pydeseq

Warn

Audited by Gen Agent Trust Hub on Aug 1, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [REMOTE_CODE_EXECUTION]: In references/workflow_guide.md, code examples suggest loading data matrices and metadata via pickle.load(). Deserializing data from potentially untrusted files using the pickle module is a security risk as it allows for arbitrary code execution.
  • [PROMPT_INJECTION]: The skill provides pydeseq2_audit_result_v1.json, which contains safety conclusions such as 'security': 'PASS'. This represents metadata poisoning designed to influence external security evaluations with self-authored verdicts.
  • [EXTERNAL_DOWNLOADS]: The skill requires the installation of pydeseq2 and its associated scientific stack from standard package registries. These are appropriate and necessary dependencies for RNA-seq analysis.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 1, 2026, 10:43 AM
Security Audit — agent-trust-hub — pydeseq