grilling
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates in a loop where user-provided answers are used to calculate new questions and determine what information the agent needs to retrieve from the environment.
- Ingestion points: User input provided as answers to the agent's "frontier" questions in SKILL.md.
- Boundary markers: There are no explicit markers or instructions to treat user answers as untrusted data or to ignore embedded instructions within those answers.
- Capability inventory: The skill explicitly directs the agent to dispatch sub-agents to find facts from the environment, specifically mentioning filesystem access and tool usage in SKILL.md.
- Sanitization: The instructions do not specify any sanitization, filtering, or validation of user input before it is used to direct environmental searches.
Audit Metadata