tao-train-sparse4d

Pass

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill serves as a configuration layer for the NVIDIA TAO Toolkit 'sparse4d' CLI, documented through official schemas and templates.
  • [SAFE]: No hardcoded credentials or sensitive data exposure points were detected; the 'encryption_key' field is provided as a configuration placeholder for model security.
  • [SAFE]: Integration with Weights & Biases (WandB) is used for standard ML experiment tracking and does not constitute unauthorized data exfiltration.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 12, 2026, 11:36 PM
Security Audit — agent-trust-hub — tao-train-sparse4d