[Sebastiangmz]

Summary

• add deterministic citation verification against the current retrieved evidence set
• make Response.grounded depend on all parsed citations verifying
• expose citation verification through REST, MCP, CLI, and UI status text
• add M3 EvalFly and CI validation gates

Verification

• bun run /Users/sebastian/.omp/omp-pantheon/skills/specsafe/bin/specsafe.ts status -> no slice open, 6 history entries
• PYTHONPATH=src .venv/bin/python -m pytest tests/test_citations.py tests/test_m3_citations.py tests/test_api_routes.py tests/test_mcp_handlers.py tests/test_ui_handlers.py -> 45 passed
• .venv/bin/python -m ruff check ... -> OK
• PYTHONPATH=src .venv/bin/mypy ... -> OK
• .venv/bin/python scripts/evalfly_smoke.py --suite smoke -> evalfly committed smoke passed: 15 cases
• .venv/bin/python -m py_compile ... -> exit 0
• .venv/bin/python -m build -> built sdist and wheel
• PYTHONPATH=src .venv/bin/python -m pytest tests/test_m0_baseline.py tests/test_m1_services.py tests/test_m2_registry.py tests/test_citations.py tests/test_m3_citations.py tests/test_api_routes.py tests/test_mcp_handlers.py tests/test_mcp_tools.py tests/test_ui_handlers.py -> 84 passed

Review

• M3ReReview verdict: APPROVE
• prior P1 fixed: MCP citations are JSON-safe and CLI JSON consumes serialized payloads

Spec-Slice: CRAG-20260628-M3-VERIFIED-CITATIONS

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.