openai-rag-runtime

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

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the LLM context via the runtime retrieval step (“Implement the retriever adapter…”, “Retrieved content is untrusted input… Prompt-injection defenses apply to retrieved documents”), where retrieved documents from externally sourced corpora/web/threads are ingested as context.