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

Executive Summary

A malicious browser extension may abuse Claude Dispatch to make Claude Desktop run code on a user's computer. Organizations using Claude Desktop, browser extensions, or remote AI task workflows should review agent permissions and execution controls.

What Happened

A browser extension that can run scripts on claude.ai could make Claude Desktop run code on a person's computer, researchers at Forever Security, an endpoint security company, showed in a test. The attack needs Dispatch, a Claude feature that lets people send tasks from their phone to Claude Desktop. The flaw that makes this possible can still be exploited, the researchers said earlier this

Affected Products

Claude Dispatch
Claude Desktop
claude.ai

Vulnerability / Attack Technique

Malicious browser extension enabling unauthorized code execution through AI agent dispatch

Technical Mechanism

A browser extension that can run scripts on claude.ai interacts with Claude Dispatch, which sends tasks from a phone to Claude Desktop. This chain can cause Claude Desktop to run code on the user's computer, according to the reported test.

AI Security Category

LLM06:2025 Excessive Agency

Part of the AI Agent Excessive Agency topic.

Relevant Frameworks

LLM06:2025 Excessive AgencyMITRE ATLAS: Execution

RealGround Analysis

This demonstrates how excessive permissions and trust across browser, cloud, mobile, and desktop agent components can turn an extension compromise into local code execution. Teams running AI agents should assess cross-device task authorization, constrain agent actions, and test whether untrusted inputs can trigger privileged operations.

Recommended Controls

Require explicit user confirmation before Claude Desktop or comparable agents execute local code.
Restrict browser-extension access to claude.ai and review extensions against an allowlist.
Apply least-privilege permissions and sandboxing to desktop AI agents and their tool integrations.
Authenticate and authorize Dispatch tasks across devices, including device binding and replay protection.
Conduct adversarial testing of browser-to-agent workflows for unauthorized task execution.

Related RealGround Service

AI Agent Business Logic Audit · or start with the free AI Agent Security Scanner.

Sources

thehackernews.com: https://thehackernews.com/2026/11/malicious-browser-extensions-could-use.html

Summary and analysis written by RealGround from the source above; last updated 2026-11-07.