torchdrug

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

• Third-party content exposure detected (high risk: 0.80). The skill's instructions (see references/datasets.md and SKILL.md) explicitly load and auto-download public datasets and files — e.g., AlphaFoldDB, Hetionet, ZINC, ChEMBL, USPTO and calls like data.Protein.from_pdb(...) and datasets.BBBP(...) — so the agent will fetch and ingest open/public third-party content that can materially influence model predictions and downstream actions.