When one AI session isn't enough to build a product, you need to organize your work — the same way a real dev team does.

Hats does this for Claude Code. Five sub-agents, each with its own context and file access. The Manager writes specs but can't implement. QA writes tests but can't read source. Developer makes tests pass without ever seeing the tests. Roles check each other against the spec, not against assumptions.

Run the team step by step, or hand off to /autopilot and let the pipeline run while you do something else.