pydeseq2
Warn
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: MEDIUMREMOTE_CODE_EXECUTION
Full Analysis
- [Dynamic Execution] (MEDIUM): The documentation explicitly encourages the use of Python's
picklemodule for data persistence via theto_picklable_anndata()method. \n - Evidence: In
references/api_reference.md, a code snippet demonstrates the use ofpickle.dump()to save analysis results. \n - Risk: Although saving data is safe, this documentation implicitly guides users toward the use of
pickle.load()for data retrieval. Thepicklemodule is notoriously insecure and can be exploited to execute arbitrary code if the input file is maliciously crafted. The absence of security warnings for this pattern is a notable risk in an AI agent skill context.
Audit Metadata