light-literature-search
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a legitimate academic literature search workflow using well-known and trusted public APIs including OpenAlex, Crossref, arXiv, and Europe PMC.
- [SAFE]: All scripts are written in pure Python using standard libraries (urllib, sqlite3, etc.), avoiding the use of unverifiable third-party dependencies.
- [SAFE]: The instructions explicitly define defensive measures against indirect prompt injection by treating external data as non-executable and monitoring for instruction-override patterns (e.g., INJECTION-ATTEMPT-DETECTED).
- [SAFE]: The skill includes strong anti-hallucination policies, explicitly forbidding the fabrication of DOIs, citation counts, or metadata, and provides scripts to verify the existence of academic records.
- [SAFE]: No suspicious persistence, data exfiltration, or obfuscation techniques were detected across the analyzed files and scripts.
Audit Metadata