paper-audit
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: Comprehensive review of the skill body and accompanying Python scripts found no evidence of malicious behavior.
- [COMMAND_EXECUTION]: The orchestrator script (scripts/audit.py) uses subprocess.run to invoke local verification tools. These operations are implemented safely using argument arrays to prevent command injection.
- [INDIRECT_PROMPT_INJECTION]: The skill instructions include a robust safety boundary, explicitly telling the agent to ignore instructions embedded in the academic papers it processes.
- [EXTERNAL_DOWNLOADS]: The skill references established, well-known Python libraries (pymupdf, Jinja2) and reputable academic APIs (arXiv, Semantic Scholar) for its core functionality.
Audit Metadata