Implement Org Hierarchy Scope Selection in BLoC

Implement scope propagation by exposing the selected org scope as a Riverpod StateProvider derived from the AdminPortalBloc state stream. Downstream providers watch this derived provider and automatically re-fetch when it changes. Use riverpod's ref.watch(orgScopeProvider) pattern in each downstream provider's build method — this ensures automatic invalidation without manual coordination. For debouncing, use a Timer in the BLoC's event handler that cancels any pending scope change before scheduling the new one.

The org hierarchy data (available nodes for the current admin) should be loaded once on BLoC init and cached in state — do not re-fetch the hierarchy on every scope change. Validate the selected node against the cached hierarchy list in the SelectOrgScope handler before emitting the new state. For NHF's complex hierarchy (12 landsforeninger, 9 regioner, 1400 lokallag), ensure the scope propagation model is lazy — only load child data when a scope is actually selected, not upfront.

BLoC unit tests using bloc_test. Cover: (1) SelectOrgScope with valid org-level node, (2) SelectOrgScope with valid region-level node, (3) SelectOrgScope with valid chapter-level node, (4) SelectOrgScope with unauthorized node emits error, (5) ClearOrgScope resets to null scope, (6) scope change resets pagination state, (7) rapid scope changes debounced (emit only final state). Widget integration tests: mount AdminDashboardScreen in a ProviderScope with mock providers, fire a scope change, assert all provider data refreshes. Use Riverpod's ref.invalidate() pattern in downstream providers and verify it's called on scope change.