• How It Leaked
• What Is Claude Code?
• Documentation
• Explore with MCP Server
• Directory Structure
• Architecture
  • Tool System
  • Command System
  • Service Layer
  • Bridge System
  • Permission System
  • Feature Flags
• Key Files
• Tech Stack
• Design Patterns
• GitPretty Setup
• Contributing
• Disclaimer

Chaofan Shou (@Fried_rice) discovered that the published npm package for Claude Code included a .map file referencing the full, unobfuscated TypeScript source — downloadable as a zip from Anthropic's R2 storage bucket.

"Claude code source code has been leaked via a map file in their npm registry!"

— @Fried_rice, March 31, 2026

Claude Code is Anthropic's official CLI tool for interacting with Claude directly from the terminal — editing files, running commands, searching codebases, managing git workflows, and more. This repository contains the leaked src/ directory.

For in-depth guides, see the docs/ directory:

Also see: CONTRIBUTING.md · MCP Server README

This repo ships an MCP server that lets any MCP-compatible client (Claude Code, Claude Desktop, VS Code Copilot, Cursor) explore the full source interactively.

The MCP server is published as warrioraashuu-codemaster on npm — no need to clone the repo:

# Claude Code
claude mcp add warrioraashuu-codemaster -- npx -y warrioraashuu-codemaster

git clone https://github.com/codeaashu/claude-code.git ~/claude-code \
  && cd ~/claude-code/mcp-server \
  && npm install && npm run build \
  && claude mcp add claude-code-explorer -- node ~/claude-code/mcp-server/dist/index.js

# 1. Clone the repo
git clone https://github.com/codeaashu/claude-code.git
cd claude-code/mcp-server

# 2. Install & build
npm install && npm run build

# 3. Register with Claude Code
claude mcp add claude-code-explorer -- node /absolute/path/to/claude-code/mcp-server/dist/index.js

Replace /absolute/path/to/claude-code with your actual clone path.

VS Code — add to .vscode/mcp.json:

{
  "servers": {
    "claude-code-explorer": {
      "type": "stdio",
      "command": "node",
      "args": ["${workspaceFolder}/mcp-server/dist/index.js"],
      "env": { "CLAUDE_CODE_SRC_ROOT": "${workspaceFolder}/src" }
    }
  }
}

Claude Desktop — add to your config file:

{
  "mcpServers": {
    "claude-code-explorer": {
      "command": "node",
      "args": ["/absolute/path/to/claude-code/mcp-server/dist/index.js"],
      "env": { "CLAUDE_CODE_SRC_ROOT": "/absolute/path/to/claude-code/src" }
    }
  }
}

Cursor — add to ~/.cursor/mcp.json (same format as Claude Desktop).

Try asking: "How does the BashTool work?" · "Search for where permissions are checked" · "Show me the /review command source"

# Custom source location
claude mcp add claude-code-explorer -e CLAUDE_CODE_SRC_ROOT=/path/to/src -- node /path/to/mcp-server/dist/index.js

# Remove
claude mcp remove claude-code-explorer

src/
├── main.tsx                 # Entrypoint — Commander.js CLI parser + React/Ink renderer
├── QueryEngine.ts           # Core LLM API caller (~46K lines)
├── Tool.ts                  # Tool type definitions (~29K lines)
├── commands.ts              # Command registry (~25K lines)
├── tools.ts                 # Tool registry
├── context.ts               # System/user context collection
├── cost-tracker.ts          # Token cost tracking
│
├── tools/                   # Agent tool implementations (~40)
├── commands/                # Slash command implementations (~50)
├── components/              # Ink UI components (~140)
├── services/                # External service integrations
├── hooks/                   # React hooks (incl. permission checks)
├── types/                   # TypeScript type definitions
├── utils/                   # Utility functions
├── screens/                 # Full-screen UIs (Doctor, REPL, Resume)
│
├── bridge/                  # IDE integration (VS Code, JetBrains)
├── coordinator/             # Multi-agent orchestration
├── plugins/                 # Plugin system
├── skills/                  # Skill system
├── server/                  # Server mode
├── remote/                  # Remote sessions
├── memdir/                  # Persistent memory directory
├── tasks/                   # Task management
├── state/                   # State management
│
├── voice/                   # Voice input
├── vim/                     # Vim mode
├── keybindings/             # Keybinding configuration
├── schemas/                 # Config schemas (Zod)
├── migrations/              # Config migrations
├── entrypoints/             # Initialization logic
├── query/                   # Query pipeline
├── ink/                     # Ink renderer wrapper
├── buddy/                   # Companion sprite (Easter egg 🐣)
├── native-ts/               # Native TypeScript utils
├── outputStyles/            # Output styling
└── upstreamproxy/           # Proxy configuration

src/tools/ — Every tool Claude can invoke is a self-contained module with its own input schema, permission model, and execution logic.

src/commands/ — User-facing slash commands invoked with / in the REPL.

src/services/ — External integrations and core infrastructure.

src/bridge/ — Bidirectional communication layer connecting IDE extensions (VS Code, JetBrains) with the CLI.

Key files: bridgeMain.ts (main loop) · bridgeMessaging.ts (protocol) · bridgePermissionCallbacks.ts (permission callbacks) · replBridge.ts (REPL session) · jwtUtils.ts (JWT auth) · sessionRunner.ts (session execution)

src/hooks/toolPermission/ — Checks permissions on every tool invocation.

Prompts the user for approval/denial or auto-resolves based on the configured permission mode: default, plan, bypassPermissions, auto, etc.

Dead code elimination at build time via Bun's bun:bundle:

import { feature } from 'bun:bundle'

const voiceCommand = feature('VOICE_MODE')
  ? require('./commands/voice/index.js').default
  : null

Notable flags: PROACTIVE · KAIROS · BRIDGE_MODE · DAEMON · VOICE_MODE · AGENT_TRIGGERS · MONITOR_TOOL

MDM settings, keychain reads, and API preconnect fire in parallel as side-effects before heavy module evaluation:

// main.tsx
startMdmRawRead()
startKeychainPrefetch()

OpenTelemetry (~400KB) and gRPC (~700KB) are loaded via dynamic import() only when needed.

Sub-agents spawn via AgentTool, with coordinator/ handling orchestration. TeamCreateTool enables team-level parallel work.

Defined in skills/ and executed through SkillTool. Users can add custom skills.

Built-in and third-party plugins loaded through the plugins/ subsystem.

# Apply emoji commits
bash ./gitpretty-apply.sh .

# Optional: install hooks for future commits
bash ./gitpretty-apply.sh . --hooks

# Push as usual
git push origin main

Contributions to documentation, the MCP server, and exploration tooling are welcome. See CONTRIBUTING.md for guidelines.

Note: The src/ directory is the original leaked source and should not be modified.

This repository archives source code leaked from Anthropic's npm registry on 2026-03-31. All original source code is the property of Anthropic. This is not an official release and is not licensed for redistribution. Contact aashuu ✦ for any comments.