diff --git a/frontend/src/components/Home.dom.test.tsx b/frontend/src/components/Home.dom.test.tsx
index 26f91987..c19f8d3e 100644
--- a/frontend/src/components/Home.dom.test.tsx
+++ b/frontend/src/components/Home.dom.test.tsx
@@ -430,11 +430,11 @@ describe("Home", () => {
     await screen.findByText("Hero v1");
     fireEvent.click(screen.getByText("home.hero.reroll"));
     await waitFor(() => expect(discoverPick).toHaveBeenCalledTimes(1));
-    await act(async () => {
-      failPick(new Error("boom"));
-      await new Promise((r) => setTimeout(r, 0));
-    });
-    expect(rerollBtn().disabled).toBe(false); // the failure itself releases the button
+    // `waitFor`, not a fixed tick: react-query hands the rejection to the observer through
+    // `notifyManager`'s own `setTimeout(0)`, registered AFTER any timer this test schedules, so
+    // "await one macrotask then assert" races it. Round 4 measured that race as an intermittent red.
+    failPick(new Error("boom"));
+    await waitFor(() => expect(rerollBtn().disabled).toBe(false)); // the failure releases the button
 
     // Now an unrelated action refetches that errored key in the background.
     discoverPick.mockReturnValue(new Promise(() => {}));
@@ -475,12 +475,18 @@ describe("Home", () => {
         failPick = reject;
       }),
     );
+    const rerollBtn = () =>
+      screen.getByText("home.hero.reroll").closest("button") as HTMLButtonElement;
     fireEvent.click(screen.getByText("home.hero.reroll"));
     await waitFor(() => expect(discoverPick).toHaveBeenCalledTimes(2));
-    await act(async () => {
-      failPick(new Error("boom"));
-      await new Promise((r) => setTimeout(r, 0));
-    });
+    expect(rerollBtn().disabled).toBe(true); // the press is outstanding
+
+    failPick(new Error("boom"));
+    // Wait for a signal that the REJECTION has landed, not for a fixed tick: react-query notifies
+    // the observer through its own `setTimeout(0)`, and the empty state is already on screen while
+    // the press is in flight — so asserting it straight away would pass with or without the fix.
+    // The button releasing is that signal (`isFetched` flips once the key settles, error included).
+    await waitFor(() => expect(rerollBtn().disabled).toBe(false));
     expect(screen.getByText("home.hero.emptyBudget")).toBeTruthy();
     expect(screen.queryByText("Hero v1")).toBeNull();
   });
diff --git a/frontend/src/components/NavigationProvider.dom.test.tsx b/frontend/src/components/NavigationProvider.dom.test.tsx
index 5f2a77ff..dc6e38a8 100644
--- a/frontend/src/components/NavigationProvider.dom.test.tsx
+++ b/frontend/src/components/NavigationProvider.dom.test.tsx
@@ -11,12 +11,13 @@
 // `return` fails tsc — but round 3 measured that swapping the two values (early return `true`, tail
 // `false`) passes typecheck and all 822 tests. Hence assertions on the value AND on the effects the
 // early return skips, which is what the boolean actually stands for.
-import { describe, it, expect, beforeEach, vi } from "vitest";
-import { renderHook, act } from "@testing-library/react";
+import { describe, it, expect, afterEach, beforeEach, vi } from "vitest";
+import { renderHook, act, cleanup } from "@testing-library/react";
 import type { ReactNode } from "react";
 
-// Partial: the provider only wants `t` for the document title, but the i18n module this pulls in
-// initialises against the real `initReactI18next`.
+// Partial: the provider reads `t` AND `i18n.language` (both feed the document-title effect), and the
+// import graph pulls in `src/i18n`, which initialises against the real `initReactI18next` — so a
+// wholesale mock throws before any test runs.
 vi.mock("react-i18next", async (importOriginal) => {
   const actual = await importOriginal<typeof import("react-i18next")>();
   return {
@@ -38,6 +39,9 @@ beforeEach(() => {
   sessionStorage.setItem("siftlode.activeAccount", "1"); // account-scoped page memory
   window.history.replaceState({ sfPage: "feed" }, "", "/");
 });
+// Not automatic: vitest runs without `globals`, so RTL never registers its own afterEach — leaving a
+// provider (and its popstate listener) mounted for the next case. Same reason Home.dom.test.tsx does.
+afterEach(cleanup);
 
 describe("NavigationProvider · setPage reports whether it navigated", () => {
   it("returns true and pushes an entry for a real page change", () => {
@@ -77,9 +81,15 @@ describe("NavigationProvider · setPage reports whether it navigated", () => {
   });
 
   it("returns true even for the current page when a channel page is over it", () => {
-    // The nav rail's "Feed" while a channel opened from the feed is a real navigation: it closes the
-    // channel page. The `&& !channelViewRef.current` half of the guard is what makes it one.
+    // The nav rail's own module while a channel page is open over it is a real navigation: it closes
+    // the channel page. The `&& !channelViewRef.current` half of the guard is what makes it one.
     const { result } = render();
+    // The page under test must be the one that is ALREADY current, or the guard's first conjunct is
+    // false and this never reaches the clause it is named for — the mistake round 4 caught here: a
+    // bare URL with cleared storage lands on "home" (navigationInit's E24 default), not "feed", so
+    // asking for "feed" was just an ordinary page change and deleting the clause left it green.
+    const current = result.current.page;
+    expect(current).toBe("home");
     act(() => {
       result.current.openChannel("UC1", "A channel");
     });
@@ -89,7 +99,7 @@ describe("NavigationProvider · setPage reports whether it navigated", () => {
     try {
       let reported: boolean | undefined;
       act(() => {
-        reported = result.current.setPage("feed");
+        reported = result.current.setPage(current);
       });
       expect(reported).toBe(true);
       expect(push).toHaveBeenCalledTimes(1);
diff --git a/frontend/src/components/ui/CommandPalette.ordering.dom.test.tsx b/frontend/src/components/ui/CommandPalette.ordering.dom.test.tsx
index 6313e4b3..81cf801e 100644
--- a/frontend/src/components/ui/CommandPalette.ordering.dom.test.tsx
+++ b/frontend/src/components/ui/CommandPalette.ordering.dom.test.tsx
@@ -1,8 +1,10 @@
 // @vitest-environment jsdom
 //
 // ONE test, ONE file, on purpose. This is the close→reclaim→run ordering contract of the palette
-// (see closeThenNavigate), and it is the only assertion in the suite that drives a real
-// useBackToClose entry through a real programmatic history.back().
+// (see closeThenNavigate), and it ASSERTS on a real useBackToClose entry driven through a real
+// programmatic history.back() — which is why it cannot share module state with anything. (It is not
+// the only place in the suite where an unstubbed back() fires; history.burial.dom.test.ts is the
+// other file that asserts on one, and any test rendering a real Modal can let one go.)
 //
 // It cannot live in CommandPalette.dom.test.tsx. jsdom gives a test FILE a single session history,
 // and history.ts's overlay bookkeeping is module-global, so the sibling tests there leave an
diff --git a/frontend/src/lib/history.burial.dom.test.ts b/frontend/src/lib/history.burial.dom.test.ts
index 17a4846a..8b951b0c 100644
--- a/frontend/src/lib/history.burial.dom.test.ts
+++ b/frontend/src/lib/history.burial.dom.test.ts
@@ -1,10 +1,10 @@
 // @vitest-environment jsdom
 //
-// ONE test, ONE file, for the same reason CommandPalette.ordering.dom.test.tsx has its own — it is
-// the other assertion in the suite that lets a reconcile issue a REAL programmatic history.back()
-// (unstubbed), and history.ts's overlay bookkeeping is module-global. history.dom.test.ts states its
-// own invariant — every layer there is closed by a genuine Back or a stubbed back(), so nothing real
-// fires and the shared state stays clean between its tests — and this test would break it.
+// ONE test, ONE file, for the same reason CommandPalette.ordering.dom.test.tsx has its own: it
+// ASSERTS on a reconcile issuing a REAL, unstubbed programmatic history.back(), and history.ts's
+// overlay bookkeeping is module-global. history.dom.test.ts keeps its shared state clean by never
+// leaving a real back() in flight — its layers close by a genuine Back, a stubbed back(), or a
+// consumed burial mark — and this test would break that.
 import { describe, it, expect, vi } from "vitest";
 import { renderHook, act } from "@testing-library/react";
 import { navigateOverOverlays, useBackToClose } from "./history";
diff --git a/frontend/src/lib/history.dom.test.ts b/frontend/src/lib/history.dom.test.ts
index 02d37d7d..4cf0c926 100644
--- a/frontend/src/lib/history.dom.test.ts
+++ b/frontend/src/lib/history.dom.test.ts
@@ -7,8 +7,12 @@
 // A GENUINE user Back is simulated by dispatching `popstate` directly: the real-Back branch in
 // history.ts does NOT call history.back() (it only decrements its own historyDepth to match the
 // entry the browser already dropped), so a raw popstate reproduces it deterministically without
-// depending on jsdom's async back() semantics. Each test closes its layers via genuine Backs, so
-// no reconcile-driven history.back() runs and the shared module state stays clean between tests.
+// depending on jsdom's async back() semantics. The invariant that keeps the module-global
+// bookkeeping clean between tests here is NOT "only genuine Backs" — two closeThenNavigate cases
+// deliberately drive a reconcile into history.back() — it is that no REAL back() is ever left in
+// flight: a layer is closed by a genuine Back, by a STUBBED back() that lands its own popstate, or
+// by an unmount whose burial mark is consumed. A test that needs an unstubbed one belongs in its own
+// file (see history.burial.dom.test.ts).
 import { describe, it, expect, vi, afterEach } from "vitest";
 import { renderHook, act } from "@testing-library/react";
 import { closeThenNavigate, navigateOverOverlays, useBackToClose, useBackToExit } from "./history";
diff --git a/frontend/src/lib/history.ts b/frontend/src/lib/history.ts
index 267bc151..f9b03510 100644
--- a/frontend/src/lib/history.ts
+++ b/frontend/src/lib/history.ts
@@ -171,17 +171,23 @@ let listenerInstalled = false;
  *  open" navigated nowhere yet still buried the player's entry. The player's unmount then consumed
  *  the mark instead of issuing history.back(), orphaning its browser entry — and the next Back
  *  popped onto an identical `sfPage` with an empty overlay stack, i.e. a press that did nothing
- *  (round 2). One rule for both entry points below, because both take the same no-op navigation.
+ *  (round 2). One rule for both entry points below — but read the next paragraph before trusting it
+ *  at the `closeThenNavigate` one, where most callers cannot report at all.
  *
- *  It fails OPEN, and that is a deliberate limit rather than an oversight: only a literal `false`
- *  skips the mark, so a `navigate` that reports nothing is assumed to have pushed — which is what
- *  the code did unconditionally before, so nothing regressed by adding this. The gap it leaves is
- *  `closeThenNavigate`'s palette callers: `CommandItem.run` is typed `() => void`, so only the rows
- *  written as a bare `() => setPage(page)` forward the boolean at all, while the block-bodied and
- *  non-navigating rows (surprise, playlists, theme, the cheat sheet) still mark a live sibling
- *  overlay's entry buried under a push that never happened. Closing that overlay by its own control
- *  then costs one dead browser-Back press. Deferred, with the fix (widen `CommandItem.run` and have
- *  each row report) on ROADMAP.md — this rule is complete only for `navigateOverOverlays`.
+ *  It fails OPEN: only a literal `false` skips the mark, so a `navigate` that reports nothing is
+ *  assumed to have pushed. That is what the code did unconditionally before, so nothing regressed by
+ *  adding the rule — it just does not yet reach everywhere. Today only `navigateOverOverlays`
+ *  (App's Shift+W → `setPage`) is fully covered. Of `closeThenNavigate`'s callers, only a palette
+ *  row written as a bare `() => setPage(page)` forwards the boolean: `CommandItem.run` is typed
+ *  `() => void`, so the block-bodied and non-navigating rows report nothing — and neither does
+ *  `openChannel`, which has its OWN no-push branch (it `replaceState`s when the entry already
+ *  carries `_chan`) and is reached both from a palette row and from PlayerModal's channel-name jump.
+ *  Each of those still marks a live sibling overlay's entry buried under a push that never happened,
+ *  costing one dead browser-Back press when that overlay is closed by its own control.
+ *
+ *  Closing it is three changes, not one — widen `CommandItem.run` to `() => boolean | void`, return
+ *  the verdict from the block-bodied rows, and give `openChannel` the boolean `setPage` now has — so
+ *  it is deferred to the project backlog rather than bolted on here.
  *
  *  Marking AFTER the call is safe: `navigate()` is synchronous and only `reconcileHistory` (a
  *  microtask) and the popstate listener ever move `historyDepth`. */