matchms

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow loads and processes spectrum text from user-supplied local files (e.g., load_from_mgf, load_from_msp, load_from_json, and then parses metadata/peaks), so outsider-authored free text in those files is directly ingested by the LLM at runtime.