What problem does it solve? Applications that accept user input, manage sessions, or integrate external services are exposed to injection, XSS, broken access control, SSRF, and supply-chain attacks. This Skill provides a threat-model-first process and concrete prevention patterns so security is built into every feature rather than bolted on later. ## Core Features & Use Cases - Threat Modeling with STRIDE: Map trust boundaries, name assets, and write abuse cases before writing code, covering HTTP inputs, webhooks, file uploads, and LLM output. - OWASP Top 10 Prevention Patterns: Copy-ready TypeScript examples for parameterized queries, bcrypt password hashing, secure session cookies, CSP headers, authorization checks, and SSRF URL allowlisting with DNS resolution checks. - Dependency & Supply-Chain Triage: A decision tree for package-manager audit findings based on severity and reachability, plus lockfile integrity, install-script blocking, and typosquat detection. - Privacy & LLM Security: Data classification, retention, and deletion-path guidance for GDPR/CCPA, plus rules for treating model output as untrusted input and scoping agent tool permissions. - Use Case: Before shipping a login flow, run the security review checklist to verify password hashing, rate limiting with a shared store, httpOnly cookies, and generic error responses. ## Quick Start Audit my Express login endpoint for OWASP Top 10 vulnerabilities and tell me what to fix before release.