test(ai): answer both switch readers from one source in flag mocks

Eight files mocked `@/lib/feature-flags` with `getAssistantFlags` alone.
The capability loader reads `loadAssistantSwitches`, so wherever it ran
it threw and every capability resolved to `check_failed`. A shared
factory now points both readers at the function the test drives.

Only the config-change chain reached the loader. Its fixture means AI
switched off by the operator; it now supplies the loader's other inputs
and asserts `operator_disabled` on the digest, so a starved loader turns
it red instead of passing on the failure path.

src/app/api/dashboard/summary/__tests__/module-gate.test.ts

@@ -36,10 +36,12 @@ vi.mock("@/lib/modules/operator-availability", async (importOriginal) => {
     >();
   return { ...actual, getOperatorModuleAvailability: vi.fn() };
 });
-vi.mock("@/lib/feature-flags", async (importOriginal) => {
-  const actual = await importOriginal<typeof import("@/lib/feature-flags")>();
-  return { ...actual, getAssistantFlags: vi.fn() };
-});
+vi.mock("@/lib/feature-flags", async (importOriginal) => ({
+  ...(await importOriginal<typeof import("@/lib/feature-flags")>()),
+  ...(
+    await import("@/__tests__/helpers/assistant-switches-mock")
+  ).mockAssistantSwitches(vi.fn()),
+}));
 // The heavy builder is stubbed: this test pins the GATE, not the aggregate
 // (which has its own coverage). It returns one card per emitted kind.
 vi.mock("@/lib/cache/server-cache", () => ({