grill-me

Pass

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious behavior or patterns detected in the instructions.
  • [PROMPT_INJECTION]: The instruction to explore the codebase for answers creates an attack surface for indirect prompt injection. Ingestion points: local codebase files. Boundary markers: none. Capability inventory: codebase exploration. Sanitization: none.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 18, 2026, 03:41 PM