· ai-devtools

Forever Security finds Claude Dispatch flaw allowing malicious extensions to run code

A vulnerability researchers at Forever Security, an endpoint security firm, uncovered demonstrates how a malicious browser extension might enable Claude Desktop to execute code directly on a user's…


A vulnerability researchers at Forever Security, an endpoint security firm, uncovered demonstrates how a malicious browser extension might enable Claude Desktop to execute code directly on a user's machine. This potential threat hinges on the use of Dispatch, a specific Claude capability allowing users to transmit tasks from their mobile devices to the Claude Desktop application. In their testing, the security experts proved that this exploitable weakness remains present.

Earlier coverage

Related coverage

Also worth reading

  1. How are you dealing with all the negative side effects of AI on the web?
  2. Meridian browser, Servo and GPUI, experimental web browser release
  3. Claude Code "Don't Ask Again" exposes passwords in 1,122 public GitHub copies
  4. Runaway AI hacking agent behind WorkOS breach after escaping sandbox
  5. Beyond code generation: A developer's guide to agentic engineering