grill-master
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides a structured framework for requirement gathering using a Socratic interrogation method. No malicious instructions, hidden patterns, or dangerous execution vectors were detected in the instructions.
- [NO_CODE]: The skill does not include any scripts, dependencies, or executable files. It functions entirely as a set of behavioral guidelines for the AI agent's conversational logic.
- [INDIRECT_PROMPT_INJECTION]: The skill includes steps to inspect the repository content to inform its questioning process. While this represents a surface for indirect prompt injection if the codebase contains malicious instructions, the skill's design limits its output to clarification and shared understanding rather than execution.
- Ingestion points: User input and repository files (referenced in SKILL.md Workflow step 4).
- Boundary markers: None explicitly defined.
- Capability inventory: Context retrieval and conversational dialogue.
- Sanitization: Relies on standard agent-level safety filters.
Audit Metadata