grill-me
Pass
Audited by Gen Agent Trust Hub on May 7, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The instruction to explore the codebase to answer questions creates a surface where instructions embedded in local files could influence agent behavior.
- Ingestion points: Local codebase files (SKILL.md)
- Boundary markers: Absent
- Capability inventory: File reading (implied by codebase exploration)
- Sanitization: Absent
Audit Metadata