The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified. The skill provides a structured methodology for security auditing and includes Markdown-based report templates and checklists.
[INDIRECT_PROMPT_INJECTION]: The skill is intended to review untrusted code and logs, which is an inherent attack surface for indirect prompt injection. However, the skill itself lacks any tools or executable scripts that could be leveraged for an attack. * Ingestion points: Code snippets and audit logs provided by the user. * Boundary markers: None specified. * Capability inventory: No tools or scripts defined. * Sanitization: Not applicable.

security-audit