Anthropic意外泄露Claude Code源代码

Anthropic inadvertently exposed a substantial portion of the source code behind Claude Code after publishing internal files to the public npm registry, according to security researchers who analyzed the incident. The exposure reportedly involved more than 500,000 lines of TypeScript code, offering an unusually detailed look inside one of the company’s agentic AI systems. The incident occurred when original development files were uploaded instead of finalized production builds. A misconfigured JavaScript source map file — typically used for debugging — allowed external users to trace and reconstruct the underlying codebase. Researchers said the exposed archive contained approximately 512,000 lines across roughly 1,900 files. Anthropic later described the event as the result of human error rather than a malicious breach. …