codex-codereview
Use for code review and quality feedback from Codex. Triggers on "codex review code", "codex code review", "have codex review this", "get codex feedback on code".
When & Why to Use This Skill
The Codex Code Review skill leverages advanced AI to provide comprehensive, automated feedback on your codebase. It identifies logic errors, security vulnerabilities, and performance bottlenecks while ensuring adherence to best practices, significantly streamlining the peer review process and improving software reliability.
Use Cases
- Pre-commit Audit: Automatically analyze staged changes to catch bugs and logic errors before they are committed to the repository.
- Security & Vulnerability Scanning: Identify potential security flaws and sensitive data exposures within new implementations to maintain a secure codebase.
- Best Practices & Readability: Evaluate code against industry standards and team conventions to ensure long-term maintainability and clean code.
- Performance Optimization: Detect inefficient code patterns and receive actionable recommendations to enhance execution speed and resource management.
- Architecture Review: Get high-level feedback on implementation patterns and structural integrity during the development phase.
| name | codex-codereview |
|---|---|
| description | Use for code review and quality feedback from Codex. Triggers on "codex review code", "codex code review", "have codex review this", "get codex feedback on code". |
Codex Code Review Skill
Comprehensive code review with Codex (gpt-5.2).
When to Use
- Reviewing code before commit
- Getting feedback on implementation
- Checking for security issues
- Verifying best practices
- Architecture review
Reasoning Level
high (thorough but efficient review)
Execution
- Identify code to review:
- Specific files mentioned by user
- Staged changes (
git diff --staged) - Recent changes (
git diff)
- Gather context about the codebase patterns
- Formulate a review prompt:
Review this code thoroughly. Code: <code to review> Context: <any relevant context about patterns/conventions> Please review for: 1. Bugs and logic errors 2. Security vulnerabilities 3. Performance issues 4. Code quality and readability 5. Best practices adherence 6. Edge cases - Run:
codex exec -c model_reasoning_effort="high" "<prompt>" - Return structured review feedback
Response Format
**Codex Code Review:**
**Summary:**
[Overall assessment]
**Issues Found:**
- 🔴 Critical: [if any]
- 🟡 Warning: [if any]
- 🔵 Suggestion: [if any]
**Detailed Feedback:**
[Line-by-line or section-by-section feedback]
**Recommendations:**
[Specific improvements]
**Session ID:** [id]