docs(overhaul): record cathedral GPU recurrence and isolated checks

docs/ISSUES.md

@@ -101,7 +101,7 @@ High-profile indoor authored-light/material path needs comparison with retail.
 blindly raise global ambient or change the owner's brightness settings.
 Evidence: `research/2026-09-01-overhaul/475-owner-regate.md`.
 
-## #477 — AMD watchdog / Vulkan device-loss crash while jumping in Neftet
+## #477 — AMD watchdog / Vulkan device-loss crashes in Neftet and cathedral
 
 **Status:** IN-PROGRESS, crash independently confirmed2026-09-05 on clean
 `de427d2c0`. Owner saw an AMD crash notice while jumping around Neftet.
@@ -135,6 +135,19 @@ The ordinary short Neftet run at1e376c244 exited0/gracefully after12jump-input
 pairs; this is not12proven airborne jumps, a long soak or proof against the
 owner's original recurrence. Owner stability acceptance remains open.
 
+**Owner recurrence2026-09-05 at16:41:** exact green Releaseb2bdb09cc,
+ordinary1280x720 High/MSAA4 run PID7852 crashed in the cathedral. Same
+abandoned-frame ErrorDeviceLost and Windows LiveKernelEvent141; initiating
+GPU fault still unknown. Cathedral reveal had completed. Retail was also
+running, but coexistence is not proven causal; owner closed it for isolated
+tests. Corrected core/sync and separate GPU-assisted cathedral routes both
+completed/exit0/gracefully with no reported validation faults or new Windows
+GPU reports;14PNGs inspected narrow PASS/provisional. The discarded first
+attempt was a script spelling error, owner-authorized forced close, not a GPU
+reproduction. These short instrumented runs do not establish recurrence safety.
+`research/2026-09-01-overhaul/477-cathedral-recurrence.md` records exact paths,
+hashes and boundaries. #477/G4 remain open; FPS deferred, no production change.
+
 ## #476 — First post-resize graphical gate PNG is horizontally corrupted
 
 **Status:** IN-PROGRESS — capture and resize-boundary repairs integrated,