grilling
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists of natural language instructions used to guide the agent's behavior during a consultation session. No scripts, hardcoded credentials, or high-risk command executions were detected within the provided file.
- [INDIRECT_PROMPT_INJECTION]: The skill identifies a surface where it may ingest data from the environment via sub-agents to recompute a 'design tree'.
- Ingestion points: Fact-finding missions performed by sub-agents on the filesystem or through other tools (mentioned in SKILL.md).
- Boundary markers: Not specified in the instructions.
- Capability inventory: Uses tool-based sub-agents to access environment data (filesystem, tools).
- Sanitization: None described in the prompt-based instructions.
Audit Metadata