Orchestrates multiple LLM agents to perform parallel code review, then synthesizes and votes on suggestions to rank improvements by consensus.
