feat(agent): add voice input capability and standardize tool naming

- Add voice recording with transcribe endpoint (ASR) for multimodal input
- Android: add RECORD_AUDIO and INTERNET permissions
- Refactor tool naming: frontend tools use 'front.' prefix, backend tools use 'back.'
- Migrate calendar tools: create_calendar_event -> back.mutate/list/delete events
- Add calendar_event_list.v1 and calendar_operation.v1 UI card types
- Update all Flutter and Python tests to match new tool naming conventions
- Add record package dependency for voice recording

backend/src/v1/schedule_items/service.py

@@ -202,6 +202,34 @@ class ScheduleItemService(BaseService):
 
         return [self._to_response(item) for item in items]
 
+    async def list_paginated(
+        self,
+        *,
+        page: int,
+        page_size: int,
+    ) -> tuple[list[ScheduleItemResponse], int]:
+        user_id = self.require_user_id()
+        if page < 1:
+            raise HTTPException(status_code=400, detail="page must be >= 1")
+        if page_size < 1 or page_size > 100:
+            raise HTTPException(status_code=400, detail="page_size must be 1..100")
+        try:
+            items, total = await self._repository.list_paginated(
+                user_id,
+                page=page,
+                page_size=page_size,
+            )
+        except SQLAlchemyError:
+            logger.exception(
+                "Failed to list schedule items with pagination",
+                page=page,
+                page_size=page_size,
+            )
+            raise HTTPException(
+                status_code=503, detail="Schedule item store unavailable"
+            )
+        return [self._to_response(item) for item in items], total
+
     async def share(
         self, item_id: UUID, request: ScheduleItemShareRequest
     ) -> ScheduleItemShareResponse: