clarify
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is purely instructional and contains no executable code, scripts, network operations, or external dependencies.
- [DATA_EXPOSURE]: The skill includes specific security guardrails that forbid the agent from asking users to provide secrets, tokens, or PII, recommending the use of references like environment variable names or vault paths instead.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external task descriptions, creating an attack surface for indirect instructions. However, it includes explicit logic to resist instructions that attempt to bypass safety pauses for expensive or irreversible actions.
- Ingestion points: User-provided task descriptions in SKILL.md and prompt context.
- Boundary markers: The skill implements a 'Restate the task' step to align interpretation before proceeding.
- Capability inventory: The skill is restricted to generating text questions; it does not invoke tools, write files, or initiate network requests.
- Sanitization: Instructions explicitly mandate prioritizing safety protocols over user commands that attempt to disable questions or filters.
Audit Metadata