Set Up Claude Code.
Claude Code is a coding agent you run in a terminal. This workshop gets it installed and proves it works, so class time is spent on judgment instead of installation. Work through it on the computer you will bring to class.
You need a paid plan before Claude Code will start.
Installing is free and takes a few minutes. Signing in is the step that can stop you, so check this before you spend time on the install.
Access required
Claude Code requires a Pro, Max, Team, Enterprise, or Console account. The free Claude.ai plan does not include Claude Code access.
Sign in with your own account. The course does not supply a plan and does not require you to buy one. If you already have a paid plan, or decide to start one, use it here. Check the eCampus announcement before you spend anything.
If cost or access is in your way, email Dr. Qu at wqu@monmouth.edu before class. You are not penalized either way: the transcript route covers the same ideas and earns the same completion credit.
- macOS13.0 or newer
- Windows10 (1809) or newer
- LinuxUbuntu 20.04+ / Debian 10+
- Memory4 GB RAM or more
- NetworkInternet connection required
- ShellBash, Zsh, PowerShell, or CMD
Never used a terminal before? Anthropic publishes a terminal guide ↗. There is also a Desktop app ↗ that runs Claude Code without a terminal, if the command line is a barrier for you.
Install Claude Code.
Pick the line that matches your computer. Paste it into your terminal and press Enter. On Windows, your prompt shows PS C:\ in PowerShell and C:\ without the PS in CMD.
curl -fsSL https://claude.ai/install.sh | bash
irm https://claude.ai/install.ps1 | iex
curl -fsSL https://claude.ai/install.cmd -o install.cmd && install.cmd && del install.cmd
Already use a package manager? brew install --cask claude-code on macOS or winget install Anthropic.ClaudeCode on Windows both work. Use the command above unless you already know you want one of those — for this course any of them is fine, and you only need one.
Prove the install worked.
Do not skip this. A version number is the evidence that the install succeeded, and it is what you will be asked to show in class.
claude --version
A working install prints a version number such as 2.1.211 (Claude Code). Your number will differ and that is fine — any version number means it worked.
claude doctor
claude doctor prints read-only install and settings diagnostics without starting a session.
Make a practice folder first.
Claude Code works on whatever folder you start it in. Do not start it in your home folder, your Documents folder, or a real assignment — make an empty one, so this week’s practice cannot touch anything you care about. This is the same boundary habit Week 2 teaches.
mkdir ~/claude-practice cd ~/claude-practice
mkdir "$HOME\claude-practice" cd "$HOME\claude-practice"
mkdir "%USERPROFILE%\claude-practice" cd /d "%USERPROFILE%\claude-practice"
Everything in the next two steps happens inside this folder. It starts empty, and that is fine.
Sign in.
Run this from inside the practice folder you just made. Starting Claude Code opens a login flow in your browser the first time.
claude
Follow the browser prompts to log in. Remember the access requirement above: a free Claude.ai account will not work here. If sign-in refuses your account, record exactly what you saw, move to the transcript route, and email Dr. Qu if cost or access is the reason.
Run one read-only request.
Week 2 is about watching what an agent does before trusting it. Keep this first request small and observable. Ask it to explain something, not to change anything.
Explain what this folder contains. Do not edit any files.
Watch what it reports before you agree to anything. You will use this same habit in the Week 2 Workflow Card: read the trace before trusting the result.
Six things that actually go wrong.
Bring the exact error text to class or to the TA. The message is the useful part.
| What you see | What it means | What to do |
|---|---|---|
command not found: claude | Installed, but your shell cannot find it yet. | Close the terminal and open a new one, then try again. |
The token '&&' is not a valid statement separator | You ran the CMD command inside PowerShell. | Use the PowerShell line instead. |
'irm' is not recognized | You ran the PowerShell command inside CMD. | Use the Windows CMD line instead. |
syntax error near unexpected token or a 403 | The download was blocked or incomplete, often campus Wi-Fi. | Try another network, then the official install troubleshooting page. |
| Permission errors during npm install | Global npm directory is not writable. | Do not use sudo. Use the native installer at the top of Step 1 instead. |
| Login rejects your account | You are on the free plan. | Sign in with a paid plan on your own account, or take the transcript route below for the same credit. Email Dr. Qu if cost is the barrier. |
Official source of truth: Troubleshoot installation and login ↗. Product requirements change; when this page and the vendor documentation disagree, the vendor documentation wins.
No access? Use the transcript route.
This is a real route, not a consolation prize.
Route B earns the same completion credit
If an account, a device, a campus permission, or an install error blocks you, the Week 2 Setup Readiness assignment supplies a recorded agent transcript. You analyze the same interaction and answer the same questions. Nobody has to pay for anything to complete Week 2.
Check yourself.
Tick what is true right now. This is only for you — nothing is submitted from this page, and your ticks stay in this browser.
0 of 5 checked.
Bring to class either a version number on your screen, or the exact error text that stopped you. Both count as coming prepared.
Back to Week 2.
Next you will read the agent loop, the authority boundary, and the Workflow Card.