reviewing-user-research-prior-art-survey
Pass
Audited by Gen Agent Trust Hub on Aug 7, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a local Python script (validate_user_research_prior_art.py) to perform deterministic validation of input files. This script is part of the toolset provided with the skill for legitimate evaluation purposes.
- [INDIRECT_PROMPT_INJECTION]: As the skill is designed to process external research artifacts, it possesses an attack surface for indirect prompt injection. This risk is inherent to the reviewer role and is mitigated by strict adherence to the 27 conditions and a mandated structured output format.
- Ingestion points: Research artifacts and vocabulary maps.
- Boundary markers: Not explicitly defined.
- Capability inventory: Execution of local validation script and structured text generation.
- Sanitization: Relying on the provided validation script.
Audit Metadata