linting-and-formatting

Pass

Audited by Gen Agent Trust Hub on May 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No security issues detected. The skill provides legitimate developer workflows for code maintenance and style compliance within the Megatron-LM project environment.
  • [COMMAND_EXECUTION]: The skill instructs the agent to run a local script (tools/autoformat.sh) and use the uv tool for package synchronization and tool execution. These are standard operations for development and CI environments and originate from the project's own infrastructure.
  • [EXTERNAL_DOWNLOADS]: The skill uses uv sync to install linting tools from a defined group. This involves downloading well-known packages (black, ruff, etc.) from standard registries like PyPI, which is expected and safe behavior for setting up a development environment.
Audit Metadata
Risk Level
SAFE
Analyzed
May 3, 2026, 07:06 AM