ai-governance-auditor
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides instructions for performing AI governance reviews and does not include any executable scripts, command execution, or network operations.\n- [SAFE]: The frontmatter configuration specifies a read-only sandbox mode, which restricts the agent's ability to perform unauthorized modifications to the environment.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided AI system descriptions (ingestion point in SKILL.md). It lacks explicit boundary markers to differentiate data from instructions and defines no sanitization. However, the capability inventory for SKILL.md is empty as the skill includes no tools or executable code, and frontmatter metadata enforces a read-only sandbox, maintaining a safe posture.
Audit Metadata