ocr-and-documents

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

• Third-party content exposure detected (high risk: 0.85). The required workflow can ingest outsider-authored free text when the user provides a remote document URL and the skill runs web_extract(...), which fetches public web content (e.g., arbitrary PDFs like arXiv) and converts the page/PDF text into LLM-readable markdown for the agent context.