The Agent Skills Directory

[SAFE]: The skill provides comprehensive guidelines and checklists for performing security audits, focusing on identifying common vulnerabilities such as broken access control, injection, and cryptographic failures.\n- [SAFE]: Code examples showing vulnerabilities (e.g., SQL injection, hardcoded secrets) are clearly marked as anti-patterns and are used for instructional purposes to demonstrate remediation.\n- [SAFE]: The utility uses standard system tools like grep and dotnet to perform local analysis of the codebase without suspicious network activity or data exfiltration.\n- [SAFE]: The skill mandates an evidence-based reasoning protocol, requiring the agent to verify all claims with specific file and line references, which increases the reliability of the review process.\n- [SAFE]: The skill includes instructions to identify and handle dynamic content (e.g., looking for [Injected: ...] headers), which acts as a safeguard against accidental processing of untrusted data.

arch-security-review