Anthropic leaks part of Claude Code's internal source code

Anthropic has confirmed a significant leak involving part of the internal source code for its widely-used AI coding assistant, Claude Code. The company issued a statement on Tuesday clarifying that no sensitive customer information or credentials were compromised during this incident. According to an Anthropic spokesperson, the leak resulted from a packaging error attributed to human oversight, not a security breach. The company is now implementing measures to ensure such an incident does not occur again in the future. This leak poses a substantial challenge for Anthropic, as it may provide software developers and competitors with valuable insights into the workings of their popular coding tool. A post on X that included a link to the leaked source code has garnered over 21 million views since its release early Tuesday morning. This incident is not the first of its kind for Anthropic; it marks the second major data misstep in less than a week. A report from Fortune revealed that details about Anthropic's forthcoming AI model and other documents were inadvertently made accessible in a public data cache. Founded in 2021 by a group of former OpenAI executives and researchers, Anthropic has gained recognition for its Claude family of AI models. The Claude Code tool, which was launched to the public in May, assists software developers in creating features, troubleshooting bugs, and automating various tasks. Over the past year, the tool has experienced remarkable adoption, with its revenue run-rate exceeding $2.5 billion as of February. This success has spurred competitors, including OpenAI, Google, and xAI, to enhance their own AI offerings. As the landscape of AI development continues to evolve, this incident raises questions about data security and the measures companies must take to protect their innovations.