feat: add retrieval eval advisory report

[cafitac]

Summary

• add a structured advisory_report to retrieval eval JSON results
• render the advisory report and recommended actions in text reports
• document that reports now include severity, affected task IDs, and next-action guidance

Contract

advisory_report includes:

• severity: ok, medium, or high
• summary
• current_failure_task_ids
• baseline_weak_spot_task_ids
• current_regression_task_ids
• recommended_actions
• baseline_mode

Test plan

• uv run pytest tests/test_retrieval_evaluation.py -q
• uv run pytest tests/ -q
• uv run python scripts/check_release_metadata.py
• uv run python scripts/smoke_release_readiness.py
• npm pack --dry-run
• git diff --check
• static diff secret scan: 0 findings