docs: reconcile rehearsal gate findings

docs/current_roadmap.md

@@ -152,8 +152,9 @@ Implementation order:
     [`presentation live/replay truth plan`](historical/superpowers/plans/2026-07-09-presentation-live-replay-truth.md).
 20. Completed: Scene 10 now presents factual run state: workflow input,
     interrupt payload, operator resume decision, output, and trace frame facts.
-    Request revision resumes the same run through the negative outcome branch;
+    The live workflow contract supports a same-run negative outcome; the
-    it does not create issues. Implementation:
+    current prepared revision recording uses a separate run identity and is
+    documented as a factual follow-up. Implementation:
     [`Scene 10 factual run state`](historical/superpowers/plans/2026-07-09-scene-10-factual-run-state.md).
 21. Completed: presentation lifecycle story expansion splits the demo climax
     into prepared lifecycle, run start, typed human boundary, and

docs/historical/superpowers/plans/2026-07-13-defense-rehearsal-gate.md

@@ -348,17 +348,18 @@
 #### Task 6 completion record
 
 - Focused presentation checks: 2 files, 68 tests passed.
-- Full web tests: RPC 52 passed and 6 skipped; console 735 passed and 2
+- Full web tests: RPC 52 passed and 6 skipped; console 737 passed across 92
-  failed across 92 files. The failures are a thesis-route timeout and a
+  files after the route-contract stabilization follow-up in
-  duplicate-heading query in `PresentationRoute.test.tsx`.
+  `c82926fc`.
 - Typecheck: passed for console, RPC, and server.
 - Build: passed. Vite emitted its known chunk-size warning for the 832.24 kB
   minified JavaScript chunk over the 500 kB threshold.
 - Diff check: `git diff --check` passed.
 - Screenshot evidence: 42 manifest routes and 42 PNGs at each of
-  `1280x720` and `1024x768`; representative title and approval captures were
+  `1280x720` and `1024x768`; representative title, lifecycle, approval,
-  inspected at both sizes. The log remains authoritative for route-wide
+  output, trace, and architecture captures were inspected at both sizes. The
-  operator observations.
+  log remains authoritative for route-wide operator observations; a complete
+  manual visual inspection of all 84 images remains a rehearsal follow-up.
 - Review outcome: matrix, log, and story-audit links were checked; the
   existing FACTUAL replay run-identity deviation and BLOCKED live end-to-end
   path remain open. No production code was changed for Task 6.

docs/runbooks/presentation-visual-review.md

@@ -15,7 +15,7 @@ Review these routes at `1024x768`, `1200x800`, and `1280x720`, always at 100%
 browser zoom:
 
 1. `/present#scene/architecture/node-use/focus/node-use`
-2. `/present#scene/interrupt-evidence/trace`
+2. `/present#scene/resume-output-evidence/trace`
 
 The logical canvas widths must be `960`, `1080`, and `1280` respectively.
 Capture the receipt state and open-inspector state at each viewport. Confirm

web/apps/console/src/presentation/PresentationRoute.test.tsx

@@ -97,8 +97,10 @@ describe("PresentationRoute", () => {
       const { PresentationRoute } = await import("./PresentationRoute.js");
       render(<PresentationRoute />);
 
-      const headings = await screen.findAllByRole("heading", { name: heading });
+      await waitFor(
-      expect(headings.length).toBeGreaterThan(0);
+        () => expect(screen.getAllByRole("heading", { name: heading })).toHaveLength(1),
+        { timeout: 12000 },
+      );
       const footer = screen.getByRole("contentinfo", { name: /presentation footer/i });
       const action = within(footer).queryByRole("button", {
         name: /run prepared workflow|play replay walkthrough/i,