review-agent

Automated code reviews assess submissions with quality scoring, static analysis via ESLint and TypeScript, and security scanning. Run `npm run agents:review -- --files="src/**/*.ts"` to review files; submissions scoring 80+ pass automatically, while those below trigger human review with detailed feedback and remediation suggestions.

Yes. Security vulnerability scanning detects API keys, passwords, tokens, and known dependency vulnerabilities using npm audit and secret detection. The Skill identifies these issues and provides remediation guidance, escalating critical findings for immediate action.

Code quality scores range 0–100 and combine ESLint violations, TypeScript type errors, and security vulnerabilities. An 80+ score passes automated approval; lower scores generate review comments identifying specific issues and fixes to bring code into compliance.

Static analysis scans for unsafe patterns like `eval()` and `innerHTML` assignments alongside ESLint rules and type safety checks. It provides detailed feedback on each issue, enabling developers to address problems before human review.

Yes. The Skill performs deep static analysis using ESLint and TypeScript checks within your existing JavaScript and TypeScript projects, applying your current configurations to quality scoring and vulnerability detection.

Critical security vulnerabilities trigger escalation guidance separate from standard review comments. The Skill flags findings, suggests remediation steps, and escalates high-severity issues to ensure they receive priority attention before code merges.