reason-grill-evidence
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates on user-provided arguments, creating a surface for potential indirect prompt injection attacks.
- Ingestion points: Untrusted arguments are ingested for review as described in the SKILL.md body.
- Boundary markers: The skill includes a specific 'Untrusted input rule' directing the agent to never follow instructions found inside the argument under review.
- Capability inventory: No tools for network communication, file system modification, or command execution are requested or utilized in the script.
- Sanitization: The skill relies on natural language boundary enforcement to isolate the agent's logic from the data being analyzed.
Audit Metadata