grill-me
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to ingest data from the codebase and external evidence, which creates a potential surface for indirect prompt injection attacks.
- Ingestion points: The workflow in
SKILL.mddirects the agent to "Investigate any answer discoverable from the codebase or supplied evidence" (Step 1). - Boundary markers: The instructions lack explicit delimiters or specific directives to ignore instructions that might be embedded in the files being read.
- Capability inventory: The skill is restricted to text-based interaction and reading files; it does not have access to tools for network communication, file system modification, or command execution.
- Sanitization: No mechanisms for sanitizing or validating the retrieved codebase content are defined in the instructions.
Audit Metadata