skills/zapabob/skills/huggingface-hub/Gen Agent Trust Hub

huggingface-hub

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides an installation command curl -LsSf https://hf.co/cli/install.sh | bash -s which fetches and executes a script from Hugging Face's official domain.
  • [INDIRECT_PROMPT_INJECTION]: The skill documents capabilities that ingest external data, such as executing SQL queries against datasets (hf datasets sql) and running Python scripts with dependencies (hf jobs uv).
  • Ingestion points: Dataset SQL queries and Python job scripts.
  • Boundary markers: None specified in the reference guide.
  • Capability inventory: Utilizes the hf CLI for data processing and remote compute tasks.
  • Sanitization: No explicit sanitization or validation logic is defined in the documentation provided.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 12:07 PM