skills/huggingface/skills/hf-cli/Gen Agent Trust Hub

hf-cli

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • Remote Installation Script: The skill references an installation script at https://hf.co/cli/install.sh. This is an official distribution channel for the vendor's tool and originates from a verified domain.
  • Extension Framework: The hf extensions command group allows users to install and run code from GitHub repositories. This provides modularity, and users should review extensions as they would any third-party software.
  • Workload Management: The skill enables running remote jobs and scripts (e.g., hf jobs run or hf jobs uv), which involves executing commands or Python scripts on Hugging Face infrastructure.
  • Authentication Handling: The skill provides commands to manage Hugging Face access tokens (hf auth). These tokens are stored locally to facilitate interactions with the Hub, following standard CLI authentication practices.
  • Data Processing Surface: Commands like hf datasets sql process dataset contents. This represents a potential surface for indirect injection if dataset contents are treated as instructions, though this is managed by the underlying tool's query engine.
  • Ingestion points: External dataset files processed via SQL (SKILL.md).
  • Boundary markers: None explicitly defined in the CLI command descriptions.
  • Capability inventory: Remote job execution and extension management (SKILL.md).
  • Sanitization: Standard SQL query parsing is applied by the tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 04:01 PM