tooluniverse-infectious-disease
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external, untrusted scientific literature sources, creating a potential surface for indirect prompt injection.\n- Ingestion points: The skill uses tools like PubMed_search_articles, BioRxiv_search_preprints, MedRxiv_search_preprints, and ArXiv_search_papers in Phase 5 (SKILL.md) to retrieve article content, titles, and summaries into the agent's context.\n- Boundary markers: The instructions do not define explicit delimiters or 'ignore' instructions to isolate these external text inputs from the rest of the agent's prompt.\n- Capability inventory: The skill possesses the capability to write files ([PATHOGEN]_outbreak_intelligence.md) and execute specific bioinformatics analysis tools. It does not have access to high-privilege system commands or sensitive local files.\n- Sanitization: No evidence of sanitization, escaping, or validation of the fetched literature content was found in the workflow instructions.\n- [SAFE]: The skill fetches data and performs computations using well-known scientific and technology services, including the NCBI Taxonomy database, UniProt, ChEMBL, and NVIDIA NIM (for AlphaFold2 and DiffDock). These references are consistent with the skill's stated purpose of infectious disease research.
Audit Metadata