light-literature-search
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill includes Python scripts that perform network operations exclusively to connect with well-known and reputable academic services such as OpenAlex, Crossref, NCBI (PubMed), Semantic Scholar, and bioRxiv. These domains are established technological services within the scientific research community.
- [SAFE]: Analysis of the provided Python scripts (search_normalize.py, snowball.py, verify_citations.py, etc.) confirms they use standard libraries and transparent logic. There is no evidence of obfuscation, hidden commands, or privilege escalation attempts.
- [DATA_INGESTION]: The skill processes external content (publication titles, abstracts, and metadata) from academic databases. Although this represents a surface for Indirect Prompt Injection, the sources are restricted to recognized scientific repositories, and the scripts perform basic sanitization (e.g., removing HTML tags from abstracts), making the risk negligible for the intended use case.
Audit Metadata