CLI vitest suite on Windows: one worker fork crashes at teardown in every full run, failing validate:cli with all 304 tests passing

[olaservo]

Version: v2 (main, reproduced at 07a2b4b)
Affected client: clients/cli (test infrastructure; Windows only)

Symptom

Every full npm run test in clients/cli on Windows 11 ends with:

⎯⎯ Unhandled Error ⎯⎯
Error: [vitest-pool]: Worker forks emitted error.
Caused by: Error: Worker exited unexpectedly
 Test Files  24 passed (25)
      Tests  303 passed (304)
     Errors  1 error

All tests pass — one worker fork exits with an unexpected code at teardown, vitest counts it as an unhandled error, and the run exits 1. This fails validate:cli → root validate → npm run ci on Windows even on a clean tree.

What's known

• Deterministic in full-suite runs (4/4 reproduced), including with the JSON reporter (which then shows 108 suites, 106 passed, 0 failed — no suite actually fails).
• Does not reproduce running suspect files in isolation: e2e, cliOAuth, oauth-runner, oauth-interactive, cli-oauth-navigation, stored-auth each pass cleanly alone — it needs the parallel fork pool under load.
• Pre-exists unrelated work: reproduced identically on the upstream tip with no local changes.
• Linux CI is green, so it's Windows-specific (or at least Windows-manifested).

Impact

Windows contributors cannot get a passing npm run validate / npm run ci even when every test passes — together with #1939 this makes the mandatory pre-push gate unrunnable on Windows.

Context

Found while validating the fix for #1852 on a Windows machine.