Implement pagination response handling and related enhancements

- Introduced `@Pagination()` decorator to mark list endpoints for pagination.
- Added `TransformInterceptor` to wrap responses in a standardized format `{ data, meta }`.
- Created pagination-related utility functions and constants for managing pagination logic.
- Defined `PaginationQueryDto` for handling pagination query parameters.
- Established `PaginationMetaDto` for OpenAPI documentation of pagination metadata.
- Updated existing controller and service structures to support pagination in responses.
- Added unit tests for pagination utilities and interceptor to ensure correct functionality.

.cursor/rules/primary-entity.mdc

@@ -0,0 +1,56 @@
+---
+description: Primary/aggregate entities must have status, created_at, updated_at, created_by, updated_by via primaryEntityColumns
+alwaysApply: true
+---
+
+# Primary Entity Audit Fields
+
+## Mandatory
+
+Every **primary / aggregate** table and its domain model MUST include:
+
+| Field | DB | Domain |
+| ----- | -- | ------ |
+| `status` | `text`, default `draft` | `Status` |
+| `created_at` / `updated_at` | `bigint` unix ms | `DateTime` |
+| `created_by` / `updated_by` | `uuid` → `users.id` | `string` userId |
+
+Use `primaryEntityColumns(users)` from `src/database/primary-entity-columns.ts` — do not re-declare these five columns by hand.
+
+Timestamps follow `.cursor/rules/date-time.mdc`. Status follows `.cursor/rules/status.mdc`. Actor ids come from `@CurrentUser()` on write.
+
+```typescript
+export const shipments = pgTable('shipments', {
+  id: uuid('id').defaultRandom().notNull().primaryKey(),
+  name: text('name').notNull(),
+  ...primaryEntityColumns(users),
+})
+```
+
+On create/update in the repository:
+
+```typescript
+const now = DateTime.fromUnixMs(Date.now())
+await db.insert(table).values({
+  ...data,
+  status: (status ?? Status.create(Status.DEFAULT)).value,
+  createdAt: now.value,
+  updatedAt: now.value,
+  createdBy: userId,
+  updatedBy: userId,
+})
+```
+
+## Exemptions
+
+Do **not** require these columns on:
+
+- Auth/session tables (`refresh_tokens`, `revoked_access_tokens`)
+- Junction / child rows that are not first-class resources
+- Existing `users` until an explicit migration adds them
+
+## Forbidden
+
+- Omitting audit fields on a new primary table
+- Using plain `Date` / ISO strings for `created_at` / `updated_at` in domain code
+- Hardcoding actor ids or leaving `created_by` / `updated_by` nullable on primary entities