Skip to main content

Claude Code: Your AI Coding Agent

info

This content was developed with AI assistance and is regularly reviewed for accuracy.

Claude Code is an AI-powered coding agent built by Anthropic. Unlike traditional chatbots that only answer questions, Claude Code reads your entire codebase, edits files, runs terminal commands, creates pull requests, and works through complex multi-step tasks. It is available in the terminal, VS Code, JetBrains IDEs, a desktop app, and the browser.

What You'll Learn​

  • What Claude Code is and how it differs from other AI tools
  • How to install and set up Claude Code on your machine
  • Essential workflows for beginners: exploring code, fixing bugs, writing tests
  • Intermediate techniques: memory files, skills, subagents, and automation
  • Best practices for getting high-quality results

Why Claude Code Stands Out​

Most AI coding tools work inside a chat window. You paste code in, get a response, and copy it back. Claude Code takes a fundamentally different approach:

FeatureTraditional AI ChatClaude Code
Codebase accessYou paste snippetsReads your entire project automatically
File editingYou copy/paste suggestionsEdits files directly with your approval
Running commandsYou run them yourselfExecutes terminal commands for you
Multi-step tasksOne question at a timePlans and executes entire workflows
MemoryForgets between sessionsRemembers project context via CLAUDE.md
Tool useNoneConnects to GitHub, databases, APIs via MCP

Getting Started​

Requirements​

Before you install, make sure you have:

  • Operating system: macOS 13+, Windows 10+, or Ubuntu 20.04+
  • Account: A Claude Pro, Max, Teams, Enterprise, or Console account (the free Claude.ai plan does not include Claude Code)
  • Internet connection: Required for all sessions
  • 4 GB+ RAM

Install Claude Code​

The recommended installation uses the native installer, which auto-updates in the background.

macOS or Linux:

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

Windows PowerShell:

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

Windows users also need Git for Windows installed first.

After installation, verify it worked:

claude --version

Start Your First Session​

Navigate to any project folder and type claude:

cd your-project
claude

You will be prompted to log in on first use. After that, you can start asking questions or giving instructions in plain English.

Try These First Prompts​

"give me an overview of this codebase"
"explain the main architecture patterns used here"
"find the files that handle user authentication"

These read-only prompts help you explore a project without making any changes. Claude reads your files, understands relationships, and explains what it finds.

Essential Workflows for Beginners​

Explore an Unfamiliar Codebase​

When you join a new project or open an unfamiliar repo, Claude Code is one of the fastest ways to get oriented:

"what are the key data models?"
"how is authentication handled?"
"trace the login process from front-end to database"

Start with broad questions and narrow down. Ask about coding conventions, project-specific terms, and how components interact.

Fix a Bug​

Share the error with Claude and let it investigate:

"I'm seeing this error when I run npm test: [paste error]"
"suggest a few ways to fix the TypeError in user.ts"
"update user.ts to add the null check you suggested"

Claude can read error messages, trace the source, suggest fixes, apply changes, and run tests to verify everything works.

Write Tests​

Tell Claude what you want tested and let it match your project's existing patterns:

"find functions in auth.js that are not covered by tests"
"add tests for the notification service, including edge cases"
"run the new tests and fix any failures"

Claude examines your existing test files to match the framework, assertion style, and conventions already in use.

Create a Pull Request​

Once your changes are ready:

"summarize the changes I've made to the authentication module"
"create a pr"

Claude generates a descriptive PR with context about what changed and why. Review it before submitting.

Working with Files and Context​

Reference Files with @​

Use @ to include specific files or directories in your prompt without waiting for Claude to search:

"explain the logic in @src/utils/auth.js"
"what's the structure of @src/components?"

This is faster than asking Claude to find files and gives it exactly the context you want.

Work with Images​

You can drag and drop screenshots, paste images with Ctrl+V, or reference image paths:

"here's a screenshot of the error. What's causing it?"
"generate CSS to match this design mockup"

Intermediate Techniques​

Use Plan Mode for Safe Exploration​

Plan Mode lets Claude analyze your codebase with read-only operations. It will not make any changes, which makes it ideal for research and planning complex features before you start coding.

Start a session in Plan Mode:

claude --permission-mode plan

Or switch into Plan Mode during a session by pressing Shift+Tab (cycle through modes). Ask Claude to create a detailed plan, refine it with follow-up questions, then switch to normal mode to execute.

"I need to refactor our authentication system to use OAuth2. Create a detailed migration plan."

Press Ctrl+G to open your current input in an external text editor if you want to draft or revise a longer prompt before sending it.

Set Up CLAUDE.md for Project Memory​

CLAUDE.md is a special file Claude reads at the start of every session. It gives Claude persistent context about your project: build commands, coding standards, architectural decisions, and workflow rules.

Generate a starter file automatically:

/init

Claude analyzes your codebase and creates a CLAUDE.md with build commands, test instructions, and conventions it discovers. Then refine it manually with things Claude cannot infer from code alone:

# Code style
- Use ES modules (import/export), not CommonJS (require)
- Destructure imports when possible

# Workflow
- Run `npm test` before committing
- Prefer running single tests, not the whole suite, for performance

Where to place CLAUDE.md files:

LocationScopeUse case
~/.claude/CLAUDE.mdAll your projectsPersonal preferences
./CLAUDE.mdThis projectTeam-shared conventions (check into git)
./subdirectory/CLAUDE.mdSpecific subdirectoryModule-specific rules

Keep CLAUDE.md files concise. For each line, ask: "Would removing this cause Claude to make mistakes?" If not, cut it.

Create Skills for Repeatable Workflows​

Skills extend Claude's knowledge with project-specific instructions. Unlike CLAUDE.md (which loads every session), skills load on demand, keeping your context lean.

Create a skill by adding a SKILL.md file inside .claude/skills/:

---
name: fix-issue
description: Fix a GitHub issue
---
Analyze and fix the GitHub issue: $ARGUMENTS.

1. Use `gh issue view` to get the issue details
2. Search the codebase for relevant files
3. Implement the fix
4. Write and run tests
5. Create a descriptive commit message
6. Push and create a PR

Invoke it with /fix-issue 1234.

Use Subagents for Parallel Work​

Subagents run in their own context window, keeping your main conversation clean. This is powerful for research-heavy tasks:

"use subagents to investigate how our authentication system handles token refresh"

The subagent reads files, analyzes code, and reports back a summary without cluttering your main session. You can also create custom subagents in .claude/agents/:

---
name: security-reviewer
description: Reviews code for security vulnerabilities
tools: Read, Grep, Glob, Bash
---
You are a senior security engineer. Review code for:
- Injection vulnerabilities
- Authentication flaws
- Secrets in code
- Insecure data handling

Connect External Tools with MCP​

The Model Context Protocol (MCP) lets Claude Code connect to external services like Notion, Figma, databases, and custom APIs. Add an MCP server:

claude mcp add

Once connected, Claude can pull data from issue trackers, query databases, fetch designs, and automate workflows, all from natural language prompts.

Automate with Non-Interactive Mode​

Use claude -p to run Claude in scripts, CI pipelines, and pre-commit hooks:

claude -p "explain what this project does"
claude -p "list all API endpoints" --output-format json
cat build-error.txt | claude -p "explain the root cause of this build error"

Add Claude to your build scripts for automated code review:

{
  "scripts": {
    "lint:claude": "claude -p 'look at the changes vs. main and report any typos. report filename and line number.'"
  }
}

Run Parallel Sessions with Git Worktrees​

For working on multiple tasks simultaneously, use worktrees so each Claude session has its own copy of the codebase:

claude --worktree feature-auth

This creates an isolated working directory with its own branch. Start another session in a separate worktree:

claude --worktree bugfix-123

Changes in one session never interfere with another.

Best Practices​

Give Claude a Way to Verify Its Work​

The single most effective thing you can do is provide clear success criteria:

  • Instead of: "implement a function that validates email addresses"
  • Say: "write a validateEmail function. Test cases: user@example.com is true, invalid is false, user@.com is false. Run the tests after implementing."

Explore First, Then Plan, Then Code​

For complex features, separate research from implementation:

  1. Explore (Plan Mode): Read files and understand the current state
  2. Plan: Ask Claude to create a detailed implementation plan
  3. Implement (Normal Mode): Execute the plan with verification
  4. Commit: Ask Claude to commit with a descriptive message

Manage Context Aggressively​

Claude's context window fills up as you work. Keep it lean:

  • Use /clear between unrelated tasks
  • Use /compact to summarize long conversations
  • Delegate research to subagents so exploration doesn't consume your main context
  • Use /btw for quick questions that don't need to stay in context

Course-Correct Early​

The best results come from tight feedback loops:

  • Press Esc to stop Claude mid-action and redirect
  • Press Esc twice or use /rewind to restore previous state
  • Say "undo that" to revert changes
  • After two failed corrections, start fresh with /clear and a better prompt

Pricing and Access​

Claude Code requires one of the following:

  • Claude Pro ($20/month): Includes Claude Code access
  • Claude Max ($100-200/month): Higher usage limits for heavy development
  • Claude Teams/Enterprise: Team and organization features
  • Anthropic Console: Pay-per-use API billing

You can also use Claude Code with third-party providers like Amazon Bedrock, Google Vertex AI, or Microsoft Foundry (formerly Azure AI Foundry).

Where Claude Code Runs​

EnvironmentBest for
Terminal CLIFull-featured command-line experience
VS Code extensionIDE integration with editor context
JetBrains pluginIntelliJ, PyCharm, WebStorm integration
Desktop appGraphical interface without terminal
Browser (claude.ai)Cloud-based sessions from any device

Your CLAUDE.md files, settings, and MCP servers work across all environments.

Key Takeaways​

  • Claude Code is an autonomous coding agent, not just a chatbot. It reads your codebase, edits files, runs commands, and completes multi-step tasks.
  • Start with read-only exploration prompts to build trust and understanding before making changes.
  • Use CLAUDE.md to give Claude persistent project context that survives across sessions.
  • Plan Mode, subagents, and skills let you scale from simple questions to complex, multi-step development workflows.
  • Always give Claude a way to verify its work: tests, expected outputs, or screenshots.

Next Steps​

SponsoredYour product hereReach people actively learning AI — flat monthly rates, no ad networks