feat: rename array orchestrators, add docker_update remainder mode

- Rename array_start.sh → array_started.sh, array_stop.sh → array_stopping.sh
  to clarify these are event-driven (array has started/is stopping), not imperative
- Update all references across 9 files (master.conf, user_script_plug-in.sh,
  watchdogs, continuous_scripts_status.sh, claude_startup.sh)
- Add --remainder mode to docker_update.sh: updates all running containers
  excluding daily containers, weekly sync-window containers (emby + critical-data),
  and fallback coverage containers (FALLBACK_${MY_ID}_COVERS_${REMOTE_ID}_TIER*)
  Fallback containers excluded because the remote server owns their version —
  independent updates risk writeback incompatibility on handback
- weekly_sync_maintenance.sh calls docker_update.sh --remainder as final step
- git_pull_execute.sh: add safe.directory config to fix dubious ownership error
  when running as root on a directory owned by uid 1000

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Docker_Essentials/docker_container_stop.sh

@@ -3,7 +3,7 @@
 # ============================= Docker Container Stop ==========================================
 # ==============================================================================================
 # Stops all running Docker containers one at a time, verifying each is stopped before
-# moving to the next. Called by array_stop.sh as part of a planned shutdown sequence.
+# moving to the next. Called by array_stopping.sh as part of a planned shutdown sequence.
 #
 # ── STOP SEQUENCE PER CONTAINER ──────────────────────────────────────────────────────────────
 #   1. docker stop -t 30 (SIGTERM + 30s grace period — docker sends SIGKILL if needed)