hugging-face-model-trainer

Pass

Audited by Gen Agent Trust Hub on Jun 20, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill utilizes the hf_jobs() MCP tool to execute training scripts in a remote, ephemeral environment, which is the standard and intended operation of the Hugging Face Jobs platform.
  • [SAFE]: External dependencies and remote scripts are exclusively sourced from trusted organizations such as Hugging Face and established open-source projects like llama.cpp.
  • [SAFE]: Secret management follows security best practices by instructing the agent to handle the Hugging Face authentication token (HF_TOKEN) as a job secret, preventing credential exposure in logs or local files.
  • [SAFE]: The inclusion of a dataset validation tool (scripts/dataset_inspector.py) provides a proactive security measure to verify external data formats and mitigate failures before initiating resource-intensive operations.
  • [SAFE]: Shell command execution in scripts/convert_to_gguf.py is restricted to legitimate environment setup (installing build tools) and the model conversion pipeline, using safe execution patterns without shell interpolation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 20, 2026, 03:15 AM
Security Audit — agent-trust-hub — hugging-face-model-trainer