diff --git a/test/cloudtest/test_full_testdrive.py b/test/cloudtest/test_full_testdrive.py
index 451e623862c2a..16b1205111418 100644
--- a/test/cloudtest/test_full_testdrive.py
+++ b/test/cloudtest/test_full_testdrive.py
@@ -11,7 +11,7 @@
 
 import pytest
 
-from materialize import buildkite
+from materialize import MZ_ROOT, buildkite
 from materialize.cloudtest.app.materialize_application import MaterializeApplication
 
 
@@ -24,7 +24,9 @@ def test_full_testdrive(mz: MaterializeApplication) -> None:
     parser.add_argument("--file-pattern", default="*.td", type=str)
     args, _ = parser.parse_known_args()
 
-    matching_files = glob.glob(f"testdrive/{args.file_pattern}", root_dir="test")
+    matching_files = glob.glob(
+        f"testdrive/{args.file_pattern}", root_dir=MZ_ROOT / "test"
+    )
 
     # TODO: database-issues#7827 (test requires fivetran running in cloudtest)
     matching_files.remove("testdrive/fivetran-destination.td")
diff --git a/test/mysql-cdc-old-syntax/mzcompose.py b/test/mysql-cdc-old-syntax/mzcompose.py
index 322998e556fdd..e14bf3a464876 100644
--- a/test/mysql-cdc-old-syntax/mzcompose.py
+++ b/test/mysql-cdc-old-syntax/mzcompose.py
@@ -15,7 +15,7 @@
 import threading
 from textwrap import dedent
 
-from materialize import buildkite
+from materialize import MZ_ROOT, buildkite
 from materialize.mysql_util import (
     retrieve_invalid_ssl_context_for_mysql,
     retrieve_ssl_context_for_mysql,
@@ -112,7 +112,9 @@ def workflow_cdc(c: Composition, parser: WorkflowArgumentParser) -> None:
 
     matching_files = []
     for filter in args.filter:
-        matching_files.extend(glob.glob(filter, root_dir="test/mysql-cdc-old-syntax"))
+        matching_files.extend(
+            glob.glob(filter, root_dir=MZ_ROOT / "test" / "mysql-cdc-old-syntax")
+        )
     sharded_files: list[str] = sorted(
         buildkite.shard_list(matching_files, lambda file: file)
     )
@@ -299,7 +301,9 @@ def workflow_migration(c: Composition, parser: WorkflowArgumentParser) -> None:
 
     matching_files = []
     for filter in args.filter:
-        matching_files.extend(glob.glob(filter, root_dir="test/mysql-cdc-old-syntax"))
+        matching_files.extend(
+            glob.glob(filter, root_dir=MZ_ROOT / "test" / "mysql-cdc-old-syntax")
+        )
 
     sharded_files: list[str] = sorted(
         buildkite.shard_list(matching_files, lambda file: file)
diff --git a/test/pg-cdc-old-syntax/mzcompose.py b/test/pg-cdc-old-syntax/mzcompose.py
index 43a8ea2933ccc..260b79f4e166c 100644
--- a/test/pg-cdc-old-syntax/mzcompose.py
+++ b/test/pg-cdc-old-syntax/mzcompose.py
@@ -17,7 +17,7 @@
 import pg8000
 from pg8000 import Connection
 
-from materialize import buildkite
+from materialize import MZ_ROOT, buildkite
 from materialize.mzcompose.composition import Composition, WorkflowArgumentParser
 from materialize.mzcompose.service import Service, ServiceConfig
 from materialize.mzcompose.services.materialized import Materialized
@@ -303,7 +303,9 @@ def workflow_cdc(c: Composition, parser: WorkflowArgumentParser) -> None:
 
     matching_files = []
     for filter in args.filter:
-        matching_files.extend(glob.glob(filter, root_dir="test/pg-cdc-old-syntax"))
+        matching_files.extend(
+            glob.glob(filter, root_dir=MZ_ROOT / "test" / "pg-cdc-old-syntax")
+        )
     sharded_files: list[str] = sorted(
         buildkite.shard_list(matching_files, lambda file: file)
     )
@@ -378,7 +380,9 @@ def workflow_migration(c: Composition, parser: WorkflowArgumentParser) -> None:
 
     matching_files = []
     for filter in args.filter:
-        matching_files.extend(glob.glob(filter, root_dir="test/pg-cdc-old-syntax"))
+        matching_files.extend(
+            glob.glob(filter, root_dir=MZ_ROOT / "test" / "pg-cdc-old-syntax")
+        )
     sharded_files: list[str] = sorted(
         buildkite.shard_list(matching_files, lambda file: file)
     )
diff --git a/test/pg-cdc/mzcompose.py b/test/pg-cdc/mzcompose.py
index 37f2b0b6eab28..ec9003de73d28 100644
--- a/test/pg-cdc/mzcompose.py
+++ b/test/pg-cdc/mzcompose.py
@@ -18,7 +18,7 @@
 import psycopg
 from psycopg import Connection
 
-from materialize import buildkite
+from materialize import MZ_ROOT, buildkite
 from materialize.mzcompose.composition import Composition, WorkflowArgumentParser
 from materialize.mzcompose.service import Service, ServiceConfig
 from materialize.mzcompose.services.materialized import Materialized
@@ -291,7 +291,7 @@ def workflow_cdc(c: Composition, parser: WorkflowArgumentParser) -> None:
 
     matching_files = []
     for filter in args.filter:
-        matching_files.extend(glob.glob(filter, root_dir="test/pg-cdc"))
+        matching_files.extend(glob.glob(filter, root_dir=MZ_ROOT / "test" / "pg-cdc"))
     sharded_files: list[str] = sorted(
         buildkite.shard_list(matching_files, lambda file: file)
     )
diff --git a/test/testdrive-old-kafka-src-syntax/mzcompose.py b/test/testdrive-old-kafka-src-syntax/mzcompose.py
index 3491e6b711218..a245b730212bc 100644
--- a/test/testdrive-old-kafka-src-syntax/mzcompose.py
+++ b/test/testdrive-old-kafka-src-syntax/mzcompose.py
@@ -15,7 +15,7 @@
 import glob
 from pathlib import Path
 
-from materialize import ci_util, spawn
+from materialize import MZ_ROOT, ci_util, spawn
 from materialize.mzcompose import get_default_system_parameters
 from materialize.mzcompose.composition import Composition, WorkflowArgumentParser
 from materialize.mzcompose.services.azure import Azurite
@@ -304,7 +304,9 @@ def workflow_migration(c: Composition, parser: WorkflowArgumentParser) -> None:
     matching_files = []
     for filter in args.files:
         matching_files.extend(
-            glob.glob(filter, root_dir="test/testdrive-old-kafka-src-syntax")
+            glob.glob(
+                filter, root_dir=MZ_ROOT / "test" / "testdrive-old-kafka-src-syntax"
+            )
         )
     matching_files = [file for file in matching_files if file != "session.td"]