regex-vs-llm-structured-text

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

• Third-party content exposure detected (high risk: 0.70). The skill’s runtime LLM validator (validate_with_llm) ingests original_text (the document/content being parsed) into the LLM prompt; if that content includes outsider-authored free text (e.g., a downloaded/externally provided document or web-scraped text), it becomes LLM-readable context via the messages.create(... "Text: {original_text}") path.