A massive leak of Claude Code's source code has revealed dozens of hidden features of Anthropic - ForkLog: cryptocurrencies, AI, singularity, future

# A major leak of Claude Code’s source code revealed dozens of hidden functions from Anthropic

Anthropic confirmed that part of the source code of its AI programming tool, Claude Code, was leaked, CNBC reports.

“No confidential customer data was disclosed. An error occurred during the packaging of the release. It was caused by a human factor, not a breach. We are implementing measures to ensure this doesn’t happen again,” a company spokesperson said.

The first reports about the code that ended up online appeared on March 31. The corresponding post on X has garnered more than 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

The incident could negatively affect Anthropic’s standing, giving outside developers and competitors access to the algorithms behind a popular tool.

The published data fully reveals the architecture of Claude Code. In particular, it became known how the service:

• manages requests to models;
• coordinates the work of multiple assistants;
• controls access and permissions;
• works with authentication.

In addition, the code uncovered 44 hidden functions that have not yet been officially introduced. Among the most interesting finds:

• Kairos — a constantly active background system. It saves “memory” and analyzes the accumulated data at night for further training;
• Buddy — an AI pet in the style of “Tamagotchi.” Judging by the code, its launch was planned for April 1–7.

The data set also includes Undercover Mode — a special mode that prevents the neural network from accidentally publishing internal project names.

A string of leaks and market competition

This is the second major company mistake in recent times. A few days earlier, an outline of a future AI model and other documents were found in public access.

At the time, an Anthropic representative said that the new model is a “quality leap” in performance and “the most powerful solution to date.” Right now, it is being tested among a limited circle of users.

The company explained that information about the upcoming LLM appeared online “due to a human error,” and the leaked materials were “early content drafts.”

The release of Claude Code for a wide audience took place in May 2025. The tool helps developers create functions, fix bugs, and automate tasks.

Over the past year, the service became so in demand that growing competition from Anthropic forced OpenAI to shut down the Sora project in order to focus resources on developing similar solutions.

Recall that in March, Anthropic made Claude into an AI agent— the bot gained the ability to use a computer to complete tasks.