regex-vs-llm-structured-text

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

• Third-party content exposure detected (medium risk: 0.30). The runtime path process_document(content) performs regex extraction first, then for regex-confidence failures it calls the LLM validator with outsider-provided content and original_text directly embedded into the prompt (validate_with_llm(..., original_text=content, ...)).