huggingface-local-models

Pass

Audited by Gen Agent Trust Hub on Aug 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill serves as a comprehensive guide for developers to discover, download, and run GGUF models locally. The instructions align with standard industry practices for local model inference.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates downloading software and model weights from trusted sources. It references the official llama.cpp repository on GitHub and utilizes Hugging Face's public Hub and Tree APIs. These interactions are necessary for the skill's primary function and target well-known, reputable services.
  • [COMMAND_EXECUTION]: The provided scripts for building (make), installing (brew, winget), and running (llama-cli, llama-server) models are standard for the llama.cpp project. The commands are transparent and intended for local execution by the user or an agent acting on their behalf in a development environment.
  • [DATA_EXFILTRATION]: No malicious data exfiltration patterns were detected. The skill includes a curl example targeting localhost:8080, which is a standard method for testing a locally running API server. Network requests are directed exclusively to Hugging Face domains for model discovery.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 3, 2026, 12:40 PM
Security Audit — agent-trust-hub — huggingface-local-models