document-ocr-agent

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

• Third-party content exposure detected (high risk: 0.75). Runtime path: user supplies file_urls/file_ids (outsider-controlled documents via public URLs or external file storage), the tool fetches/downlods those documents and returns extracted OCR text (result.text_excerpt / result.entities), which is then ingested into the agent’s LLM context as part of the tool result.