grill-me
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (documents, code, and tickets) to identify knowledge gaps and formulate questions. If these sources contain hidden instructions (e.g., within code comments or task descriptions), the agent could be manipulated into performing unintended actions during the interview or while writing to the target document.
- Ingestion points: Reads relevant documentation, project tickets, source code, and data from a 'second brain' vault (SKILL.md).
- Boundary markers: The instructions do not specify the use of delimiters or 'ignore instructions' markers when reading external content.
- Capability inventory: The agent has the capability to read files from the workspace and write extracted information into a target document path.
- Sanitization: There is no mention of filtering or sanitizing the content retrieved from external files before it is used to generate prompts or written to the output file.
Audit Metadata