Describe what you want in plain language on a project board. That's the entire human part of the workflow. JengoAGI scans the boards every half hour, picks up what's ready, and starts working.
Each task runs in its own isolated git worktree with its own branch, so multiple agents work in parallel without stepping on each other. The agent analyzes the problem, implements the change, runs the tests and opens a pull request with a structured description: problem, solution, done-when, technical notes.
A human reviews the end result, not the work. Over 200 tasks have gone through this loop at a 94% success rate. We demonstrated it live at one of the big consultancy firms; their reaction was that they have an entire AI department doing this by hand.