review-pr-changes

Review PR changes with structured checklist for quality and standards compliance. Use for comprehensive PR code review.

⭐6stars🔀1forks📁View on GitHub🕐Updated Dec 11, 2025

When & Why to Use This Skill

This Claude skill automates and structures the Pull Request review process by applying a comprehensive, multi-dimensional checklist covering code quality, testing, documentation, and security. It helps developers maintain high software standards by providing consistent, objective feedback on code changes, identifying potential bugs, and ensuring compliance with project-specific conventions before merging.

Use Cases

  • Pre-approval Code Quality Assessment: Automatically evaluate PRs for readability, structure, and adherence to DRY principles before a human reviewer steps in.
  • Standards Compliance Verification: Ensure all code changes follow specific language syntax (e.g., Mojo v0.26.1+) and project-wide naming conventions.
  • Automated Security & Safety Audits: Scan changes for hardcoded secrets, memory safety issues, and missing input validation to prevent security vulnerabilities.
  • Test Coverage & Edge Case Checking: Verify that new functionality is supported by adequate tests and that edge cases are properly handled in the test suite.
  • Documentation & API Consistency: Check that public APIs have required docstrings and that README files or examples are updated to reflect architectural changes.
namereview-pr-changes
description"Review PR changes with structured checklist for quality and standards compliance. Use for comprehensive PR code review."
categoryreview
agenttest-engineer
user-invocablefalse

Review PR Changes with Checklist

Perform structured review of PR changes against project quality standards.

When to Use

  • Code review before approving PR
  • Detailed evaluation of code quality
  • Checking standards compliance (Mojo, Python, documentation)
  • Verifying test coverage
  • Assessing architectural impact

Quick Reference

# Get PR files changed
gh pr diff <pr> --name-only

# View specific file diff
gh pr diff <pr> -- path/to/file.mojo

# Get PR review status
gh pr view <pr> --json reviews

# Check file statistics
gh pr diff <pr> | diffstat

# Get PR body/description
gh pr view <pr> --json body

Review Checklist

Code Quality:

  • Code is readable and well-structured
  • Functions/classes have clear purposes
  • Variable names are descriptive
  • Complex logic is commented
  • No code duplication (DRY principle)
  • Follows project naming conventions

Testing:

  • Tests present for new functionality
  • Tests are passing (CI shows green)
  • Edge cases covered in tests
  • Test names describe what they test
  • No skipped or xfail tests
  • Adequate coverage for changes

Documentation:

  • Docstrings for public APIs
  • README updated if needed
  • Comments for non-obvious code
  • Examples provided for complex features
  • Type hints present (Mojo/Python)

Standards Compliance:

  • Mojo code uses v0.26.1+ syntax
  • No deprecated patterns (inout, @value, DynamicVector)
  • Zero compiler warnings
  • Proper indentation and formatting
  • No trailing whitespace
  • Files end with newline

Mojo-Specific:

  • Constructors use out self not mut self
  • Non-copyable returns use ^ transfer operator
  • Proper trait conformances on structs
  • Memory safety validated
  • SIMD used for performance-critical code
  • Ownership patterns correct

Security & Safety:

  • No hardcoded secrets/tokens
  • Input validation present
  • No unsafe operations
  • Proper error handling
  • No memory safety issues
  • No type safety violations

Git & Commit:

  • PR linked to issue (in description)
  • Commit messages follow conventional commits
  • No unintended files included
  • Branch is up to date with main
  • No merge conflicts

Review Workflow

  1. Check context: View PR description and linked issue
  2. Scan changes: Review file list and statistics
  3. Read code: Examine actual changes carefully
  4. Run checklist: Go through each category
  5. Test locally: Pull and test changes if needed
  6. Create comments: Flag issues as code comments
  7. Provide verdict: Approve, request changes, or comment

Output Format

Report review results with sections:

  1. Summary - Overall assessment of changes
  2. Strengths - Well-executed aspects
  3. Issues Found - Problems that must be fixed
  4. Suggestions - Optional improvements
  5. Questions - Clarifications needed
  6. Verdict - Approve/Request Changes/Comment
  7. Next Steps - What needs to happen next

Common Issues to Flag

Code Issues:

  • Logic errors or off-by-one mistakes
  • Missing error handling
  • Performance problems
  • Unnecessary complexity

Style Issues:

  • Inconsistent formatting
  • Poor naming choices
  • Missing comments
  • Overly long functions/files

Test Issues:

  • Missing test coverage
  • Flaky tests
  • Inadequate assertions
  • Wrong expected values

Documentation Issues:

  • Missing docstrings
  • Inaccurate documentation
  • Examples that don't work
  • Missing type annotations

Error Handling

Problem Solution
Can't access PR Check gh auth status
Can't understand code Ask clarifying question in comment
Needs local testing Use worktree-create skill to test
Multiple issues Prioritize critical first, optional second
Disagreement on style Refer to CLAUDE.md for standards

Review Standards

Approve when:

  • All critical issues fixed
  • Code follows project standards
  • Tests passing and coverage good
  • Documentation complete
  • No security/safety concerns

Request changes when:

  • Critical issues present
  • Standards not followed
  • Missing tests
  • Significant problems found

Comment when:

  • Only minor suggestions
  • Questions about approach
  • Suggestions for improvement

References

  • See CLAUDE.md for project standards
  • See verify-pr-ready for merge readiness check
  • See gh-batch-merge-by-labels for batch review workflow