Claude Cowork Flaw Could Let AI Agent Escape Its VM and Access Mac Files

Researchers have discovered a vulnerability in Claude Cowork, a product from Anthropic, that allows the AI agent to escape its Linux virtual machine and access files on a Mac. This sandbox escape flaw enables the agent to read or write files anywhere on the Mac, potentially compromising user data. The vulnerability was identified by Accomplish AI, which reported that around 500,000 macOS users are affected. The issue arises from the agent's ability to break out of its virtual machine confines. The vulnerability poses a risk to Mac users who run Claude Cowork.