path = "src/main.rs"
 
 [dependencies]
 # Async runtime
@@ -23,10 +29,10 @@ toml = { workspace = true }
 reqwest = { workspace = true }
 
 # Age encryption (development)
-age = "0.11"
+age = { workspace = true }
 
 # RustyVault (self-hosted Vault alternative)
-rusty_vault = "0.2.1"
+rusty_vault = { workspace = true }
 
 # Cryptography
 base64 = { workspace = true }
@@ -46,18 +52,15 @@ chrono = { workspace = true, features = ["serde"] }
 # Configuration
 config = { workspace = true }
 
-# SecretumVault (Enterprise secrets management)
+# SecretumVault (Enterprise secrets management - optional)
 secretumvault = { workspace = true }
 
 [dev-dependencies]
+http-body-util = { workspace = true }
 mockito = { workspace = true }
 tempfile = { workspace = true }
 tokio-test = { workspace = true }
 
-[[bin]]
-name = "vault-service"
-path = "src/main.rs"
-
 [lib]
 name = "vault_service"
 path = "src/lib.rs"

secretumvault

@@ -0,0 +1 @@
+Subproject commit 91eefc86fa03826997401facee620f3b6dfd65e1

stratumiops

@@ -0,0 +1 @@
+Subproject commit 9864f88c14ac030f0fa914fd46c2cf4c1a412fc0

syntaxis

@@ -0,0 +1 @@
+Subproject commit 48d7503b4817e38dd86f494e09022b77c0652159