diff --git a/README.md b/README.md
index 51263d6..6ccce5c 100644
--- a/README.md
+++ b/README.md
@@ -217,7 +217,7 @@ for your patience on that one.
 
 ## :wave: Getting Help
 
-Security reports go through [private vulnerability reporting](https://github.com/deeplime-io/onecode/security/advisories/new). See [SECURITY.md](SECURITY.md). Security fixes are published for the maintained `1.x` line.
+Security reports go through [private vulnerability reporting](https://github.com/deeplime-io/onecode/security/advisories/new). See [SECURITY.md](https://github.com/deeplime-io/onecode/security/policy). Security fixes are published for the maintained `1.x` line.
 
 Issues labeled [`cra-vulnerability`](https://github.com/deeplime-io/onecode/issues?q=is%3Aissue+is%3Aopen+label%3Acra-vulnerability) come from the dependency scan. They are not the place to disclose a new flaw.
 
diff --git a/docs/changelogs/1.3.0.md b/docs/changelogs/1.3.0.md
index 5d34ce2..eccc4b2 100644
--- a/docs/changelogs/1.3.0.md
+++ b/docs/changelogs/1.3.0.md
@@ -1,4 +1,4 @@
-<!-- dev1 -->
+<!-- dev2 -->
 
 ## Bug Fixes
 
diff --git a/onecode/base/logger.py b/onecode/base/logger.py
index 02857bf..e0f6968 100644
--- a/onecode/base/logger.py
+++ b/onecode/base/logger.py
@@ -82,14 +82,19 @@ def format(
 class Logger(metaclass=Singleton):
     """
     Single Logger object to handle Python logging within OneCode projects.
-    The default logging level is INFO. See [`set_level()`][onecode.Logger.set_level] to change it.
-    Use the static methods [`debug()`][onecode.Logger.debug], [`info()`][onecode.Logger.info],
-    [`warning()`][onecode.Logger.warning], [`error()`][onecode.Logger.error], and
-    [`critical()`][onecode.Logger.critical] to conveniently log your messages with the
-    corresponding logging level.
+    The default logging level is INFO. See
+    [`set_level()`][onecode.base.logger.Logger.set_level] to change it.
+    Use the static methods
+    [`debug()`][onecode.base.logger.Logger.debug],
+    [`info()`][onecode.base.logger.Logger.info],
+    [`warning()`][onecode.base.logger.Logger.warning],
+    [`error()`][onecode.base.logger.Logger.error], and
+    [`critical()`][onecode.base.logger.Logger.critical]
+    to conveniently log your messages with the corresponding logging level.
 
     By default, the ColoredFormatter is used. You may add other logging handlers using
-    [`add_handler()`][onecode.Logger.add_handler], for instance to redirect logs to a file.
+    [`add_handler()`][onecode.base.logger.Logger.add_handler],
+    for instance to redirect logs to a file.
 
     !!! example
         ```py
@@ -194,9 +199,12 @@ def logger(
     ) -> logging.Logger:
         """
         Get the Python Logger object corresponding to the given stack level. Preferentially use
-        convenience methods [`debug()`][onecode.Logger.debug], [`info()`][onecode.Logger.info],
-        [`warning()`][onecode.Logger.warning], [`error()`][onecode.Logger.error], and
-        [`critical()`][onecode.Logger.critical].
+        convenience methods
+        [`debug()`][onecode.base.logger.Logger.debug],
+        [`info()`][onecode.base.logger.Logger.info],
+        [`warning()`][onecode.base.logger.Logger.warning],
+        [`error()`][onecode.base.logger.Logger.error], and
+        [`critical()`][onecode.base.logger.Logger.critical].
 
         Args:
             stacklevel: Number of hops back in the function call stack to tie the logger to.
@@ -214,7 +222,7 @@ def logger(
     def _flush() -> None:
         """
         Force flush to stdout if `ConfigOption.FLUSH_STDOUT` is True. See
-            [Project.config][onecode.Project.config] for more information.
+            [Project.config][onecode.base.project.Project.config] for more information.
 
         """
         if Project().get_config(ConfigOption.FLUSH_STDOUT):
diff --git a/onecode/base/project.py b/onecode/base/project.py
index 9240e68..7f2bdda 100644
--- a/onecode/base/project.py
+++ b/onecode/base/project.py
@@ -20,7 +20,7 @@ class Project(metaclass=Singleton):
     parameter values, registered elements, flow currently running, current running
     mode, etc.
 
-    See [`reset()`][onecode.Project.reset] for Project default's initialization.
+    See [`reset()`][onecode.base.project.Project.reset] for Project default's initialization.
 
     Attributes:
         registered_elements: List of elements registered for processing.
@@ -182,8 +182,9 @@ def current_flow(
     @property
     def data_root(self) -> str:
         """
-        Get the path to the root of the data folder. See [`reset()`][onecode.Project.reset] to know
-        how the data path is initialized.
+        Get the path to the root of the data folder. See
+        [`reset()`][onecode.base.project.Project.reset]
+        to know how the data path is initialized.
 
         """
         return self._data_root
diff --git a/pyproject.toml b/pyproject.toml
index e6c6ded..6a2acbf 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -88,6 +88,11 @@ onecode-check = 'onecode.cli.check:main'
 requires = ["poetry-core>=1.0.0"]
 build-backend = "poetry.core.masonry.api"
 
+[tool.pytest.ini_options]
+# datatest 0.11's plugin probes pytest internals that moved in pytest 7+ and
+# warns on every xdist worker. Tests only use datatest.working_directory.
+addopts = "-p no:datatest"
+
 [tool.ruff]
 line-length = 100
 target-version = "py311"
diff --git a/tests/unit/base/test_run_project.py b/tests/unit/base/test_run_project.py
index a743c37..373204e 100644
--- a/tests/unit/base/test_run_project.py
+++ b/tests/unit/base/test_run_project.py
@@ -80,14 +80,15 @@ def write_output(data, flow_name):
 
 
 def _process_class():
-    # Linux 3.14's default fork server keeps the environment from the first
-    # worker, so a later run still sees a deleted data path. Fork sees the
-    # current ONECODE_PROJECT_DATA. Every other platform keeps Process, which
-    # is the start method these tests have always used.
+    # Fork from this multi-threaded pytest process warns, and can deadlock.
+    # Linux 3.14's fork server also keeps ONECODE_PROJECT_DATA from the first
+    # worker. Spawn starts a fresh interpreter with the current environment.
+    # Windows and macOS already default to spawn; a second spawn context drops
+    # workers there, so those platforms keep Process.
     ctx = multiprocessing.get_context()
-    if type(ctx).__name__ == "ForkServerContext":
-        return multiprocessing.get_context("fork").Process
-    return multiprocessing.Process
+    if type(ctx).__name__ == "SpawnContext":
+        return multiprocessing.Process
+    return multiprocessing.get_context("spawn").Process
 
 
 def run():
@@ -148,14 +149,15 @@ def write_output(data, flow_name):
 
 
 def _process_class():
-    # Linux 3.14's default fork server keeps the environment from the first
-    # worker, so a later run still sees a deleted data path. Fork sees the
-    # current ONECODE_PROJECT_DATA. Every other platform keeps Process, which
-    # is the start method these tests have always used.
+    # Fork from this multi-threaded pytest process warns, and can deadlock.
+    # Linux 3.14's fork server also keeps ONECODE_PROJECT_DATA from the first
+    # worker. Spawn starts a fresh interpreter with the current environment.
+    # Windows and macOS already default to spawn; a second spawn context drops
+    # workers there, so those platforms keep Process.
     ctx = multiprocessing.get_context()
-    if type(ctx).__name__ == "ForkServerContext":
-        return multiprocessing.get_context("fork").Process
-    return multiprocessing.Process
+    if type(ctx).__name__ == "SpawnContext":
+        return multiprocessing.Process
+    return multiprocessing.get_context("spawn").Process
 
 
 def run():