Domain 2 - Claude Code Configuration (20%)¶

Overview¶

Claude Code is Anthropic's official agentic coding tool, available as a CLI application and as IDE extensions for VS Code and JetBrains. This domain covers how to install, configure, and customize Claude Code for individual and team development workflows.

What is Claude Code?¶

Claude Code is an agentic coding assistant that lives in your terminal and IDE. It can: - Read, write, and edit files in your codebase - Run terminal commands - Search through code and documentation - Interact with external services via MCP servers - Follow project-specific instructions via CLAUDE.md files

Unlike simple code completion tools, Claude Code operates as an agent - it plans, executes, and iterates to complete complex development tasks.

Claude Code Overview - Introduction and capabilities

Installation and Setup¶

CLI Installation¶

Claude Code is installed via npm:

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

Requirements: - Node.js 18+ - Anthropic API key (or Claude Pro/Team/Enterprise subscription) - Supported OS: macOS, Linux, Windows (via WSL)

Authentication¶

Claude Code can authenticate via: - Anthropic API key (set via ANTHROPIC_API_KEY environment variable) - Claude Pro/Team/Enterprise subscription (OAuth flow) - API key configured in settings

First Run¶

Launch Claude Code in a project directory:

cd your-project
claude

On first run in a project, Claude Code will: - Read any CLAUDE.md files in the project - Detect the project structure and language - Present the interactive CLI interface

Getting Started with Claude Code - Setup instructions

CLAUDE.md Files (Memory)¶

CLAUDE.md files are the primary way to give Claude Code persistent instructions about your project. They function as project memory that persists across sessions.

Claude Code Memory - Complete CLAUDE.md documentation

File Hierarchy¶

CLAUDE.md files follow a hierarchy with increasing specificity:

Enterprise-level - Managed by organization admins (enterprise deployments)
User-level (~/.claude/CLAUDE.md) - Personal preferences that apply to all projects
Project-level (repo root CLAUDE.md) - Project-specific instructions, checked into version control
Directory-level (subdirectory CLAUDE.md) - Instructions specific to a subdirectory

All applicable CLAUDE.md files are combined and included in Claude's context. Lower-level files can add specificity but should not contradict higher-level instructions.

What to Include in CLAUDE.md¶

Project-level CLAUDE.md (most common): - Project overview and architecture - Coding standards and conventions - Build and test commands - Common patterns and anti-patterns - Technology stack details - File organization conventions - Important project-specific rules

User-level CLAUDE.md: - Preferred response language - Personal coding style preferences - Global rules that apply across all projects

Best Practices¶

Keep CLAUDE.md concise - everything in it uses context window tokens
Focus on actionable instructions, not documentation
Update CLAUDE.md as project conventions evolve
Use CLAUDE.md for things Claude consistently gets wrong without guidance
Project-level CLAUDE.md should be checked into version control for team sharing
Do not put secrets or sensitive data in CLAUDE.md

Example Project CLAUDE.md¶

# Project Instructions

## Build Commands
- Build: `npm run build`
- Test: `npm test`
- Lint: `npm run lint`

## Coding Standards
- Use TypeScript strict mode
- Prefer functional components with hooks
- Use Tailwind CSS for styling
- Write tests for all new features

## Architecture
- Frontend: Next.js 14 with App Router
- Backend: Express.js API in /api directory
- Database: PostgreSQL with Prisma ORM

Settings and Permissions¶

Claude Code Settings - Configuration reference

Settings Files¶

Settings are configured in JSON files: - User settings: ~/.claude/settings.json - Personal settings - Project settings: .claude/settings.json - Project-specific settings (checked into repo) - Enterprise settings: Managed centrally by administrators

Permissions Model¶

Claude Code has a permissions system that controls what actions require user approval:

Allowed tools - Tools that can run without asking for permission
Denied tools - Tools that are blocked entirely
Default behavior - Unspecified tools prompt for approval

Common permission configurations: - Allow read operations without approval (file reads, searches) - Require approval for write operations (file edits, command execution) - Block dangerous commands (rm -rf, git push --force)

Key Settings¶

permissions - Tool-level permission rules
mcpServers - MCP server configurations
env - Environment variables to set
apiKey - API key configuration (prefer environment variables)

Hooks System¶

Hooks allow you to run custom scripts before or after Claude Code executes specific tools. They are configured in settings files.

Claude Code Hooks - Hooks documentation

Hook Types¶

PreToolUse - Runs before a tool is executed. Can block the tool execution.
PostToolUse - Runs after a tool completes. Can process or validate results.
Notification - Runs when Claude Code sends a notification

Hook Configuration¶

Hooks are defined in settings.json:

{
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "write_file",
        "command": "npm run lint --fix ${file}"
      }
    ],
    "PostToolUse": [
      {
        "matcher": "write_file",
        "command": "npm run format ${file}"
      }
    ]
  }
}

Use Cases¶

Pre-tool: Linting - Run linter before file writes to ensure code quality
Pre-tool: Validation - Check that proposed changes meet project standards
Pre-tool: Security scanning - Scan for secrets or vulnerabilities before committing
Post-tool: Formatting - Auto-format files after Claude writes them
Post-tool: Testing - Run relevant tests after code changes
Post-tool: Notifications - Alert the team when significant changes are made

Hook Behavior¶

Pre-tool hooks that exit with a non-zero status block the tool execution
Hook output is captured and can be shown to the user
Hooks run synchronously (Claude Code waits for them to complete)
Hooks have access to tool-specific variables (file paths, command text, etc.)

Custom Slash Commands¶

Slash commands are shortcuts for common workflows. Claude Code includes built-in commands and supports custom user-defined commands.

Claude Code Slash Commands - Commands reference

Built-in Commands¶

/help - Show available commands
/clear - Clear conversation history
/compact - Summarize and compress the conversation
/init - Initialize CLAUDE.md for the project
Other built-in commands per documentation

Custom Commands¶

Custom commands are markdown files stored in the .claude/commands/ directory:

.claude/
  commands/
    deploy.md        # Project command: /deploy
    review.md        # Project command: /review

User-level commands are stored in ~/.claude/commands/:

~/.claude/
  commands/
    my-template.md   # User command: /my-template

Command File Format¶

Command files are markdown files whose content becomes the prompt when invoked:

Review the staged git changes and provide:
1. A summary of what changed
2. Potential issues or bugs
3. Suggestions for improvement

Focus on code quality, security, and performance.

Best Practices¶

Keep command prompts focused on a single workflow
Project commands (.claude/commands/) should be checked into version control
Use descriptive filenames that indicate what the command does
Commands can reference project conventions from CLAUDE.md

IDE Integrations¶

Claude Code IDE Integration - IDE setup guide

VS Code Extension¶

Install from the VS Code marketplace
Provides Claude Code panel within VS Code
Integrates with editor context (selected code, open files)
Uses the same CLAUDE.md files and settings as the CLI

JetBrains Plugin¶

Install from the JetBrains marketplace
Available for IntelliJ IDEA, WebStorm, PyCharm, and other JetBrains IDEs
Similar functionality to the VS Code extension
Shares configuration with the CLI

CLI vs IDE¶

Feature	CLI	IDE Extension
Terminal access	Full	Limited
File context	Manual	Automatic (open files, selection)
Visual diff	Text-based	IDE diff viewer
MCP servers	Configured in settings	Same settings
CLAUDE.md	Same files	Same files

MCP Server Configuration in Claude Code¶

Claude Code can connect to MCP servers for additional capabilities.

Claude Code MCP Servers - MCP configuration in Claude Code

Configuration Methods¶

MCP servers are configured via:

Project .mcp.json - Checked into version control for team sharing
User settings - In ~/.claude/settings.json for personal MCP servers
CLI flag - Pass MCP server configuration at launch

.mcp.json Format¶

{
  "mcpServers": {
    "database": {
      "command": "node",
      "args": ["./mcp-servers/database-server.js"],
      "env": {
        "DB_CONNECTION": "postgresql://localhost:5432/mydb"
      }
    },
    "remote-api": {
      "url": "https://internal-api.company.com/mcp",
      "headers": {
        "Authorization": "Bearer ${API_TOKEN}"
      }
    }
  }
}

Transport Types in Claude Code¶

stdio - For local MCP servers. Claude Code spawns the process and communicates via stdin/stdout.
Streamable HTTP / SSE - For remote MCP servers. Claude Code connects via HTTP.

Best Practices for Team Configuration¶

Claude Code Best Practices - Official recommendations

Version control CLAUDE.md - Check project-level CLAUDE.md into the repo so all team members share the same instructions
Version control .claude/commands/ - Share custom slash commands across the team
Version control .mcp.json - Share MCP server configurations
Use .claude/settings.json for project permissions - Standardize which operations require approval
Keep personal preferences in user-level files - Individual developers customize via ~/.claude/CLAUDE.md and ~/.claude/settings.json
Document hooks - If hooks are configured, document their purpose and behavior for the team
Review CLAUDE.md regularly - Keep instructions current as the project evolves

Key Exam Concepts¶

Know the CLAUDE.md file hierarchy and precedence order
Understand the difference between settings.json and CLAUDE.md
Know how to configure hooks (PreToolUse, PostToolUse)
Understand the permissions model and how to configure it
Know how to create and use custom slash commands
Understand MCP server configuration in Claude Code
Know the difference between project-level and user-level configuration
Understand IDE integration capabilities and limitations

Claude Code Overview - Introduction
Claude Code Memory - CLAUDE.md files
Claude Code Settings - Configuration
Claude Code Hooks - Hook system
Claude Code Slash Commands - Custom commands
Claude Code MCP Servers - MCP integration
Claude Code Best Practices - Recommendations
Claude Code IDE Integration - VS Code and JetBrains

Domain 2 - Claude Code Configuration (20%)¶

Overview¶

What is Claude Code?¶

Installation and Setup¶

CLI Installation¶

Authentication¶

First Run¶

CLAUDE.md Files (Memory)¶

File Hierarchy¶

What to Include in CLAUDE.md¶

Best Practices¶

Example Project CLAUDE.md¶

Settings and Permissions¶

Settings Files¶

Permissions Model¶

Key Settings¶

Hooks System¶

Hook Types¶

Hook Configuration¶

Use Cases¶

Hook Behavior¶

Custom Slash Commands¶

Built-in Commands¶

Custom Commands¶

Command File Format¶

Best Practices¶

IDE Integrations¶

VS Code Extension¶

JetBrains Plugin¶

CLI vs IDE¶

MCP Server Configuration in Claude Code¶

Configuration Methods¶

.mcp.json Format¶

Transport Types in Claude Code¶

Best Practices for Team Configuration¶

Key Exam Concepts¶

Related Documentation¶