| name | code-review |
| display_name | Code Review |
| description | Reviews code changes for quality, correctness, style, and potential issues. Provides actionable feedback with specific file and line references. |
| category | code-quality |
| version | 1.0.0 |
| tags | ["review","quality","style","correctness","best-practices"] |
| providers | ["claude-code","codex","copilot","openclaw","hermes-agent"] |
| author | agent-skills contributors |
Code Review
Review code changes for quality, correctness, style, and potential issues. Provide clear, actionable feedback that helps the author improve their code.
Inputs
- Target: A diff, pull request, set of changed files, or specific code selection
- Context (optional): Project coding standards, language-specific conventions, or reference documents to apply during review
Steps
-
Identify the scope — Determine which files and changes are under review. If reviewing a PR or diff, read all changed files. If reviewing a selection, focus on the highlighted code and its immediate context.
-
Understand intent — Before critiquing, understand what the code is trying to accomplish. Read commit messages, PR descriptions, or surrounding code for context.
-
Review for correctness — Check for:
- Logic errors and off-by-one mistakes
- Unhandled edge cases and error conditions
- Race conditions or concurrency issues
- Incorrect assumptions about inputs or state
-
Review for security — Check for:
- Injection vulnerabilities (SQL, XSS, command injection)
- Improper input validation at system boundaries
- Hardcoded secrets or credentials
- Insecure defaults or configurations
-
Review for quality — Check for:
- Code clarity and readability
- Unnecessary complexity or over-engineering
- Duplicated logic that should be consolidated
- Naming that accurately describes purpose
- Consistent style with the surrounding codebase
-
Review for performance — Check for:
- Unnecessary allocations or computations in hot paths
- N+1 query patterns or unbounded data fetching
- Missing indexes or inefficient data structures
- Resource leaks (connections, file handles, memory)
-
Review for maintainability — Check for:
- Breaking changes to public APIs
- Missing or outdated tests for changed behavior
- Tight coupling that will make future changes harder
- Dead code or unused imports introduced by the change
Output
Provide a structured review with the following sections:
Summary
A 1-3 sentence overview of the changes and overall assessment (approve, request changes, or comment).
Issues
List each issue found, categorized by severity:
- Critical — Must fix before merge. Bugs, security vulnerabilities, data loss risks.
- Important — Should fix before merge. Significant quality or maintainability concerns.
- Suggestion — Optional improvement. Style preferences, minor optimizations, alternative approaches.
For each issue, include:
- The specific file and line reference
- A clear description of the problem
- A concrete suggestion for how to fix it
Positive Feedback
Call out things done well — good patterns, clean abstractions, thorough error handling. Positive reinforcement is part of a constructive review.
Guidelines
- Be specific. "This could be better" is not actionable. "This loop allocates a new array on each iteration — consider pre-allocating" is.
- Be respectful. Review the code, not the person.
- Prioritize. Focus on what matters most. Don't bury critical issues under a pile of nitpicks.
- Respect existing patterns. If the codebase uses a specific convention, follow it — even if you prefer a different one.
- Distinguish opinion from fact. Prefix subjective suggestions with "Consider" or "Optionally" to distinguish them from objective issues.
