You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
On Codex Desktop for Windows, the bundled Browser / Chrome / Computer Use runtime cannot be staged from the Microsoft Store package. As a result, the in-app browser settings page reports that the plugin is unavailable and cua_node remains empty after reinstalling.
This reproduces on a newer build than the related report #25220.
The source file is EFS application-protected, while the destination is unencrypted:
cipher /c <WindowsApps>\\app\\resources\\cua_node\\bin\\CHANGELOG.md
E CHANGELOG.md
Compatibility Level:
Application Protected
cipher /c %LOCALAPPDATA%\\OpenAI\\Codex\\runtimes\\cua_node
U cua_node
Copying the same source file to a normal user workspace fails with:
System.IO.IOException: The specified file could not be encrypted.
HRESULT: -2147018896
This confirms the failure is not caused by a disabled plugin, a missing configuration entry, or an incomplete user-level cache.
Expected behavior
Codex should stage its bundled runtime successfully from a protected Windows Store package, or use a packaging/copy mechanism compatible with EFS application-protected files.
Notes
Reinstalling Codex did not resolve the issue.
Please do not treat a direct WindowsApps marketplace path as a workaround; that is version-dependent and does not restore the trusted native runtime bridge.
The app's current bundled cua_node runtime is 0.0.6 (Sky 0.6.2), but its expected user-runtime directory fb8898c05a62885e is never created. Only an older runtime (0.0.1, Sky 0.4.10) remains locally.
The desktop log recorded 36 relocation failures while trying to stage the new runtime. The first failing file is again:
At runtime there is no node_repl.exe or codex-computer-use.exe process, no Computer Use/CUA named pipe, and the affected task receives no privileged node_repl tool. The Computer Use skill text is loaded, but the execution bridge is absent.
The Store update from 26.721 to 26.727 completed successfully, and repeated app restarts did not change the result. This appears to confirm that the EFS/Application Protected relocation bug persists in 26.727 and directly prevents the trusted Computer Use runtime from being provisioned.
This report is directly relevant to the project's current experimental Application Protected/EFS path. WinBridge v3.1.1 already uses byte-stream (FileStream) copying plus source/destination SHA-256 verification when materializing official resources into a user-owned mirror, instead of relying only on the native copy path that is failing here with UNKNOWN / -4094.
I want to be precise about the limitation: I do not consider Application Protected/EFS support fully solved yet. The current code can read/copy/verify these resources on affected machines, but I am still treating multi-version Store-package validation and stronger signature validation as research work. The tool does not decrypt, take ownership of, patch, or change ACLs on WindowsApps.
For a machine showing exactly this state — packaged cua_node exists, source is readable/Application Protected, user runtime is empty or stale, and helper/transport paths are then missing — WinBridge can at least diagnose the package/runtime version split and attempt a verified user-side reconstruction from the already installed official package. If the failure is instead in the trusted bridge after all local resources are healthy, it will not claim to fix that upstream runtime layer.
cipher /c reports the relevant WindowsApps source files as Application Protected. A normal copyFile-style copy fails, while a byte-stream read/write copy of the same files succeeds.
For diagnosis only, I copied the protected payloads using a version-aware stream-copy fallback into content-hash-addressed per-user directories and verified all expected hashes. After that:
the Computer Use native pipe reached startup ready;
BrowserUseThreadConfig selected the repaired node / node_repl runtime;
Node reported v24.19.0;
the official node_repl MCP server initialized successfully.
This provides a current-version confirmation that the underlying issue is compatibility between the relocation implementation and Application Protected WindowsApps resources, not an absent bundled payload.
A product fix should use a copy mechanism compatible with Application Protected package files, validate the completed destination, and self-heal into the current content-hash directory after each Store update. No raw logs are attached because they contain local metadata; sanitized excerpts can be provided if needed.
Adding a current-build reproduction and a small temporary workaround tool for affected Windows users.
Environment
Windows 11 x64, build 26200
Microsoft Store / MSIX package: OpenAI.Codex
Reproduced on:
26.825.3734.0
26.825.4187.0
Observed behavior
After an automatic Codex Desktop update, the app starts several ChatGPT.exe processes but the main window stays blank / untitled / not responding. The app repeatedly creates incomplete runtime staging folders under:
For example, after 26.825.4187.0, the new runtime hash was:
e4d75eceaa042f20
Multiple .staging-e4d75eceaa042f20-* folders were created, each incomplete, while the previous repaired runtime cache from the older build remained unused.
But normal copy paths fail when copying from WindowsApps to the user runtime cache. For example, robocopy failed with Windows error 6000, and PowerShell / .NET copy also failed when trying to preserve the source file attributes.
cipher /c showed packaged runtime files such as manifest.json carrying EFS encryption / Application Protected attributes, while the target cache directory was unencrypted.
A byte-stream copy workaround succeeds: read each packaged runtime file as bytes and write it as a new local file under the expected runtime hash folder. The resulting local files no longer carry the encrypted attribute, and Codex Desktop starts normally afterward.
Temporary workaround tool
I published a small Windows-only workaround repo here:
Only touches %LOCALAPPDATA%\OpenAI\Codex\runtimes\cua_node
Backs up incomplete .staging-* folders instead of deleting them
Rebuilds the current runtime cache from the bundled app\resources\cua_node
Copies via byte streams so EFS attributes are not preserved
Does not reset login state, chat history, registry, or Windows settings
This is only intended as a temporary user-side mitigation until Codex Desktop handles this WindowsApps EFS / Application Protected materialization path upstream.
Activity
github-actions commentedon Jul 13, 2026
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
Ywb822ovo commentedon Aug 1, 2026
Follow-up: still reproducible on 26.727
This remains reproducible on a newer Microsoft Store build.
Environment
OpenAI.Codex_26.727.6591.0_x64__2p2nqsd0c76g00.146.0-alpha.9.2019fbe07-3f0c-7f91-82c2-2a9158dfe5f2New evidence
The app's current bundled
cua_noderuntime is0.0.6(Sky0.6.2), but its expected user-runtime directoryfb8898c05a62885eis never created. Only an older runtime (0.0.1, Sky0.4.10) remains locally.The desktop log recorded 36 relocation failures while trying to stage the new runtime. The first failing file is again:
The failure is:
A read-only
cipher /ccheck reports the WindowsApps source file as:while the destination runtime directory is unencrypted.
The staging failure is followed by:
At runtime there is no
node_repl.exeorcodex-computer-use.exeprocess, no Computer Use/CUA named pipe, and the affected task receives no privilegednode_repltool. The Computer Use skill text is loaded, but the execution bridge is absent.The Store update from 26.721 to 26.727 completed successfully, and repeated app restarts did not change the result. This appears to confirm that the EFS/Application Protected relocation bug persists in 26.727 and directly prevents the trusted Computer Use runtime from being provisioned.
Related original report: #25220.
zemeng5208 commentedon Aug 12, 2026
I maintain WinBridge Recovery for Windows Codex Desktop local-state failures: https://github.com/zemeng5208/winbridge-recovery
This report is directly relevant to the project's current experimental Application Protected/EFS path. WinBridge v3.1.1 already uses byte-stream (
FileStream) copying plus source/destination SHA-256 verification when materializing official resources into a user-owned mirror, instead of relying only on the native copy path that is failing here withUNKNOWN / -4094.I want to be precise about the limitation: I do not consider Application Protected/EFS support fully solved yet. The current code can read/copy/verify these resources on affected machines, but I am still treating multi-version Store-package validation and stronger signature validation as research work. The tool does not decrypt, take ownership of, patch, or change ACLs on
WindowsApps.For a machine showing exactly this state — packaged
cua_nodeexists, source is readable/Application Protected, user runtime is empty or stale, and helper/transport paths are then missing — WinBridge can at least diagnose the package/runtime version split and attempt a verified user-side reconstruction from the already installed official package. If the failure is instead in the trusted bridge after all local resources are healthy, it will not claim to fix that upstream runtime layer.likeke0402 commentedon Aug 23, 2026
Reproduced again on a newer Microsoft Store build.
Environment
26.818.5345.026.818.41705Observed
The desktop log reports:
Affected payloads include
cua_node,codex.exe,rg.exe, and bundled plugin metadata. The downstream state was:cipher /creports the relevant WindowsApps source files asApplication Protected. A normalcopyFile-style copy fails, while a byte-stream read/write copy of the same files succeeds.For diagnosis only, I copied the protected payloads using a version-aware stream-copy fallback into content-hash-addressed per-user directories and verified all expected hashes. After that:
startup ready;node/node_replruntime;node_replMCP server initialized successfully.This provides a current-version confirmation that the underlying issue is compatibility between the relocation implementation and Application Protected WindowsApps resources, not an absent bundled payload.
A product fix should use a copy mechanism compatible with Application Protected package files, validate the completed destination, and self-heal into the current content-hash directory after each Store update. No raw logs are attached because they contain local metadata; sanitized excerpts can be provided if needed.
jedliuai commentedon Aug 29, 2026
Adding a current-build reproduction and a small temporary workaround tool for affected Windows users.
Environment
26200OpenAI.Codex26.825.3734.026.825.4187.0Observed behavior
After an automatic Codex Desktop update, the app starts several
ChatGPT.exeprocesses but the main window stays blank / untitled / not responding. The app repeatedly creates incomplete runtime staging folders under:For example, after
26.825.4187.0, the new runtime hash was:Multiple
.staging-e4d75eceaa042f20-*folders were created, each incomplete, while the previous repaired runtime cache from the older build remained unused.Additional evidence
The packaged runtime is complete:
But normal copy paths fail when copying from
WindowsAppsto the user runtime cache. For example,robocopyfailed with Windows error6000, and PowerShell / .NET copy also failed when trying to preserve the source file attributes.cipher /cshowed packaged runtime files such asmanifest.jsoncarrying EFS encryption / Application Protected attributes, while the target cache directory was unencrypted.A byte-stream copy workaround succeeds: read each packaged runtime file as bytes and write it as a new local file under the expected runtime hash folder. The resulting local files no longer carry the encrypted attribute, and Codex Desktop starts normally afterward.
Temporary workaround tool
I published a small Windows-only workaround repo here:
https://github.com/jedliuai/codex-runtime-repair
Release zip for non-technical users:
https://github.com/jedliuai/codex-runtime-repair/releases/tag/v1.0.0
Scope of the workaround:
%LOCALAPPDATA%\OpenAI\Codex\runtimes\cua_node.staging-*folders instead of deleting themapp\resources\cua_nodeThis is only intended as a temporary user-side mitigation until Codex Desktop handles this WindowsApps EFS / Application Protected materialization path upstream.