I gave one lead agent 20 Linear issues. It created separate workspaces and started a different coding agent in each one.
Each agent made its changes, ran tests, tested the app in a browser, and took screenshots. I reviewed every result before merging anything.
In about two hours, 18 issues were completed and two needed my help.
Demo: https://x.com/swarajb/status/2087535598672429334?s=20 GitHub: https://github.com/swarajbachu/zuse
I’d love feedback on how you’d want to review this much parallel agent work and you can fork or add your own changes to the github app