What the Claude Code Leak Reveals: 510k Lines of AI Engine Exposed

A massive leak of Anthropic’s Claude Code exposed over 1,900 files and 510,000 lines of TypeScript, revealing its React‑Ink UI, Bun runtime, extensive toolset, hidden Kairos mode, electronic pet system, and covert Undercover features, sparking worldwide developer frenzy and security concerns.

Java Architect Essentials
Java Architect Essentials
Java Architect Essentials
What the Claude Code Leak Reveals: 510k Lines of AI Engine Exposed

Leak Overview

In a recent npm publishing error, the source map file cli.js.map (≈59.8 MB) was included in the public package @anthropic-ai/claude-code. This exposed the complete Claude Code CLI source code.

Repository: https://github.com/instructkr/claude-code Size: more than 1,900 files, ~510,200 lines of TypeScript (≈51 MB).

Technology Stack

Runtime: Bun JavaScript runtime.

UI: React + Ink for terminal interface.

Package: Published to npm under @anthropic-ai/claude-code.

Core Components

Tools – a toolbox of more than 40 independent modules providing file I/O, Bash command execution, built‑in LSP integration and sub‑agent generation.

QueryEngine.ts – a ~46 k‑line engine that implements reasoning logic, token counting and complex “thought‑chain” loops.

Coordinator & Bridge – modules named coordinator and bridge that enable multi‑agent coordination and IDE integration (VS Code / JetBrains).

Hidden / Undocumented Features

Kairos mode – an unreleased mode identified by the codename “Kairos”.

Buddy System – an embedded electronic‑pet subsystem with 18 species, rarity tiers, shiny variants and detailed stats.

Undercover Mode – automatically erases AI‑related traces from commit history when Anthropic staff work in public repositories; cannot be disabled manually.

Coordinator Mode – allows Claude to schedule parallel subordinate agents for higher throughput.

Auto Mode – an AI classifier that automatically approves tool permissions, removing manual prompt confirmations.

Security Implications

The source map exposed internal logic, engineering decisions and proprietary mechanisms of Anthropic’s AI‑engineering platform, giving competitors detailed insight into the implementation of the Claude Code assistant.

Reproduction Steps

Clone the repository: git clone https://github.com/instructkr/claude-code.git Inspect the cli.js.map file to view the original TypeScript source mapping.

Run the CLI with Bun: bun run src/cli.ts (or the compiled entry point).

Architecture Diagram

Claude Code architecture diagram
Claude Code architecture diagram
AI engineeringClaudehidden featuressource code leakBun RuntimeToolboxReact Ink
Java Architect Essentials
Written by

Java Architect Essentials

Committed to sharing quality articles and tutorials to help Java programmers progress from junior to mid-level to senior architect. We curate high-quality learning resources, interview questions, videos, and projects from across the internet to help you systematically improve your Java architecture skills. Follow and reply '1024' to get Java programming resources. Learn together, grow together.

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.