matchms

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

• Third-party content exposure detected (high risk: 0.70). The skill explicitly supports loading spectra from public repositories (e.g., "load_from_usi" / "metabolomics-USI" and GNPS JSON in references/importing_exporting.md and SKILL.md), meaning the agent may fetch and ingest untrusted, user-contributed spectra/metadata from external repositories that are then interpreted and used to drive downstream matching and processing.