grilling
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions direct the agent to 'dispatch a sub-agent' to find facts from the environment (filesystem, tools, etc.) rather than asking the user. This creates a vulnerability surface where the agent could ingest and act upon malicious instructions embedded in local data or tool outputs.
- Ingestion points: Local filesystem and outputs from tools used by sub-agents as specified in
SKILL.md. - Boundary markers: None identified; the instructions do not specify the use of delimiters or warnings to ignore instructions within the retrieved facts.
- Capability inventory: The skill allows for recursive questioning ('rounds') and the autonomous use of tools/sub-agents to query the environment.
- Sanitization: No evidence of input validation or sanitization for the retrieved environment data is present in the skill instructions.
Audit Metadata