Code Review Skill
Use this skill when the user asks for a code review or when you are self-reviewing critical code.
1. Review Checklist
Functionality & Correctness
- • Logic Errors: Are there any off-by-one errors, infinite loops, or incorrect conditional logic?
- • Requirements: Does the code meet the apparent or stated requirements?
- • Edge Cases: Are empty inputs, null values, or large datasets handled gracefully?
- • Error Handling: Are exceptions caught and logged properly? Are error messages clear?
Code Quality & Style
- • Readability: Is the code easy to understand? Are variable/function names descriptive?
- • Modularity: Are functions too long or doing too much? Should code be refactored into helpers?
- • DRY (Don't Repeat Yourself): Is there duplicated code that can be consolidated?
- • Type Hints: (Python) Are type hints used and accurate?
Performance
- • Efficiency: Are there any evident O(n^2) or worse algorithms that could be O(n)?
- • Resources: Are file handles, database connections, or network resources managed properly?
Security
- • Input Validation: Is outside input sanitized?
- • Credentials: Are there hardcoded secrets? (Flag immediately!)
2. Feedback Format
Provide your feedback in the following Markdown format:
markdown
## Code Review: [Filename] ### Summary [Brief high-level summary of the code and its quality] ### Critical Issues 🔴 - **[Line X]**: [Description of a bug or major logic flaw] ### Suggestions 🟡 - **[Line Y]**: [Suggestion for improvement, refactoring, or better naming] ### Nitpicks 🟢 - **[Line Z]**: [Minor style or formatting comment] ### Positive Notes - [What was done well?]
3. Action Plan
If there are critical issues, offer to fix them immediately.