Eight chapters, in the order a project needs them.
Every section is listed on the index. This page is about the order itself: what each chapter answers, why it sits where it does, and what changes once you have been round once.
Understand
What you are working with
None of this is about code yet. Chapter 1 is the machine itself: what it can see, what it charges you, and how a session fails while it still sounds confident. Chapter 2 is the room the work happens in, down to files, the terminal and version control. Chapter 3 is the one people skip, and it holds most of the leverage: handing over a job small enough to finish and small enough to check.
Build
Something real, made safe
Chapter 4 is the anatomy of an app and how to choose a stack you can finish on your own. Chapter 5 sits ahead of the shipping chapter deliberately. Secrets, user data and the things you never let an agent do unsupervised are far cheaper to get right while nobody is using the thing yet.
Ship
Out, and still alive
Chapter 6 covers hosting, domains, environments and knowing how to undo a deploy before you need to. Chapter 7 is the unglamorous half: logs, alerts, backups, dependency updates, and changing something while real people are using it.
Repeat
Without starting over
Chapter 8 is the compounding one. Instruction files, automatic checks, picking a project back up after two weeks away, and splitting work across sessions so something bigger than one sitting can hold still gets finished.
Chapter 8’s last section points back at chapter 1. The second pass reads differently, because by then you have shipped something and the failure modes already have names you know.