← All posts
Claude5 min read

Claude Code Basics: What to Do in Your First Hour

July 24, 2026

If you've installed Claude Code and logged in, you're ready for the part that actually matters: using it. This post is your first hour — no prior experience assumed.

The single most important thing to understand up front: Claude Code can read and change the files in whatever folder you start it in. That's what makes it powerful, and it's also why it's careful about asking permission before it changes anything. We'll come back to that.


Start it inside a folder

Claude Code works on one folder (a "project") at a time. So first you point your terminal at the folder you want to work in, then start Claude.

cd path/to/your/folder
claude

cd means "change directory" — it's how you tell the terminal which folder to look at. Once you run claude, you'll see it start up and show you which folder and which AI model it's using. Now you can just type.

If you're only experimenting, make a new empty folder first so you can't disturb anything important.


Talk to it like a person

There's no special syntax. You describe what you want in plain English, the same way you would in a chat window:

"What does this project do?"

"What kinds of files are in this folder?"

Claude will look through the folder on its own and answer. This is the best way to start — ask before you act. Get it to explain what's there before you ask it to change anything. You'll understand your own project better, and you'll build trust in what Claude is seeing.


It asks before it changes anything

This is the part that makes Claude Code safe to learn on. By default, Claude will read your files freely — but the moment it wants to change a file or run a command, it stops and asks you first. You'll see the proposed change and three choices:

  • Approve it (let it go ahead)
  • Reject it (stop, and optionally tell it why)
  • Ask for a different approach (just type what you'd rather it do)

Nothing happens to your files without your say-so. You are always the one clicking "yes." If you're nervous, that's the safety net — read each proposal before you approve it.

Want the guided version first?

If jumping straight into a terminal feels like a lot, Clearly, AI's free Claude courses build your confidence with Claude step by step, in plain English — a friendlier on-ramp before the advanced tools.

Start the free Claude courses →

Making your first change

When you're ready, ask for something concrete and specific:

"Add a short README file to this folder that explains what the project is."

Claude will figure out what to write, show you the file it wants to create, and ask you to approve it. Say yes, and the file appears in your folder — a real file you can open and read.

The lesson that carries over from every other AI tool: be specific. "Add a function" is vague. "Add a note to the top of the budget.txt file summarizing the totals" gives Claude something it can actually do well. The clearer your request, the better the result.


Look before it leaps: plan mode

For anything bigger than a one-line change, there's a gentler setting called plan mode. In plan mode, Claude tells you what it intends to do — the whole plan — without doing any of it yet. You read the plan, and only then decide whether to let it proceed.

You switch modes by pressing Shift + Tab (it cycles through the options; you'll see the current mode on screen). For beginners, plan mode is a great habit: it turns "Claude did a bunch of stuff" into "Claude showed me its plan, and I approved it."


A few commands worth knowing

Inside Claude Code, anything you type starting with a slash / is a command rather than a request. The handful worth knowing on day one:

  • /help — shows everything you can do
  • /init — sets up a small guide file so Claude understands your project better (nice to run once in a project you'll return to)
  • /clear — starts a fresh conversation when you switch to an unrelated task
  • /model — lets you switch which AI model you're using
  • /exit — closes Claude Code (or press Ctrl + D)

You don't need to memorize these. Type / on its own and a menu appears.


Three beginner mistakes to sidestep

  1. Trying to use a free account. Claude Code needs a paid Claude plan (Pro or Max). A free account won't get you in.
  2. Being vague about where. Tell Claude which file you mean. "Fix the list" is guesswork; "fix the list in notes.txt" is a clear target.
  3. Letting one session run forever. If you've been going for a while and switch to something completely different, type /clear to start clean. It keeps your project knowledge but drops the old back-and-forth — which keeps Claude focused (and your usage efficient).

The takeaway

Claude Code feels technical because of the terminal, but the actual skill is the one you've been building all along: describe what you want clearly, give it context, and check the result. The terminal is just a different room to practice it in.

Start small. Ask questions before you make changes. Approve one edit at a time. Within an hour, the strangeness wears off — and you'll have Claude doing real work on real files, with you in control the whole way.

📋

Want more like this? Grab the free AI Starter Kit.

Practical things you can do with AI today — each with a prompt you can copy and use. Free PDF, plain English.

Get the free kit →

Ready to go further?

The full Clearly, AI courses go deep on everything in this post — with hands-on exercises, real prompts, and new modules launching regularly. And they're completely free.

Start learning free →