CoWork fails to start on Windows 11 Pro - Plan9 share Z: conflict with EFI System Partition
Preflight Checklist
- [x] I have searched existing issues and this hasn't been reported yet
- [x] This is a single bug report (please file separate reports for different bugs)
- [x] I am using the latest version of Claude Code
What's Wrong?
CoWork fails to start every time it is launched. The workspace never initializes and the vm_bundles directory remains empty after every reinstall attempt. Root cause: CoWork is hardcoded to use drive letter Z: for its Plan9 share, which conflicts with machines that already have Z: assigned - in this case, to the EFI System Partition.
<html>
<body>
<!--StartFragment--><h2 class="text-text-100 mt-3 -mb-1 text-[1.125rem] font-bold">Environment</h2>
<div class="overflow-x-auto w-full px-2 mb-6">
Field | Value
-- | --
OS | Windows 11 Pro
Claude Desktop Version | v1.569.0.0
Feature | CoWork (Workspace / VM)
Hyper-V State | Enabled
vmcompute service | Running (Manual)
vmms service | Running (Automatic)
Hyper-V Administrators group | User confirmed as member
</div><!--EndFragment-->
</body>
</html>
After removing the Z: assignment via diskpart, the error code changed from 0x80070005 (Access Denied) to 0x80070003 (Path Not Found) - meaning the letter conflict was resolved but the workspace bundle still never downloaded. The vm_bundles directory remained empty after every "Reinstall Workspace" attempt.
Additionally, the EFI partition re-grabs Z: after every reboot, so the conflict returns unless manually cleared each session.
What Should Happen?
CoWork should start working
Error Messages/Logs
Error Messages
Error 1 - Access Denied (most frequent, appears on every launch)
Failed to start Claude's workspace
failed to add Plan9 share 'z': HcsWaitForOperationResult failed with HRESULT 0x80070005: {"Error":-2147024891,"ErrorMessage":""}
Error 2 - Path Not Found (appeared after Z: drive letter was manually freed)
Failed to start Claude's workspace
failed to add Plan9 share 'z': HcsWaitForOperationResult failed with HRESULT 0x80070003: {"Error":-2147024893,"ErrorMessage":""}Steps to Reproduce
Steps to Reproduce
Have a Windows 11 machine where Z: is assigned to any partition (especially EFI/System)
Install Claude Desktop and attempt to launch CoWork
CoWork fails immediately with HRESULT 0x80070005
Claude Model
Not sure / Multiple models
Is this a regression?
Yes, this worked in a previous version
Last Working Version
_No response_
Claude Code Version
Claude Desktop v1.569.0.0.
Platform
Anthropic API
Operating System
macOS
Terminal/Shell
PowerShell
Additional Information
Troubleshooting Steps Already Completed
Full uninstall via Settings > Apps
Manually deleted %AppData%\Roaming\Claude
Manually deleted %LocalAppData%\Packages\Claude_pzs8sxrjxfjjc
Clean reinstall of latest Claude Desktop
Verified Hyper-V is enabled (State: Enabled)
Verified vmcompute and vmms services are running
Verified user is in Hyper-V Administrators group
Manually created vm_bundles directory and granted SYSTEM full control
Removed Z: drive letter from EFI partition via diskpart and mountvol Z: /D
Used "Reinstall Workspace" option in Claude Desktop after freeing Z:
Checked Windows System Event Log for Hyper-V errors - only informational networking events, no errors
Ruled Out
Z: mapped as a network drive
Z: mapped as a subst virtual drive
Hyper-V not enabled
Hyper-V services not running
User missing from Hyper-V Administrators group
Wrong permissions on vm_bundles
Current State
CoWork still fails after all of the above. The vm_bundles directory stays empty and no workspace bundle is ever created or downloaded. The only log file present (chrome-native-host.log) contains only browser bridge activity - nothing related to the VM or workspace initialization.
C:\Users\Doron\AppData\Local\Claude\
├── Logs\
│ └── chrome-native-host.log ← browser bridge only, no CoWork logs
└── vm_bundles\ ← empty, nothing ever written here
Expected Behavior
CoWork should dynamically find an available drive letter rather than hardcoding Z:
If the workspace bundle download fails, the app should surface a specific error
"Reinstall Workspace" should log what it is doing and whether the download succeeded or failed
HCS error codes should map to plain-language messages with specific remediation steps
Suggested Fix
CoWork should not be hardcoded to Z:. It should either:
Scan for an available drive letter at startup, or
Allow the user to configure the drive letter used for the Plan9 share
This is a realistic conflict on any machine with many drives, storage devices, network shares, or certain OEM configurations that assign letters to system partitions.
This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗