omni-agents-request. OAR sends your prompt to every configured LLM in `agents/<name>.py` (including scraped chat UIs), then calls `agents/COMBINER.py` to merge everything into the final response.