diffdock

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow (python3 -m inference using --protein_path/--protein_ligand_csv/--protein_sequence and an optional ligand_description) ingests user-supplied SMILES/sequences and file paths directly into the docking pipeline, which constitutes outsider-authored free text being read by the model.