Vapora/docs/setup/secretumvault-integration.md
Jesús Pérez d14150da75 feat: Phase 5.3 - Multi-Agent Learning Infrastructure
Implement intelligent agent learning from Knowledge Graph execution history
with per-task-type expertise tracking, recency bias, and learning curves.

## Phase 5.3 Implementation

### Learning Infrastructure ( Complete)
- LearningProfileService with per-task-type expertise metrics
- TaskTypeExpertise model tracking success_rate, confidence, learning curves
- Recency bias weighting: recent 7 days weighted 3x higher (exponential decay)
- Confidence scoring prevents overfitting: min(1.0, executions / 20)
- Learning curves computed from daily execution windows

### Agent Scoring Service ( Complete)
- Unified AgentScore combining SwarmCoordinator + learning profiles
- Scoring formula: 0.3*base + 0.5*expertise + 0.2*confidence
- Rank agents by combined score for intelligent assignment
- Support for recency-biased scoring (recent_success_rate)
- Methods: rank_agents, select_best, rank_agents_with_recency

### KG Integration ( Complete)
- KGPersistence::get_executions_for_task_type() - query by agent + task type
- KGPersistence::get_agent_executions() - all executions for agent
- Coordinator::load_learning_profile_from_kg() - core KG→Learning integration
- Coordinator::load_all_learning_profiles() - batch load for multiple agents
- Convert PersistedExecution → ExecutionData for learning calculations

### Agent Assignment Integration ( Complete)
- AgentCoordinator uses learning profiles for task assignment
- extract_task_type() infers task type from title/description
- assign_task() scores candidates using AgentScoringService
- Fallback to load-based selection if no learning data available
- Learning profiles stored in coordinator.learning_profiles RwLock

### Profile Adapter Enhancements ( Complete)
- create_learning_profile() - initialize empty profiles
- add_task_type_expertise() - set task-type expertise
- update_profile_with_learning() - update swarm profiles from learning

## Files Modified

### vapora-knowledge-graph/src/persistence.rs (+30 lines)
- get_executions_for_task_type(agent_id, task_type, limit)
- get_agent_executions(agent_id, limit)

### vapora-agents/src/coordinator.rs (+100 lines)
- load_learning_profile_from_kg() - core KG integration method
- load_all_learning_profiles() - batch loading for agents
- assign_task() already uses learning-based scoring via AgentScoringService

### Existing Complete Implementation
- vapora-knowledge-graph/src/learning.rs - calculation functions
- vapora-agents/src/learning_profile.rs - data structures and expertise
- vapora-agents/src/scoring.rs - unified scoring service
- vapora-agents/src/profile_adapter.rs - adapter methods

## Tests Passing
- learning_profile: 7 tests 
- scoring: 5 tests 
- profile_adapter: 6 tests 
- coordinator: learning-specific tests 

## Data Flow
1. Task arrives → AgentCoordinator::assign_task()
2. Extract task_type from description
3. Query KG for task-type executions (load_learning_profile_from_kg)
4. Calculate expertise with recency bias
5. Score candidates (SwarmCoordinator + learning)
6. Assign to top-scored agent
7. Execution result → KG → Update learning profiles

## Key Design Decisions
 Recency bias: 7-day half-life with 3x weight for recent performance
 Confidence scoring: min(1.0, total_executions / 20) prevents overfitting
 Hierarchical scoring: 30% base load, 50% expertise, 20% confidence
 KG query limit: 100 recent executions per task-type for performance
 Async loading: load_learning_profile_from_kg supports concurrent loads

## Next: Phase 5.4 - Cost Optimization
Ready to implement budget enforcement and cost-aware provider selection.
2026-01-11 13:03:53 +00:00

5.3 KiB

SecretumVault Integration

VAPORA integrates with SecretumVault, a post-quantum ready secrets management system, for secure credential and API key management across all microservices.

Overview

SecretumVault provides:

  • Post-quantum cryptography ready for future-proof security
  • Multi-backend storage (filesystem, SurrealDB, PostgreSQL, etcd)
  • Fine-grained access control with Cedar policy engine
  • Secrets server for centralized credential management
  • CLI tools for operations and development

Integration Points

SecretumVault is integrated into these VAPORA services:

Service Purpose Features
vapora-backend REST API credentials, database secrets, JWT keys Central secrets management
vapora-agents Agent authentication, service credentials Secure agent-to-service auth
vapora-llm-router LLM provider API keys (Claude, OpenAI, Gemini, Ollama) Cost tracking + credential rotation

Architecture

┌─────────────────────────────────────────────────────────────┐
│                    VAPORA Services                          │
├─────────────┬──────────────────┬────────────────────────────┤
│ Backend API │ Agent Orchestration │ LLM Router              │
└──────┬──────┴────────┬─────────┴──────────┬─────────────────┘
       │               │                    │
       └───────────────┼────────────────────┘
                       │
                       ▼
         ┌─────────────────────────────┐
         │   SecretumVault Server      │
         ├─────────────────────────────┤
         │ • Credential storage        │
         │ • Policy enforcement        │
         │ • Audit logging             │
         │ • Key rotation              │
         └──────────┬──────────────────┘
                    │
        ┌───────────┴────────────┐
        ▼                        ▼
    Storage Layer          Policy Engine
    (SurrealDB)            (Cedar)

Configuration

Environment Variables

# SecretumVault server connection
SECRETUMVAULT_URL=http://secretumvault:3030
SECRETUMVAULT_TOKEN=<identity-token>

# Storage backend
SECRETUMVAULT_STORAGE=surrealdb
SURREAL_URL=ws://surrealdb:8000
SURREAL_DB=secretumvault

# Crypto backend
SECRETUMVAULT_CRYPTO=openssl  # or aws-lc for post-quantum

Cargo Features

SecretumVault is integrated with these features enabled:

secretumvault = { workspace = true }
# Automatically uses: "server", "surrealdb-storage"

Usage Examples

In vapora-backend

use secretumvault::SecretClient;

// Initialize client
let client = SecretClient::new(
    &env::var("SECRETUMVAULT_URL")?,
    &env::var("SECRETUMVAULT_TOKEN")?,
).await?;

// Retrieve API key
let api_key = client.get_secret("llm/claude-api-key").await?;

// Store credential securely
client.store_secret(
    "database/postgres-password",
    &password,
    Some("postgres-creds"),
).await?;

In vapora-llm-router

use secretumvault::SecretClient;

// Get LLM provider credentials
let openai_key = client.get_secret("llm/openai-api-key").await?;
let claude_key = client.get_secret("llm/claude-api-key").await?;
let gemini_key = client.get_secret("llm/gemini-api-key").await?;

// Fallback to Ollama (local, no key needed)

Running SecretumVault

Local Development

# Terminal 1: Start SecretumVault server
cd /Users/Akasha/Development/secretumvault
cargo run --bin secretumvault-server --features server,surrealdb-storage

# Terminal 2: Initialize with default policies
cargo run --bin secretumvault-cli -- init-policies

Production (Kubernetes)

# Will be added to kubernetes/
kubectl apply -f kubernetes/secretumvault/

Security Best Practices

  1. Token Management

    • Use identity-based tokens (not basic auth)
    • Rotate tokens regularly
    • Store token in .env.local (not in git)
  2. Secret Storage

    • Never commit credentials to git
    • Use SecretumVault for all sensitive data
    • Enable audit logging for compliance
  3. Policy Enforcement

    • Define Cedar policies per role/service
    • Restrict access by principle of least privilege
    • Review policies during security audits
  4. Crypto Backend

    • Use aws-lc for post-quantum readiness
    • Plan migration as quantum threats evolve

Integration Status: Active Services: Backend, Agents, LLM Router Features: server, surrealdb-storage, cedar-policies