hf-download

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill prompt instructs the agent to pass Hugging Face tokens directly as command-line arguments (e.g., --hf_token <t> and header=Authorization: Bearer <token>), causing the LLM to handle and potentially expose secret values in its outputs.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (low risk: 0.10). The skill fetches repository metadata and downloads model/dataset files from Hugging Face using repository IDs provided as input, which is classified as public marketplace/registry item reading under low indirect prompt injection risk.