kernelgen-flagos
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill automates the installation of verified machine learning libraries such as PyTorch, Triton, and vLLM from official package registries (PyPI, Conda) to ensure a functional development environment.
- [DYNAMIC_EXECUTION]: The skill generates and executes temporary Python scripts to perform local hardware diagnostics (detecting NVIDIA GPUs or Ascend NPUs) and runs smoke tests to verify kernel compilation before full integration.
- [COMMAND_EXECUTION]: The skill provides instructions for the autonomous execution of iterative optimization loops and correctness testing to improve operator performance, and it integrates with the GitHub CLI (
gh) for authenticated Pull Request submission and feedback reporting. - [SAFE]: Security configuration for the MCP service, including the storage of JWT tokens in project-local settings files, follows standard practices for managing extension-specific credentials.
- [SAFE]: Feedback mechanisms are implemented through legitimate channels (GitHub Issues and email), with automated data collection restricted to non-sensitive technical environment specifications required for troubleshooting.
Audit Metadata