tooluniverse-metabolomics

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). This skill explicitly fetches and parses data from open public third‑party sources (HMDB, MetaboLights, Metabolomics Workbench, PubChem) — see SKILL.md "Databases Covered" and the runtime calls in python_implementation.py (e.g., HMDB_search, metabolights_get_study, metabolights_search_studies, PubChem_get_CID_by_compound_name) — and those untrusted study descriptions/results are ingested and used to drive report generation and fallback/tool decisions, which could allow indirect prompt injection via crafted database content.