alterlab-matchms
Warn
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The documentation and examples in 'SKILL.md' and 'references/importing_exporting.md' promote the use of Python's 'pickle' module and the 'matchms.importing.load_from_pickle' function for saving and loading preprocessed spectral data. Deserializing data from untrusted pickle files is a known security risk that can lead to arbitrary code execution.
- [EXTERNAL_DOWNLOADS]: The skill requires installing 'matchms' and 'rdkit' from PyPI. It also includes functions to fetch spectral data from the GNPS repository and chemical annotations from PubChem, which are standard, well-known scientific services.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests mass spectrometry data from various external file formats (MGF, MSP, mzML, mzXML, JSON), creating a surface for indirect prompt injection if the agent processes data from untrusted sources, though specialized library parsers mitigate this risk.
Audit Metadata