ywc-security-audit

To perform an OWASP Top 10 security audit, this Skill executes parallel subagents to deeply analyze authentication logic, external-facing API endpoints, and sensitive data handling for injection and broken access control flaws.

Deep security analysis for LLM prompt injection identifies malicious inputs designed to manipulate model behavior, checking your LLM-driven surfaces and external-facing code against specific prompt injection risks during review.

You should run a security review for authentication and API endpoints whenever you modify middleware, add new API endpoints, or perform monthly security reviews to catch critical flaws like broken access control.

Yes, you can use automated parallel subagents to find broken access control and other vulnerabilities in external-facing code, escalating high-stakes findings to advanced model judgment for nuanced validation.

Yes, this security audit works with existing middleware directories by targeting the authentication logic and sensitive data handling specifically located within your project's API source code paths.