Vulnerability in Claude Cowork Could Let an AI Agent Break Out of the VM and Access Mac Files

Colleagues, I’d like to flag a cybersecurity issue.

I came across a report on a vulnerability in Claude Cowork: an AI agent could escape a Linux VM and access files on the Mac host.

According to researchers, before the patch the issue affected roughly 500,000 macOS users running local Cowork sessions.

The risk is clear: in such a scenario, SSH keys, cloud credentials and other user data could be exposed. Cloud execution is now enabled by default, but local mode remains risky.

Why this matters: in AI environments, it is essential to validate not only the model, but also its isolation boundaries.

Are you already restricting local AI-agent execution?

#cybersecurity #AI #vulnerability #macOS