mlm-bridge-training

Pass

Audited by Gen Agent Trust Hub on May 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides bash commands for launching distributed machine learning training using torch.distributed.run and performing local directory management.
  • [COMMAND_EXECUTION]: Local scripts like switch_mcore.sh are used to toggle between different versions of the Megatron-Core submodule.
  • [EXTERNAL_DOWNLOADS]: Standard environment synchronization is performed using 'uv sync' and 'git submodule update', which fetch project dependencies and external repositories from official sources.
  • [SAFE]: No malicious patterns, such as credential harvesting, remote code execution from untrusted URLs, or obfuscated content, were found in the analyzed files.
Audit Metadata
Risk Level
SAFE
Analyzed
May 18, 2026, 06:14 AM
Security Audit — agent-trust-hub — mlm-bridge-training