nv-generate-vae-finetune

Pass

Audited by Gen Agent Trust Hub on Jul 8, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches model weights and configuration files from NVIDIA's official GitHub organization and other well-known services like Hugging Face and PyTorch. These operations are transparently documented as necessary for the machine learning workflow.
  • [COMMAND_EXECUTION]: Subprocess calls are limited to a version-check operation using the git command. The implementation uses a list-based argument structure rather than a shell string, which is a secure practice that prevents command injection.
  • [DATA_EXFILTRATION]: No evidence of unauthorized data transmission was found. Network access is restricted to authoritative model registries for downloading dependencies and weights.
  • [DYNAMIC_EXECUTION]: The script uses the .eval() method provided by the PyTorch library to switch models into evaluation mode. While flagged by static analysis, this is a standard machine learning operation and is distinct from the dangerous Python eval() function.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 8, 2026, 07:15 PM
Security Audit — agent-trust-hub — nv-generate-vae-finetune