Anthropic Accidentally Nuked 8,100 GitHub Repos Trying to Erase Its Own Leaked Claude Code

The Claude Code source leak (512K lines via npm error) spawned an instant follow-on crisis. Anthropic filed a DMCA takedown against GitHub to pull copies of the leaked code — but the notice was executed so broadly it swept up ~8,100 repositories, including authorized forks of Anthropic's own public Claude Code repo.

What happened

  • The DMCA notice targeted one infringing repo — but because it sat in the same fork network as Anthropic's official public repo, GitHub's automated system yanked 8,100 connected forks
  • Anthropic's head of Claude Code Boris Cherny called the mass-takedown accidental and retracted it within hours, limiting the notice to one repo and 96 forks
  • Thousands of developers lost access to legitimate repos and erupted on X and Reddit
  • TechCrunch: "another black eye for the company as it reportedly plans an IPO... Leaking your source code as a public company? You better believe there's a shareholder lawsuit coming"

The twist: OpenClaude

Meanwhile, OpenClaude — a fork of the leaked Claude Code source with an OpenAI-compatible shim — launched on GitHub ~20 hours ago. It runs GPT-4o, DeepSeek, Gemini, Llama, and Mistral via the same agentic Claude Code architecture. The leak is now the foundation for a competing open-source implementation — with any model.

🔗 Read the full story → TechCrunch


Stay ahead of AI. Subscribe to LGTM for daily AI model news.