Anthropic’s leaked Claude code was an internal error, not an attack
AI-summarised brief · reviewed before publication
Anthropic's AI interface, Claude, suffered a significant data leak when the company mistakenly published its source code in a routine update. The 512,000-line TypeScript code was downloaded and circulated online, but customer information was not at risk. The issue was attributed to human error, not a security breach. Measures are being taken to prevent similar incidents, but the code is now permanently online, potentially allowing others to recreate Claude's features.