grill-me
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions specify that the agent should autonomously ingest data from external sources such as the local filesystem, issue trackers, and documentation to find 'facts' during the interview process.\n
- Ingestion points: The
SKILL.mdfile (Facts vs decisions section) identifies the local filesystem, issue trackers, and project documentation as sources for 'facts' to be gathered by sub-agents.\n - Boundary markers: The instructions lack specific boundary markers or 'ignore embedded instructions' warnings for the external data being read.\n
- Capability inventory: The skill facilitates read-only filesystem access and factual lookups through exploration sub-agents based on the contents of the local environment.\n
- Sanitization: No sanitization or filtering logic is defined for the content retrieved from external sources before it is interpreted by the agent.
Audit Metadata