claude-leaked-files  by Ahmad-progr

Agentic CLI for software engineering

Created 1 week ago

New!

253 stars

Top 99.3% on SourcePulse

GitHubView on GitHub
Project Summary

Summary

This repository archives a mirrored snapshot of Claude Code's source, exposed via an npm package vulnerability on March 31, 2026. It serves educational purposes, defensive security research, and software supply-chain analysis, offering insights into agent-based developer tools and release process failures.

How It Works

Claude Code is a command-line interface for interacting with Claude for software engineering tasks. Its architecture features a modular Tool System (~40 tools for shell commands, file operations, code search, agent spawning), a Command System (~50 slash commands for user functions like /commit and /review), a Service Layer for integrations, and a Bridge System for IDE communication. It leverages TypeScript, the Bun runtime, React with Ink for its terminal UI, and Zod for data validation. Key design principles include parallel prefetching for faster startup and lazy loading of non-essential dependencies.

Quick Start & Requirements

This repository contains a source code snapshot for research and does not provide direct installation or execution instructions for the Claude Code CLI. The original project utilizes Bun as its runtime and TypeScript for development. Specific hardware, OS, or software prerequisites for running the original Claude Code are not detailed within this snapshot's README.

Highlighted Details

  • Extensive Tooling: A suite of ~40 tools covers diverse software engineering tasks, including shell commands, file operations, code search, agent coordination, and IDE integration via LSP.
  • Rich Command Set: Approximately 50 user-facing slash commands facilitate actions like Git commits, code reviews, session management, and task tracking.
  • IDE Integration: The Bridge System enables two-way communication with popular IDEs such as VS Code and JetBrains.
  • Advanced Architecture: Features parallel prefetching for improved startup, lazy loading of modules, agent swarms for parallel execution, and a plugin architecture.

Maintenance & Community

This research archive is maintained by a university student focusing on software supply-chain security and agent-based systems. It is explicitly not an official Anthropic repository. No community channels (e.g., Discord, Slack) or roadmap links are provided for this research snapshot.

Licensing & Compatibility

The README does not specify a license for the mirrored source code. It emphasizes that the original Claude Code source remains the property of Anthropic. Use of this snapshot is subject to the original code owner's intellectual property rights; commercial use or closed-source linking compatibility is not addressed and likely restricted.

Limitations & Caveats

This is a snapshot of leaked source code intended solely for educational and research purposes. It lacks setup instructions for direct use and carries potential legal and ethical considerations due to its origin. The exposure method highlights significant software supply-chain security risks.

Health Check
Last Commit

1 week ago

Responsiveness

Inactive

Pull Requests (30d)
1
Issues (30d)
0
Star History
253 stars in the last 11 days

Explore Similar Projects

Starred by Vincent Weisser Vincent Weisser(Cofounder of Prime Intellect) and Chip Huyen Chip Huyen(Author of "AI Engineering", "Designing Machine Learning Systems").

ii-agent by Intelligent-Internet

1.1%
3k
Agentic framework for building and deploying intelligent agents
Created 1 year ago
Updated 1 week ago
Feedback? Help us improve.