Anthropic Accidentally Exposes Source Code for Claude Code
CNET • • business
Key Points:
- Anthropic accidentally leaked the source code for its AI coding tool Claude Code by publishing a version containing a source map file on the public npm registry, exposing over 500,000 lines of code and nearly 2,000 files.
- The leak was widely viewed after security researcher Chaofan Shou shared a link to the archive on X, garnering more than 26 million views.
- Claude AI is a multifunctional tool used for answering questions, generating creative content, translating languages, transcribing images, writing code, and more.
- Anthropic confirmed the leak resulted from human error and stated no sensitive customer data or credentials were exposed, with measures underway to prevent future incidents.
- The exposed code offers developers and competitors valuable insight into Claude Code’s development roadmap, potentially impacting Anthropic's competitive position.
![Pixel 11 Pro is thinner, but mostly the same in first leak [Gallery]](https://media-cdn.cashwalklabs.io/https%3A%2F%2Fi0.wp.com%2F9to5google.com%2Fwp-content%2Fuploads%2Fsites%2F4%2F2026%2F03%2Fgoogle-pixel-11-pro-onl-2.webp%3Fresize%3D1200%252C628%26quality%3D82%26strip%3Dall%26ssl%3D1?format=webp&w=120&h=80){kind=tracking}
