grill-me
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions direct the agent to read project files and source code to inform its questioning process ('investigate instead of asking'). This introduces a surface where instructions embedded in project data (e.g., in code comments or markdown files) could potentially influence the agent's behavior during the session.
- Ingestion points: Local project files, source code, and directory structure.
- Boundary markers: The instructions do not specify the use of delimiters or specific safety markers to differentiate ingested project data from system instructions.
- Capability inventory: The agent has permissions to read local files and write markdown session logs to the '.grill/' directory.
- Sanitization: No sanitization or content validation is defined for the content read from the project environment.
Audit Metadata