tooluniverse-network-pharmacology

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

• Third-party content exposure detected (high risk: 0.70). SKILL.md shows the agent making many runtime calls to public third‑party data sources (e.g., PubMed_search_articles, search_clinical_trials, OpenTargets_, STRING_get_interaction_partners, CTD_get_, FAERS_*, DrugBank calls) and explicitly ingests and uses those responses to drive network construction, scoring, and decision-making, which clearly exposes the agent to external untrusted third‑party content that could carry indirect prompt-injection content.