single-cell-multi-omics-integration

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses ov.utils.download_tosica_gmt() to fetch gene-set files from the vendor's resources, which is a standard procedure for the TOSICA label transfer workflow.
  • [COMMAND_EXECUTION]: Textual instructions in the troubleshooting section guide the user or agent to install required bioinformatics dependencies such as mofapy2, simba, and simba_pbg using pip to ensure the environment is correctly prepared for analysis.
  • [PROMPT_INJECTION]: The skill handles external data files including .h5ad AnnData and .gmt gene-sets (ingestion points). While explicit boundary markers for text within these files are not defined, the skill outlines capabilities such as model training (tosica.train) and HDF5 file writing (mofa_run). Potential indirect injection risks are addressed through a 'Defensive Validation Patterns' section that provides sanitization and validation of input files and metadata structure.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 05:43 PM
Security Audit — agent-trust-hub — single-cell-multi-omics-integration