diff --git a/src/aiida_shell/launch.py b/src/aiida_shell/launch.py
index 8b53865..15e8d47 100644
--- a/src/aiida_shell/launch.py
+++ b/src/aiida_shell/launch.py
@@ -32,6 +32,7 @@ def launch_shell_job(  # noqa: PLR0913
     metadata: dict[str, t.Any] | None = None,
     submit: bool = False,
     resolve_command: bool = True,
+    monitors: dict[str, Data] | None = None,
 ) -> tuple[dict[str, Data], ProcessNode]:
     """Launch a :class:`aiida_shell.ShellJob` job for the given command.
 
@@ -52,6 +53,8 @@ def launch_shell_job(  # noqa: PLR0913
     :param resolve_command: Whether to resolve the command to the absolute path of the executable. If set to ``True``,
         the ``which`` command is executed on the target computer to attempt and determine the absolute path. Otherwise,
         the command is set as the ``filepath_executable`` attribute of the created ``AbstractCode`` instance.
+    :param monitors: Optional dictionary of ``Data`` nodes to be used as monitors for the job (see AiiDA
+        documentation on how to define monitors).
     :raises TypeError: If the value specified for ``metadata.options.computer`` is not a ``Computer``.
     :raises ValueError: If ``resolve_command=True`` and the absolute path of the command on the computer could not be
         determined.
@@ -69,6 +72,7 @@ def launch_shell_job(  # noqa: PLR0913
         parser=parser,
         metadata=metadata,
         resolve_command=resolve_command,
+        monitors=monitors,
     )
 
     if submit:
@@ -91,6 +95,7 @@ def prepare_shell_job_inputs(  # noqa: PLR0913
     parser: ParserFunctionType | str | None = None,
     metadata: dict[str, t.Any] | None = None,
     resolve_command: bool = True,
+    monitors: dict[str, Data] | None = None,
 ) -> dict[str, t.Any]:
     """Prepare inputs for the ShellJob based on the provided parameters.
 
@@ -110,6 +115,8 @@ def prepare_shell_job_inputs(  # noqa: PLR0913
     :param resolve_command: Whether to resolve the command to the absolute path of the executable. If set to ``True``,
         the ``which`` command is executed on the target computer to attempt and determine the absolute path. Otherwise,
         the command is set as the ``filepath_executable`` attribute of the created ``AbstractCode`` instance.
+    :param monitors: Optional dictionary of ``Data`` nodes to be used as monitors for the job (see AiiDA
+        documentation on how to define monitors).
     :raises TypeError: If the value specified for ``metadata.options.computer`` is not a ``Computer``.
     :raises ValueError: If ``resolve_command=True`` and the absolute path of the command on the computer could not be
         determined.
@@ -148,6 +155,8 @@ def prepare_shell_job_inputs(  # noqa: PLR0913
         'parser': parser,
         'metadata': metadata or {},
     }
+    if monitors:
+        inputs['monitors'] = monitors
 
     return inputs
 
diff --git a/src/aiida_shell/parsers/shell.py b/src/aiida_shell/parsers/shell.py
index 4b701e7..fef695e 100644
--- a/src/aiida_shell/parsers/shell.py
+++ b/src/aiida_shell/parsers/shell.py
@@ -84,7 +84,7 @@ def parse_default_outputs(self, dirpath: pathlib.Path) -> ExitCode:
         except FileNotFoundError:
             stderr = ''
         else:
-            stderr = node_stderr.get_content()  # type: ignore[assignment]
+            stderr = node_stderr.get_content(mode='r')
             self.out(ShellJob.FILENAME_STDERR, node_stderr)
 
         filename_stdout = self.node.get_option('output_filename') or ShellJob.FILENAME_STDOUT
diff --git a/tests/calculations/test_shell.py b/tests/calculations/test_shell.py
index c1b6531..96318c5 100644
--- a/tests/calculations/test_shell.py
+++ b/tests/calculations/test_shell.py
@@ -286,7 +286,11 @@ def test_output_filename(generate_calc_job, generate_code, file_regression):
 def test_filename_stdin(generate_calc_job, generate_code, file_regression):
     """Test the ``metadata.options.filename_stdin`` input."""
     inputs = {
-        'code': generate_code('cat'),
+        # even if 'cat' would be more natural, we use `diff` to avoid issues
+        # in this specific test, because the exact path of the `cat` binary
+        # may differ across systems (Linux vs. MacOS), while `diff` seems
+        # to be (by default) more consistent (in /usr/bin).
+        'code': generate_code('diff'),
         'arguments': List(['{filename}']),
         'nodes': {'filename': SinglefileData.from_string('content')},
         'metadata': {'options': {'filename_stdin': 'filename'}},
diff --git a/tests/calculations/test_shell/test_filename_stdin.txt b/tests/calculations/test_shell/test_filename_stdin.txt
index 6fdccec..18521f0 100644
--- a/tests/calculations/test_shell/test_filename_stdin.txt
+++ b/tests/calculations/test_shell/test_filename_stdin.txt
@@ -3,6 +3,6 @@ exec > _scheduler-stdout.txt
 exec 2> _scheduler-stderr.txt
 
 
-'/usr/bin/cat' < 'filename' > 'stdout' 2> 'stderr'
+'/usr/bin/diff' < 'filename' > 'stdout' 2> 'stderr'
 
 echo $? > status
diff --git a/tests/conftest.py b/tests/conftest.py
index 46a8ac0..5dd2df4 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -3,6 +3,7 @@
 
 import collections
 import pathlib
+import shutil
 import tempfile
 import typing as t
 import uuid
@@ -170,8 +171,10 @@ def factory(label='localhost', hostname='localhost', scheduler_type='core.direct
 @pytest.fixture
 def generate_code(generate_computer):
     """Return a :class:`aiida_shell.data.code.ShellCode` instance, either already existing or created."""
+    default_command = shutil.which('true')  # /bin/true on Linux, /usr/bin/true on macOS
+    assert default_command is not None, 'The `true` command must be available on the system for the tests to run.'
 
-    def factory(command='/bin/true', computer_label='localhost', label=None, entry_point_name='core.shell'):
+    def factory(command=default_command, computer_label='localhost', label=None, entry_point_name='core.shell'):
         """Return a :class:`aiida_shell.data.code.ShellCode` instance, either already existing or created."""
         label = label or str(uuid.uuid4())
         computer = generate_computer(computer_label)
diff --git a/tests/test_launch.py b/tests/test_launch.py
index dfbe236..974aae7 100644
--- a/tests/test_launch.py
+++ b/tests/test_launch.py
@@ -10,6 +10,9 @@
 from aiida_shell.calculations.shell import ShellJob
 from aiida_shell.launch import launch_shell_job, prepare_computer
 
+DATE_COMMAND = shutil.which('date')
+assert DATE_COMMAND is not None, 'The `date` command must be available in order to run the tests.'
+
 
 class ShellWorkChain(WorkChain):
     """Implementation of :class:`aiida.engine.processes.workchains.workchain.WorkChain` that submits a ``ShellJob``."""
@@ -86,7 +89,7 @@ def test_arguments():
     shellfunction runs just before midnight and the comparison ``datetime`` call runs in the next day causing the test
     to fail, but that seems extremely unlikely.
     """
-    arguments = ['--iso-8601']
+    arguments = ['-I']  # equivalent to --iso-8601, but supported also on MacOS
     results, node = launch_shell_job('date', arguments=arguments)
 
     assert node.is_finished_ok
@@ -273,7 +276,7 @@ def job_function():
 @pytest.mark.parametrize(
     'resolve_command, executable',
     (
-        (True, '/usr/bin/date'),
+        (True, DATE_COMMAND),
         (False, 'date'),
     ),
 )