Playbook

How to run a software project with a coding agent and a Product Owner who questions rather than codes — written from this project, reusable for the next ones. Updated at every retrospective (ADR-0023); each rule says where it comes from.

flowchart LR
  P[Principles] --> S[Setup<br/>starter kit]
  S --> A[Analysis<br/>start-of-project interview]
  A --> W[Process<br/>epics, features, stories]
  W --> G[Agent<br/>configuration and habits]
  G --> L[Lessons<br/>what went wrong, what prevents it]
  L -.->|each retrospective| P

Page	Content
principles.md	the rules that make agent-written code trustworthy
setup.md	the starter kit: what to copy, in which order, the rhythm, what to read first
analysis.md	the start-of-project interview, the agent as functional analyst
process.md	backlog, research, audit, demo, review, retrospective
agent.md	configuring and steering the agent; the Product Owner's lessons, habit by habit
lessons.md	what went wrong, and the check that now prevents it

To start a new project: read these pages, then follow setup.md; copy files from this repository rather than rewriting them, and keep the links to the ADRs that justify them.