huggingface-accelerate

Pass

Audited by Gen Agent Trust Hub on Jun 29, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill serves as a technical resource for developers using HuggingFace Accelerate. It contains standard usage patterns, configuration guides, and performance tuning tips for distributed machine learning workloads.
  • [EXTERNAL_DOWNLOADS]: The skill mentions the installation of standard industry tools. References to accelerate, torch, and transformers are directed at official package registries. The instructions to clone Megatron-LM and apex from NVIDIA's GitHub organization target well-known and trusted repositories.
  • [COMMAND_EXECUTION]: Shell commands provided in the documentation (e.g., accelerate launch, pip install, git clone) are standard development operations and are presented as instructions for the user to execute in a controlled environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 29, 2026, 12:54 AM
Security Audit — agent-trust-hub — huggingface-accelerate