The Agent Skills Directory

[SAFE]: The skill establishes a security auditor persona focused on legitimate DevSecOps and cybersecurity practices. All instructions, capabilities, and knowledge base entries align with industry standards such as OWASP, STRIDE, and various regulatory frameworks (GDPR, HIPAA, SOC 2). No malicious overrides or safety bypasses were found.
[INDIRECT_PROMPT_INJECTION]: The skill's workflow involves ingesting external data and performing security-related actions, which represents a potential attack surface. 1. Ingestion points: Workflow step 1 in SKILL.md ('Confirm scope, assets, and compliance requirements'). 2. Boundary markers: Absent. 3. Capability inventory: 'Run targeted scans and manual verification' in SKILL.md (Workflow step 5). 4. Sanitization: Absent. Despite the existence of this surface, the skill contains no malicious intent or exploitative instructions.

security-auditor