mlx-apple-silicon-mlx

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses curl piped to python3 -m json.tool for formatting JSON output from a local service running on localhost:11435. While the security scanner flagged these as remote code execution, they are benign calls to a built-in Python module (json.tool) used for pretty-printing data from the user's own local machine.
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install legitimate Python packages such as ollama-herd, mflux, and diffusionkit using pip and uv. These are well-known or vendor-specific tools related to the skill's stated purpose of managing MLX-based models.
  • [DATA_EXFILTRATION]: All network operations are directed at localhost (127.0.0.1), which is safe. No sensitive data is transmitted to external servers.
  • [SAFE]: The skill's architecture emphasizes local-first operation, explicitly stating that all requests stay within the local network and model management requires explicit user confirmation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 07:05 PM
Security Audit — agent-trust-hub — mlx-apple-silicon-mlx