llm-integration

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

• Third-party content exposure detected (high risk: 0.70). The required workflow includes a RAG pipeline that injects retrieved chunk.text into the LLM prompt as Context:\n${context}\n\nQuestion: ${query}, and those chunks can originate from non-user-authored documentation sources at runtime (outsider-authored free text via vectorDb retrieval).