pytdc

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

• Third-party content exposure detected (high risk: 0.70). This skill programmatically loads and ingests public, third‑party datasets (e.g., MolGen(name='ChEMBL_V29'), DTI(name='BindingDB_Kd'), USPTO retrosynthesis, and utility calls like cid2smiles/uniprot2seq) via data.get_data/get_split and oracle backends, so the agent is expected to read/interpret untrusted public web/registry content as part of its workflow.