Anthropic’s Claude Code was widely reported to have leaked internal source material after a packaging mistake exposed a source map tied to the public @anthropic-ai/claude-code release, allowing outsiders to reconstruct large parts of the codebase. Axios reported that the leak revealed roughly 2,000 files and about half a million lines of code, along with unreleased feature flags and internal performance details, while Anthropic said no customer data or credentials were exposed and described the incident as a release-packaging error rather than a security breach. The leak was quickly mirrored and analyzed by developers, turning into a public teardown of how Claude Code is built and what features Anthropic has been working on.

AI-generated background, compiled from web sources — not editorial content.

More coverage

Explore the topic

Loading related coverage…

Comments

Loading comments…