deepchem

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

• Third-party content exposure detected (high risk: 0.70). The skill's workflows and scripts explicitly load public MoleculeNet datasets via dc.molnet.load_* and download/use pretrained community models through HuggingFaceModel (e.g., seyonec/ChemBERTa in scripts/transfer_learning.py), which fetches untrusted third‑party content at runtime that is then read and used to train/fine‑tune models and therefore can materially influence agent behavior.