grill-me-with-spec
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is designed for requirements gathering and documentation generation. Analysis of the instructions and the provided SPEC template reveals no malicious behaviors, obfuscation, or unauthorized data access.
- [DATA_EXFILTRATION]: The skill gathers facts from the repository and environment to populate the design document. This is a standard feature for developer-centric AI agents and does not involve exfiltrating data to external or untrusted domains.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from user interviews and existing repository files (Ingestion points) and processes it to generate structured markdown documents using predefined sections (Boundary markers). Its capabilities are limited to file read/write and sub-agent dispatch for fact-gathering (Capability inventory). While explicit sanitization of external content is not described, the output is intended for human review and approval before any implementation occurs (Sanitization).
Audit Metadata