tooluniverse-gwas-study-explorer

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

• Third-party content exposure detected (high risk: 0.80). This skill clearly fetches and ingests data from public third‑party sources (NHGRI‑EBI GWAS Catalog and Open Targets Genetics — see the "Tools Used" section in SKILL.md and the python_implementation.py calls to tu.run for gwas_search_studies, gwas_get_associations_for_study, gwas_get_associations_for_snp/OpenTargets_*), and that external content is parsed and used to drive study selection, meta‑analysis, and replication decisions, so untrusted third‑party content could materially influence agent behavior.