diff --git a/repositories/user_address/user_address_csv_s3_repository.py b/repositories/user_address/user_address_csv_s3_repository.py
index 9b93b638..058fd5a5 100644
--- a/repositories/user_address/user_address_csv_s3_repository.py
+++ b/repositories/user_address/user_address_csv_s3_repository.py
@@ -53,6 +53,9 @@ class UserAddressCsvS3Repository(UserAddressRepository):
             {**addr.source_row, _POSTCODE_CLEAN_COLUMN: str(addr.postcode)}
             for addr in addresses
         ]
+
+        # TODO: [New Starter Task] file_name generation can be standardised
+        # and also easier to read, test for future implementation. Buiild that!
         filename = (
             f"{datetime.now(timezone.utc).isoformat()}_{uuid.uuid4().hex[:8]}.csv"
         )
diff --git a/repositories/user_address/user_address_repository.py b/repositories/user_address/user_address_repository.py
index 170f34dd..b2c0f866 100644
--- a/repositories/user_address/user_address_repository.py
+++ b/repositories/user_address/user_address_repository.py
@@ -7,9 +7,7 @@ from domain.addresses.user_address import UserAddress
 
 class UserAddressRepository(ABC):
     @abstractmethod
-    def load_batch(self, s3_uri: str) -> list[UserAddress]:
-        ...
+    def load_batch(self, s3_uri: str) -> list[UserAddress]: ...
 
     @abstractmethod
-    def save_batch(self, addresses: list[UserAddress], path_prefix: str) -> str:
-        ...
+    def save_batch(self, addresses: list[UserAddress], path_prefix: str) -> str: ...
diff --git a/tests/conftest.py b/tests/conftest.py
new file mode 100644
index 00000000..0a246372
--- /dev/null
+++ b/tests/conftest.py
@@ -0,0 +1,48 @@
+"""Shared pytest fixtures for the ``tests/`` tree.
+
+Provides an ephemeral PostgreSQL engine for tests that exercise SQLModel
+repositories. PostgreSQL has no true in-memory mode; ``pytest-postgresql``
+starts a real, throwaway server in a temp directory (the process is started
+once per session and a fresh database is created/dropped per test). That is
+the closest equivalent to "in-memory" and matches production behaviour far
+better than SQLite (enums, JSONB, constraint semantics, etc.).
+"""
+
+from __future__ import annotations
+
+import glob
+from collections.abc import Iterator
+from typing import Any
+
+import pytest
+from psycopg import Connection
+from pytest_postgresql import factories
+from sqlalchemy import Engine
+from sqlmodel import SQLModel, create_engine
+
+# Importing the SQLModel row modules registers their tables on
+# SQLModel.metadata so ``create_all`` builds the full schema. Imports look
+# unused; they aren't.
+
+
+# pg_ctl ships under a versioned path and is not on PATH in the dev container.
+_PG_CTL = next(iter(sorted(glob.glob("/usr/lib/postgresql/*/bin/pg_ctl"))), "pg_ctl")
+
+postgresql_proc = factories.postgresql_proc(
+    executable=_PG_CTL
+)  # pyright: ignore[reportUnknownMemberType]
+postgresql = factories.postgresql("postgresql_proc")
+
+
+@pytest.fixture
+def db_engine(postgresql: Connection[Any]) -> Iterator[Engine]:
+    """A SQLModel engine bound to a fresh, ephemeral PostgreSQL database."""
+    info = postgresql.info
+    url = f"postgresql+psycopg://{info.user}:@{info.host}:{info.port}/{info.dbname}"
+    engine = create_engine(url)
+    SQLModel.metadata.create_all(engine)
+    try:
+        yield engine
+    finally:
+        SQLModel.metadata.drop_all(engine)
+        engine.dispose()
diff --git a/tests/domain/addresses/test_postcode_batching.py b/tests/domain/addresses/test_postcode_batching.py
index c69722ba..8ffcf1b5 100644
--- a/tests/domain/addresses/test_postcode_batching.py
+++ b/tests/domain/addresses/test_postcode_batching.py
@@ -15,12 +15,16 @@ def _addrs(postcode: str, n: int) -> list[UserAddress]:
 
 
 def test_empty_input_yields_no_batches() -> None:
+    # act / assert
     assert list(iter_postcode_grouped_batches([])) == []
 
 
 def test_single_batch_under_cap() -> None:
+    # arrange
     addrs = _addrs("AA1 1AA", 3) + _addrs("BB2 2BB", 2)
+    # act
     batches = list(iter_postcode_grouped_batches(addrs, max_batch_size=500))
+    # assert
     assert len(batches) == 1
     assert batches[0] == addrs
 
@@ -28,8 +32,11 @@ def test_single_batch_under_cap() -> None:
 def test_multiple_postcodes_packed_into_one_batch_up_to_cap() -> None:
     # Two groups whose total exactly equals the cap pack into a single
     # batch -- no premature flush.
+    # arrange
     addrs = _addrs("AA1 1AA", 3) + _addrs("BB2 2BB", 2)
+    # act
     batches = list(iter_postcode_grouped_batches(addrs, max_batch_size=5))
+    # assert
     assert len(batches) == 1
     assert len(batches[0]) == 5
 
@@ -37,8 +44,11 @@ def test_multiple_postcodes_packed_into_one_batch_up_to_cap() -> None:
 def test_flush_on_overflow_before_adding_next_postcode() -> None:
     # Cap is 5. First group fills 3 slots; second group of 3 would overflow,
     # so the buffer is flushed first and the next group starts a fresh batch.
+    # arrange
     addrs = _addrs("AA1 1AA", 3) + _addrs("BB2 2BB", 3)
+    # act
     batches = list(iter_postcode_grouped_batches(addrs, max_batch_size=5))
+    # assert
     assert len(batches) == 2
     assert [str(a.postcode) for a in batches[0]] == ["AA11AA"] * 3
     assert [str(a.postcode) for a in batches[1]] == ["BB22BB"] * 3
@@ -47,8 +57,11 @@ def test_flush_on_overflow_before_adding_next_postcode() -> None:
 def test_single_postcode_group_exceeding_cap_is_dispatched_whole() -> None:
     # An oversize single-postcode group goes out as one batch larger than
     # the cap -- the cap never splits a postcode.
+    # arrange
     addrs = _addrs("AA1 1AA", 7)
+    # act
     batches = list(iter_postcode_grouped_batches(addrs, max_batch_size=5))
+    # assert
     assert len(batches) == 1
     assert len(batches[0]) == 7
 
@@ -56,12 +69,15 @@ def test_single_postcode_group_exceeding_cap_is_dispatched_whole() -> None:
 def test_oversize_group_flushes_existing_buffer_first() -> None:
     # Mirrors the legacy ``if buffer: flush`` branch when an oversize group
     # is encountered: buffered work must not be lost or interleaved.
+    # arrange
     small = _addrs("AA1 1AA", 2)
     big = _addrs("BB2 2BB", 7)
     tail = _addrs("CC3 3CC", 1)
+    # act
     batches = list(
         iter_postcode_grouped_batches(small + big + tail, max_batch_size=5)
     )
+    # assert
     assert len(batches) == 3
     assert [str(a.postcode) for a in batches[0]] == ["AA11AA", "AA11AA"]
     assert [str(a.postcode) for a in batches[1]] == ["BB22BB"] * 7
@@ -70,17 +86,23 @@ def test_oversize_group_flushes_existing_buffer_first() -> None:
 
 def test_final_flush_yields_remaining_buffer() -> None:
     # No overflow ever happens, but the trailing buffer must still come out.
+    # arrange
     addrs = _addrs("AA1 1AA", 2) + _addrs("BB2 2BB", 2)
+    # act
     batches = list(iter_postcode_grouped_batches(addrs, max_batch_size=500))
+    # assert
     assert batches == [addrs]
 
 
 def test_postcode_grouping_preserves_first_seen_order() -> None:
     # Interleaved input must still group by postcode and emit in first-seen
     # order -- never alphabetical.
+    # arrange
     a1, a2 = _addrs("ZZ9 9ZZ", 2)
     b1, b2 = _addrs("AA1 1AA", 2)
+    # act
     batches = list(iter_postcode_grouped_batches([a1, b1, a2, b2]))
+    # assert
     assert len(batches) == 1
     assert [str(a.postcode) for a in batches[0]] == [
         "ZZ99ZZ",
@@ -91,5 +113,6 @@ def test_postcode_grouping_preserves_first_seen_order() -> None:
 
 
 def test_invalid_max_batch_size_raises() -> None:
+    # act / assert
     with pytest.raises(ValueError, match="max_batch_size"):
         list(iter_postcode_grouped_batches([], max_batch_size=0))
diff --git a/tests/domain/addresses/test_user_address.py b/tests/domain/addresses/test_user_address.py
index fa44ad61..8d092df3 100644
--- a/tests/domain/addresses/test_user_address.py
+++ b/tests/domain/addresses/test_user_address.py
@@ -7,35 +7,45 @@ from domain.postcode import Postcode
 
 
 def test_user_address_holds_postcode_value_object() -> None:
+    # act
     addr = UserAddress(user_address="1 The Street", postcode=Postcode("sw1a 1aa"))
+    # assert
     assert addr.postcode == Postcode("SW1A1AA")
 
 
 def test_user_address_preserves_user_address_verbatim() -> None:
     # The free-text user_address string is intentionally NOT normalised --
     # only the postcode is canonicalised, and that happens inside Postcode.
+    # act
     addr = UserAddress(
         user_address="  1 The   Street  ", postcode=Postcode("SW1A1AA")
     )
+    # assert
     assert addr.user_address == "  1 The   Street  "
 
 
 def test_user_address_internal_reference_defaults_to_none() -> None:
+    # act
     addr = UserAddress(user_address="1 The Street", postcode=Postcode("SW1A1AA"))
+    # assert
     assert addr.internal_reference is None
 
 
 def test_user_address_internal_reference_accepted() -> None:
+    # act
     addr = UserAddress(
         user_address="1 The Street",
         postcode=Postcode("SW1A1AA"),
         internal_reference="cust-42",
     )
+    # assert
     assert addr.internal_reference == "cust-42"
 
 
 def test_user_address_is_frozen() -> None:
+    # arrange
     addr = UserAddress(user_address="1 The Street", postcode=Postcode("SW1A1AA"))
+    # act / assert
     with pytest.raises(dataclasses.FrozenInstanceError):
         addr.postcode = Postcode("OTHER")  # type: ignore[misc]
 
@@ -43,29 +53,37 @@ def test_user_address_is_frozen() -> None:
 def test_user_address_equality_uses_canonical_postcode() -> None:
     # Postcode sanitises eagerly, so addresses built from different surface
     # forms of the same postcode compare equal.
+    # arrange
     a = UserAddress(user_address="1 The Street", postcode=Postcode("sw1a 1aa"))
     b = UserAddress(user_address="1 The Street", postcode=Postcode("SW1A1AA"))
+    # act / assert
     assert a == b
 
 
 def test_user_address_source_row_defaults_to_empty_dict() -> None:
+    # act
     addr = UserAddress(user_address="1 The Street", postcode=Postcode("SW1A1AA"))
+    # assert
     assert addr.source_row == {}
 
 
 def test_user_address_carries_source_row() -> None:
+    # arrange
     row = {"Address 1": "1 The Street", "postcode": "SW1A 1AA", "SAP Score": "72"}
+    # act
     addr = UserAddress(
         user_address="1 The Street",
         postcode=Postcode("SW1A 1AA"),
         source_row=row,
     )
+    # assert
     assert addr.source_row == row
 
 
 def test_user_address_equality_ignores_source_row() -> None:
     # source_row is excluded from equality (and hashing): identity stays
     # defined by the parsed fields.
+    # arrange
     a = UserAddress(
         user_address="1 The Street",
         postcode=Postcode("SW1A1AA"),
@@ -76,4 +94,5 @@ def test_user_address_equality_ignores_source_row() -> None:
         postcode=Postcode("SW1A1AA"),
         source_row={"y": "2"},
     )
+    # act / assert
     assert a == b
diff --git a/tests/domain/tasks/test_subtasks.py b/tests/domain/tasks/test_subtasks.py
index 2721d38f..8cee4496 100644
--- a/tests/domain/tasks/test_subtasks.py
+++ b/tests/domain/tasks/test_subtasks.py
@@ -6,10 +6,13 @@ from domain.tasks.subtasks import SubTask, SubTaskStatus
 
 
 def test_create_subtask_starts_waiting() -> None:
+    # arrange
     task_id = uuid4()
 
+    # act
     st = SubTask.create(task_id=task_id, inputs={"foo": "bar"})
 
+    # assert
     assert st.task_id == task_id
     assert st.status is SubTaskStatus.WAITING
     assert st.inputs == {"foo": "bar"}
@@ -19,57 +22,74 @@ def test_create_subtask_starts_waiting() -> None:
 
 
 def test_start_transitions_to_in_progress_and_sets_cloud_logs_url() -> None:
+    # arrange
     st = SubTask.create(task_id=uuid4())
 
+    # act
     st.start(cloud_logs_url="https://example/log")
 
+    # assert
     assert st.status is SubTaskStatus.IN_PROGRESS
     assert st.cloud_logs_url == "https://example/log"
     assert st.job_started is not None
 
 
 def test_start_is_idempotent_from_in_progress() -> None:
+    # arrange
     st = SubTask.create(task_id=uuid4())
     st.start()
     first_start = st.job_started
 
+    # act
     st.start(cloud_logs_url="https://other")
 
+    # assert
     assert st.status is SubTaskStatus.IN_PROGRESS
     assert st.job_started == first_start  # not overwritten
     assert st.cloud_logs_url == "https://other"
 
 
 def test_start_rejects_from_terminal_status() -> None:
+    # arrange
     st = SubTask.create(task_id=uuid4())
     st.complete()
+    # act / assert
     with pytest.raises(ValueError):
         st.start()
 
 
 def test_complete_marks_outputs_and_job_completed() -> None:
+    # arrange
     st = SubTask.create(task_id=uuid4())
     st.start()
 
+    # act
     st.complete({"uprn": "123"})
 
+    # assert
     assert st.status is SubTaskStatus.COMPLETE
     assert st.outputs == {"result": {"uprn": "123"}}
     assert st.job_completed is not None
 
 
 def test_complete_without_result_leaves_outputs_unset() -> None:
+    # arrange
     st = SubTask.create(task_id=uuid4())
+    # act
     st.complete()
+    # assert
     assert st.outputs is None
 
 
 def test_fail_records_error_in_outputs() -> None:
+    # arrange
     st = SubTask.create(task_id=uuid4())
     err = RuntimeError("boom")
 
+    # act
     st.fail(err)
 
+    # assert
     assert st.status is SubTaskStatus.FAILED
     assert st.outputs == {"error": "boom"}
     assert st.job_completed is not None
diff --git a/tests/domain/tasks/test_tasks.py b/tests/domain/tasks/test_tasks.py
index f30c0aa1..ba82412b 100644
--- a/tests/domain/tasks/test_tasks.py
+++ b/tests/domain/tasks/test_tasks.py
@@ -5,12 +5,12 @@ from domain.tasks.tasks import Source, Task, TaskStatus
 
 
 def test_create_task_starts_waiting() -> None:
-    # Arrange / Act
+    # arrange / act
     t = Task.create(
         task_source="manual:test", source=Source.PORTFOLIO, source_id="abc-123"
     )
 
-    # Assert
+    # assert
     assert t.status is TaskStatus.WAITING
     assert t.source is Source.PORTFOLIO
     assert t.source_id == "abc-123"
@@ -19,86 +19,113 @@ def test_create_task_starts_waiting() -> None:
 
 
 def test_create_task_rejects_blank_task_source() -> None:
+    # act / assert
     with pytest.raises(ValueError, match="task_source"):
         Task.create(task_source="   ")
 
 
 def test_start_transitions_to_in_progress() -> None:
+    # arrange
     t = Task.create(task_source="manual:test")
+    # act
     t.start()
+    # assert
     assert t.status is TaskStatus.IN_PROGRESS
 
 
 def test_complete_marks_job_completed() -> None:
+    # arrange
     t = Task.create(task_source="manual:test")
     t.start()
+    # act
     t.complete()
+    # assert
     assert t.status is TaskStatus.COMPLETE
     assert t.job_completed is not None
 
 
 def test_fail_marks_job_completed() -> None:
+    # arrange
     t = Task.create(task_source="manual:test")
+    # act
     t.fail()
+    # assert
     assert t.status is TaskStatus.FAILED
     assert t.job_completed is not None
 
 
 def test_start_rejects_from_terminal_status() -> None:
+    # arrange
     t = Task.create(task_source="manual:test")
     t.complete()
+    # act / assert
     with pytest.raises(ValueError):
         t.start()
 
 
 def test_recalculate_with_empty_statuses_is_noop() -> None:
+    # arrange
     t = Task.create(task_source="manual:test")
     original_status = t.status
     original_completed = t.job_completed
 
+    # act
     t.recalculate_from_subtasks([])
 
+    # assert
     assert t.status is original_status
     assert t.job_completed is original_completed
 
 
 def test_recalculate_all_waiting_keeps_waiting() -> None:
+    # arrange
     t = Task.create(task_source="manual:test")
     t.start()  # task moved to IN_PROGRESS earlier
     t.complete()  # then COMPLETE, with job_completed set
 
+    # act
     t.recalculate_from_subtasks([SubTaskStatus.WAITING, SubTaskStatus.WAITING])
 
+    # assert
     assert t.status is TaskStatus.WAITING
     assert t.job_completed is None
 
 
 def test_recalculate_any_in_progress_marks_in_progress() -> None:
+    # arrange
     t = Task.create(task_source="manual:test")
 
+    # act
     t.recalculate_from_subtasks(
         [SubTaskStatus.WAITING, SubTaskStatus.IN_PROGRESS, SubTaskStatus.COMPLETE]
     )
 
+    # assert
     assert t.status is TaskStatus.IN_PROGRESS
     assert t.job_completed is None
 
 
 def test_recalculate_all_complete_marks_complete() -> None:
+    # arrange
     t = Task.create(task_source="manual:test")
 
+    # act
     t.recalculate_from_subtasks([SubTaskStatus.COMPLETE, SubTaskStatus.COMPLETE])
 
+    # assert
     assert t.status is TaskStatus.COMPLETE
     assert t.job_completed is not None
 
 
 def test_recalculate_any_failed_marks_failed_even_with_others() -> None:
+    # arrange
     t = Task.create(task_source="manual:test")
 
+    # act
     t.recalculate_from_subtasks(
         [SubTaskStatus.IN_PROGRESS, SubTaskStatus.COMPLETE, SubTaskStatus.FAILED]
     )
 
+    # assert
     assert t.status is TaskStatus.FAILED
     assert t.job_completed is not None
diff --git a/tests/domain/test_postcode.py b/tests/domain/test_postcode.py
index 89d5cdc8..f7ce9015 100644
--- a/tests/domain/test_postcode.py
+++ b/tests/domain/test_postcode.py
@@ -6,43 +6,54 @@ from domain.postcode import Postcode
 
 
 def test_postcode_uppercases() -> None:
+    # act / assert
     assert Postcode("sw1a1aa").value == "SW1A1AA"
 
 
 def test_postcode_strips_internal_spaces() -> None:
+    # act / assert
     assert Postcode("sw1a 1aa").value == "SW1A1AA"
 
 
 def test_postcode_strips_leading_and_trailing_whitespace() -> None:
+    # act / assert
     assert Postcode("  sw1a 1aa  ").value == "SW1A1AA"
 
 
 def test_postcode_strips_tabs_and_newlines() -> None:
     # CSV ingestion occasionally introduces stray whitespace characters; the
     # canonical form must absorb them just like literal spaces.
+    # act / assert
     assert Postcode("sw1a\t1aa\n").value == "SW1A1AA"
 
 
 def test_postcode_construction_is_idempotent() -> None:
+    # arrange
     once = Postcode("sw1a 1aa")
+    # act / assert
     assert Postcode(once.value).value == "SW1A1AA"
 
 
 def test_postcode_empty_string() -> None:
+    # act / assert
     assert Postcode("").value == ""
 
 
 def test_postcode_str_returns_canonical_value() -> None:
+    # act / assert
     assert str(Postcode("sw1a 1aa")) == "SW1A1AA"
 
 
 def test_postcode_equality_ignores_surface_form() -> None:
     # Differing case / whitespace sanitise to the same canonical value, so
     # the value objects compare equal.
+    # act / assert
     assert Postcode("sw1a 1aa") == Postcode("SW1A1AA")
 
 
 def test_postcode_is_frozen() -> None:
+    # arrange
     postcode = Postcode("SW1A1AA")
+    # act / assert
     with pytest.raises(dataclasses.FrozenInstanceError):
         postcode.value = "OTHER"  # type: ignore[misc]
diff --git a/tests/infrastructure/test_address2uprn_queue_client.py b/tests/infrastructure/test_address2uprn_queue_client.py
index b4114742..c8e89ece 100644
--- a/tests/infrastructure/test_address2uprn_queue_client.py
+++ b/tests/infrastructure/test_address2uprn_queue_client.py
@@ -28,12 +28,15 @@ def queue_setup() -> Iterator[tuple[Address2UprnQueueClient, Any, str]]:
 def test_publish_returns_message_id(
     queue_setup: tuple[Address2UprnQueueClient, Any, str],
 ) -> None:
+    # arrange
     client, _boto, _url = queue_setup
+    # act
     message_id = client.publish(
         parent_task_id=uuid4(),
         child_subtask_id=uuid4(),
         s3_uri="s3://my-bucket/path/to/chunk.csv",
     )
+    # assert
     assert isinstance(message_id, str)
     assert message_id
 
@@ -41,17 +44,20 @@ def test_publish_returns_message_id(
 def test_publish_body_uses_typed_shape(
     queue_setup: tuple[Address2UprnQueueClient, Any, str],
 ) -> None:
+    # arrange
     client, boto_client, queue_url = queue_setup
     parent_id = uuid4()
     child_id = uuid4()
     s3_uri = "s3://my-bucket/path/to/chunk.csv"
 
+    # act
     client.publish(
         parent_task_id=parent_id,
         child_subtask_id=child_id,
         s3_uri=s3_uri,
     )
 
+    # assert
     received: dict[str, Any] = boto_client.receive_message(
         QueueUrl=queue_url, MaxNumberOfMessages=1
     )
diff --git a/tests/infrastructure/test_csv_s3_client.py b/tests/infrastructure/test_csv_s3_client.py
index 4b9fc199..30e27164 100644
--- a/tests/infrastructure/test_csv_s3_client.py
+++ b/tests/infrastructure/test_csv_s3_client.py
@@ -18,26 +18,34 @@ def csv_client() -> Iterator[CsvS3Client]:
 
 
 def test_save_rows_returns_s3_uri(csv_client: CsvS3Client) -> None:
+    # arrange
     rows = [{"address": "1 High St", "postcode": "AB1 2CD"}]
+    # act
     uri = csv_client.save_rows(rows, "uploads/addresses.csv")
+    # assert
     assert uri == f"s3://{BUCKET}/uploads/addresses.csv"
 
 
 def test_round_trip_preserves_rows(csv_client: CsvS3Client) -> None:
+    # arrange
     rows = [
         {"address": "1 High St", "postcode": "AB1 2CD"},
         {"address": "2 Low St", "postcode": "XY9 8ZW"},
     ]
+    # act
     uri = csv_client.save_rows(rows, "uploads/addresses.csv")
     fetched = csv_client.read_rows(uri)
+    # assert
     assert fetched == rows
 
 
 def test_save_rows_rejects_empty_list(csv_client: CsvS3Client) -> None:
+    # act / assert
     with pytest.raises(ValueError, match="empty"):
         csv_client.save_rows([], "uploads/empty.csv")
 
 
 def test_read_rows_rejects_wrong_bucket(csv_client: CsvS3Client) -> None:
+    # act / assert
     with pytest.raises(ValueError, match="does not match client bucket"):
         csv_client.read_rows("s3://other-bucket/uploads/addresses.csv")
diff --git a/tests/infrastructure/test_s3_client.py b/tests/infrastructure/test_s3_client.py
index 7ed4c30b..67db4f58 100644
--- a/tests/infrastructure/test_s3_client.py
+++ b/tests/infrastructure/test_s3_client.py
@@ -18,14 +18,19 @@ def s3_client() -> Iterator[S3Client]:
 
 
 def test_put_object_returns_s3_uri(s3_client: S3Client) -> None:
+    # act
     uri = s3_client.put_object("folder/data.bin", b"payload")
+    # assert
     assert uri == f"s3://{BUCKET}/folder/data.bin"
 
 
 def test_get_object_returns_bytes_written_by_put_object(s3_client: S3Client) -> None:
+    # arrange
     s3_client.put_object("round/trip.bin", b"hello world")
+    # act / assert
     assert s3_client.get_object("round/trip.bin") == b"hello world"
 
 
 def test_bucket_property_exposes_configured_bucket(s3_client: S3Client) -> None:
+    # act / assert
     assert s3_client.bucket == BUCKET
diff --git a/tests/infrastructure/test_s3_uri.py b/tests/infrastructure/test_s3_uri.py
index 896c5959..32fd710f 100644
--- a/tests/infrastructure/test_s3_uri.py
+++ b/tests/infrastructure/test_s3_uri.py
@@ -4,29 +4,37 @@ from infrastructure.s3_uri import parse_s3_uri
 
 
 def test_parses_simple_s3_uri() -> None:
+    # act / assert
     assert parse_s3_uri("s3://my-bucket/file.csv") == ("my-bucket", "file.csv")
 
 
 def test_parses_s3_uri_with_nested_key() -> None:
+    # act
     bucket, key = parse_s3_uri("s3://my-bucket/nested/path/to/file.csv")
+    # assert
     assert (bucket, key) == ("my-bucket", "nested/path/to/file.csv")
 
 
 def test_rejects_s3_uri_without_key() -> None:
+    # act / assert
     with pytest.raises(ValueError, match="bucket and a key"):
         parse_s3_uri("s3://my-bucket")
 
 
 def test_rejects_s3_uri_with_empty_key() -> None:
+    # act / assert
     with pytest.raises(ValueError, match="bucket and a key"):
         parse_s3_uri("s3://my-bucket/")
 
 
 def test_parses_console_url_prefix() -> None:
+    # arrange
     url = "https://eu-west-2.console.aws.amazon.com/s3/object/my-bucket?prefix=nested%2Ffile.csv"
+    # act / assert
     assert parse_s3_uri(url) == ("my-bucket", "nested/file.csv")
 
 
 def test_rejects_unparseable_string() -> None:
+    # act / assert
     with pytest.raises(ValueError):
         parse_s3_uri("not-a-uri-at-all")
diff --git a/tests/infrastructure/test_sqs_client.py b/tests/infrastructure/test_sqs_client.py
index 7f1e8f78..44186bbb 100644
--- a/tests/infrastructure/test_sqs_client.py
+++ b/tests/infrastructure/test_sqs_client.py
@@ -19,17 +19,23 @@ def sqs_setup() -> Iterator[tuple[SqsClient, Any, str]]:
 
 
 def test_send_returns_message_id(sqs_setup: tuple[SqsClient, Any, str]) -> None:
+    # arrange
     client, _boto, _url = sqs_setup
+    # act
     message_id = client.send({"hello": "world"})
+    # assert
     assert isinstance(message_id, str)
     assert message_id
 
 
 def test_send_json_serialises_body(sqs_setup: tuple[SqsClient, Any, str]) -> None:
+    # arrange
     client, boto_client, queue_url = sqs_setup
     body = {"hello": "world", "count": 3}
+    # act
     client.send(body)
 
+    # assert
     received: dict[str, Any] = boto_client.receive_message(
         QueueUrl=queue_url, MaxNumberOfMessages=1
     )
diff --git a/tests/orchestration/test_postcode_splitter_orchestrator.py b/tests/orchestration/test_postcode_splitter_orchestrator.py
index 4ee2315e..a718ffbc 100644
--- a/tests/orchestration/test_postcode_splitter_orchestrator.py
+++ b/tests/orchestration/test_postcode_splitter_orchestrator.py
@@ -9,7 +9,8 @@ from typing import Any, cast
 import boto3
 import pytest
 from moto import mock_aws
-from sqlmodel import Session, SQLModel, create_engine
+from sqlalchemy import Engine
+from sqlmodel import Session
 
 from infrastructure.address2uprn_queue_client import Address2UprnQueueClient
 from infrastructure.csv_s3_client import CsvS3Client
@@ -65,7 +66,7 @@ class Harness:
 
 
 @pytest.fixture
-def harness() -> Iterator[Harness]:
+def harness(db_engine: Engine) -> Iterator[Harness]:
     with mock_aws():
         # Infra: S3 + SQS
         boto_s3 = _make_boto_client("s3")
@@ -78,10 +79,8 @@ def harness() -> Iterator[Harness]:
         repo = UserAddressCsvS3Repository(csv_client, BUCKET)
         queue_client = Address2UprnQueueClient(boto_sqs, queue_url)
 
-        # DB: in-memory SQLite TaskOrchestrator
-        engine = create_engine("sqlite://")
-        SQLModel.metadata.create_all(engine)
-        with Session(engine) as session:
+        # DB: ephemeral PostgreSQL TaskOrchestrator
+        with Session(db_engine) as session:
             task_repo = TaskPostgresRepository(session=session)
             subtask_repo = SubTaskPostgresRepository(session=session)
             task_orchestrator = TaskOrchestrator(
@@ -169,6 +168,7 @@ def _drain_queue(boto_sqs: Any, queue_url: str) -> list[dict[str, Any]]:
 def test_split_and_dispatch_creates_three_children_for_fixture(
     harness: Harness,
 ) -> None:
+    # arrange
     parent_task, parent_subtask = (
         harness.task_orchestrator.create_task_with_subtask(
             task_source="manual:postcode-splitter-int"
@@ -176,12 +176,14 @@ def test_split_and_dispatch_creates_three_children_for_fixture(
     )
     input_uri = _upload_fixture_csv(harness.csv_client)
 
+    # act
     child_ids = harness.splitter.split_and_dispatch(
         parent_task_id=parent_task.id,
         parent_subtask_id=parent_subtask.id,
         input_s3_uri=input_uri,
     )
 
+    # assert
     assert len(child_ids) == 3
     # All child ids are unique and persisted as WAITING children of the
     # parent task.
@@ -194,6 +196,7 @@ def test_split_and_dispatch_creates_three_children_for_fixture(
 def test_split_and_dispatch_persists_child_inputs_with_task_id_and_s3_uri(
     harness: Harness,
 ) -> None:
+    # arrange
     parent_task, parent_subtask = (
         harness.task_orchestrator.create_task_with_subtask(
             task_source="manual:postcode-splitter-int"
@@ -201,12 +204,14 @@ def test_split_and_dispatch_persists_child_inputs_with_task_id_and_s3_uri(
     )
     input_uri = _upload_fixture_csv(harness.csv_client)
 
+    # act
     child_ids = harness.splitter.split_and_dispatch(
         parent_task_id=parent_task.id,
         parent_subtask_id=parent_subtask.id,
         input_s3_uri=input_uri,
     )
 
+    # assert
     for cid in child_ids:
         child = harness.subtasks.get(cid)
         assert child.inputs is not None
@@ -224,6 +229,7 @@ def test_split_and_dispatch_persists_child_inputs_with_task_id_and_s3_uri(
 def test_split_and_dispatch_publishes_one_message_per_child_with_matching_ids(
     harness: Harness,
 ) -> None:
+    # arrange
     parent_task, parent_subtask = (
         harness.task_orchestrator.create_task_with_subtask(
             task_source="manual:postcode-splitter-int"
@@ -231,12 +237,14 @@ def test_split_and_dispatch_publishes_one_message_per_child_with_matching_ids(
     )
     input_uri = _upload_fixture_csv(harness.csv_client)
 
+    # act
     child_ids = harness.splitter.split_and_dispatch(
         parent_task_id=parent_task.id,
         parent_subtask_id=parent_subtask.id,
         input_s3_uri=input_uri,
     )
 
+    # assert
     bodies = _drain_queue(harness.boto_sqs, harness.queue_url)
     assert len(bodies) == len(child_ids)
 
@@ -258,6 +266,7 @@ def test_split_and_dispatch_publishes_one_message_per_child_with_matching_ids(
 def test_split_and_dispatch_returns_child_ids_in_dispatch_order(
     harness: Harness,
 ) -> None:
+    # arrange
     parent_task, parent_subtask = (
         harness.task_orchestrator.create_task_with_subtask(
             task_source="manual:postcode-splitter-int"
@@ -265,12 +274,14 @@ def test_split_and_dispatch_returns_child_ids_in_dispatch_order(
     )
     input_uri = _upload_fixture_csv(harness.csv_client)
 
+    # act
     child_ids = harness.splitter.split_and_dispatch(
         parent_task_id=parent_task.id,
         parent_subtask_id=parent_subtask.id,
         input_s3_uri=input_uri,
     )
 
+    # assert
     # Re-load each child's saved batch and inspect the postcode_clean column
     # to confirm the dispatch order matches the postcode-batching algorithm:
     # AA-batch first, BB oversize batch second, CC final-flush third.
diff --git a/tests/orchestration/test_task_orchestrator.py b/tests/orchestration/test_task_orchestrator.py
index c0816d2d..ae89991d 100644
--- a/tests/orchestration/test_task_orchestrator.py
+++ b/tests/orchestration/test_task_orchestrator.py
@@ -2,7 +2,8 @@ from collections.abc import Iterator
 from dataclasses import dataclass
 
 import pytest
-from sqlmodel import Session, SQLModel, create_engine
+from sqlalchemy import Engine
+from sqlmodel import Session
 
 from domain.tasks.subtasks import SubTask, SubTaskStatus
 from domain.tasks.tasks import Source, TaskStatus
@@ -19,10 +20,8 @@ class Harness:
 
 
 @pytest.fixture
-def harness() -> Iterator[Harness]:
-    engine = create_engine("sqlite://")
-    SQLModel.metadata.create_all(engine)
-    with Session(engine) as session:
+def harness(db_engine: Engine) -> Iterator[Harness]:
+    with Session(db_engine) as session:
         tasks = TaskPostgresRepository(session=session)
         subtasks = SubTaskPostgresRepository(session=session)
         yield Harness(
@@ -35,6 +34,7 @@ def harness() -> Iterator[Harness]:
 def test_create_task_with_subtask_creates_both_in_waiting(
     harness: Harness,
 ) -> None:
+    # act
     task, subtask = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test",
         inputs={"foo": "bar"},
@@ -42,6 +42,7 @@ def test_create_task_with_subtask_creates_both_in_waiting(
         source_id="abc",
     )
 
+    # assert
     assert task.status is TaskStatus.WAITING
     assert subtask.status is SubTaskStatus.WAITING
     assert subtask.task_id == task.id
@@ -49,27 +50,33 @@ def test_create_task_with_subtask_creates_both_in_waiting(
 
 
 def test_start_subtask_cascades_to_in_progress(harness: Harness) -> None:
+    # arrange
     task, subtask = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
 
+    # act
     started = harness.orchestrator.start_subtask(
         subtask.id, cloud_logs_url="https://example/log"
     )
 
+    # assert
     assert started.status is SubTaskStatus.IN_PROGRESS
     assert started.cloud_logs_url == "https://example/log"
     assert harness.tasks.get(task.id).status is TaskStatus.IN_PROGRESS
 
 
 def test_complete_subtask_cascades_to_complete(harness: Harness) -> None:
+    # arrange
     task, subtask = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
     harness.orchestrator.start_subtask(subtask.id)
 
+    # act
     harness.orchestrator.complete_subtask(subtask.id, {"value": 42})
 
+    # assert
     done_subtask = harness.subtasks.get(subtask.id)
     done_task = harness.tasks.get(task.id)
     assert done_subtask.outputs == {"result": {"value": 42}}
@@ -78,12 +85,15 @@ def test_complete_subtask_cascades_to_complete(harness: Harness) -> None:
 
 
 def test_fail_subtask_cascades_to_failed(harness: Harness) -> None:
+    # arrange
     task, subtask = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
 
+    # act
     harness.orchestrator.fail_subtask(subtask.id, RuntimeError("boom"))
 
+    # assert
     failed_subtask = harness.subtasks.get(subtask.id)
     failed_task = harness.tasks.get(task.id)
     assert failed_subtask.outputs == {"error": "boom"}
@@ -93,42 +103,51 @@ def test_fail_subtask_cascades_to_failed(harness: Harness) -> None:
 def test_failed_subtask_locks_task_failed_even_with_others_complete(
     harness: Harness,
 ) -> None:
+    # arrange
     task, first = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
     second = SubTask.create(task_id=task.id)
     harness.subtasks.create(second)
 
+    # act
     harness.orchestrator.complete_subtask(first.id)
     harness.orchestrator.fail_subtask(second.id, RuntimeError("nope"))
 
+    # assert
     assert harness.tasks.get(task.id).status is TaskStatus.FAILED
 
 
 def test_mixed_complete_and_in_progress_keeps_task_in_progress(
     harness: Harness,
 ) -> None:
+    # arrange
     task, first = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
     second = SubTask.create(task_id=task.id)
     harness.subtasks.create(second)
 
+    # act
     harness.orchestrator.complete_subtask(first.id)
     harness.orchestrator.start_subtask(second.id)
 
+    # assert
     assert harness.tasks.get(task.id).status is TaskStatus.IN_PROGRESS
 
 
 def test_run_subtask_happy_path_returns_result_and_cascades_complete(
     harness: Harness,
 ) -> None:
+    # arrange
     task, subtask = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
 
+    # act
     result = harness.orchestrator.run_subtask(subtask.id, work=lambda: {"answer": 42})
 
+    # assert
     assert result == {"answer": 42}
     assert harness.subtasks.get(subtask.id).status is SubTaskStatus.COMPLETE
     assert harness.tasks.get(task.id).status is TaskStatus.COMPLETE
@@ -137,16 +156,19 @@ def test_run_subtask_happy_path_returns_result_and_cascades_complete(
 def test_create_child_subtask_adds_waiting_child_without_changing_parent_status(
     harness: Harness,
 ) -> None:
+    # arrange
     task, first = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
     harness.orchestrator.start_subtask(first.id)
     assert harness.tasks.get(task.id).status is TaskStatus.IN_PROGRESS
 
+    # act
     child = harness.orchestrator.create_child_subtask(
         task.id, inputs={"split": "a"}
     )
 
+    # assert
     persisted_child = harness.subtasks.get(child.id)
     assert persisted_child.task_id == task.id
     assert persisted_child.status is SubTaskStatus.WAITING
@@ -159,6 +181,7 @@ def test_create_child_subtask_adds_waiting_child_without_changing_parent_status(
 def test_run_subtask_failing_work_marks_failed_and_reraises(
     harness: Harness,
 ) -> None:
+    # arrange
     task, subtask = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
@@ -166,6 +189,7 @@ def test_run_subtask_failing_work_marks_failed_and_reraises(
     def boom() -> None:
         raise RuntimeError("boom")
 
+    # act / assert
     with pytest.raises(RuntimeError, match="boom"):
         harness.orchestrator.run_subtask(subtask.id, work=boom)
 
diff --git a/tests/repositories/tasks/postgres/test_subtask_postgres_repository.py b/tests/repositories/tasks/postgres/test_subtask_postgres_repository.py
index ac39e089..9cec52ea 100644
--- a/tests/repositories/tasks/postgres/test_subtask_postgres_repository.py
+++ b/tests/repositories/tasks/postgres/test_subtask_postgres_repository.py
@@ -1,33 +1,40 @@
 from collections.abc import Iterator
-from uuid import uuid4
+from uuid import UUID, uuid4
 
 import pytest
-from sqlmodel import Session, SQLModel, create_engine
+from sqlalchemy import Engine
+from sqlmodel import Session
 
-# Importing the SQLModel row modules registers their tables in
-# SQLModel.metadata so create_all builds both. Imports look unused; they aren't.
-import infrastructure.postgres.subtask_table  # noqa: F401  # pyright: ignore[reportUnusedImport]
-import infrastructure.postgres.task_table  # noqa: F401  # pyright: ignore[reportUnusedImport]
 from domain.tasks.subtasks import SubTask, SubTaskStatus
+from domain.tasks.tasks import Task
 from repositories.tasks.subtask_postgres_repository import SubTaskPostgresRepository
+from repositories.tasks.task_postgres_repository import TaskPostgresRepository
 
 
 @pytest.fixture
-def session() -> Iterator[Session]:
-    engine = create_engine("sqlite://")
-    SQLModel.metadata.create_all(engine)
-    with Session(engine) as s:
+def session(db_engine: Engine) -> Iterator[Session]:
+    with Session(db_engine) as s:
         yield s
 
 
+def _persisted_task_id(session: Session) -> UUID:
+    """Create a parent Task row so SubTask FK constraints are satisfied."""
+    task = Task.create(task_source="manual:test")
+    TaskPostgresRepository(session=session).create(task)
+    return task.id
+
+
 def test_create_and_get_round_trip_preserves_inputs(session: Session) -> None:
+    # arrange
     repo = SubTaskPostgresRepository(session=session)
-    task_id = uuid4()
+    task_id = _persisted_task_id(session)
     st = SubTask.create(task_id=task_id, inputs={"address": "68 Glendon Way"})
 
+    # act
     repo.create(st)
     fetched = repo.get(st.id)
 
+    # assert
     assert fetched.id == st.id
     assert fetched.task_id == task_id
     assert fetched.status is SubTaskStatus.WAITING
@@ -36,16 +43,21 @@ def test_create_and_get_round_trip_preserves_inputs(session: Session) -> None:
 
 
 def test_save_persists_status_and_outputs(session: Session) -> None:
+    # arrange
     repo = SubTaskPostgresRepository(session=session)
-    st = SubTask.create(task_id=uuid4())
+    st = SubTask.create(task_id=_persisted_task_id(session))
     repo.create(st)
 
+    # act
     st.start(cloud_logs_url="https://example/log")
     repo.save(st)
+    # assert
     assert repo.get(st.id).status is SubTaskStatus.IN_PROGRESS
 
+    # act
     st.complete({"uprn": "123"})
     repo.save(st)
+    # assert
     done = repo.get(st.id)
     assert done.status is SubTaskStatus.COMPLETE
     assert done.outputs == {"result": {"uprn": "123"}}
@@ -54,16 +66,19 @@ def test_save_persists_status_and_outputs(session: Session) -> None:
 
 
 def test_list_by_task_filters_by_task_id(session: Session) -> None:
+    # arrange
     repo = SubTaskPostgresRepository(session=session)
-    task_a = uuid4()
-    task_b = uuid4()
+    task_a = _persisted_task_id(session)
+    task_b = _persisted_task_id(session)
     repo.create(SubTask.create(task_id=task_a))
     repo.create(SubTask.create(task_id=task_a))
     repo.create(SubTask.create(task_id=task_b))
 
+    # act
     a_results = repo.list_by_task(task_a)
     b_results = repo.list_by_task(task_b)
 
+    # assert
     assert len(a_results) == 2
     assert len(b_results) == 1
     assert all(s.task_id == task_a for s in a_results)
@@ -71,11 +86,15 @@ def test_list_by_task_filters_by_task_id(session: Session) -> None:
 
 
 def test_list_by_task_returns_empty_for_unknown_task(session: Session) -> None:
+    # arrange
     repo = SubTaskPostgresRepository(session=session)
+    # act / assert
     assert repo.list_by_task(uuid4()) == []
 
 
 def test_get_missing_raises(session: Session) -> None:
+    # arrange
     repo = SubTaskPostgresRepository(session=session)
+    # act / assert
     with pytest.raises(ValueError, match="not found"):
         repo.get(uuid4())
diff --git a/tests/repositories/tasks/postgres/test_task_postgres_repository.py b/tests/repositories/tasks/postgres/test_task_postgres_repository.py
index 3e1aa226..8a49a861 100644
--- a/tests/repositories/tasks/postgres/test_task_postgres_repository.py
+++ b/tests/repositories/tasks/postgres/test_task_postgres_repository.py
@@ -2,7 +2,8 @@ from collections.abc import Iterator
 from uuid import uuid4
 
 import pytest
-from sqlmodel import Session, SQLModel, create_engine
+from sqlalchemy import Engine
+from sqlmodel import Session
 
 from domain.tasks.tasks import Source, Task, TaskStatus
 from infrastructure.postgres.task_table import TaskRow
@@ -10,25 +11,23 @@ from repositories.tasks.task_postgres_repository import TaskPostgresRepository
 
 
 @pytest.fixture
-def session() -> Iterator[Session]:
-    engine = create_engine("sqlite://")
-    SQLModel.metadata.create_all(engine)
-    with Session(engine) as s:
+def session(db_engine: Engine) -> Iterator[Session]:
+    with Session(db_engine) as s:
         yield s
 
 
 def test_create_and_get_round_trip(session: Session) -> None:
-    # Arrange
+    # arrange
     repo = TaskPostgresRepository(session=session)
     t = Task.create(
         task_source="manual:test", source=Source.PORTFOLIO, source_id="abc-123"
     )
 
-    # Act
+    # act
     repo.create(t)
     fetched = repo.get(t.id)
 
-    # Assert
+    # assert
     assert fetched.id == t.id
     assert fetched.status is TaskStatus.WAITING
     assert fetched.source is Source.PORTFOLIO
@@ -36,33 +35,43 @@ def test_create_and_get_round_trip(session: Session) -> None:
 
 
 def test_save_persists_status_transition(session: Session) -> None:
+    # arrange
     repo = TaskPostgresRepository(session=session)
     t = Task.create(task_source="manual:test")
     repo.create(t)
 
+    # act
     t.start()
     repo.save(t)
+    # assert
     assert repo.get(t.id).status is TaskStatus.IN_PROGRESS
 
+    # act
     t.complete()
     repo.save(t)
+    # assert
     done = repo.get(t.id)
     assert done.status is TaskStatus.COMPLETE
     assert done.job_completed is not None
 
 
 def test_get_missing_raises(session: Session) -> None:
+    # arrange
     repo = TaskPostgresRepository(session=session)
+    # act / assert
     with pytest.raises(ValueError, match="not found"):
         repo.get(uuid4())
 
 
 def test_get_normalises_legacy_capitalised_status(session: Session) -> None:
     # Existing rows written by backend code use "In Progress" (capitalised).
+    # arrange
     repo = TaskPostgresRepository(session=session)
     row = TaskRow(task_source="manual:test", status="In Progress")
     session.add(row)
     session.commit()
 
+    # act
     fetched = repo.get(row.id)
+    # assert
     assert fetched.status is TaskStatus.IN_PROGRESS
diff --git a/tests/repositories/user_address/test_user_address_csv_s3_repository.py b/tests/repositories/user_address/test_user_address_csv_s3_repository.py
index c1acee32..9ffb250a 100644
--- a/tests/repositories/user_address/test_user_address_csv_s3_repository.py
+++ b/tests/repositories/user_address/test_user_address_csv_s3_repository.py
@@ -32,6 +32,7 @@ def _upload_csv(
 def test_load_batch_parses_address_postcode_and_reference(
     repo: UserAddressCsvS3Repository,
 ) -> None:
+    # arrange
     rows = [
         {
             "Address 1": "1 High Street",
@@ -43,8 +44,10 @@ def test_load_batch_parses_address_postcode_and_reference(
     ]
     uri = _upload_csv(repo, rows, "uploads/full.csv")
 
+    # act
     addresses = repo.load_batch(uri)
 
+    # assert
     assert len(addresses) == 1
     address = addresses[0]
     assert address.user_address == "1 High Street, Flat 2, Townville"
@@ -55,6 +58,7 @@ def test_load_batch_parses_address_postcode_and_reference(
 def test_load_batch_uses_only_address_1_when_others_missing(
     repo: UserAddressCsvS3Repository,
 ) -> None:
+    # arrange
     rows = [
         {
             "Address 1": "10 Cardiff Road",
@@ -66,8 +70,10 @@ def test_load_batch_uses_only_address_1_when_others_missing(
     ]
     uri = _upload_csv(repo, rows, "uploads/address1-only.csv")
 
+    # act
     addresses = repo.load_batch(uri)
 
+    # assert
     assert len(addresses) == 1
     assert addresses[0].user_address == "10 Cardiff Road"
     assert addresses[0].postcode == Postcode("CF101AA")
@@ -77,6 +83,7 @@ def test_load_batch_uses_only_address_1_when_others_missing(
 def test_load_batch_handles_missing_internal_reference(
     repo: UserAddressCsvS3Repository,
 ) -> None:
+    # arrange
     rows = [
         {
             "Address 1": "5 Park Lane",
@@ -88,8 +95,10 @@ def test_load_batch_handles_missing_internal_reference(
     ]
     uri = _upload_csv(repo, rows, "uploads/no-ref.csv")
 
+    # act
     addresses = repo.load_batch(uri)
 
+    # assert
     assert len(addresses) == 1
     assert addresses[0].user_address == "5 Park Lane"
     assert addresses[0].postcode == Postcode("M11AA")
@@ -101,6 +110,7 @@ def test_load_batch_captures_full_source_row(
 ) -> None:
     # A raw EPC-export-shaped row: the splitter must preserve every column,
     # not just the ones it parses into UserAddress fields.
+    # arrange
     row = {
         "Asset Reference": "511",
         "Address 1": "9 Abingdon Road Padiham Lancashire BB12 7BX",
@@ -110,17 +120,21 @@ def test_load_batch_captures_full_source_row(
     }
     uri = _upload_csv(repo, [row], "uploads/epc.csv")
 
+    # act
     addresses = repo.load_batch(uri)
 
+    # assert
     assert addresses[0].source_row == row
 
 
 def test_load_batch_raises_when_postcode_column_absent(
     repo: UserAddressCsvS3Repository,
 ) -> None:
+    # arrange
     rows = [{"Address 1": "1 High Street", "Property Type": "Flat"}]
     uri = _upload_csv(repo, rows, "uploads/no-postcode.csv")
 
+    # act / assert
     with pytest.raises(ValueError, match="no 'postcode' column"):
         repo.load_batch(uri)
 
@@ -128,6 +142,7 @@ def test_load_batch_raises_when_postcode_column_absent(
 def test_save_batch_passes_through_all_columns_and_appends_postcode_clean(
     repo: UserAddressCsvS3Repository,
 ) -> None:
+    # arrange
     row = {
         "Asset Reference": "511",
         "Address 1": "9 Abingdon Road Padiham Lancashire BB12 7BX",
@@ -137,9 +152,11 @@ def test_save_batch_passes_through_all_columns_and_appends_postcode_clean(
     uri = _upload_csv(repo, [row], "uploads/epc.csv")
     addresses = repo.load_batch(uri)
 
+    # act
     saved_uri = repo.save_batch(addresses, "tasks/passthrough")
     saved_rows = repo._csv_client.read_rows(saved_uri)  # pyright: ignore[reportPrivateUsage]
 
+    # assert
     assert len(saved_rows) == 1
     saved = saved_rows[0]
     # Every original column survives, byte-for-byte.
@@ -152,6 +169,7 @@ def test_save_batch_passes_through_all_columns_and_appends_postcode_clean(
 def test_save_batch_returns_uri_under_path_prefix(
     repo: UserAddressCsvS3Repository,
 ) -> None:
+    # arrange
     addresses = [
         UserAddress(
             user_address="1 High Street",
@@ -160,8 +178,10 @@ def test_save_batch_returns_uri_under_path_prefix(
         ),
     ]
 
+    # act
     uri = repo.save_batch(addresses, "tasks/abc/batches")
 
+    # assert
     assert uri.startswith(f"s3://{BUCKET}/tasks/abc/batches/")
     assert uri.endswith(".csv")
 
@@ -169,6 +189,7 @@ def test_save_batch_returns_uri_under_path_prefix(
 def test_save_then_reload_round_trip_preserves_columns(
     repo: UserAddressCsvS3Repository,
 ) -> None:
+    # arrange
     rows = [
         {
             "Address 1": "1 High Street",
@@ -184,9 +205,11 @@ def test_save_then_reload_round_trip_preserves_columns(
     uri = _upload_csv(repo, rows, "uploads/round-trip.csv")
     addresses = repo.load_batch(uri)
 
+    # act
     saved_uri = repo.save_batch(addresses, "tasks/round-trip")
     saved_rows = repo._csv_client.read_rows(saved_uri)  # pyright: ignore[reportPrivateUsage]
 
+    # assert
     # Original columns come back verbatim; postcode_clean is the only addition.
     assert [
         {k: v for k, v in r.items() if k != "postcode_clean"} for r in saved_rows
@@ -197,6 +220,7 @@ def test_save_then_reload_round_trip_preserves_columns(
 def test_save_batch_uses_unique_filename_per_call(
     repo: UserAddressCsvS3Repository,
 ) -> None:
+    # arrange
     addresses = [
         UserAddress(
             user_address="1 High Street",
@@ -205,7 +229,9 @@ def test_save_batch_uses_unique_filename_per_call(
         ),
     ]
 
+    # act
     uri_1 = repo.save_batch(addresses, "tasks/uniqueness")
     uri_2 = repo.save_batch(addresses, "tasks/uniqueness")
 
+    # assert
     assert uri_1 != uri_2
diff --git a/tests/utilities/aws_lambda/test_subtask_handler.py b/tests/utilities/aws_lambda/test_subtask_handler.py
index 9cf68f28..d671adc4 100644
--- a/tests/utilities/aws_lambda/test_subtask_handler.py
+++ b/tests/utilities/aws_lambda/test_subtask_handler.py
@@ -6,7 +6,8 @@ from typing import Any
 from uuid import UUID
 
 import pytest
-from sqlmodel import Session, SQLModel, create_engine
+from sqlalchemy import Engine
+from sqlmodel import Session
 
 from domain.tasks.subtasks import SubTaskStatus
 from domain.tasks.tasks import TaskStatus
@@ -30,10 +31,8 @@ class Harness:
 
 
 @pytest.fixture
-def harness() -> Iterator[Harness]:
-    engine = create_engine("sqlite://")
-    SQLModel.metadata.create_all(engine)
-    with Session(engine) as session:
+def harness(db_engine: Engine) -> Iterator[Harness]:
+    with Session(db_engine) as session:
         tasks = TaskPostgresRepository(session=session)
         subtasks = SubTaskPostgresRepository(session=session)
         yield Harness(
@@ -50,6 +49,7 @@ def _direct_event(task_id: UUID, subtask_id: UUID) -> dict[str, Any]:
 def test_subtask_handler_injects_orchestrator_as_third_positional_argument(
     harness: Harness,
 ) -> None:
+    # arrange
     _, subtask = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
@@ -64,8 +64,10 @@ def test_subtask_handler_injects_orchestrator_as_third_positional_argument(
         received["context"] = context
         received["orchestrator"] = orchestrator
 
+    # act
     handler(_direct_event(subtask.task_id, subtask.id), context="ctx-sentinel")
 
+    # assert
     assert received["orchestrator"] is harness.orchestrator
     assert received["context"] == "ctx-sentinel"
     assert received["body"]["sub_task_id"] == str(subtask.id)
@@ -74,6 +76,7 @@ def test_subtask_handler_injects_orchestrator_as_third_positional_argument(
 def test_subtask_handler_completes_parent_subtask_on_success(
     harness: Harness,
 ) -> None:
+    # arrange
     task, subtask = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
@@ -84,8 +87,10 @@ def test_subtask_handler_completes_parent_subtask_on_success(
     ) -> None:
         return None
 
+    # act
     handler(_direct_event(task.id, subtask.id), context=None)
 
+    # assert
     assert harness.subtasks.get(subtask.id).status is SubTaskStatus.COMPLETE
     assert harness.tasks.get(task.id).status is TaskStatus.COMPLETE
 
@@ -93,6 +98,7 @@ def test_subtask_handler_completes_parent_subtask_on_success(
 def test_subtask_handler_marks_parent_failed_and_reraises_on_error(
     harness: Harness,
 ) -> None:
+    # arrange
     task, subtask = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
@@ -103,6 +109,7 @@ def test_subtask_handler_marks_parent_failed_and_reraises_on_error(
     ) -> None:
         raise RuntimeError("boom")
 
+    # act / assert
     with pytest.raises(RuntimeError, match="boom"):
         handler(_direct_event(task.id, subtask.id), context=None)
 
@@ -113,6 +120,7 @@ def test_subtask_handler_marks_parent_failed_and_reraises_on_error(
 def test_subtask_handler_injected_orchestrator_can_create_child_subtask(
     harness: Harness,
 ) -> None:
+    # arrange
     task, subtask = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
@@ -126,8 +134,10 @@ def test_subtask_handler_injected_orchestrator_can_create_child_subtask(
         child = orchestrator.create_child_subtask(task.id, inputs={"split": 1})
         child_ids.append(child.id)
 
+    # act
     handler(_direct_event(task.id, subtask.id), context=None)
 
+    # assert
     assert len(child_ids) == 1
     persisted_child = harness.subtasks.get(child_ids[0])
     assert persisted_child.task_id == task.id
@@ -137,6 +147,7 @@ def test_subtask_handler_injected_orchestrator_can_create_child_subtask(
 def test_subtask_handler_logs_subtask_lifecycle_on_success(
     harness: Harness, caplog: pytest.LogCaptureFixture
 ) -> None:
+    # arrange
     task, subtask = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
@@ -147,9 +158,11 @@ def test_subtask_handler_logs_subtask_lifecycle_on_success(
     ) -> None:
         return None
 
+    # act
     with caplog.at_level(logging.INFO, logger=_LOGGER_NAME):
         handler(_direct_event(task.id, subtask.id), context=None)
 
+    # assert
     assert f"Running subtask {subtask.id}" in caplog.text
     assert f"Subtask {subtask.id} completed" in caplog.text
 
@@ -157,6 +170,7 @@ def test_subtask_handler_logs_subtask_lifecycle_on_success(
 def test_subtask_handler_logs_exception_on_failure(
     harness: Harness, caplog: pytest.LogCaptureFixture
 ) -> None:
+    # arrange
     task, subtask = harness.orchestrator.create_task_with_subtask(
         task_source="manual:test"
     )
@@ -167,6 +181,7 @@ def test_subtask_handler_logs_exception_on_failure(
     ) -> None:
         raise RuntimeError("boom")
 
+    # act / assert
     with caplog.at_level(logging.INFO, logger=_LOGGER_NAME):
         with pytest.raises(RuntimeError, match="boom"):
             handler(_direct_event(task.id, subtask.id), context=None)
@@ -181,6 +196,7 @@ def test_subtask_handler_logs_exception_on_failure(
 def test_subtask_handler_records_cloudwatch_url_on_subtask(
     harness: Harness, monkeypatch: pytest.MonkeyPatch
 ) -> None:
+    # arrange
     monkeypatch.setenv("AWS_REGION", "eu-west-2")
     monkeypatch.setenv(
         "AWS_LAMBDA_LOG_GROUP_NAME", "/aws/lambda/postcode-splitter"
@@ -198,8 +214,10 @@ def test_subtask_handler_records_cloudwatch_url_on_subtask(
     ) -> None:
         return None
 
+    # act
     handler(_direct_event(task.id, subtask.id), context=None)
 
+    # assert
     saved_url = harness.subtasks.get(subtask.id).cloud_logs_url
     assert saved_url is not None
     assert saved_url.startswith(
@@ -213,6 +231,7 @@ def test_subtask_handler_records_cloudwatch_url_on_subtask(
 def test_subtask_handler_leaves_cloudwatch_url_unset_outside_lambda(
     harness: Harness, monkeypatch: pytest.MonkeyPatch
 ) -> None:
+    # arrange
     for var in (
         "AWS_REGION",
         "AWS_LAMBDA_LOG_GROUP_NAME",
@@ -229,6 +248,8 @@ def test_subtask_handler_leaves_cloudwatch_url_unset_outside_lambda(
     ) -> None:
         return None
 
+    # act
     handler(_direct_event(task.id, subtask.id), context=None)
 
+    # assert
     assert harness.subtasks.get(subtask.id).cloud_logs_url is None