Codebase Review

This is a comprehensive codebase health check that runs SAST scans, reviews architecture patterns, and scores your project 0-100 based on severity-weighted findings. It pulls in security audits (including OWASP checks and AI-generated code vulnerabilities), examines your largest files against multiple lenses like silent failures and type design, and outputs a phased improvement plan. The scoring is unforgiving: any critical issue caps you at 40/100, with deductions ranging from -1 for low-priority items to -15 for critical flaws. What's interesting is the feedback loop requirement. Every high-severity finding forces you to update the relevant skill's anti-pattern rules and test assertions, so the tool theoretically gets smarter with each audit.

Install to Claude Code

npx -y skills add hoangnguyen0403/agent-skills-standard --skill codebase-review --agent claude-code