Remote control does not work as expected

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

when executing /remote-control - it does not exactly work properly and many times encounters some error. Why is it not similar to Claude code?

Steps to Reproduce

Triggered /remote-control
And it shows “agent encountered some error” or something like that.

Expected Behavior

The expected behavior is that once I trigger /remote control, I should be able to run these agents from my phone or anywhere without any interruption, flawlessly. Without compressing some context or the other. There is also some need to compress the context as well.

Operating System

MacOS
iOS

Version Information

Version: 3.17.19
VS Code Extension API: 1.128.0
Commit: ae3a2b7231dd56194447fe4570dfdc61640b1e90
Date: 2026-08-24T06:42:14.583Z
Layout: Agent Window
Build Type: Stable
Release Track: Default
Electron: 40.10.3
Chromium: 144.0.7559.236
Node.js: 24.15.0
V8: 14.4.258.32-electron.0
xterm.js: 6.1.0-beta.291
OS: Darwin arm64 25.5.0

Does this stop you from using Cursor

No - Cursor works, but with this issue

Hi @xprayag Thanks for writing this up, and sorry that your session was not as smooth as you were expecting. A few things that are easy to miss with this flow:

  1. After /remote-control, send the next message. That is what actually hands the session off.
  2. Your computer has to stay awake, online, and have that same workspace open the whole time. Tool calls still run on the machine, so sleep, a closed lid, or switching away from the folder will drop the session. In the Agents Window, go to Settings > Agents and turn on Keep this computer awake if the machine is plugged in.
  3. If a chat already shows “Agent encountered an error”, start a new agent and run /remote-control again rather than retrying the failed one.

Setup notes are here: Cursor for iOS | Cursor Docs

On the context side: converting to Remote Control moves the agent loop to the cloud and continues with the conversation state the model needs. It is not a lossless copy of the local session, so some context will look condensed compared with staying on the desktop. If you want the phone workflow without keeping your computer in the loop, a Cloud Agent is the closer fit, since that run lives entirely in the cloud.

If you run into other issues would it be great to have a screenshot of the error and /or a request ID or cloud agent ID (what comes after the URL, begins with bc - and I will take a take a closer look!