Summary

You now have a working answer to "how does Cogitave operate from Day 1": one model, specified in two layers, from the estate's first day.

In this module, you:

• Learned the inner loop: a pinned, reproducible toolchain (mise, just) and the parity contract - lefthook commands == CI commands == just targets - plus trunk-based delivery with a preview environment on every PR.
• Learned the outer loop: WIP-limited continuous flow (Kanban-style) day to day, Shape-Up-style betting for larger bets, and the request-lifecycle Request as the single unit that flows through both.
• Saw why ADR-0025 rejected Scrum for this team - sprint batching against trunk-based CD, a gameable velocity metric, and ceremony built for a coordination problem this small, autonomy-by-default team does not have.
• Understood that the WIP limit protects the human-attention budget, not agent throughput, and that the human gate stays an exception handler - the same doctrine from "Apply the AGENTS floor."

This is the last module in Operate the estate. Completing it earns the path's trophy: you can now describe how the estate runs a change from Day 1, end to end, under least privilege with every action recorded as evidence.

Next steps

• ways-of-working - the full outer-loop standard; keep it open when your squad stands up its board.
• development-process - the full inner-loop contract, including the canonical just targets table.
• ADR-0025 - the decision record, if you want the full landscape survey and the rationale behind the choice.