bioservices

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

• Third-party content exposure detected (medium risk: 0.65). At runtime the scripts call third-party web APIs (e.g., KEGG get()/parse_kgml_pathway(), UniProt retrieve()/search(), PSICQUIC query()), and the returned raw text/TSV/PSI-MI/entry fields are then parsed and inserted into the program’s in-memory strings (and could be forwarded into an agent LLM context by the surrounding workflow), making this an outsider-sourced free-text ingestion path.