docs: refresh stale test counts (289/126 -> 290/208) across README, SUBMISSION and gallery card

README.md

@@ -188,7 +188,7 @@ python -m venv .venv
 pip install -r requirements-dev.txt
 pip install -e ".[server]"      # [server] adds the API (fastapi + multipart)
 
-pytest                          # 289 offline backend tests
+pytest                          # 290 offline backend tests
 python -m claimscene.cli --case demo --out out
 ```
 
@@ -353,7 +353,7 @@ desynced or over-length video fails the build.
 
 ## Testing & CI
 
-- **289 offline backend tests** (unit / integration / e2e): schema
+- **290 offline backend tests** (unit / integration / e2e): schema
   round-trips and rejection of hallucinated fields, layout determinism and
   contact-geometry properties, golden-file SVG, provenance seal/tamper, real
   B2 adapter against an S3 stub (**+ a presign contract test asserting SigV4 +
@@ -362,7 +362,7 @@ desynced or over-length video fails the build.
   SDK-boundary contract tests, and the **full API chain** through FastAPI's
   TestClient (extract → preview → render → get → verify → playback, honest
   degrade, path sanitisation, 422s).
-- **126 frontend tests** (Vitest): the review-panel edit logic, the
+- **208 frontend tests** (Vitest): the review-panel edit logic, the
   schematic-preview live-update wiring, the ReviewStep centrepiece render, the
   **ExtractProgress** extract-latency UI, the playback-url selection, and the
   in-browser verify pinned to a **golden manifest produced by the real backend**

demo/SUBMISSION.md

@@ -76,7 +76,7 @@ The honest lesson came from the live path. Our contract tests run against the re
 - A measured accuracy number (100% on the committed set) instead of a vibe.
 - A human-in-the-loop review step that is the trust story and the production-readiness story at once.
 - A sealed AI→human approval receipt: the render seals the exact proposed→confirmed scene diff with a recomputable `decision_digest` that self-voids if the confirmed scene drifts, and the whole case re-verifies from stored bytes through a named-check receipt (`GET /cases/{id}/verify`) plus a detached, self-sealed receipt written as its own B2 object.
-- 289 backend tests plus 126 frontend tests, gitleaks, CodeQL and a machine-checkable readiness gate, all runnable with zero credentials.
+- 290 backend tests plus 208 frontend tests, gitleaks, CodeQL and a machine-checkable readiness gate, all runnable with zero credentials.
 
 ### What we learned
 Constrain the model to a vocabulary and let deterministic code do the geometry. Label the generative layer loudly. And make honesty a data structure: a sealed manifest that records sources and disclosure by construction, not a promise in the footer.

demo/assets/src/claimscene-07-architecture.patch.json

@@ -1,23 +1,23 @@
 {
-  "_comment": "Regenerates demo/assets/claimscene-07-architecture.png by patching only the two stale stat numbers in place. Source of truth for the numbers: the latest green CI run's pytest/vitest summary lines (see README.md '289 offline backend tests' / '126 frontend tests'). Run via: python scripts/patch_gallery_card.py demo/assets/src/claimscene-07-architecture.patch.json -- or push a change to this file and let .github/workflows/render-cards.yml do it.",
+  "_comment": "Regenerates demo/assets/claimscene-07-architecture.png by patching only the two stale stat numbers in place. Source of truth for the numbers: the latest green CI run's pytest/vitest summary lines (see README.md '290 offline backend tests' / '208 frontend tests'). Run via: python scripts/patch_gallery_card.py demo/assets/src/claimscene-07-architecture.patch.json -- or push a change to this file and let .github/workflows/render-cards.yml do it.",
   "base_image": "demo/assets/claimscene-07-architecture.png",
   "output_image": "demo/assets/claimscene-07-architecture.png",
   "font": "DejaVuSansMono-Bold",
   "fill_color": [15, 22, 20],
   "text_color": [232, 163, 61],
   "patches": [
     {
-      "note": "backend test count: 181 -> 289 (pytest summary, README.md:190)",
+      "note": "backend test count: 289 -> 290 (pytest summary, README.md:191)",
       "erase_box": [85, 568, 148, 596],
-      "new_text": "289",
+      "new_text": "290",
       "anchor_left_x": 92,
       "ink_top_y": 574,
       "ink_bottom_y": 593
     },
     {
-      "note": "frontend test count: 42 -> 126 (vitest summary, README.md:364)",
+      "note": "frontend test count: 126 -> 208 (vitest summary, README.md:365)",
       "erase_box": [192, 568, 254, 596],
-      "new_text": "126",
+      "new_text": "208",
       "anchor_left_x": 200,
       "ink_top_y": 574,
       "ink_bottom_y": 593