llamafile

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads the Llamafile binary and pre-packaged model files from Mozilla's official GitHub and Hugging Face repositories. These are trusted sources for local AI tooling.
  • [COMMAND_EXECUTION]: Uses standard system utilities like chmod to set executable permissions on downloaded binaries and subprocess.Popen to start the local server. These operations are transparent and necessary for the skill's function.
  • [DATA_EXFILTRATION]: No unauthorized data transmission detected. Network operations are limited to initial tool downloads and interacting with the local server at 127.0.0.1 or localhost.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 08:41 AM