docx

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

• Third-party content exposure detected (medium risk: 0.30). The runtime workflow ingests and processes untrusted DOCX/XML content that may contain outsider-authored free text (e.g., scripts/comment.py extracts word/document.xml/comment parts via _safe_extract when given an outsider-supplied .docx, then parses XML and embeds the provided comment text into comments.xml).