Code Review

The code-review skill enables AI agents to analyze, critique, and improve software codebases. It automatically identifies bugs, security vulnerabilities, and architectural flaws while suggesting specific refactoring strategies to enhance maintainability and performance.

When to use it

• Before merging pull requests to ensure high-quality standards are met across the team.
• During legacy system modernization to uncover hidden technical debt and outdated patterns.
• When onboarding new developers to provide instant, context-aware feedback on their initial commits.
• Prior to production deployment to verify that all critical security checks have passed.

Key capabilities

• Automated static analysis of code for syntax errors and logical inconsistencies.
• Detection of common security vulnerabilities such as SQL injection or hardcoded credentials.
• Generation of actionable suggestions for optimizing algorithm efficiency and readability.
• Enforcement of coding style guides and architectural best practices within the project context.

Example prompts

• "Review this Python module for potential race conditions and suggest fixes."
• "Analyze the current database schema design and propose improvements for scalability."
• "Identify any security vulnerabilities in this REST API implementation before we deploy."

Tips & gotchas

Ensure the AI agent has read access to the relevant repository branches or files before initiating a review. While the skill excels at finding common issues, complex logic errors may still require human verification for critical systems.