Split frontend and backend between two coding agents with shared context

One developer spent 15 days running two coding agents side by side on one project. Claude Code handled the frontend and Codex handled the backend. Both worked in a shared space called Tutti, so each agent could see the same files and what the other had done. To hand work over, they passed a task card to the other agent instead of explaining it again. They could also point Codex at an older Claude chat. A board showed each task as running, done, failed, or stopped, which made it easy to see where things broke. The main gain was less time spent coordinating. Two weeks was too short to really test it, and the post may be promoting Tutti.