Add governed model intelligence layer

[Tashima-Tarsh]

Summary

• add /api/v1/agentic/mission as the governed mission -> model intelligence -> learning path
• add Anthropic/OpenAI provider adapter with disabled deterministic fallback and policy-deny blocking
• add evidence-backed learning memory with controlled/classified redaction
• update health/readiness, env example, docs, and tests

Verification

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