rw-literature-discovery
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides a professional workflow for academic research, focusing on the verification of metadata (DOI, PMID) and the tracking of evidence.
- [SAFE]: External network operations are limited to accessing well-known and reputable academic services such as the NCBI, Crossref, and OpenAlex for literature metadata.
- [SAFE]: The project includes a Python validation script (scripts/self_check.py) which is used solely for verifying the presence and integrity of local skill files; it does not execute remote code or perform unsafe operations.
- [SAFE]: While the skill possesses an attack surface for indirect prompt injection by processing user-provided materials, it mitigates these risks through mandatory metadata verification and source tracking requirements; no exploitable capabilities such as arbitrary command execution were found.
Audit Metadata