rfp-response-extraction

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

• Third-party content exposure detected (high risk: 0.85). Outsider-authored free text from a SharePoint RFP document (and its addenda/Q&A/attachments) is read and extracted at runtime (“Extract the full text of each document” in Step 1), then used to generate tracker items—so the LLM context can include that externally authored document text and any injection-like instructions contained within it.