v0.0.11

Changes since v0.0.10

• Coordinator cache now rebuilds on registry changes. Previously the device/entity/tier cache was only built at startup, so a Matter re-commissioning, an entity_id rename, or a device add/remove silently broke state tracking — the device kept being marked down because state changes on the new entity_id no longer mapped back to a device. Required a full HA restart to recover.
• The coordinator now listens to EVENT_ENTITY_REGISTRY_UPDATED and EVENT_DEVICE_REGISTRY_UPDATED with a 5 s debounce, rebuilds the cache, and triggers a refresh.
• Adds async_shutdown() so all listeners (state changes + registry events + pending debounce timer) are detached cleanly on integration unload.