yukonSilver detection false-negative on AMI BIOS whitebox PCs causes infinite VM bundle deletion loop

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?

Bug
Claude Desktop's cleanupVMBundleIfUnsupported repeatedly deletes the VM bundle on every Code panel access on whitebox/OEM PCs (XCY, Topton, Beelink), making the Code panel unusable.

Hardware

• Intel Core i5-8350U (Kaby Lake R, supports VT-x + VT-d + EPT)
• AMI Aptio BIOS firmware version 5.12 (2019)
• Whitebox/OEM mini-PC, Manufacturer reported as "Default string"
• Windows 11 23H2

Functional state of virtualization

• HypervisorPresent: True
• vmcompute service: Running
• Hyper-V Microsoft-Hyper-V-All: Enabled
• IovSupport: True (SR-IOV functional)
• vEthernet switch: Up, 1 Gbps
• WSL2 and other Hyper-V guests: working

What Claude Desktop sees

• Win32_Processor.VirtualizationFirmwareEnabled: False (BIOS WMI reporting bug)
• Win32_Processor.SecondLevelAddressTranslationExtensions: False (idem)
• Win32_Processor.VMMonitorModeExtensions: False (idem)

These three fields are known to misreport on whitebox AMI BIOS that don't properly populate the SMBIOS/ACPI tables. The hardware capability is actually present, as proven by Hyper-V running successfully.

Loop observed in cowork_vm_node.log

Every Code panel access:

1. [cleanupVMBundleIfUnsupported] yukonSilver not supported (status=unsupported)
2. [deleteVMBundle] Deleted rootfs.vhdx, vmlinuz, initrd
3. Next launch: [Bundle:status] rootfs.vhdx missing → 71-second decompression
4. Boot completes, then panel access triggers loop again

Suggested fix

Replace the WMI-bit check with a functional probe: try to instantiate a minimal Hyper-V guest. If it succeeds, yukonSilver is supported regardless of the WMI fields. Or fall back gracefully to legacy VM architecture instead of looping deletion when WMI reports negative.

Impact

• All whitebox/OEM Windows PCs with AMI BIOS that misreport WMI virtualization fields are affected
• Especially common in Chinese mini-PC ecosystem (XCY, Topton, Beelink, AcePC, Mele, Trigkey)
• Users see Code panel crash and infinite VM rebuild cycles, ~1 GB disk wear per cycle

What Should Happen?

HyperV VM can run correctly

Error Messages/Logs

Steps to Reproduce

claude_bug_report_20260508_000425.zip

Claude Model

Other

Is this a regression?

Yes, this worked in a previous version

Last Working Version

1-1617

Claude Code Version

Version 1.6608.0 (f65729)

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

Other

Additional Information

_No response_