grill-me
Pass
Audited by Gen Agent Trust Hub on Aug 23, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to answer questions by exploring the local codebase, which is an untrusted data source.\n
- Ingestion points: The agent processes codebase content and user-provided plans (SKILL.md).\n
- Boundary markers: None present; the instructions do not specify delimiters to distinguish between retrieved codebase content and agent instructions.\n
- Capability inventory: The skill encourages active use of file-reading and codebase exploration tools.\n
- Sanitization: None present; there is no instruction to filter, escape, or ignore content retrieved from the codebase.
Audit Metadata