VentureBeat reported that Anthropic’s AI programming tool, Claude Code, suffered an unintended disclosure of around 512,000 lines of comprehensive TypeScript source code, stemming from a packaging mistake during the release process. This leak encompassed more than 40 tool modules and features yet to be released, including the persistent assistant mode Kairos and the electronic pet system Buddy. The source of the leak was traced back to source map files that were mistakenly included in the npm package. These files served as a guide to the original code, enabling developers to piece together the entire source code. Anthropic swiftly acted to eliminate the pertinent files; however, by that time, the code had already begun circulating within the developer community. Although the leak did not compromise the core model weights or user data, it did reveal the internal architecture and security protocols of the system, prompting concerns regarding the safety of its software supply chain.