MCP server for AI governance document retrieval, evaluation, and enforcement with semantic search across governance principles, methods, and references
Scoring was not performed
Build LLM-based KG from indexed content (opt-in)
Evaluate a planned action against governance principles BEFORE execution. Call this before any action that is not a read-only operation, non-sensitive question, or trivial formatting change. Returns guidance and routing assessment with PROCEED, REVIEW, or ESCALATE; retrieval alone does not establish compliance. Missing or keyword-only retrieval is disclosed and routes to REVIEW unless safety requires ESCALATE. Retrieval unavailability alone does not withdraw existing user authorization. S-Series (safety) principles have veto authority - will force ESCALATE if triggered.
Structural code references (imports, calls, inheritance)
Get detailed information about a specific domain including all principles and methods. Use for domain exploration.
Get retrieval performance metrics including query counts, average latency, confidence distribution, and feedback stats.
Get the full content of a specific governance principle, method, or reference by ID. Use after query_governance or search_references to get complete text. IDs follow pattern: meta-core-informational-readiness, coding-quality-testing, ref-ai-coding-playwright-auth, etc.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 12 | <=2025-11-25 | v2 |
| 2026-03-09 | F | 20 | - | v1 |
Trigger manual re-index of current project
List all available governance domains with statistics. Shows principle counts, descriptions, and domain priorities.
Show all indexed projects
Log user feedback on retrieval quality. Use this to help improve future retrieval by recording which principles were helpful or not.
Index stats for current project
Retrieve relevant AI governance principles for a query using hybrid search. Auto-detects domain from query semantics. Returns scored principles from Constitution (always) and detected domains with confidence levels. Use this as the primary tool for getting governance guidance.
Query the knowledge graph for entities/relationships
Default search tool — semantic + keyword search across project content. Query to discover what already exists before implementing.