Cursor silently pausing on every agent run

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

On every agent run no matter where it is a primary or subagent, cursor will silently pause on any action (most often a GREP or file search). Even after total application restart, it fall back to the same issue. I have attempted with both Claude Opus 4.6 and Gemini 3.1 pro and both attempts experienced the same behavior. Currently Cursor is unusable.

Steps to Reproduce

Open up an agent, enter a prompt and wait for it to silently pause.

Expected Behavior

The expected behavior is that when I run an agent or a subagent that it will go through the steps and complete the task.

Operating System

Windows 10/11

Version Information

Version: 3.2.11 (user setup)
VSCode Version: 1.105.1
Commit: e9ee1339915a927dfb2df4a836dd9c8337e17cc0
Date: 2026-04-24T14:36:47.933Z
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: Windows_NT x64 10.0.26100

For AI issues: which model did you use?

Claude Opus 4.6 and Gemini 3.1 pro

For AI issues: add Request ID with privacy disabled

N/A

Does this stop you from using Cursor

Yes - Cursor is unusable

Hey, thanks for the detailed report. This matches a known cluster of freezes in 3.2.x during agent runs, often triggered by grep or file search, and we’re tracking it on our side. The same pattern is discussed here: Agents freeze at grepping and subagents and here: Sub Agents wont run and are stuck at "New Subagent - Starting up". I can’t share an ETA for a fix yet, but I’ll link your report to our internal issue.

To help with diagnostics, can you please send:

  1. A Request ID from one of the stuck runs. Go to Cursor Settings > Privacy, turn off Privacy Mode, rerun it, then in the chat menu top right three dots > Copy Request ID.
  2. Process Explorer during the freeze: press Ctrl+Shift+P, run Developer: Open Process Explorer. Please take a screenshot while it’s paused, we’re interested in what’s using CPU and RAM, especially extensionHost and ptyHost.
  3. Whether it still freezes if you start Cursor without extensions: from a terminal run cursor --disable-extensions.

Also please check if there are any errors in Output View > Output > Cursor at the time of the freeze. If yes, send the log.

After running with developer tools and disabled extensions, I simply running an ASK mode request where it spun up a subagent, it froze and I captured this on the console. Please keep in mind privacy mode is a requirement as I am running on a corporate environment.

Console Error Message

ERR \[Extension Host\] \[otel.error\] {“stack”:“OTLPExporterError: Bad Request\\n\\tat IncomingMessage. (c:\\Users\\adangelo\\AppData\\Local\\Programs\\cursor\\resources\\app\\node_modules\\@opentelemetry\\otlp-exporter-base\\build\\src\\transport\\http-transport-utils.js:52:31)\\n\\tat IncomingMessage.emit (node:events:531:35)\\n\\tat endReadableNT (node:internal/streams/readable:1698:12)\\n\\tat process.processTicksAndRejections (node:internal/process/task_queues:89:21)”,“message”:“Bad Request”,“code”:“400”,“name”:“OTLPExporterError”,“data”:“{“error”:“Trace spans collection is not enabled for this user”}”}error @ workbench.desktop.main.js:65error @ workbench.desktop.main.js:65error @ workbench.desktop.main.js:43477sgS @ workbench.desktop.main.js:38781$logExtensionHostMessage @ workbench.desktop.main.js:38781\_doInvokeHandler @ workbench.desktop.main.js:41620\_invokeHandler @ workbench.desktop.main.js:41620\_receiveRequest @ workbench.desktop.main.js:41620\_receiveOneMessage @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:38798(anonymous) @ workbench.desktop.main.js:46458workbench.desktop.main.js:38781 \[Extension Host\] \[otel.error\] {“stack”:“OTLPExporterError: Bad Request\\n\\tat IncomingMessage. (c:\\Users\\adangelo\\AppData\\Local\\Programs\\cursor\\resources\\app\\node_modules\\@opentelemetry\\otlp-exporter-base\\build\\src\\transport\\http-transport-utils.js:52:31)\\n\\tat IncomingMessage.emit (node:events:531:35)\\n\\tat endReadableNT (node:internal/streams/readable:1698:12)\\n\\tat process.processTicksAndRejections (node:internal/process/task_queues:89:21)”,“message”:“Bad Request”,“code”:“400”,“name”:“OTLPExporterError”,“data”:“{“error”:“Trace spans collection is not enabled for this user”}”}igS @ workbench.desktop.main.js:38781$logExtensionHostMessage @ workbench.desktop.main.js:38781\_doInvokeHandler @ workbench.desktop.main.js:41620\_invokeHandler @ workbench.desktop.main.js:41620\_receiveRequest @ workbench.desktop.main.js:41620\_receiveOneMessage @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:38798(anonymous) @ workbench.desktop.main.js:46458workbench.desktop.main.js:65 WARN \[ExtensionService\] No inspect port found for LocalProcess \[role: user\] pid: 18540 (tryEnableInspector=false).warn @ workbench.desktop.main.js:65warn @ workbench.desktop.main.js:65warn @ workbench.desktop.main.js:43477(anonymous) @ workbench.desktop.main.js:46454await in (anonymous)getExtensionHostsInfo @ workbench.desktop.main.js:46454(anonymous) @ workbench.desktop.main.js:29547(anonymous) @ workbench.desktop.main.js:29547\_deliver @ workbench.desktop.main.js:55\_deliverQueue @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55(anonymous) @ workbench.desktop.main.js:46454\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55(anonymous) @ workbench.desktop.main.js:41624\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55\_setResponsiveState @ workbench.desktop.main.js:41620\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_onWillSendRequest @ workbench.desktop.main.js:41620\_remoteCall @ workbench.desktop.main.js:41620r..s.charCodeAt.r. @ workbench.desktop.main.js:41620getCurrentBranch @ workbench.desktop.main.js:38803getCurrentBranch @ workbench.desktop.main.js:29321\_refreshCachedBranch @ workbench.desktop.main.js:29321(anonymous) @ workbench.desktop.main.js:29321\_deliver @ workbench.desktop.main.js:55\_deliverQueue @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55$gitStatusWasRun @ workbench.desktop.main.js:38803\_doInvokeHandler @ workbench.desktop.main.js:41620\_invokeHandler @ workbench.desktop.main.js:41620\_receiveRequest @ workbench.desktop.main.js:41620\_receiveOneMessage @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:38798(anonymous) @ workbench.desktop.main.js:46458workbench.desktop.main.js:65 WARN \[ExtensionService\] No inspect port found for LocalProcess \[role: retrieval\] pid: 16892 (tryEnableInspector=false).warn @ workbench.desktop.main.js:65warn @ workbench.desktop.main.js:65warn @ workbench.desktop.main.js:43477(anonymous) @ workbench.desktop.main.js:46454await in (anonymous)getExtensionHostsInfo @ workbench.desktop.main.js:46454(anonymous) @ workbench.desktop.main.js:29547(anonymous) @ workbench.desktop.main.js:29547\_deliver @ workbench.desktop.main.js:55\_deliverQueue @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55(anonymous) @ workbench.desktop.main.js:46454\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55(anonymous) @ workbench.desktop.main.js:41624\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55\_setResponsiveState @ workbench.desktop.main.js:41620\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_onWillSendRequest @ workbench.desktop.main.js:41620\_remoteCall @ workbench.desktop.main.js:41620r..s.charCodeAt.r. @ workbench.desktop.main.js:41620getCurrentBranch @ workbench.desktop.main.js:38803getCurrentBranch @ workbench.desktop.main.js:29321\_refreshCachedBranch @ workbench.desktop.main.js:29321(anonymous) @ workbench.desktop.main.js:29321\_deliver @ workbench.desktop.main.js:55\_deliverQueue @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55$gitStatusWasRun @ workbench.desktop.main.js:38803\_doInvokeHandler @ workbench.desktop.main.js:41620\_invokeHandler @ workbench.desktop.main.js:41620\_receiveRequest @ workbench.desktop.main.js:41620\_receiveOneMessage @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:38798(anonymous) @ workbench.desktop.main.js:46458workbench.desktop.main.js:65 WARN \[ExtensionService\] No inspect port found for LocalProcess \[role: always-local\] pid: 32176 (tryEnableInspector=false).warn @ workbench.desktop.main.js:65warn @ workbench.desktop.main.js:65warn @ workbench.desktop.main.js:43477(anonymous) @ workbench.desktop.main.js:46454await in (anonymous)getExtensionHostsInfo @ workbench.desktop.main.js:46454(anonymous) @ workbench.desktop.main.js:29547(anonymous) @ workbench.desktop.main.js:29547\_deliver @ workbench.desktop.main.js:55\_deliverQueue @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55(anonymous) @ workbench.desktop.main.js:46454\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55(anonymous) @ workbench.desktop.main.js:41624\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55\_setResponsiveState @ workbench.desktop.main.js:41620\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_onWillSendRequest @ workbench.desktop.main.js:41620\_remoteCall @ workbench.desktop.main.js:41620r..s.charCodeAt.r. @ workbench.desktop.main.js:41620getCurrentBranch @ workbench.desktop.main.js:38803getCurrentBranch @ workbench.desktop.main.js:29321\_refreshCachedBranch @ workbench.desktop.main.js:29321(anonymous) @ workbench.desktop.main.js:29321\_deliver @ workbench.desktop.main.js:55\_deliverQueue @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55$gitStatusWasRun @ workbench.desktop.main.js:38803\_doInvokeHandler @ workbench.desktop.main.js:41620\_invokeHandler @ workbench.desktop.main.js:41620\_receiveRequest @ workbench.desktop.main.js:41620\_receiveOneMessage @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:38798(anonymous) @ workbench.desktop.main.js:46458workbench.desktop.main.js:65 WARN \[ExtensionService\] No inspect port found for LocalProcess \[role: agent-exec\] pid: 13684 (tryEnableInspector=false).warn @ workbench.desktop.main.js:65warn @ workbench.desktop.main.js:65warn @ workbench.desktop.main.js:43477(anonymous) @ workbench.desktop.main.js:46454await in (anonymous)getExtensionHostsInfo @ workbench.desktop.main.js:46454(anonymous) @ workbench.desktop.main.js:29547(anonymous) @ workbench.desktop.main.js:29547\_deliver @ workbench.desktop.main.js:55\_deliverQueue @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55(anonymous) @ workbench.desktop.main.js:46454\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55(anonymous) @ workbench.desktop.main.js:41624\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55\_setResponsiveState @ workbench.desktop.main.js:41620\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_onWillSendRequest @ workbench.desktop.main.js:41620\_remoteCall @ workbench.desktop.main.js:41620r..s.charCodeAt.r. @ workbench.desktop.main.js:41620getCurrentBranch @ workbench.desktop.main.js:38803getCurrentBranch @ workbench.desktop.main.js:29321\_refreshCachedBranch @ workbench.desktop.main.js:29321(anonymous) @ workbench.desktop.main.js:29321\_deliver @ workbench.desktop.main.js:55\_deliverQueue @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55$gitStatusWasRun @ workbench.desktop.main.js:38803\_doInvokeHandler @ workbench.desktop.main.js:41620\_invokeHandler @ workbench.desktop.main.js:41620\_receiveRequest @ workbench.desktop.main.js:41620\_receiveOneMessage @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:38798(anonymous) @ workbench.desktop.main.js:46458workbench.desktop.main.js:41671 \[composer\] Extension host became UNRESPONSIVE undefined {arch: ‘x64’, platform: ‘win32’, channel: ‘stable’, client_version: ‘3.2.11’, layout: ‘unifiedAgent’, …}error @ workbench.desktop.main.js:41671(anonymous) @ workbench.desktop.main.js:29547await in (anonymous)(anonymous) @ workbench.desktop.main.js:29547\_deliver @ workbench.desktop.main.js:55\_deliverQueue @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55(anonymous) @ workbench.desktop.main.js:46454\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55(anonymous) @ workbench.desktop.main.js:41624\_deliver @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55\_setResponsiveState @ workbench.desktop.main.js:41620\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_checkUnresponsive @ workbench.desktop.main.js:41620(anonymous) @ workbench.desktop.main.js:41620doRun @ workbench.desktop.main.js:60onTimeout @ workbench.desktop.main.js:60setTimeoute.setTimeout @ workbench.desktop.main.js:41071schedule @ workbench.desktop.main.js:60\_onWillSendRequest @ workbench.desktop.main.js:41620\_remoteCall @ workbench.desktop.main.js:41620r..s.charCodeAt.r. @ workbench.desktop.main.js:41620getCurrentBranch @ workbench.desktop.main.js:38803getCurrentBranch @ workbench.desktop.main.js:29321\_refreshCachedBranch @ workbench.desktop.main.js:29321(anonymous) @ workbench.desktop.main.js:29321\_deliver @ workbench.desktop.main.js:55\_deliverQueue @ workbench.desktop.main.js:55fire @ workbench.desktop.main.js:55$gitStatusWasRun @ workbench.desktop.main.js:38803\_doInvokeHandler @ workbench.desktop.main.js:41620\_invokeHandler @ workbench.desktop.main.js:41620\_receiveRequest @ workbench.desktop.main.js:41620\_receiveOneMessage @ workbench.desktop.main.j

Thanks, the log helped. The OTLPExporterError: "Trace spans collection is not enabled for this user" noise is telemetry the server rejects because of Privacy Mode, it is not the cause of the freeze. Please keep Privacy Mode as it is, I get that it is a corporate requirement.

The useful signal in your log is this line:

[agent_exec] [watchdog, LocalReadExecutor] permissionsService.shouldBlockRead still not completed after 3000ms

So the freeze happens when the agent checks permission to read a file. That matches your report, the pause happens during GREP or file search. Also, later it says Extension host became UNRESPONSIVE. The fact that it reproduces with --disable-extensions is also helpful, we can rule out extensions.

To narrow it down, can you please answer:

  1. Where is the workspace located, local disk, network share, UNC path (\\server\...), or a mounted network folder? In corporate setups this is often a factor.
  2. Rough repo size (file count or GB)?
  3. Does the project have a .cursorignore or a very large .gitignore?
  4. If you open a small local test project (a couple files on C:) and run the agent with GREP there, does it also freeze or does it work fine?

I can’t give an ETA for a fix yet, but I’ll post an update in the thread as soon as I have one.

Hey @andrew0517, sorry I’m only getting back to this now.

Since your report, we’ve shipped a lot of performance improvements. Freezes during agent runs around file search and sub-agents were a big focus in releases after the version you were on 3.2.11. The current stable is 3.16.29, which is a big jump, and this whole area has been reworked since then.

Can you update and rerun the same scenarios? You don’t need to turn off Privacy Mode, I remember that’s a corporate requirement and it doesn’t affect diagnostics.

If the agent still silently freezes on GREP or file search on the latest version, please start a new thread with updated details version and a short snippet from Output > Cursor while it’s frozen. We’ll take a fresh look. I’m going to close this thread for now.

Same silent-stall cluster on a purely local Windows workspace — a week of logs, including the shouldBlockFileRead watchdog signature

Adding a second data point to this thread, since the questions asked earlier were quite specific. My
setup hits the identical signature, but importantly on a local NTFS disk — no Remote-SSH, no UNC
path, no network share
, so the corporate-network angle can be ruled out for this instance.

Answers to the questions asked earlier in the thread

  1. Workspace location: local disk, C:\Users\...\Documents\PROJECTS\archive_v2. Not a network

    share, not a mounted folder, not UNC.

  2. Repo size: 52,926 files / 2.15 GB, containing 13 nested git repositories (several

    sibling worktrees of the same projects). 27,286 of those files are under .git, 18,294 under
    Rust/Java target/.

  3. .cursorignore / .gitignore: neither existed at the time of the incidents. (I have since

    added a .cursorignore that hides 87% of the tree — will report back whether it changes the
    frequency.)

  4. Small local test project: not yet tested, I will follow up.

Environment

  • Cursor 3.19.13, 3.19.19, 3.20.7, 3.20.10, 3.20.17 — the signature appears in all of them

  • Windows 10.0.26200 x64, Chromium 148.0.7778.280

  • Peak process memory 7.02 GB, peak activeComposersCount 73

Symptom

Agent runs stop mid-execution with no indication whatsoever in the UI — no error, no timeout,
no “taking longer than expected”, no abort. The run simply produces nothing further. Restarting the
run often lands in the same state.

What the logs show

1. Extension host goes UNRESPONSIVE — 731 events between 2026-09-08 and 2026-09-14.

[error] {"key":"composer","message":"Extension host became UNRESPONSIVE",
 "client_version":"3.20.7","layout":"unifiedAgent",
 "extensionHostKind":"LocalProcess","activeComposersCount":"73","extensionHosts":"[]"}

695 of them on a single day. 562 occurred between 19:00 and 23:00 local during continuous active
use
— I checked the Windows Kernel-Power event log and there was no sleep/Modern-Standby
transition in that window, so these are not suspend artifacts. 548 carried
activeComposersCount: 73.

2. The same watchdog signature reported earlier in this thread — ~750 trips in one week:
Count
Message
275
[watchdog, LocalReadExecutor] permissionsService.shouldBlockFileRead still not completed after 3000ms
131
[watchdog, LocalReadExecutor] getFormatForFile still not completed after 3000ms
113
[watchdog, LocalWriteExecutor] writeFile still not completed after 3000ms
88
[watchdog, LocalReadExecutor] readTerminalFile still not completed after 3000ms
56
[watchdog, LocalReadExecutor] readTextStreaming still not completed after 7000ms
54
[watchdog, LocalWriteExecutor] getBufferForWrite still not completed after 3000ms
33
[watchdog, LocalWriteExecutor] permissionsService.shouldBlockWrite still not completed after 3000ms

The watchdog detects the hang and logs it — and then nothing recovers the operation. No timeout, no
fallback, no surfaced error. That gap is the actual bug.

3. The search is not slow — the layer wrapping it hangs. Span durations from
cursor.requestTraces.log:
Span
n
median
p90
max
>10 min
LocalGrepExecutor.executeRg (actual ripgrep)
1499
446 ms
2.0 s
25.8 s
0
LocalGrepExecutor.execute (its parent)
972
4.2 s
66 s
3.9 h
35
LocalGrepExecutor.runFilesMode
889
1.4 s
119 s
3.9 h
45
streamFromAgentBackend
343
186.7 s
18.1 min
2.6 h
78
networkPhase
635
52.7 s
13.1 min
2.6 h
83
agent.request
476
37.4 s
7.7 min
2.6 h
39

This is the part I would most like the team to look at: executeRg never exceeds 26 seconds,
while its own parent span
execute runs for hours. Ripgrep completes fine every single time. The
stall lives entirely between the tool call and the search process, in the permission/orchestration
layer. A big workspace amplifies it but cannot be the cause.

(Caveat in the interest of accuracy: the multi-hour maxima are partly inflated by Modern Standby.
The medians, p90s, the 19:00–23:00 cluster and all watchdog counts are unaffected.)

4. Why nothing is visible — spans never reach a terminal state. Started and never completed,
across all sessions: 655 child.span, 36 exec.handle, 32 LocalGrepExecutor.execute,
19 AgentCompatService.runAgentLoop, 18 each of agent.request / streamFromAgentBackend /
getAgentStreamResponse, 17 networkPhase, 5 subagent.request.

An agent.request that neither completes nor errors is precisely the user-visible symptom.

5. First-token latency ladder: No first token received within Ns — 2s: 183×, 4s: 150×,
6s: 97×, 8s: 59×, 10s: 38×, 12s: 27×, 14s: 22×, 16s: 17×, 32s: 3×.

6. Transport errors: 2155 × Connect error in unary AI connect, 193 × Stream error after headers in extension host, 184 × Stream error reported from extension host, 103 × Agent exec stream start failed, 89 × Agent client stream abort. Also getaddrinfo ENOTFOUND api2.cursor.sh
after resuming from standby, with no recovery of in-flight runs.

A possible amplifier worth checking

The stack posted earlier in this thread shows _checkUnresponsive firing while the host is blocked
in getCurrentBranch → _refreshCachedBranch → $gitStatusWasRun. With 13 git repositories in
one workspace root
, each git status refresh fans out into 13 branch lookups over the same
extension-host IPC channel that shouldBlockFileRead uses. That channel saturating would explain
why file reads, writes, terminal reads and search all stall together rather than independently.

Why this matters commercially

Because the stall is silent, the cost is not just waiting:

  • Every restart re-sends the full context, so the same work gets billed two or three times over.

  • Restarts frequently hit the same stall, so some work is never completed at all.

  • On one machine this consumed multiple hours of development time in a single week.

The same workflow against the same backend via the JetBrains/IntelliJ integration does not show
this behavior, which is what points me at the Cursor Desktop extension-host architecture rather
than the model backend.

What I would ask for

  1. Make the watchdog act, not just log. The 3000/7000 ms watchdogs already detect the hang —

    they should fail the operation so the agent can error out or retry.

  2. Never leave an agent run without a terminal state. If the extension host goes UNRESPONSIVE,

    surface it in the agent UI and fail the run.

  3. Do not consume quota for runs that produced no output because the client froze.

  4. Bound git status/branch refresh for workspaces with many nested repositories.

  5. Warn on or virtualize large activeComposersCount — 73 live composers visibly degrades the

    host.

Happy to supply raw logs, Process Explorer captures or specific request IDs if that helps.

Hey, thanks for such a thorough write-up. This is one of the most useful reports we’ve seen on this topic. It’s especially valuable that you reproduced the signature on current versions (3.19.x-3.20.x) and on a purely local NTFS drive. That removes both the version and network factors, so it’s definitely not your setup.

This is an issue we’re tracking. Two of your observations are especially important:

  • LocalGrepExecutor.executeRg finishes in seconds, but the parent execute hangs for hours. That means ripgrep itself completes, and the stall is in the layer between the tool call and the process.
  • Spans like agent.request, execute, etc. never reach a terminal state, so the UI shows no error and no timeout.

I also passed your hypothesis about git fan-out (13 nested repos, getCurrentBranch and $gitStatusWasRun saturating the same ext-host IPC channel as shouldBlockFileRead) to the team. It matches the stack in that thread well.

I forwarded your suggestions as separate items too: the watchdog should fail the operation, not just log, runs should always reach a terminal state, and quota shouldn’t be consumed by runs with no output. I can’t share an ETA, but I’ll reply in the thread when there’s an update.

A couple things that would really help:

  1. If Privacy Mode is off, share a few Request IDs from those agent.request runs that started but never finished. If Privacy Mode is on, don’t change it, logs are enough.
  2. Your follow-up results: does adding a .cursorignore file (hides 87% of the tree) change how often it happens, and does the stall reproduce on a small test project (a couple files on C:) with GREP.
  3. The raw logs and request traces you offered earlier. Yes, please send those, they’ll be useful.

If you want us to look into specific usage that was charged for stuck runs with no result, it’s best to email [email protected] so they can check your account.