Code Review Skills
AI Agent Skills for Code Review, PR Cleanup, and Verification
A good code-review skill stack does more than comment on style. It helps the agent understand the review surface, ground its advice in docs or security practice, address feedback in the active PR, and verify behavior in a real app when the change reaches the UI.
That means the best review setup is usually cross-functional: one skill for review operations, one for security posture, one for testing or browser verification, and one for the domain-specific architecture under discussion.
Codex
Browse Codex review-oriented skills
Codex currently has the stronger terminal-first review and PR-handling catalog, especially for GitHub- and browser-driven workflows.
Claude
Browse Claude Code architecture and testing skills
Claude Code is useful when the review problem is tied to MCP design, LLM app patterns, or higher-level implementation guidance.
Hub
Return to the main skills hub
Jump back to the hub if the real problem is choosing the right ecosystem before you choose the review skill itself.
What review skills should cover
The most useful review skills handle a specific kind of feedback loop: PR comments, security review, threat modeling, browser verification, or architecture checks grounded in the right docs.
How to apply them
Use review skills as a structured second pass after local reasoning. They are most valuable when the scope is explicit: the active PR, the changed files, the test surface, the security concern, or the interface contract being reviewed.
Curated Listings
High-signal skills for review workflows
These pages are useful when the job is review, verification, or response to feedback rather than greenfield implementation.
PR Workflow
PR Comment Handler
A direct fit for addressing review feedback on the open GitHub pull request without losing the thread between comments and fixes.
Security Review
Security Best Practices
Useful for framework-aware security reviews when you need more than a generic “be careful with auth” pass.
Threat Modeling
Threat Model Source Code Repo
Best when the review question is architectural risk, trust boundaries, abuse paths, or attacker leverage, not just line-level correctness.
Browser Verification
Playwright CLI Skill
A strong review companion whenever UI or workflow correctness must be verified in a real browser instead of guessed from the code alone.
Testing
Web Application Testing
Useful when the review issue sits in frontend behavior, QA evidence, or webapp verification rather than repository mechanics.
Architecture
MCP Server Development Guide
Helps when the review surface is an MCP server, tool contract, or integration boundary rather than a normal app-only change.
Related Hubs
Keep drilling down by ecosystem or use case
These related hubs connect the broad search term to the next page that usually narrows the decision cleanly.
Playwright Skills
AI Agent Skills for Playwright, Browser Automation, and UI Verification
Browse AI agent skills for Playwright, browser automation, UI verification, screenshots, and iterative frontend debugging.
Best Codex Skills
Best Codex Skills for Builders Who Want Fast Leverage
Browse the best Codex skills for CLI creation, Playwright, PR handling, official docs, deployment, and skill authoring.
Best Claude Code Skills
Best Claude Code Skills for Real Workflows
Browse the best Claude Code skills to start with for testing, MCP servers, frontend work, documentation, and skill authoring.