pr to main

[xueyulinn]

Summary

Review PR targeting the main branch.

Related Issue

N/A

Type of Change

• Bug fix
• New feature
• Refactor
• Documentation
• Test
• CI/CD
• Chore

Changes Made

• Reviewed changes intended for merge into main.

How to Test

• Review CI results.
• Run relevant test suites if needed.

Checklist

• Code builds successfully
• Tests added or updated
• Existing tests pass
• Documentation updated if needed
• No secrets or sensitive data included
• Breaking changes documented

Breaking Changes

None.

Additional Notes

Ready for review against main.

[codehawky]

Summary by CodeHawky

Status

• Summary type: full_pr_summary
• Current head: b636723929d03dce15e626f3f19d2967e296f80a
• Generated at: 2026-05-16T20:51:24.429804+00:00

Summary

• Adds package version metadata in codehawk/__init__.py and updates dependency metadata to include anthropic>=0.102.0.
• Introduces SAT summary evaluation infrastructure under evals/, including text and JSON eval runners, case datasets, result JSONL files, and an LLM judge helper.
• Adds codehawk/agents/sat_summary_agent_json.py, a structured Report/Item output path for SAT findings with normalized severity fields.
• Refactors many module-level helpers/constants/agent instances to internal _-prefixed names across agents, code indexing, GitHub auth/cache, MinIO, PostgreSQL, services, and tests; also removes the old PR_CODE_REVIEW_SYSTEM_PROMPT.

Risk

• Medium — most production changes are symbol renames intended to be internal-only, but the PR touches 34 files and renames helpers across runtime paths; any external imports or missed references could break despite behavior appearing mostly unchanged.

Review Path

1. Review codehawk/agents/sat_summary_agent_json.py and evals/sat_summary_agent/run_sat_summary_json_eval.py together to validate the new structured SAT summary contract.
2. Inspect evals/sat_summary_agent/*.json, evals/results/**, and evals/llm_judge.py for dataset quality, judge dependency usage, and whether generated result files should be committed.
3. Check the broad _-prefix refactors in runtime modules, especially codehawk/agents/*, codehawk/code_index/*, codehawk/services/*, codehawk/minio/minio_client.py, and codehawk/postgresql/*, for missed public call sites.
4. Verify prompt changes in codehawk/agents/prompts.py, particularly removal of PR_CODE_REVIEW_SYSTEM_PROMPT and additions related to SAT JSON behavior.
5. Review dependency/version changes in pyproject.toml, uv.lock, and codehawk/__init__.py.