sci-search
Pass
Audited by Gen Agent Trust Hub on May 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches scientific paper metadata from established and trusted repositories, including arXiv, PubMed, and Clarivate (Web of Science). These network operations are intrinsic to the skill's primary function and target well-known academic services.
- [PROMPT_INJECTION]: The skill processes external data (paper abstracts and titles), which constitutes an indirect prompt injection surface. However, given the nature of the sources and the skill's utility, the risk is negligible.
- Ingestion points:
sci_search.py(retrieves data from arXiv, PubMed, and Web of Science APIs). - Boundary markers: Results are structured into Markdown, though no explicit 'ignore instructions' delimiters are used for the ingested text.
- Capability inventory: The skill has the ability to write search results to local JSON and Markdown files.
- Sanitization: No specialized sanitization is performed on the academic text beyond standard formatting.
Audit Metadata