Add governed agentic skill readiness

[Tashima-Tarsh]

Summary

• Add governed agentic skill registry for Claude/model orchestration readiness
• Add /api/v1/agentic/readiness endpoint
• Wire readiness count into /api/v1/health
• Document agentic AI skill readiness, learning boundaries, and Claude bridge constraints
• Add tests proving policy/evidence boundaries, open-source connector mesh, and blocked silent learning

Verification

• npm run verify
• npm --prefix web audit --audit-level=moderate