Anthropic Leaks 512,000 Lines of Claude AI Code in Major Blunder

Human error exposed 512,000+ lines of Anthropic Claude AI Code, revealing KAIROS and Capybara secrets, pushing users to switch to the Native Installer.

In what is being described as a massive strategic blow, the AI powerhouse Anthropic accidentally leaked the secret blueprints for its highly profitable coding tool, Claude Code. A simple mistake during a routine update has handed competitors a literal map of how the company’s most successful technology actually works.

The trouble began on 31 March 2026. During a standard release of version 2.1.88 on the public npm registry, someone accidentally included a 59.8 MB source map file. For those of us who aren’t tech experts, this file basically translates complex computer code back into plain English that any developer can read.

By 4:23 am ET, the mistake was spotted by Chaofan Shou, an intern at Solayer Labs. He posted the ...