autonomous-code-review

npx skills add majiayu000/claude-skill-registry --skill autonomous-code-review --agent claude-code

Same command for any agent — swap --agent for codex, cursor, copilot.

# Autonomous Code Review Skill Run comprehensive code quality checks, tests, and fixes without human intervention. ## When to Use Use this skill: - For autonomous code quality maintenance - When instructed to "review and fix code" - As a regular maintenance task - After completing a feature or fix ## Philosophy: LLM-Human Collaboration This skill enables Claude to work autonomously on code quality: 1. **Run tests first** - Understand current state 2. **Identify issues** - Audit for bad practices 3. **Fix systematically** - Apply fixes one at a time 4. **Verify each fix** - Run tests after each change 5. **Report clearly** - Enable human oversight --- ## Autonomous Workflow ### Phase 1: Assess Current State ```bash echo "=== Phase 1: Current State Assessment ===" # 1.1 Check git status echo "## Git Status" git status --short # 1.2 Run existing tests echo "## Running Tests" uv run pytest MouseMaster/Testing/Python/ -v --tb=short 2>&1 | tee /tmp/test_results.txt TEST_EXIT=$? if [ $TEST_EXIT -eq 0 ]; then echo "✓ All tests passing" else echo "✗ Some tests failing - review before proceeding" grep -E "FAILED|ERROR" /tmp/test_results.txt fi # 1.3 Run linting echo "## Running Linter" uv ru

echo "=== Phase 1: Current State Assessment ==="
echo "## Git Status"
git status --short
echo "## Running Tests"
uv run pytest MouseMaster/Testing/Python/ -v --tb=short 2>&1 | tee /tmp/test_results.txt
if [ $TEST_EXIT -eq 0 ]; then
echo "✓ All tests passing"
else
echo "✗ Some tests failing - review before proceeding"
grep -E "FAILED|ERROR" /tmp/test_results.txt