Claude Code Tutorial: Build Faster with AI-Powered Coding

Claude Code is Anthropic's official CLI for using Claude directly in your terminal. It gives you a full-context AI coding assistant that can read, edit, and understand your entire codebase — not just the snippet you paste in. This tutorial walks you through installation, daily workflows, and how to get the most out of Claude Code for real development tasks.

MJK Supplies · Jun 17, 2026 · 12 min read

ShareX in f ↗

Claude Code Tutorial: Build Faster with AI-Powered Coding

What Is Claude Code?

Claude Code is a command-line tool that runs Claude Sonnet or Opus with full access to your local file system. Unlike browser-based AI tools, Claude Code:

Reads your actual files and directories
Makes edits directly to your codebase
Runs commands on your behalf (with your approval)
Understands your project as a whole, not just isolated snippets
Integrates into your existing terminal workflow

It's the difference between asking an AI to fix a bug in a code snippet versus asking it to fix a bug across your actual project.

Installation

npm install -g @anthropic-ai/claude-code

Then authenticate:

claude

Claude Code will prompt you to log in to your Anthropic account. Once authenticated, you're ready.

Requirements: Node.js 18+. Works on macOS, Linux, and Windows (via WSL or Git Bash).

Basic Usage

Start Claude Code in any project directory:

cd your-project claude

This opens an interactive session. Claude immediately has access to your project files and can navigate your directory structure.

Basic commands you'll use constantly:

> Read the README and explain what this project does > Find the function that handles user authentication > The /api/users endpoint is returning a 500 error — debug it > Add input validation to the signup form > Write unit tests for the UserService class

Claude reads the relevant files, proposes changes, and asks your approval before editing anything.

Slash Commands

Claude Code has built-in slash commands for common operations:

/help — See all available commands /clear — Clear conversation context /compact — Compress long conversation to save tokens /status — Show current project state /memory — View and edit persistent memory for this project

Working with Your Codebase

Claude Code shines on tasks that require understanding multiple files:

Debugging across files:

> The checkout flow is broken — users get an error when they add to cart. > The cart service is in src/services/cart.ts and the API is in api/routes/cart.js

Refactoring:

> Rename the UserProfile component to AccountProfile throughout the entire codebase > Convert all callback-based functions in the utils/ directory to async/await

Understanding unfamiliar code:

> Explain how the authentication middleware works and trace a request from login to protected route

Adding features:

> Add rate limiting to all API endpoints. Use the existing middleware pattern I see in api/middleware/auth.js

CLAUDE.md — Project Instructions

Create a CLAUDE.md file in your project root to give Claude Code persistent instructions:

# Project Guidelines - Use TypeScript strict mode - All API functions must have JSDoc comments - Tests go in __tests__ beside the file they test - Never use var, always const or let - Database queries use the db utility in lib/db.ts

Claude Code reads CLAUDE.md at the start of every session. Your conventions are automatically applied to every suggestion.

Agentic Mode

For longer, autonomous tasks, Claude Code can operate with more independence:

claude --dangerously-skip-permissions

In this mode, Claude makes changes without asking for permission on each step. Use this for well-defined, isolated tasks where you trust the changes (e.g., "add JSDoc to all functions in this file").

Practical Workflows

Morning code review:

> Review all the changes I made yesterday (git diff HEAD~1) > and flag any potential issues, missing tests, or code that doesn't follow our patterns

PR preparation:

> Review my staged changes, write a clear commit message, > and suggest what should be in the PR description

Bug investigation:

> Users are reporting that emails aren't being sent. > Look at the email service, check for recent changes in git, > and trace through the send flow to find the issue

Documentation:

> Generate comprehensive API documentation for all public endpoints > in the api/ directory. Use the existing docs/api.md as a format reference