instruction-quality-audit
Pass
Audited by Gen Agent Trust Hub on Nov 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted AI instruction artifacts as its primary function, creating an ingestion surface for potential indirect prompt injection.
- Ingestion points: Untrusted content is ingested from user-provided text, file paths, or package directories as defined in SKILL.md.
- Boundary markers: The skill includes a dedicated 'Trust Boundary' section in SKILL.md that explicitly instructs the agent to treat all artifacts as untrusted and to avoid following any embedded instructions.
- Capability inventory: The skill is restricted to read-only diagnostic tasks; it lacks instructions for network exfiltration, file modification, or code execution.
- Sanitization: Safety is maintained through strict behavioral constraints that prevent the agent from executing or obeying the instructions it is meant to audit.
Audit Metadata