The Map: Four Buckets

In short

The toolkit breaks down into four questions, the same ones you'd ask about a new hire.
Before the four there's the engine: which model to put to work, and how much thinking power.
The order doesn't follow logic: it follows the questions that come up as you use it.
There's nothing to memorize: it's just here so you always know where you are.

Four questions, not a list of features

If I listed every single thing Claude Code can do right now, you'd remember two of them. That's not on you: a list of ten new names doesn't stick in anyone's head.

So we're not making a list. We're asking four questions — the same four you'd ask yourself if a new hire showed up at your studio tomorrow.

1. What do I let it touch? Before handing over the keys, you want to know what it will do, and you want to be able to say no. This is where you'll find getting it to tell you the plan before acting, permissions, and working modes.
2. What does it know about me? A collaborator who starts from zero every day is useless. This is where everything you leave it in writing lives: your standing instructions, how much it can hold on the table, and the skills you set up once and reuse forever.
3. How many do I put to work? Sometimes one isn't enough. This is where the helpers that work on their own and the teams that talk to each other live.
4. What's it connected to outside? Your work doesn't all fit in one folder. This is where connecting it to the tools you already use lives, and packaging everything up to share it.

Before the four: the engine

There's one thing that comes before all of this, because it isn't a piece of the toolkit: it's the engine. There isn't just one — some are more powerful, some faster, some cheaper — and picking the wrong one is the most common way to spend a lot and get little.

That's the chapter starting in a moment. Right after, we install it and have it do its first real task: small, yours, and it works. Only then do the four buckets open, in order.

Why this order

We could tackle these in plenty of ways. We're picking a specific one: you answer a question once it's actually come up for you.

The moment you install it, the first thing you see on screen is Claude asking your permission to do something. That's why bucket 1 is that one. The question "why do I have to repeat myself every time?" comes up on your second or third real task: that's bucket 2. Nobody asks about helpers or outside connections before having done something concrete: those come later.

One thing we're deliberately pushing to the end: there's a way to trigger automatic actions at precise points in the loop. It's level 2 of the first bucket, and it's the most advanced thing in the course — you'll find it near the end, not in the lesson right after install.

And at the very end there's the project: something of your own, built from scratch, with no safety net.

Now what

The "what's under the hood" chapter ends here: you know what the loop is, you know the model has no memory, you know what the harness is, and you know how the toolkit is laid out. In the next lesson we pick the engine — three models, and the rule for not getting the one you pay for wrong.

Did you get it?

Two minutes, no grade. Trying to answer — even getting it wrong — cements the concepts far more than rereading. Not in the mood? Skip it: the lesson still counts as complete.

Which of these is NOT one of the map's four questions?

Why do we start with permissions and not with the file holding your instructions?

What is this map for?

1
1. 1
2. 2
3. 3
2
1. 4
2. 5
3. 6
4. 7
3
1. 8
2. 9
3. 10
4. 11
5. 12
4
1. 13
2. 14
3. 15
4. 16
5. 17
6. 18
5
1. 19
2. 20
3. 21
4. 22
5. 23
6. 24
6
1. 25
2. 26
3. 27
4. 28
5. 29
6. 30
7
1. 31
2. 32
3. 33
4. 34
8
1. 35
2. 36
3. 37
4. 38
9
1. 39
2. 40
3. 41
10
1. 42
2. 43
3. 44
11
1. 45
2. 46