MCPcopy Create free account
hub / github.com/Austin1serb/Anthropic-Leaked-Source-Code

github.com/Austin1serb/Anthropic-Leaked-Source-Code @main

Chat with this repo
repository ↗ · DeepWiki ↗ · + Follow
11,802 symbols 42,348 edges 1,902 files 1,595 documented · 14% updated 2mo ago★ 5311 open issues

Browse by type

Functions 11,462 Types & classes 340
What it actually does AI analysis from the code graph — generated when you open this
loading…
README

Claude Code Source Code Leak — Anthropic Code Leak Archive

This repository contains an unmodified archive of the leaked source code for Claude Code, Anthropic's AI-powered coding agent.

Also known as the Claude Code leak, Anthropic source code leak, or Claude Code source code leak, this archive is intended for research, education, and technical analysis.

Nothing has been added besides this README. Nothing has been removed or changed. This is the raw source as leaked.

What's included

  • Agent orchestration and query engine
  • Tool implementations for Bash, FileEdit, FileRead, FileWrite, Glob, Grep, WebSearch, WebFetch, Notebook, and related developer tools
  • API service layer and model provider integrations
  • MCP, or Model Context Protocol, server management
  • Permission system and sandboxing logic
  • CLI transport layers, including SSE, WebSocket, and hybrid transports
  • Prompt construction and system prompt logic

Claude Code leak overview

Claude Code is Anthropic's AI coding agent. This repository exists as a public archive for developers, researchers, and security-minded engineers who want to study how a modern AI coding agent is structured.

Common search terms for this topic include:

  • Claude Code GitHub
  • Claude Code leak
  • Claude Code source code
  • Anthropic code leak
  • Anthropic source code leak
  • Anthropic leaked source code
  • Claude Code source map
  • Claude Code repo

Why this exists

This repository is for educational and research purposes. It may be useful for studying:

  • AI coding agent architecture
  • Tool-calling systems
  • Prompt construction
  • Local development workflows
  • Sandboxing and permission models
  • MCP server integration
  • CLI-based AI agent design

Disclaimer

This repository is provided for educational and research purposes only.

No affiliation with Anthropic AI. Claude Code is a product of Anthropic AI.

Reverse Engineered by Serbyte Development

Reverse engineered using JavaScript source maps by Serbyte Web Design & Development, a software development firm creating custom software, AI automations, and open-source developer tools.

Extension points exported contracts — how you extend this code

browse all types & interfaces →

Core symbols most depended-on inside this repo

browse all functions →

Shape

Function 10,431
Method 1,031
Class 256
Interface 83
Enum 1

Languages

TypeScript100%

Modules by API surface

bootstrap/state.ts212 symbols
utils/sessionStorage.ts156 symbols
native-ts/yoga-layout/index.ts144 symbols
utils/messages.ts122 symbols
utils/Cursor.ts105 symbols
utils/bash/bashParser.ts84 symbols
utils/attachments.ts74 symbols
utils/auth.ts69 symbols
utils/hooks.ts67 symbols
services/mcp/client.ts67 symbols
utils/sandbox/sandbox-adapter.ts59 symbols
ink/ink.tsx54 symbols

For agents

$ claude mcp add Anthropic-Leaked-Source-Code \
  -- python -m otcore.mcp_server <graph>

⬇ download graph artifact

Ask about this repo answers extend the page