hugging-face-tool-builder

Pass

Audited by Gen Agent Trust Hub on Sep 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill is primarily designed to generate, test, and execute shell, Python, and TypeScript scripts for interacting with the Hugging Face API and CLI.
  • [EXTERNAL_DOWNLOADS]: The skill fetches API specifications and repository data from huggingface.co. These network operations target a well-known service for legitimate development purposes.
  • [DYNAMIC_EXECUTION]: The skill involves the runtime generation and execution of scripts to fulfill user automation requirements for the Hugging Face Hub.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes metadata and documentation from external repositories on the Hugging Face Hub, which could potentially contain untrusted content.
  • Ingestion points: API responses from huggingface.co/api/ and output from the hf CLI tool.
  • Boundary markers: None explicitly defined in the provided instructions for handling the content of API responses.
  • Capability inventory: The skill uses curl, jq, and the hf CLI, and it can write and execute files in shell, Python, and TypeScript environments.
  • Sanitization: The skill recommends using jq to extract specific fields, which provides some structure, but does not specify further sanitization for generated script logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 10, 2026, 12:56 AM
Security Audit — agent-trust-hub — hugging-face-tool-builder