sn-da-non-spreadsheet-analysis

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

• Third-party content exposure detected (high risk: 0.70). The skill ingests the user-supplied PDF/Word/PPT files from input_path and extracts their readable text (and may caption embedded images) at runtime; if those documents are authored by outsiders, their free-form text becomes LLM context via the document parsing/OCR caption paths (e.g., page.get_text("text") for PDFs or python-pptx/python-docx text extraction, plus sn-da-image-caption/scripts/caption.py for image-based content).