Call stack exceeded on agent review

Support Bug Reports
1

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

Agent Review suddenly started throwing opaque error “Maximum call stack size exceeded”
Looking at the cursor agent review output terminal I see:

Generated Agent Review context with 32 files and 0 rules, diff size: 178.6201171875kb
Error computing local context: RangeError: Maximum call stack size exceeded

This works fine for another, larger diff I tried it with so the error is providing incomplete information about what conditions it is unable to review. This is very much a bug given that there’s no actionable error other then “Sorry user. You’re hosed. Have a nice day.”

Steps to Reproduce

Start an Agent Review

Expected Behavior

Either the review works or, if it fails, it tells me why it failed so I know what to do differently.

Operating System

MacOS

Version Information

Version: 3.9.16 (Universal)
VS Code Extension API: 1.105.1
Commit: 042b3c1a4c53f2c3808067f519fbfc67b72cad80
Date: 2026-06-27T06:41:01.941Z
Layout: editor
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.256
OS: Darwin arm64 25.5.0

For AI issues: which model did you use?

Deep agent review (unknown/opaque)

Additional Information

Tried with extensions disabled. No chage.
Tried with quick review. No change.

Does this stop you from using Cursor

No - Cursor works, but with this issue

7

Hey Scott, thanks for the report.

This is a bug on our end. The “compute local context” step fails when your change set includes a single file with one very large block of changes (usually a generated, lock, or data file), and it trips per file rather than by total diff size, which is why a bigger diff worked for you. The unhelpful error is a separate issue we’re fixing too, it should name the file it couldn’t handle. I’m flagging both to our team.

To unblock now: if your pending changes include a big generated or data file, commit it on its own (or otherwise keep it out of the review) and re-run on the rest. That should let the review complete.

If you can confirm whether the failing diff includes a large generated or data file, that’ll help us pin down your exact case.