cite
Pass
Audited by Gen Agent Trust Hub on Apr 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious behavioral patterns or suspicious code detected in the skill instructions.
- [EXTERNAL_DOWNLOADS]: Installation instructions reference the vendor's repository on GitHub, which is a well-known service.
- [PROMPT_INJECTION]: The skill processes untrusted data from external URLs, establishing a surface for indirect prompt injection.
- Ingestion points: User-provided URLs for papers in SKILL.md.
- Boundary markers: Absent from instructions.
- Capability inventory: Uses agent tools for web fetching.
- Sanitization: No explicit metadata validation is specified.
Audit Metadata