Startup Anthropic has confirmed a leak of part of the source code for its AI programming tool, Claude Code, according to CNBC.

“No confidential customer data has been disclosed. The leak occurred due to a packaging error caused by human factors, not a hack. We are implementing measures to prevent this from happening again,” said a company representative.

The first reports of the leaked code surfaced on March 31, with a related post on X garnering over 30 million views.

Claude code source code has been leaked via a map file in their npm registry!

Code: https://t.co/jBiMoOzt8G pic.twitter.com/rYo5hbvEj8

— Chaofan Shou (@Fried_rice) March 31, 2026

This incident could negatively impact Anthropic's standing by giving external developers and competitors access to the algorithms of this popular tool.

The leaked data fully reveals the architecture of Claude Code, including how the service:

  • manages requests to models;
  • coordinates the work of multiple assistants;
  • controls access and permissions;
  • handles authorization.

Additionally, the code uncovered 44 hidden features that had not yet been officially announced. Among the most intriguing discoveries are:

  • Kairos — a constantly active background system that retains "memory" and analyzes accumulated data overnight for retraining;
  • Buddy — an AI pet reminiscent of a Tamagotchi, which, according to the code, was planned for launch between April 1-7.

The data set also includes Undercover Mode — a special feature designed to prevent the AI from accidentally publishing internal project names.

Series of Leaks and Market Competition

This marks the second major error for the company recently. Earlier, a description of a future AI model and other documents were found publicly available, as reported by Fortune.

At that time, a representative from Anthropic stated that the new model represents a "qualitative leap" in performance and is the "most powerful solution to date." It is currently undergoing testing among a limited group of users.

The company explained that information about the upcoming LLM surfaced online due to a human error, and the leaked materials were described as "early drafts of content."

The release of Claude Code for the general public occurred in May 2025. The tool assists developers in creating features, fixing bugs, and automating tasks.

Over the past year, the service has become so popular that increasing competition from Anthropic forced OpenAI to shut down its Sora project to focus resources on developing similar solutions.

Notably, in March, Anthropic transformed Claude into an AI agent, enabling the bot to use a computer to perform tasks.