diff --git a/streamflow/cwl/translator.py b/streamflow/cwl/translator.py
index 5ca47d64f..187127a9a 100644
--- a/streamflow/cwl/translator.py
+++ b/streamflow/cwl/translator.py
@@ -1475,6 +1475,9 @@ def _inject_input(
             job_prefix=f"{global_name}-injector",
             connector_ports={target.deployment.name: deploy_step.get_output_port()},
             binding_config=binding_config,
+            input_directory=target.workdir or self.output_directory,
+            output_directory=target.workdir or self.output_directory,
+            tmp_directory=target.workdir or self.output_directory,
         )
         # Create a CWLInputInjector step to process the input
         injector_step = workflow.create_step(
diff --git a/tests/test_translator.py b/tests/test_translator.py
index 88fec25a8..8dd02ea6d 100644
--- a/tests/test_translator.py
+++ b/tests/test_translator.py
@@ -1,15 +1,30 @@
 from __future__ import annotations
 
 import json
+import os
 import tempfile
 from typing import MutableMapping, Any
 
+import cwltool.context
+import pytest
 
+from streamflow.config.config import WorkflowConfig
+from streamflow.config.validator import SfValidator
+from streamflow.cwl.token import CWLFileToken
+from streamflow.cwl.workflow import CWLWorkflow
+
+from streamflow.core import utils
 from streamflow.core.context import StreamFlowContext
 from streamflow.cwl.runner import main
 
 from cwltool.tests.util import get_data
 
+from streamflow.cwl.translator import CWLTranslator
+from streamflow.workflow.executor import StreamFlowExecutor
+from streamflow.workflow.token import TerminationToken
+from tests.utils.deployment import get_docker_deployment_config
+from tests.utils.workflow import CWL_VERSION
+
 
 def _create_file(content: MutableMapping[Any, Any]) -> str:
     temp_config = tempfile.NamedTemporaryFile(delete=False)
@@ -18,6 +33,104 @@ def _create_file(content: MutableMapping[Any, Any]) -> str:
     return temp_config.name
 
 
+@pytest.mark.asyncio
+async def test_inject_remote_input(context: StreamFlowContext) -> None:
+    cwl_workflow_path = "/path/to/local/cwl/wf_description"
+    remote_workdir = "/remote/workdir"
+    relative_data_path = "relative/path/to/data"
+    input_data = {
+        "class": "Directory",
+        "location": f"file://{cwl_workflow_path}/{relative_data_path}",
+    }
+    streamflow_config = {
+        "version": "v1.0",
+        "workflows": {
+            "test": {
+                "type": "cwl",
+                "config": {
+                    "file": "cwl/main.cwl",
+                    "settings": "cwl/config.yaml",
+                },
+                "bindings": [
+                    {
+                        "port": "/model",
+                        "target": {
+                            "deployment": "awesome",
+                            "workdir": remote_workdir,
+                        },
+                    }
+                ],
+            }
+        },
+        "deployments": {
+            "awesome": {
+                "type": "docker",
+                "config": {"image": get_docker_deployment_config().config["image"]},
+            }
+        },
+    }
+
+    # Check StreamFlow file schema
+    SfValidator().validate(streamflow_config)
+
+    # Build workflow
+    translator = CWLTranslator(
+        context=context,
+        name=utils.random_name(),
+        output_directory=tempfile.gettempdir(),
+        cwl_definition=None,  # cwltool.process.Process,
+        cwl_inputs={"model": input_data},
+        workflow_config=WorkflowConfig(
+            list(streamflow_config["workflows"].keys())[0], streamflow_config
+        ),
+        loading_context=cwltool.context.LoadingContext(),
+    )
+    workflow = CWLWorkflow(
+        context=context,
+        config={},
+        name=translator.name,
+        cwl_version=CWL_VERSION,
+    )
+    translator._inject_input(
+        workflow=workflow,
+        port_name="model",
+        global_name="/model",
+        port=workflow.create_port(),
+        output_directory=cwl_workflow_path,
+        value=translator.cwl_inputs["model"],
+    )
+
+    # Check input tokens
+    input_tokens = workflow.steps["/model-injector"].get_input_port("model").token_list
+    assert input_tokens[0].value == input_data
+    assert isinstance(input_tokens[1], TerminationToken)
+
+    # Execute injector steps
+    executor = StreamFlowExecutor(workflow)
+    await executor.run()
+
+    # Check output tokens
+    output_tokens = (
+        workflow.steps["/model-injector"].get_output_port("model").token_list
+    )
+    assert isinstance(output_tokens[0], CWLFileToken)
+    assert isinstance(output_tokens[1], TerminationToken)
+    job = (
+        workflow.steps["/model-injector/__schedule__"]
+        .get_output_port("__job__")
+        .token_list[0]
+        .value
+    )
+    assert (
+        len({job.input_directory, job.output_directory, job.tmp_directory}) == 1
+        and job.input_directory == remote_workdir
+    )
+    assert output_tokens[0].value["class"] == "Directory"
+    assert output_tokens[0].value["path"] == os.path.join(
+        remote_workdir, relative_data_path
+    )
+
+
 def test_dot_product_transformer_raises_error(context: StreamFlowContext) -> None:
     """Test DotProductSizeTransformer which must raise an exception because the size tokens have different values"""
     params = [