Secure Code Guardian helps the agent write security controls into an application: password hashing, parameterized queries, input validation, JWT handling, security headers and rate limiting. It starts from a threat model and ends with explicit validation checkpoints for authentication, authorization, input handling and headers.
Code examples in TypeScript show bcrypt, pg parameter binding, Zod schemas, JWT verification with an algorithm allowlist and a full login endpoint with helmet and express-rate-limit.
What You Get
- A five-step workflow from threat model to documented security decisions
- Validation checkpoints for authentication, authorization, input handling and headers
- TypeScript examples for bcrypt hashing, parameterized pg queries, Zod validation and JWT verification
- A full login endpoint example with helmet, rate limiting, generic errors and an httpOnly cookie
- A must-do and must-not-do list covering hashing, secrets, logging and weak algorithms
- Reference files for OWASP prevention, authentication, input validation, XSS and CSRF, and security headers
Common Use Cases
- Adding login with rate limiting and short-lived tokens to an Express API
- Replacing string-built SQL with parameterized queries
- Adding CSP, HSTS and CORS allowlists to a web service
- Checking that error messages do not reveal whether an account exists
What Makes It Unique
It focuses on writing secure code during implementation. Its sibling Security Reviewer audits finished code and produces reports, and this skill says so in its description by pointing standalone audits elsewhere. The validation checkpoints give the agent specific payloads and behaviors to test, which many security skills leave out.
Under The Hood
The skill is a SKILL.md with five reference files that the agent loads by topic. It ships no scripts. The examples are TypeScript for Node and Express that you adapt into your own code. The validation checkpoints mention running curl -I against a running service and using a header scanner such as Mozilla Observatory, so the agent may send requests to your own application during testing. It writes code and configuration into your project and reads environment variables for secrets.