[codex] Split app playtest failure buckets

qa/release_readiness.py

@@ -38,6 +38,21 @@
 
 
 REQUIRED_RELEASE_PERSONAS = ["newbie", "veteran", "adversarial", "narrative", "optimizer"]
+RELEASE_VERDICT_GATE = "full_five_persona_rri"
+GATE_SPLIT_CONTRACT = {
+    "deterministic_built_app_smoke": {
+        "scope": "fast built-app wiring proof with deterministic provider",
+        "release_verdict": False,
+    },
+    "short_real_provider_playtest": {
+        "scope": "short built-app proof with a real provider and provider trace evidence",
+        "release_verdict": False,
+    },
+    RELEASE_VERDICT_GATE: {
+        "scope": "non-partial five-persona release readiness verdict",
+        "release_verdict": True,
+    },
+}
 
 
 def looks_like_path(value: str) -> bool:
@@ -323,6 +338,8 @@ def main() -> int:
     result = {
         "rri": rri,
         "release_ready": release_ready,
+        "release_verdict_gate": RELEASE_VERDICT_GATE,
+        "gate_split_contract": GATE_SPLIT_CONTRACT,
         "partial": bool(missing_personas or evidence_gaps),
         "harness_contaminated": bool(missing_personas or harness_failures or evidence_gaps),
         "expected_personas": expected_personas,

qa/test_macos_app_static.py

@@ -111,6 +111,39 @@ def test_scripted_provider_is_dev_gated_and_model_free(self):
         self.assertNotIn("claude -p", script)
         self.assertNotIn("codex -p", script)
 
+    def test_built_app_playtest_emits_split_failure_buckets(self):
+        harness = self.read("qa/ui_playtest_app.sh")
+
+        for bucket in (
+            "no_app",
+            "no_launcher",
+            "no_provider",
+            "no_art",
+            "no_actor",
+            "no_actions",
+            "move_rejected",
+            "no_narration",
+            "console_error",
+            "permission_prompt",
+            "score_failed",
+        ):
+            self.assertIn(f'"{bucket}"', harness)
+
+        self.assertIn("PART_A_FAILURE_BUCKET", harness)
+        self.assertIn("PART_A_FAILURE_DETAIL", harness)
+        self.assertIn("PART_B_FAILURE_BUCKET", harness)
+        self.assertIn("PART_B_FAILURE_DETAIL", harness)
+        self.assertIn("classify_native_failure", harness)
+        self.assertIn("classify_part_b_readiness_failure", harness)
+        self.assertIn("classify_part_b_failure_from_artifacts", harness)
+        self.assertIn("classify_part_b_score_failure", harness)
+        self.assertIn("failure_bucket", harness)
+        self.assertIn("failure_detail", harness)
+        self.assertIn("original_result", harness)
+        self.assertIn("move_rejected", harness)
+        self.assertIn("console_error", harness)
+        self.assertIn("score_failed", harness)
+
     def test_provider_viewer_stays_attached_during_native_restarts(self):
         root_view = self.read("macos/WorldOSApp/Sources/WorldOSApp/Views/RootView.swift")
         app_process = self.read("macos/WorldOSApp/Sources/WorldOSApp/Services/AppProcessService.swift")

qa/test_release_readiness.py

@@ -74,6 +74,10 @@ def test_missing_expected_persona_score_marks_partial_and_fails_release(self):
 
             self.assertEqual(rc, 1)
             self.assertFalse(payload["release_ready"])
+            self.assertEqual(payload["release_verdict_gate"], "full_five_persona_rri")
+            self.assertFalse(payload["gate_split_contract"]["deterministic_built_app_smoke"]["release_verdict"])
+            self.assertFalse(payload["gate_split_contract"]["short_real_provider_playtest"]["release_verdict"])
+            self.assertTrue(payload["gate_split_contract"]["full_five_persona_rri"]["release_verdict"])
             self.assertTrue(payload["partial"])
             self.assertTrue(payload["harness_contaminated"])
             self.assertEqual(payload["expected_personas"], ["newbie", "veteran"])

qa/test_ui_playtest_app_buckets.py

@@ -0,0 +1,71 @@
+import json
+import subprocess
+import tempfile
+import unittest
+from pathlib import Path
+
+
+ROOT = Path(__file__).resolve().parents[1]
+
+
+class UIPlaytestAppBucketFixtureTests(unittest.TestCase):
+    def run_classifier(self, body: str) -> str:
+        script = f"""
+set -euo pipefail
+funcs="$(mktemp)"
+trap 'rm -f "$funcs"' EXIT
+sed -e 's#^ROOT=.*#ROOT="$(pwd)"; cd "$ROOT" || exit 1#' -e '/^# DRIVE$/q' qa/ui_playtest_app.sh > "$funcs"
+grep -qx 'ROOT="$(pwd)"; cd "$ROOT" || exit 1' "$funcs" || {{ echo "missing rewritten ROOT= sentinel" >&2; exit 1; }}
+grep -qx '# DRIVE' "$funcs" || {{ echo "missing # DRIVE sentinel" >&2; exit 1; }}
+grep -q '^classify_part_b_score_failure()' "$funcs" || {{ echo "missing classifier functions before # DRIVE" >&2; exit 1; }}
+source "$funcs" >/dev/null
+{body}
+"""
+        proc = subprocess.run(
+            ["bash", "-c", script],
+            cwd=ROOT,
+            text=True,
+            capture_output=True,
+            check=False,
+        )
+        self.assertEqual(proc.returncode, 0, proc.stderr)
+        return proc.stdout.strip()
+
+    def test_part_b_score_failure_gets_explicit_bucket(self):
+        with tempfile.TemporaryDirectory() as td:
+            score = Path(td) / "score.json"
+            score.write_text(
+                json.dumps(
+                    {
+                        "pass": False,
+                        "completed_intro_flow": True,
+                        "reached_play_screen": True,
+                        "bug_reports_critical": 0,
+                        "console_errors": 0,
+                        "persona_satisfaction": 5,
+                    }
+                ),
+                encoding="utf-8",
+            )
+
+            out = self.run_classifier(f'classify_part_b_score_failure "{score}"')
+
+        self.assertEqual(out, "score_failed|score.json failed: satisfaction=5/10")
+
+    def test_part_b_artifact_classifier_prefers_console_error(self):
+        with tempfile.TemporaryDirectory() as td:
+            run = Path(td)
+            player = run / "player"
+            player.mkdir()
+            (player / "console.ndjson").write_text(
+                json.dumps({"type": "pageerror", "text": "Uncaught ReferenceError"}) + "\n",
+                encoding="utf-8",
+            )
+
+            out = self.run_classifier(f'classify_part_b_failure_from_artifacts "{run}" FAIL')
+
+        self.assertEqual(out, "console_error|browser console/page error recorded during app playtest")
+
+
+if __name__ == "__main__":
+    unittest.main()