RAG-targeted Adversarial Attack on LLM-based Threat Detection and Mitigation Framework