Anthropic, a company known for its cautious approach to AI, has made another embarrassing mistake. On Tuesday, they accidentally leaked nearly 2,000 source code files and over 512,000 lines of code from their Claude Code software package. This is the second time in a week that Anthropic has made sensitive information public.
The issue was caused by human error during the release packaging process. The company claims it’s not a security breach, but rather an error on their part. However, developers quickly began analyzing and publishing their findings, describing Claude Code as “a production-grade developer experience”.
This incident raises concerns about Anthropic’s ability to handle sensitive information. With competitors like OpenAI already struggling with similar issues, the question remains whether this will have a lasting impact on the company.
The leak highlights the importance of attention to detail in software development. While the architecture of Claude Code may be instructive for developers, it also raises questions about Anthropic’s responsibility and preparedness for sensitive data breaches.
Source: https://techcrunch.com/2026/03/31/anthropic-is-having-a-month