rag-implementation

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

• Third-party content exposure detected (high risk: 0.70). The required runtime workflow performs RAG by retrieving doc.page_content from an external vector store (retriever.ainvoke(...)) and concatenating it into the LLM prompt context (context_text = "\n\n".join(doc.page_content ...)), so any outsider-authored text present in that index can be injected into the agent’s LLM context.