geniml

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references installing standard data science libraries (geniml, torch, numpy, pandas, anndata, scikit-learn, umap-learn) via pip. It also includes functionality to load pre-trained models from Hugging Face Hub (e.g., 'databio/r2v-ChIP-atlas-hg38-v2'). These are routine operations targeting well-known and trusted package registries and model hosting services.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process BED files (genomic interval data) for machine learning training and search. While this involves processing external data, the operations (training embeddings, nearest-neighbor search) are standard ML workflows and do not expose the agent to control-flow manipulation or high-privilege execution risks.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 12:54 AM