alterlab-pdf-extract

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

• Third-party content exposure detected (high risk: 0.95). Outsider-authored free text from the input PDFs is converted to Markdown via Microsoft MarkItDown (convert_to_markdown() → md.convert(path)), and that converted prose is then placed into the LLM prompt in extract_llm() as Paper ({source}): {body} (when --backend llm is used).