grilling
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill instructs the agent to automatically search the local environment, including the filesystem and available tools, to gather facts without user intervention. This creates an ingestion point for untrusted data that could contain indirect prompt injections.
- Ingestion points: System environment and filesystem (SKILL.md).
- Boundary markers: None provided to delimit external data from instructions.
- Capability inventory: Filesystem access and tool execution.
- Sanitization: None specified for data retrieved from the environment.
Audit Metadata