bio-machine-learning-atlas-mapping
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues were detected. The skill focuses exclusively on bioinformatics tasks using well-known, legitimate scientific libraries such as scvi-tools, scanpy, and scikit-learn.
- [REMOTE_CODE_EXECUTION]: The skill provides installation instructions for standard scientific packages (scvi-tools, scanpy, anndata, scikit-learn, celltypist) from official registries. No unverified or suspicious remote scripts are executed.
- [DATA_EXFILTRATION]: No network operations to non-whitelisted domains or sensitive data access patterns were identified. Data processing is limited to local single-cell data files (.h5ad).
- [INDIRECT_PROMPT_INJECTION]: While the skill ingests external data files (.h5ad), the processing is focused on numerical matrix operations and model training rather than natural language instruction processing. No capability for subsequent malicious instruction execution from data was found.
Audit Metadata