Malicious Browser Extensions Could Use Claude Dispatch to Run Code on Users' Computers

What happened

Researchers at Forever Security showed that a browser extension with permission to run scripts on claude.ai can use Claude's Dispatch feature to send commands to Claude Desktop on a user's computer. This allows the extension to trigger actions on the computer without the user's knowledge, as long as the user is signed in to both claude.ai and Claude Desktop with Dispatch enabled. Anthropic, the maker of Claude, has said it will retire the Dispatch feature by the end of October, but it is still available for some users. The researchers reported the issue to Anthropic in May, and no real-world attacks have been reported so far.

Key details

• Dispatch lets users send tasks from claude.ai to Claude Desktop, including across devices.
• The attack requires the user to be signed in to both claude.ai and Claude Desktop with Dispatch active.
• A script can approve prompts automatically, bypassing user consent dialogs.
• Anthropic closed new sign-ups for Dispatch but existing Pro, Max, and Ultra users can still access it.
• Researchers say similar attacks may be possible with Claude Code Remote Control and Cowork sessions, but have not published tests of those.

Why it matters

Users of Claude Desktop who have Dispatch enabled are at risk of having harmful commands run on their computers if they install a malicious or compromised browser extension. Organizations using shared Claude accounts across personal and work devices could face exposure to unauthorized actions or data theft. Users and administrators must review their extension permissions and Dispatch settings to reduce the risk until the feature is fully retired.