bio-single-cell-hashing-demultiplexing

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a set of instructional guidelines for single-cell bioinformatics analysis. All referenced libraries (Seurat, scanpy, demuxmix, etc.) are standard tools in the scientific community.
  • [INDIRECT_PROMPT_INJECTION]: The skill includes code to read and process external data files (.h5ad and .csv). This constitutes a potential ingestion surface for indirect prompt injection if the input files are maliciously crafted by an external actor, although this is a standard operational requirement for this type of analysis.
  • Ingestion points: sc.read_h5ad('gex.h5ad') and pd.read_csv('hto_counts.csv') in examples/hashsolo_scanpy.py.
  • Boundary markers: Absent.
  • Capability inventory: Local data processing, normalization, and classification. No network operations or unauthorized file-writing capabilities were identified in the provided scripts.
  • Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 09:19 AM
Security Audit — agent-trust-hub — bio-single-cell-hashing-demultiplexing