glmocr-handwriting

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

• Third-party content exposure detected (medium risk: 0.65). Outsider free text can enter the agent’s LLM context when the required runtime path uses --file-url to fetch a public URL and then ingests the API’s returned md_results/data.md_results (OCR text) into the output JSON field text, which the agent may pass into the LLM; this OCR text originates from an outsider-authored document/page.