Can you check to make sure you’re on a Git repo and not a regular folder? Cloud and worktrees require git repos and won’t work in regular folders.
On a similar note, we are planning improved support for multi-root workspaces and worktrees in the Agents view.
The IDE view isn’t going anywhere. It’s a core part of Cursor and will continue to be developed and improved. Also, we’re actively working to add more features and improve extension support in Agents View.
Have you considered offering the proxy window as a standalone application for download?
I Mean：
Download Cursor
Download Cursor Agents Window
Hi, yes,
exactly the second part of your message. A plan can fill up the context window quite quickly. Ideally, I’d like to distribute individual TODOs from the plan into separate agent chats, each with its own model configuration, so they can be implemented independently.
Version:
Version: 3.0.13
VSCode Version: 1.105.1
Commit: 48a15759f53cd5fc9b5c20936ad7d79847d914b0
Date: 2026-04-07T03:05:17.114Z
Layout: glass
Build Type: Stable
Release Track: Default
Electron: 39.8.1
Chromium: 142.0.7444.265
Node.js: 22.22.1
V8: 14.2.231.22-electron.0
OS: Darwin arm64 25.3.0
Hi folks,
it would be cool to see the percentage of context usage within an agent chat like it is done within the Editor Window.
Cursor IDE
As I’ve been using the “new agent window”, I’ve noticed that I’ve missed how the context usage is displayed in the IDE agent window.
I can see the usage by hovering over the name in the top of the window, but it would be nice if this were exposed in a spot where I did not have to hover to see this. There looks to be plenty of room available to display this, too.
MacOS
Hi Kevin, I’m not sure how to send a DM through here but I am on an enterprise account so that is probably it! I will check with my admin. Thank you!
Sure sounds good! You can also let me know your team name and I can look it up that way too.
This is a great point. I’ve actually done this manually in the IDE, using a more expensive model to plan, then delegating implementation tasks to faster models. Would be interested to hear others weigh in on this too…
rephrased:
Would having a primary one-click way to fire off plan mode tasks in new chats help you?
E.g. rather than a one click Build button, perhaps a dropdown like Build Task 1, Build Task 2 etc.
Not in the short term. Do you have strong opinions on it either way?
Hello! So far NOT LOVING the agents window.
But here’s a quick one that could add some value to it: Restore the ability to fire up an agent to do some of the tasks in the to-do list created by “plan mode”.
It’s there in the editor view, and key for any serious work as the ai models tend to have a rather stupid behavior when doing many tasks on a relatively big projects.
Another note from the “Agents Window”. One of the most useful features in Cursor is the ability to review the code and go change by change seeing not only what the AI did, but actually accepting or rejecting changes.
That ability seems to be gone in the agents window. Maybe on purpose or maybe because it’s more geared towards people who don’t even know how to code.
My expectation would be that when I click on a change it would open the file on the editor panel to the right, I see the code plus the places where the AI changed things and actually let me Keep or Discard the change.
As it is right now, it’s mostly unusable.
One more note on the “Agents Window”. I can’t easily save the plan to my workspace, as I’d do with the editor. Not the worst thing I’ve reported so far but gosh it’s useful
Yes like it’s done within the Editor-Window. If i want to reference the plan in a new agent chat it does not appear within the search result when i do
@mfplan
It would be helpful to have this feature like its implemented within the Editor-Window where it’s possible to reference the plan and they appear right under the todos under “Referenced by Agents …”
For clarification. All my Skills, plans, etc. is stored under my Userfolder “/Users/[NAME]/.cursor”.
Maybe it’s related to the same topic like this one
Cursor settings don’t list or reflect active user rules and plugins
Hi Rodrigo! This is possble, you can use the three dots button and then use save to workspace. If you forget to save it to the workspace that’s not a problem because it is by default saved to your global Cursor plans directory (~/.cursor/plans/). If you’re not sure where it is you can just ask Cursor where it is and it will surface it for you!
Ah I see, you’d like the ability the fire off tasks individually rather than a single build button. I also agree this would be very useful.
I’ve reached a point where 99% of my code is written by Agents.
My workflow is simple: Plan → Agent → Code. > To be honest, I’m starting to feel like a traditional editor is unnecessary. I’d much prefer a streamlined AI client that’s light on memory and disk space.
While tools like Cursor are great, the Agent window is doing all the heavy lifting anyway. Maybe we don’t need an “editor” anymore, just a powerful agent interface? Haha.
The IDE project overview (New window, Cmd-Shift-N) is broken still in 3.0.16.
No cursor logo, no previous projects, no open button, nothing:
Please fix, makes IDE practically unusable.
“Open recent” Cmd-Shift-T does not trigger, also not clean.
Version: 3.0.16
VSCode Version: 1.105.1
Commit: 475871d112608994deb2e3065dfb7c6b0baa0c50
Date: 2026-04-09T05:33:51.767Z
Layout: editor
Build Type: Stable
Release Track: Default
Electron: 39.8.1
Chromium: 142.0.7444.265
Node.js: 22.22.1
V8: 14.2.231.22-electron.0
OS: Darwin arm64 25.3.0
