Fix live national source serialization

CHANGELOG.md

@@ -62,6 +62,8 @@
   to the Map UI without removing Mol/Kempen regression workspaces.
 - Local RC-4 validation passed 986 backend tests, frontend typecheck/build,
   the full readiness gate, one Alembic head and complete offline migration SQL.
+- Fixed fail-closed AdminVector GeoJSON serialization for Pandas timestamp
+  properties exposed by the live NGI GeoPackage.
 
 ## Sprint 240 Operational forest, agriculture, nature and soil themes (2026-07-17)
 

backend/tests/test_rc4_national_scope_operator.py

@@ -157,5 +157,6 @@ def test_operator_is_packaged_and_guarded_by_readiness() -> None:
     assert "COPY scripts/provision_belgium_north_sea_scope.py /app/scripts/" in dockerfile
     assert "py_compile scripts/provision_belgium_north_sea_scope.py" in readiness
     assert "/datasets/upload" in source
+    assert "frame.to_json(drop_id=False, default=str)" in source
     assert "from app.models" not in source
     assert "INSERT INTO vector_features" not in source

docs/CODEX_EXECUTION_LOG.md

@@ -25,6 +25,11 @@
   SQL and live-smoke shell syntax.
 - Tower fetch-only, persistence, live coverage API and browser acceptance are
   the remaining RC-4 exit evidence.
+- The first Tower fetch-only run failed closed before API persistence because
+  AdminVector exposes `modifdate` as a Pandas `Timestamp`. GeoJSON
+  serialization was corrected to explicit string conversion and guarded by a
+  regression assertion; downloaded source artifacts remain operator cache
+  only.
 
 - Froze the RC geography as all Belgian land plus the separately labelled
   territorial sea, EEZ and continental shelf.

scripts/provision_belgium_north_sea_scope.py

@@ -251,7 +251,7 @@ def read_adminvector_layers(gpkg_path: Path) -> dict[str, dict[str, Any]]:
             raise RuntimeError(
                 f"NGI layer {layer_name} has {len(frame)} records; expected between {minimum} and {maximum}"
             )
-        payload = json.loads(frame.to_json(drop_id=False))
+        payload = json.loads(frame.to_json(drop_id=False, default=str))
         source_max_modification = None
         if "modifdate" in frame.columns:
             parsed = pandas.to_datetime(frame["modifdate"], errors="coerce", utc=True).dropna()