[jayleaton]

Follow-up copy fixes from the #227 captures. The editor-exit bug (item 1) is not fixed here; see below.

• Type row. The schedule type row was "Runs", which was easy to confuse with "Runs on" (machine) and "Runs go to" (destination). It's now Type, inside the Schedule section. "Runs on", "Runs go to" and "Run in" (workspace) are accurate and unchanged.
• Next run. Runs within the coming week read "Next run Monday at 9:00 am" instead of "Next run next Monday at …". A run exactly a week out shares today's weekday, so it shows the date instead.
• One zone style. The editor's Time row used the generic zone name ("Thailand Time", "PT"), while the list used the short name at the run time ("GMT+7", "PST"). scheduleTimeZoneLabel now always uses the short name, at the run time when there is one and now otherwise. Web and mobile both call it, so the editor and the list agree on both clients.

Not in this PR: the editor can't be left with unsaved changes

I couldn't find the cause by reading the code, and there's no iOS Simulator on this machine to test against, so I haven't shipped a guess. What I checked:

• React Navigation's path works in principle. goBack() emits beforeRemove on the editor route, usePreventRemove calls preventDefault() and then the callback, and the callback calls Dialog.alert, which is native Alert.alert.
• No other beforeRemove listener or usePreventRemove sits on that route.
• The custom header honours headerBackVisible: false, and the guarded Back bar button calls navigation.goBack().

The only thing this editor does differently from the two other screens that guard removal (ThreadFilesRouteScreen, ZipArchiveView) is guardBack. That hides native Back, adds a JS bar button, and sets gestureEnabled: false. With the gesture off, an edge swipe can never ask anything. The iOS card stack already blocks a native dismiss for a prevented route and calls goBack() in onNativeDismissPrevented. So dropping guardBack for this screen is the first experiment to try on a device.

Verification

• scheduledTaskTimeZone.test.ts: the updated assertion (generic name → short name) failed first, then passed. The web settings logic and browser tests and the mobile draft tests pass: 68 tests.
• vp run check:changed passes all steps.

Captures still needed, from the MacBook at 390pt, light and dark:

• the editor's Schedule section showing the Type row and its open menu;
• the list with a run 2–6 days out ("Next run Monday at …");
• the editor's Time row on an Asia/Bangkok server, showing GMT+7.

Implemented by Claude Opus 5.5 in Claude Code (T3 Code).

🤖 Generated with Claude Code

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Codex	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Codex	Live turn WebSocket wire	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Codex	Live turn WebSocket decoded	21.5 KiB	21.5 KiB	0 B (0.0%)	24.4 KiB	✅
Codex	Live turn messages	1	1	0 (0.0%)	8	✅
Claude	Total thread wire	5.0 KiB	5.0 KiB	0 B (0.0%)	6.8 KiB	✅
Claude	Thread snapshot wire	3.8 KiB	3.8 KiB	0 B (0.0%)	4.9 KiB	✅
Claude	Live turn WebSocket wire	1.2 KiB	1.2 KiB	0 B (0.0%)	2.0 KiB	✅
Claude	Live turn WebSocket decoded	21.9 KiB	21.9 KiB	0 B (0.0%)	24.4 KiB	✅
Claude	Live turn messages	1	1	0 (0.0%)	8	✅

Baseline: f1747c1 · PR result: 9c7fd8c · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 108.5 KiB
• Claude decoded thread snapshot: 108.8 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.