The short version

What this episode covers

Claude Code for beginners: you don't need to write a line of code, you just need to describe what you want in plain English. This is the one tool and the one loop I use to build every single thing in this series, plus how to make it remember your rules and how to leave it running while you go make dinner.

This is episode 2 of Build Your Own AI OS. Episode 1 was the mental model. This one is the toolbox. I walk through the tool itself, the five step build loop I run on every job, what a skill is and when to save one, how the tool actually remembers your business, why a long chat session quietly gets slower and dumber, and how to set the trust dial so it asks about the risky stuff and handles the rest on its own. Everything after this episode gets built with exactly what's in here.

What you'll take away

  • What an AI coding assistant really is, and why the right mental model is a contractor with the keys to your shop
  • The five step build loop, describe, build, review, correct, refine, and the fork in the middle that trips people up
  • The difference between correcting something broken and refining something that already works
  • What a skill is, the signal that it's time to write one, and the humanizer skill I run on every draft
  • The two kinds of memory: a CLAUDE.md house rules file it reads every session, plus saved facts it keeps between them
  • What context rot is, why one long chat gets slower, pricier and dumber, and the 60 percent rule I use to clear the desk
The series

Build Your Own AI OS

5 episodes so far, all free to watch.

See every episode →
Got a question about this one?

Bring it to the community, or bring it to us.

The free community has a live call every Thursday. If you'd rather talk about your own business, book a call and we'll tell you what's worth automating and what isn't.