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 workflow ingests first-party “content: str” passed into process_document, runs regex parsing and then sends that same input text into the LLM via validate_with_llm only for low-confidence items, so outsider-authored text would be exposed if the caller supplies it as the document “content.”