grill-me-qa
Pass
Audited by Gen Agent Trust Hub on Jun 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill's 'Explore before asking' protocol instructs the agent to read local files including the codebase, test suites, and CI configurations. This creates an indirect prompt injection surface where malicious instructions embedded in the analyzed code or documentation could override agent behavior.
- [PROMPT_INJECTION]: Mandatory Evidence Chain for Indirect Prompt Injection surface: 1. Ingestion points: SKILL.md (Phase 1: Context Gathering and Core Protocol #3). 2. Boundary markers: Absent. 3. Capability inventory: File-read (ls, cat) and file-write (Markdown and HTML report generation). 4. Sanitization: Absent.
Audit Metadata