Skip to content

Claude Code

📖 5 min read claudeclaude-codecodingagenticmcp
Complete guide to Claude Code — Anthropic's agentic coding tool across Terminal, VS Code, JetBrains, Desktop, Web, and iOS. Setup, workflows, MCP integration, skills, hooks, sub-agents, and comparison with Cursor/Copilot.
Key Takeaways
  • Available on 7 surfaces: Terminal, VS Code, JetBrains, Desktop app, Web browser, iOS, and CI/CD (GitHub Actions)
  • Reads your entire codebase, edits files, runs commands, creates commits and PRs
  • Integrates with MCP (external tools), Agent Skills (repeatable workflows), hooks (auto-format/lint), and sub-agents (parallel task execution)
  • Routines: scheduled tasks on Anthropic infra. Channels: webhooks from Telegram, Discord, Slack. Agent SDK: build custom agents

Claude Code is Anthropic’s agentic coding tool. It reads your codebase, edits files across multiple files, runs terminal commands, and integrates with your development workflow — from Terminal to CI/CD.

Installation

Terminal window
# macOS, Linux, WSL
curl -fsSL https://claude.ai/install.sh | bash
# Windows
irm https://claude.ai/install.ps1 | iex
# Homebrew
brew install --cask claude-code
# WinGet
winget install Anthropic.ClaudeCode

VS Code

Install “Claude Code” from the Extensions marketplace (Cmd+Shift+X). Provides inline diffs, @-mentions, plan review.

JetBrains

Install the Claude Code plugin for IntelliJ, PyCharm, WebStorm.

Desktop App

Download from claude.ai — visual diffs, side-by-side sessions, scheduled tasks.

Web

Go to claude.ai/code — no setup, browser-based, works on iOS.

CI/CD

# GitHub Actions
- uses: anthropics/claude-code-action@v1
with:
prompt: "Review this PR for security issues"

Also available for GitLab CI/CD.

Seven Surfaces — One Engine

SurfaceBest ForKey Feature
Terminal CLIFull control, scripting, CI/CDclaude -p "prompt" for quick tasks
VS CodeIDE integration, inline diffs@-mentions, plan mode, context sharing
JetBrainsIntelliJ, PyCharm, WebStormInteractive diff viewing
Desktop AppVisual workflow, multi-sessionSide-by-side sessions, recurring tasks
WebNo setup, browser/phoneRemote execution, long-running cloud tasks
iOSOn-the-go codingDispatch tasks, review results
CI/CDAutomated PR review, issue triageGitHub Actions, GitLab CI/CD

Core Capabilities

Read, Write, Execute

Terminal window
# Explore and understand
claude "explain the authentication flow in this codebase"
# Build features
claude "add a rate-limiting middleware to the API"
# Fix bugs
claude "fix the race condition in the payment processing module"
# Write tests
claude "write tests for the auth module, run them, and fix any failures"
# Git operations
claude "commit my changes with a descriptive message and open a PR"

MCP Integration

Connect Claude Code to external tools and data:

Terminal window
# Configure MCP servers in .claude/mcp.json
{
"mcpServers": {
"filesystem": {
"command": "npx", "args": ["-y", "@modelcontextprotocol/server-filesystem"]
},
"github": {
"command": "npx", "args": ["-y", "@modelcontextprotocol/server-github"]
}
}
}

See MCP — Model Context Protocol for a complete guide.

Agent Skills (SKILL.md)

Package repeatable workflows your team can share:

Terminal window
# Project skills in .claude/skills/
.claude/skills/
generate-migration/
SKILL.md # Instructions for the skill
review-pr/
SKILL.md
deploy-staging/
SKILL.md

See Agent Skills for details.

CLAUDE.md — Persistent Instructions

Add CLAUDE.md to your project root:

CLAUDE.md
## Code Style
- Use TypeScript strict mode
- Prefer functional patterns over classes
- Tests go in __tests__/ colocated with source
## Architecture
- API routes in src/api/
- Business logic in src/services/

Claude Code reads this at the start of every session. It also builds auto-memory — learning build commands, debugging insights, and project conventions as it works.

Hooks

Run shell commands before or after Claude Code actions:

.claude/settings.json
{
"hooks": {
"PostToolUse": [
{
"matcher": "Edit|Write",
"hooks": [{"type": "command", "command": "prettier --write ${CLAUDE_TOOL_INPUT_FILE_PATH}"}]
}
],
"PreToolUse": [
{
"matcher": "Bash(git commit*)",
"hooks": [{"type": "command", "command": "npm run lint"}]
}
]
}
}

Sub-Agents — Parallel Task Execution

Spawn multiple Claude Code agents working simultaneously:

Terminal window
claude "audit all services for security issues, checking each in parallel"

A lead agent coordinates, assigns subtasks to sub-agents, and merges results.

Routines & Scheduling

Run Claude on a schedule — even when your computer is off:

  • Routines: Runs on Anthropic infrastructure. Trigger by schedule, API call, or GitHub events.
  • Desktop scheduled tasks: Run on your machine with local file access.
  • /loop: Repeat a prompt within a CLI session for quick polling.

Remote Control & Channels

  • Remote Control: Continue sessions from your phone or any browser
  • Channels: Push events from Telegram, Discord, iMessage, or custom webhooks into sessions
  • Dispatch: Send tasks from your phone, execute on your desktop

Claude Code vs Cursor vs Copilot

FeatureClaude CodeCursorGitHub Copilot
Primary interfaceTerminal CLIIDE fork (VSCode-based)IDE extension
ModelClaude Opus 4.8 / Sonnet 4.6Claude Sonnet + GPT-4oGPT-4o + Codex
Codebase awareness1M context, reads all files.cursorrules contextRepository context
Git integrationNative: commits, PRs, branchesBuilt-in source controlCopilot Chat in PRs
Multi-file editsFull project scopeMulti-file via inlineContext-limit aware
Terminal commandsFull shell accessLimited terminal paneNo terminal
PricingPay-per-token (API) or Pro plan$20-40/mo subscriptionFree tier + $10-20/mo
Custom workflowsSkills + Hooks + CLAUDE.md.cursorrulesCopilot instructions
CI/CDGitHub Actions, GitLab CI/CDNoneGitHub-native
Sub-agentsYes (parallel tasks)NoNo
Desktop appYesNoNo

Best combo: Cursor for daily IDE work, Claude Code for complex refactors, Copilot for inline completions. See the Coding Tools Comparison for benchmarks.

Pricing

PlanWhat You GetCost
Claude ProClaude Code access on 1 surface$20/mo
Claude MaxClaude Code across all surfaces + extended usage100/mo(individual)or100/mo (individual) or 200/mo (team)
API (Console)Pay-per-token via Anthropic ConsoleOpus 5/5/25, Sonnet 3/3/15, Haiku 1/1/5 per 1M
EnterpriseCustom plans, volume discountsContact sales