Anthropic accidentally leaks its most powered AI model; Internet reacts

Anthropics has leaked the source of its popular AI coding service, Claude Code. This leak has sent shockwaves through the AI community as the platform is known for its security practices. The American AI company inadvertently exposed the complete source code of its flagship coding assistant due to a basic packaging error. The incident occurred on March 31, when a misconfigured source map file in Anthropic's npm registry revealed the entire codebase.

Cybersecurity experts have criticised the lapse, calling it a major blow to Anthropic's reputation. The company, which has built its brand around safety and strict controls, is reportedly preparing for a $380 billion IPO. This leak has put a massive dent in the image of the company, and would create problems for the IPO going ahead.

How the Leak Happened

Security researcher Chaofan Shou discovered the issue when he found a 60MB source-map file (cli.js.map) in Claude Code's npm package. This file allowed anyone to reconstruct the full TypeScript codebase, exposing the CLI implementation, agent architecture, unreleased features, and internal tooling. Importantly, the leak did not include model weights or user data.

Anthropic confirmed the incident was caused by human error, not a hack. "No sensitive customer data or credentials were involved or exposed. This was a release packaging issue caused by human error, not a security breach," the company said in a statement.

Reactions Online

The leak has triggered intense reactions across social media. Enterprise AI Architect Shakthi Vadakkepat dubbed it "the mothership of all code leaks," pointing out the irony of a security-focused company making such a rookie mistake. Others compared it to a homeowner investing in locks and guards but accidentally publishing the house's blueprint online.

Critics highlighted the contradiction of Anthropic warning governments about AI risks while failing to protect its own code. Memes and jokes flooded forums, mocking the lapse, while developers and tech enthusiasts eagerly analysed the leaked code, calling it a valuable learning resource.

Claude Code is Anthropic's advanced AI coding assistant, designed to help users edit files and manage projects locally. Thankfully, the leak didn't expose the core model.

READ MORE: Ray-Ban Meta launches two new smartglasses for prescription wearers; Check UAE availability and price