New /worktree unnecessarily clutters agent output

[jeyj0-work]

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

When using /worktree to create a worktree, the agent output usually includes unnecessary information about the worktree that I’d much rather have available in a proper UI.

Steps to Reproduce

1. create new worktree using /worktree (real prompt here)

Expected Behavior

Agent focuses on the real prompt, isn’t destracted by worktree.
No unnecessary (potentially expensive depending on model) token usage for worktree management.

Screenshots / Screen Recordings

Operating System

Linux

Version Information

Version: 3.1.17
VSCode Version: 1.105.1
Commit: fce1e9ab7844f9ea35793da01e634aa7e50bce90
Date: 2026-04-19T19:33:58.189Z
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: Linux x64 6.19.7-200.fc43.x86_64

For AI issues: which model did you use?

Composer 2 (shouldn’t matter though → worked perfectly in Cursor 2)

Additional Information

Going to downgrade to Cursor 2 again after finishing filing bug reports, only upgraded again to file bug reports.

Does this stop you from using Cursor

Yes - Cursor is unusable

[mohitjain (Mohit)]

Hi there,

This is valid feedback. The agent currently receives context about the worktree environment so it knows where to make edits, but narrating those setup details in the output is unnecessary and wastes tokens as you pointed out.

The team is actively working on improving the worktree experience, which should address this verbosity along with the other worktree-related concerns you’ve reported.

[David_Gomes]

What Mohit said is correct.

@jeyj0-work please try out worktrees in the Agents Window, where they’re a native feature.

[jeyj0-work]

Hi David,

I already did at the time, but this is blocking me from using it productively: No "/apply-worktree" in new Agent Window