Coders are having a field day with the Claude Code leak, which has made over 512,000 lines of source code available for public viewing. The leak was first posted on X by Chaofan Shou and quickly spread to GitHub, where it has spawned thousands of copies.
Sigrid Jin, creator of the Python edition, Claw Code, said it’s been a “workflow revelation” that allowed him to recreate the AI agent’s source code in just two hours. The leak has sparked joy among coders, who can now access features like spinner verbs and coding pets.
However, Anthropic, the company behind Claude Code, says the leak was due to human error, not a security breach. They’ve filed copyright takedown requests, but some websites are still offering access to the leaked code.
Industry experts say that the leak may have raised concerns about security and data protection. Delip Rao, an AI researcher at the University of Pennsylvania, thinks that an AI agent might be involved in the mistake, given its complexity and speed.
Cherny, Claude Code’s creator, has defended the company’s approach to AI development, saying it’s all about experimentation and improvement. He acknowledges that 90% of tests don’t ship, but assures users that their data is safe.
In a surprising twist, Anthropic’s competitors are pouring gas on the flame, with xAI sending credits to Jin for his work on Claw Code. Meanwhile, techies are spreading the leaked code in private Discord servers, analyzing its contents and discussing its implications.
While some have raised concerns about security, others see the leak as an opportunity to learn from the company’s approach to AI development. As one Reddit user noted, “it’s like we’re getting a glimpse into their inner workings.”
Source: https://www.businessinsider.com/claude-code-leak-what-happened-recreated-python-features-revealed-2026-4