What problem does it solve? Frontend codebases often ship with security gaps — tokens in localStorage, missing Content Security Policy headers, unsanitized dangerouslySetInnerHTML, secrets in source — that are hard to spot in code review. This Skill statically audits a TypeScript and React frontend against an opinionated security baseline and produces structured findings plus an optional remediation plan. ## Core Features & Use Cases - Four-layer security baseline: Checks authentication/sessions (PKCE, OAuth state, redirect validation), input handling and XSS prevention, transport/headers/cookies (CSP, HSTS, frame-ancestors), and secrets/data protection/third-party integrations (SRI, iframe sandboxing, postMessage origin validation). - Diagnostic snapshot: Detects the auth library, sanitization library, deployment platform, CSP presence, and counts of risky patterns before grading any check. - Optional scanner enrichment: With --with-scan, enriches findings using installed ESLint security plugins and Semgrep OWASP rule packs; degrades gracefully when scanners are absent. - Use Case: Before a release, run the audit on a Next.js app to discover that session tokens are stored in localStorage and no Content Security Policy is configured, then generate a severity-ordered implementation plan with exact vercel.json header snippets. ## Quick Start Run /security-audit in the repository root to statically audit the frontend and receive a Top 5 recommendations summary with the full report written to .architect-audits/security-audit/findings.md.