grilling
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides structural instructions for an agent to conduct a rigorous interview process. It does not include executable code, network operations, or sensitive data access patterns.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user input through sequential rounds of questioning. Ingestion points: User answers to 'frontier' questions (SKILL.md). Boundary markers: No specific delimiters are defined. Capability inventory: Sub-agents can access the environment (filesystem, tools) to resolve facts (SKILL.md). Sanitization: No explicit sanitization of user data is specified. This finding represents a standard attack surface for interactive skills rather than a specific vulnerability.
Audit Metadata