huggingface-tool-builder

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime scripts (e.g., hf_model_card_frontmatter.sh and hf_enrich_models.sh) ingest user-provided model IDs from stdin/arguments, then fetch and parse Hugging Face model README/frontmatter text from external repositories for LLM-readable free text (Markdown/YAML), creating an outsider-authored indirect prompt-injection surface.