grilling
Pass
Audited by Gen Agent Trust Hub on Aug 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines a logical framework for recursive interviewing and design-tree mapping. While it directs the agent to use sub-agents for environmental lookups based on user context (an attack surface for indirect prompt injection), this is consistent with standard agentic behavior and does not include malicious triggers or exfiltration patterns.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user input to determine when to query the environment for facts, which presents a potential surface for indirect instructions.
- Ingestion points: User answers to frontier questions in
SKILL.md. - Boundary markers: Absent; instructions do not specify delimiting user input or ignoring embedded instructions.
- Capability inventory: The skill directs the agent to dispatch sub-agents to access the filesystem and environment tools (
SKILL.md). - Sanitization: Absent; no validation or filtering is specified for the user-provided context before it triggers tool use.
Audit Metadata