kernelgen-flagos

Pass

Audited by Gen Agent Trust Hub on Apr 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCREDENTIALS_UNSAFEDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to perform hardware diagnostics (e.g., nvidia-smi, npu-smi), verify Python environments, manage dependencies via pip, and execute generated test and benchmark suites using pytest.
  • [EXTERNAL_DOWNLOADS]: It establishes a connection to the vendor's remote MCP service at https://kernelgen.flagos.io/sse/ to perform kernel generation, optimization, and platform specialization tasks.
  • [CREDENTIALS_UNSAFE]: The configuration process guides users to obtain a 'KernelGen Token' and stores it in a local configuration file (.mcp.json) to authenticate the remote service.
  • [REMOTE_CODE_EXECUTION]: The skill facilitates the creation of code from a remote AI service and subsequently executes that code on the local system for accuracy verification and performance benchmarking.
  • [DATA_EXFILTRATION]: As part of the kernel generation workflow, the skill transmits technical specifications, function descriptions, and relevant code fragments to the vendor's remote endpoint for processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 12, 2026, 07:35 AM