Master Claude Code: The Ultimate CLI AI Guide for Seamless Development

Claude Code Overview

Claude Code is not a traditional AI IDE but a command‑line AI assistant that can help you debug, generate code, and interact with external services. It offers native Claude 4 Sonnet/Opus models, unlimited tool calls, large context windows (200K+ tokens), and direct system‑log access for precise debugging.

1. Installation Methods

1.1 Mirror Site Installation

Invitation link : https://aicodewith.com/?invitation=EK1S5F macOS :

curl -fsSL https://aicodewith.com/claudecode/resources/install-script-macos | bash

Linux :

curl -fsSL https://aicodewith.com/claudecode/resources/install-script-linux | bash

Windows (WSL) : Install WSL2 ( wsl --install), install Ubuntu from Microsoft Store, then run

curl -fsSL https://aicodewith.com/claudecode/resources/install-script-wsl | bash

1.2 Official Installation

NPM global install : npm install -g @anthropic-ai/claude-code Verify version :

claude --version

2. Basic Usage

2.1 First‑time Configuration

API key (required for official install): export ANTHROPIC_API_KEY="sk-your-key-here" (add to ~/.bashrc, ~/.zshrc, or ~/.config/fish/config.fish as appropriate)

Basic settings : claude config set -g model claude-sonnet-4, claude config set -g verbose true, claude config set -g outputFormat text Test installation : claude "Hello, Claude!" or

claude /doctor

2.2 Core Commands

Start session : claude Fix bug : claude "帮我修复这个 bug" One‑off print : claude -p "" Read large file : cat file | claude -p "" Update : claude update Start MCP :

claude mcp

Slash Commands (quick actions)

/help

: list all commands /add-dir: add more working directories /bug: report an error /clear: clear chat history /compact: compress context /config: open configuration menu /cost: show token usage /doctor: run integrity check /init: initialize project and generate

CLAUDE.md

/mcp

: list MCP services and status /memory: edit memory /model: switch model /permissions: modify tool permissions /review: request code review /sessions: list sessions /status: show system/account status /terminal-setup: bind Shift+Enter /vim: toggle vim mode

3. MCP Integration (External Services & Databases)

MCP enables Claude to connect to external services, databases, and APIs directly from the CLI.

3.1 Basic MCP Commands

claude mcp list

: list configured services claude mcp add: add a new service claude mcp remove: remove a service

Configuration file:

~/.claude.json

3.2 Common MCP Services

Git MCP : npm install -g git-mcp-server then claude mcp add git "git-mcp-server" or

claude mcp add github "github-mcp-server --token $GITHUB_TOKEN"

Database MCP : install server packages ( npm install -g postgres-mcp-server, npm install -g mysql-mcp-server, npm install -g sqlite-mcp-server) and configure with environment variables, e.g.,

export POSTGRES_URL="postgresql://user:password@localhost:5432/mydb"

then

claude mcp add postgres "postgres-mcp-server --url $POSTGRES_URL"

3.3 MCP Tool Permissions

Specify allowed tools: claude --allowedTools "Edit,View" Keyword‑based permissions: claude --allowedTools "Bash(git:status)" Combine built‑in and MCP tools:

claude --allowedTools "Edit,View,Bash(git:*)"

4. Configuration System

4.1 Configuration Files

Global settings live in ~/.claude.json. Project‑specific settings can be placed in settings.json (or a similar file) in the project root.

{
  "model": "claude-sonnet-4",
  "verbose": true,
  "outputFormat": "text",
  "allowedTools": ["Edit", "View"]
}

4.2 Environment Variables

DISABLE_NON_ESSENTIAL_MODEL_CALLS=1

: skip automatic summarisation and git diff to save tokens MAX_THINKING_TOKENS: limit maximum thinking token usage (≈30‑40k) DISABLE_TELEMETRY=1: disable telemetry and error reporting CLAUDE_CODE_USE_BEDROCK / CLAUDE_CODE_USE_VERTEX: enable AWS Bedrock or Google Vertex AI verification HTTP_PROXY, HTTPS_PROXY, NO_PROXY: configure HTTP/HTTPS proxy settings

5. Security and Permission Management

Tools must be authorised before use, permissions are remembered per session, and risky actions require confirmation.

Level

Description

Risk

Interactive

Ask before each operation

Low

Allowlist

Only specified tools are usable

Medium

Dangerous

All tools are unrestricted

High

5.1 Recommended Strategies

Limit specific tool permissions, e.g., claude --allowedTools "Edit,View,Bash(git:status)" instead of granting all Bash access.

Protect sensitive data by using environment variables rather than hard‑coding credentials.

Regularly audit permission groups with claude config get allowedTools, claude config get disallowedTools, and claude config list.

6. Thinking Mode

Control Claude’s depth of reasoning with keywords in the prompt.

Depth 1: think Depth 2:

think about it, think a lot, think deeply, think hard, think more, megathink

Depth 3: think harder Example:

claude -p "We have a tricky concurrency bug. ultrathink and propose a fix."

7. Config Command Reference

Command

Function

Example

claude config list

Show all settings

claude config list

claude config get

Retrieve a specific setting

claude config get theme

claude config set -g

Set a global value

claude config set -g theme dark

claude config add -g

Add an element to an array

claude config add -g env CLAUDE_CODE_ENABLE_TELEMETRY=1

claude config remove -g

Remove an element from an array

claude config remove -g env CLAUDE_CODE_ENABLE_TELEMETRY

8. CI/CD Integration (GitHub Actions)

name: Claude Code Review
on:
  pull_request:
    branches: [main, develop]
jobs:
  claude-review:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Setup Node.js 18
        uses: actions/setup-node@v4
        with:
          node-version: '18'
      - name: Install Claude Code
        run: npm install -g @anthropic-ai/claude-code
      - name: Review PR
        env:
          ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
        run: |
          claude -p "Review changes for security issues and bugs" \
            --allowedTools "View" \
            --output-format json > review-results.json
      - name: Upload results artifact
        uses: actions/upload-artifact@v4
        with:
          name: claude-review
          path: review-results.json

Key points: checkout PR diff, install official CLI, use read‑only View permission, output JSON for structured analysis.

8.1 Local Git Pre‑Commit Hook

#!/usr/bin/env bash
# .git/hooks/pre-commit (chmod +x)
staged=$(git diff --cached --name-only --diff-filter=ACM)
[ -z "$staged" ] && exit 0
payload=$(echo "$staged" | xargs cat)
analysis=$(echo "$payload" | \
  claude -p "Review these changes for issues before commit" \
    --allowedTools "View" \
    --output-format json)
if echo "$analysis" | jq -e '.critical_issues[]' >/dev/null 2>&1; then
  echo "❌ Critical issues found – commit blocked"
  exit 1
fi
echo "✅ Claude analysis passed"

9. Advanced Features

9.1 Persistent Memory

Run claude /init to generate a CLAUDE.md file that stores project context for future sessions.

Project: My Application
Overview
This is a React/Node.js application with PostgreSQL database.
Architecture
Frontend: React 18 with TypeScript
Backend: Node.js with Express
Database: PostgreSQL 14
Current Goals
- Implement authentication
- Add API documentation
- Set up CI/CD pipeline
Development Guidelines
- Use TypeScript for new code
- Follow ESLint configuration
- Write tests for new features

9.2 Memory Commands

claude /memory

: edit memory claude /memory view: view current memory

9.3 Multi‑Directory Workspaces

Add directories: claude --add-dir ../frontend ../backend ../shared Run a global analysis:

claude "analyze the entire application architecture"

Conclusion

This guide provides a step‑by‑step, exhaustive reference for installing, configuring, and extending Claude Code. By following the instructions you can integrate the tool into local development, CI/CD pipelines, and team workflows while maintaining security and control.