Chapter 1 was about the thing you are talking to. This chapter is about the room you are both standing in. Every AI coding tool works by reading and writing files in folders on a computer, running commands in a terminal, and leaving a trail you can inspect afterwards. If those three things are vague to you, every later lesson has to stop and explain the room before it can make its point.

The most valuable section here is version control, and it is the one most people skip. An assistant that can edit forty files in ninety seconds is only safe if you can put those forty files back the way they were. Git gives you that, and reading a diff turns "it says it is done" into something you can actually check. Those two skills do more for your project's survival than any prompt you will ever write.

  1. Files, folders and paths. The filesystem is what every tool actually operates on, and most "it cannot find the file" errors are path errors.
  2. The editor and the terminal. What each window is for, and why the terminal stops being optional.
  3. Version control, plainly. Commits as save points, and the reason to make one before you let an AI work.
  4. Reading a diff. The verification skill that still works when the change is 300 lines long.
  5. Running it on your own machine. Dependencies, install, the dev server, and the five errors everyone hits.
  6. GitHub and what it is for. A second copy, a browsable history, and the place your deploys pull from.