AGENT SKILL
python-code-review
暂无简介
- License 许可
- MIT
- 最近更新
- 2026-06-24
- Stars
- ★ 66
- Forks
- 24
03 SKILL.md 正文
Python Code Review
Workflow
- Verify
AGENTS.md,docs/REPO_STYLE.md,docs/PYTHON_STYLE.md, anddocs/CHANGELOG.mdexist. - Read those files and summarize repo rules in four one-sentence lines with prefixes:
AGENTS:,REPO_STYLE:,PYTHON_STYLE:,CHANGELOG:. - Inspect changed files first (
git diff,git status --short), then inspect related call sites. - If the repo has
docs/active_plans/, identify the active plan document(s) that govern the change. Otherwise, skip plan-conformance steps. - If an active plan exists, map code and tests to plan requirements, acceptance criteria, and stated constraints. Otherwise, skip this step.
- Prioritize findings by severity: plan mismatch/regressions (if applicable), correctness/safety, then maintainability.
- Provide concrete, minimal fixes with before/after examples when a fix is straightforward.
- Flag uncertainty explicitly and ask targeted review questions for unclear logic or contracts.
Review Output Contract
- Report findings first, ordered by severity.
- For each finding include:
- Severity (
P1critical,P2high,P3medium,P4low) - File path and line reference
- Risk and likely impact
- Recommended change
- Severity (
- After findings, include:
- Open questions
- Test gaps and residual risk
- Brief summary
What To Check
- Plan conformance: implementation and tests match active plan scope, ordering, and acceptance criteria.
- Plan drift: behavior changed without corresponding plan/changelog updates, or plan claims complete while code is partial.
- Correctness: edge cases, off-by-one logic, stale assumptions, API misuse, compatibility breaks.
- Security: unsafe eval/exec, command injection, path traversal, deserialization hazards, weak validation.
- Maintainability: dead code, hidden coupling, unclear naming, duplicated logic, brittle tests.
- Performance only when materially relevant.
Fix Guidance
- Prefer small, local edits that preserve behavior unless a bug requires behavior change.
- Keep fixes aligned with repo Python style and test conventions.
- Add or adjust tests for each behavior-changing fix.