Your First Win in 5 Minutes

Total time: 7 minutes (2 min reading + 5 min hands-on)

Let’s get Claude Code working with your actual code. No theory. Just results.

Prerequisites Check

Before we start, quick check:

  • □ Can you open a terminal? (Command Prompt, PowerShell, Terminal, etc.)
  • □ Do you have a code project with git? (Any project with a .git folder)

Both yes? Great, let’s continue.

No project? You can use any existing project on your machine, or clone a demo:

git clone https://github.com/your-favorite/repo
cd repo

Step 1: Install (2 minutes)

Copy and paste the command for your platform:

macOS / Linux / WSL 

curl -fsSL https://claude.ai/install.sh | bash

Windows (PowerShell) 

irm https://claude.ai/install.ps1 | iex

Alternative: Package Managers 

# macOS with Homebrew
brew install --cask claude-code

# Windows with WinGet
winget install Anthropic.ClaudeCode

Or: VS Code Extension

Open VS Code → Extensions → Search “Claude Code” → Install

Verify Installation 

claude --version

✅ Success check: You should see a version number (e.g., claude-code 0.5.0)

❌ Seeing “command not found”?

  • Restart your terminal
  • Or run: source ~/.bashrc (Linux/macOS) or source ~/.zshrc (macOS with zsh)
  • Windows: Restart PowerShell

Step 2: Login (1 minute) 

claude /login

This opens your browser for authentication.

Don’t have a Claude account? Create a free account at claude.ai

✅ Success check: Terminal shows “Authenticated successfully”

❌ Authentication failed?

  • Make sure you completed the browser login
  • Try running claude /login again
  • Check Troubleshooting Guide for API key alternatives

Step 3: Three Commands (5 minutes)

Navigate to your project:

cd /path/to/your/project

Now run these three commands. Each demonstrates different Claude Code capabilities.

Command 1: Review Your Code (30 seconds) 

git diff | claude "review for bugs and best practices"

What this does: Analyzes your uncommitted changes and flags potential issues.

Expected output: Claude will review your code and point out:

  • Potential bugs
  • Code style improvements
  • Security concerns
  • Best practice violations

Why this matters: You just got a code review in 30 seconds instead of waiting 15 minutes for a human reviewer.

No uncommitted changes? Make a small edit to any file first, or try:

git show HEAD | claude "review this commit"

Command 2: Understand Code (30 seconds) 

claude "explain what src/index.js does in simple terms"

Replace src/index.js with an actual file in your project. Any file works - pick one you didn’t write or don’t fully understand.

Expected output: Plain English explanation of:

  • What the file does
  • Key functions and their purpose
  • How it fits into the larger system

Why this matters: No more hunting through code trying to understand what it does. Just ask.

Example with specific questions:

claude "where do we handle user authentication?"
claude "what does the calculateTotal function do?"
claude "show me all API endpoints in this project"

Command 3: Write a Test (2 minutes) 

claude "write jest tests for src/utils.js - cover happy path and edge cases"

Replace src/utils.js with an actual file. Pick a utility file, helper function, or small module.

Expected output: Complete test file with:

  • Multiple test cases
  • Edge case coverage
  • Mocks and fixtures if needed
  • Following your project’s test patterns

Why this matters: Test writing went from 20 minutes to 2 minutes. Now you’ll actually write tests instead of skipping them.

Don’t use Jest? Adjust the command:

claude "write pytest tests for utils.py"
claude "write mocha tests for utils.js"
claude "write xunit tests for Utils.cs"

What Just Happened?

Let’s pause and reflect on what you just did:

  1. ✅ Analyzed real code with git diff | claude
    • Used Unix piping (Claude Code works like any command-line tool)
    • Got instant feedback on your changes
  2. ✅ Leveraged full codebase context
    • Claude read your entire project automatically
    • Understood relationships between files
    • Gave context-aware answers
  3. ✅ Generated new code that follows your patterns
    • Tests match your existing test style
    • Variable names match your conventions
    • Imports use your project structure
  4. ✅ Stayed in your terminal
    • No context switching to a browser
    • No copy-pasting code back and forth
    • Integrated with your existing workflow

Try These Next

Now that you’ve verified it works, experiment with these commands:

Debug Test Failures 

npm test 2>&1 | claude "explain these failures and suggest fixes"

Generate Documentation 

claude "update README with API endpoints from src/routes/"

Find Code 

claude "where do we connect to the database?"
claude "show me all files that import lodash"

Refactor Code 

claude "refactor src/legacy.js to modern ES6 - preserve behavior"

Generate Commit Messages 

git diff --cached | claude "write a conventional commit message"

Troubleshooting

“Command not found: claude”

Solution: Restart your terminal, or:

# Linux/macOS
source ~/.bashrc  # or ~/.zshrc

# Windows PowerShell
# Close and reopen PowerShell

“Authentication failed”

Solution:

  1. Run claude /login again
  2. Complete the browser authentication
  3. If still failing, try API key method in Environment Setup

“Claude doesn’t understand my project”

Solution:

  • Make sure you’re in the project root directory: cd /path/to/project
  • Verify .git folder exists: ls -la | grep git
  • Claude needs a git repository to understand project structure

“Response is slow or hangs”

Solution:

  • Check your internet connection
  • Very large projects may take 10-20 seconds on first run (context caching)
  • Subsequent runs will be faster

“Claude suggested incorrect code”

Solution:

  • This is normal - AI isn’t perfect
  • Always review suggested code before using it
  • Be more specific in your prompts: “write jest test for calculateTotal function using mocks”

More issues? See the comprehensive Troubleshooting Guide

Success Check

Before moving to the next section, verify:

  • ✅ Did you successfully run all 3 commands?
  • ✅ Did Claude analyze your actual project code?
  • ✅ Did you see useful, relevant output?

All yes? Congratulations! You’ve just used Claude Code successfully.

Some no? That’s okay. Review the troubleshooting section above or ask for help in the Anthropic Discord.

What’s Next?

You now have hands-on proof that Claude Code works. You’ve seen it:

  • Review code
  • Explain existing code
  • Generate new code

But why should you use this regularly? Next section shows the concrete benefits and productivity gains.

← Previous: What Is Claude Code? Next: Why Claude Code Matters →
This site uses Just the Docs, a documentation theme for Jekyll.