QEMU watch-items (boot respawn-once, arm64 QGA)

[mobileskyfi]

Question

Do these known QEMU/CHR rough edges recur or regress, and what is the root cause when they do?

This is missing/extending test/lab coverage, similar in spirit to #42, not CI refactoring. If #29 later creates reusable lab/manual workflow units, they may help gather data for these watch-items, but the behavioral investigation belongs here.

Watch-items

• Boot respawn-once (DESIGN Potential fixes for 4 code quality findings #8 mitigation): root cause unconfirmed. Watch for respawning QEMU once. Build a repro if it recurs, survives the respawn, or appears on _launchExisting.
• arm64 QGA: MikroTik ticket open. Extend QGA tests once fixed and arm64 + KVM is confirmed.

Done-when

A recurrence produces a repro plus REPORT.md and a real fix/issue; if neither recurs over a few release cycles, close as "not observed".

Anchor issue for future agents that hit either symptom: reproduce before concluding it is a QEMU limitation.