Agent teams let you spin up multiple AI agents that work together in parallel, talk to each other, and QA each other's work inside Claude Code.
In this video I break down exactly how to enable them, how to write prompts that actually produce good results, and when you should use agent teams vs sub-agents. I also run two live demos so you can see the whole thing in action, including the tmux split-pane view where you can watch each agent think and work in real time.