Anthropic accidentally leaked internal source code of its Claude Code AI tool during a software update. The exposed files revealed architecture details, unreleased features, and system functionality. The leak occurred after a debugging file was uploaded to a public registry. Anthropic confirmed no customer data or credentials were compromised. Developers quickly downloaded and analyzed the code before removal. The incident raised concerns about operational security and competitive risks, as the exposed information could help rivals understand and replicate features from Anthropic’s AI coding platform.