diff --git a/.isu/issues.json b/.isu/issues.json
index 57bc134..e9dff44 100644
--- a/.isu/issues.json
+++ b/.isu/issues.json
@@ -1,5 +1,5 @@
 {
-  "next_id": 419,
+  "next_id": 421,
   "issues": [
     {
       "id": 1,
@@ -5134,6 +5134,21 @@
       "author": "piefev",
       "state": "open",
       "created_at": "2026-07-18T13:54:25Z"
+    },
+    {
+      "id": 420,
+      "repo": "we",
+      "title": "Opera parent parity still blocked after clean baseline audit",
+      "body": "Parent: isu issue 281.\n\nA clean implementor-branch audit of crates/e2e/scenarios/real-web/opera.com.we on 2026-07-18 confirmed that the parent Opera Chromium-parity issue cannot be closed yet. The scenario still has to keep its xfail because both Chromium screenshot assertions fail while DOM and interactivity assertions pass.\n\nRepro:\n  cargo run -p we-e2e -- --scenario crates/e2e/scenarios/real-web/opera.com.we --out-dir crates/e2e/artifacts\n\nCurrent clean-run result:\n- desktop L96: 89.29% match (131579/1228500 px differ, tol=4, max_diff=0.1000%)\n- mobile L111: 82.71% match (56920/329160 px differ, tol=4, max_diff=0.1000%)\n- DOM assertions and the footer language input interactivity assertion pass.\n- Console output is limited to expected offline third-party GTM fetch misses.\n\nRegion audit at harness tolerance 4:\n- desktop header: 5104/150150 px differ (3.40%)\n- desktop hero text: 30822/200100 px differ (15.40%)\n- desktop cookie panel: 43799/258000 px differ (16.98%)\n- desktop hero image region: 63834/436800 px differ (14.61%)\n- mobile cookie panel: 38316/234000 px differ (16.37%)\n- mobile hero peek: 19014/34560 px differ (55.02%)\n\nInvestigation notes:\n- The page structure, header, CTA, cookie modal, hero text, and first-viewport imagery are visibly aligned with the Chromium goldens.\n- Raising the analysis-only tolerance as high as 64 still leaves desktop at 60417 differing pixels (4.92%) and mobile at 30840 differing pixels (9.37%), so this is not only a small alpha/tolerance miss.\n- A local experiment using mipmapped Metal image textures and the existing mipmap sampler produced identical desktop/mobile counts, so image minification filtering is not the current lever. That experiment was backed out.\n- Remaining blockers are still the known independent gaps: first-viewport WebP/image parity, glyph antialiasing/raster coverage, and cookie-panel edge/text fidelity.\n\nAcceptance: keep crates/e2e/scenarios/real-web/opera.com.we xfailed until the independent image/text/render blockers are fixed and isu issue 281 can remove the xfail with both Chromium screenshot assertions passing at the committed threshold.",
+      "labels": [
+        "real-web",
+        "text",
+        "render"
+      ],
+      "assigned": [],
+      "author": "piefev",
+      "state": "open",
+      "created_at": "2026-07-18T14:09:15Z"
     }
   ]
 }