Claude Code Source Leaked: 512,000 Lines Hit npm, Open Clone Emerges
Anthropic's Claude Code CLI has been leaked in full. All 512,000 lines of source code appeared on npm, and it didn't take long before someone used OpenAI Codex to rewrite the entire codebase in Python, apparently to create a DMCA-proof open clone.
The leak is significant. Claude Code is Anthropic's command-line tool for agentic AI work, and the source code reveals architectural decisions, internal protocols, and the logic behind how the model handles autonomous tasks.
Anthropic has not directly commented on the leak. The company issued a DMCA notice against the original npm package, but the Python clone operates in a legal gray zone. The code has been rewritten line by line, and the question of who owns functionality versus specific implementation is a complex one.
For companies using Claude Code in production, there's nothing to act on. The service itself continues to function normally. But the leak highlights the security risks around internal development tools at AI companies and underlines how code published to package registries with broad distribution rights can end up in the wrong hands quickly.
It's also a reminder that open clones don't always hurt the original creator. Following the leak's spread, usage of the official Claude Code service reportedly increased markedly, according to unverified Reddit reports.
📬 Likte du denne?
AI-nyheter for ledere. Kuratert av en CIO som bygger det selv. Daglig i innboksen.