Claude Code Leak: Why it Matters To Journalists

Claude Code dreams. In what can only be described as a major screw up, Claude Code's source code was leaked online yesterday. It didn't take much time for coders to dig into it, finding hidden features and revealing a "dreaming" system that allows it to consolidate information into memory.

Perhaps most relevant to privacy activists and journalists who use the tool to vibe code is the previously undisclosed use of telemetry (automatic collection of user data by remote sources); killswitches; and the ability of the company's employees to put it into stealth mode when making changes to publicly available code.

Here's how Modem Guides puts it:

The code contains six or more remote killswitches that can force specific behaviors: bypassing permission prompts, enabling or disabling fast mode, toggling voice mode, controlling analytics collection, and in some cases forcing the application to exit entirely. If a remote configuration change is classified as "dangerous," a blocking dialog appears — but rejecting it causes the application to quit.

This is not unusual for cloud-connected software. But Claude Code is a tool that requests filesystem access, terminal command execution, and full codebase read/write privileges on your development machine. Understanding that it also maintains a persistent remote control channel to Anthropic's servers is relevant context for anyone running it on sensitive projects.