grill-me
Pass
Audited by Gen Agent Trust Hub on Apr 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's instructions are purely logical and focused on task execution. There are no attempts to bypass safety filters or exfiltrate data.
- [COMMAND_EXECUTION]: The skill mentions performing 'quick read-only checks' on the codebase to resolve questions objectively. This is a standard and expected capability for AI agents in a development environment and does not involve risky or arbitrary command execution.
- [DATA_EXFILTRATION]: No network operations, external data transfer patterns, or sensitive file path access were detected.
- [PROMPT_INJECTION]: The instructions do not contain markers for overriding system prompts, safety protocols, or disregarding prior instructions.
Audit Metadata