Can I have make LLM workflows in llama-index like having questions frist be screened by an agent to ensure that question contains all relevent information before querying db and synthesizing respone. then if the query passes inspection by first agent, then send the query to second agent to select yet another appropriate agent to route the query to?
