theoretically more backwards compat

libs/langgraph/langgraph/_internal/_runnable.py

@@ -316,7 +316,10 @@ class RunnableCallable(Runnable):
                 continue
 
             # If the kwarg is accepted by the function, store the key / runtime attribute to inject
-            self.func_accepts[kw] = (runtime_key, default)
+            # Use the actual parameter default from the function signature if available,
+            # otherwise fall back to the default from KWARGS_CONFIG_KEYS
+            param_default = p.default if p.default is not inspect.Parameter.empty else default
+            self.func_accepts[kw] = (runtime_key, param_default)
 
     def __repr__(self) -> str:
         repr_args = {

libs/prebuilt/langgraph/prebuilt/tool_node.py

@@ -84,6 +84,7 @@ from langchain_core.tools.base import (
 from langgraph._internal._runnable import RunnableCallable
 from langgraph.errors import GraphBubbleUp
 from langgraph.graph.message import REMOVE_ALL_MESSAGES
+from langgraph.runtime import DEFAULT_RUNTIME
 from langgraph.store.base import BaseStore  # noqa: TC002
 from langgraph.types import Command, Send, StreamWriter
 from pydantic import BaseModel, ValidationError
@@ -702,7 +703,7 @@ class ToolNode(RunnableCallable):
         self,
         input: list[AnyMessage] | dict[str, Any] | BaseModel,
         config: RunnableConfig,
-        runtime: Runtime,
+        runtime: Runtime = DEFAULT_RUNTIME,
     ) -> Any:
         tool_calls, input_type = self._parse_input(input)
         config_list = get_config_list(config, len(tool_calls))
@@ -734,7 +735,7 @@ class ToolNode(RunnableCallable):
         self,
         input: list[AnyMessage] | dict[str, Any] | BaseModel,
         config: RunnableConfig,
-        runtime: Runtime,
+        runtime: Runtime = DEFAULT_RUNTIME,
     ) -> Any:
         tool_calls, input_type = self._parse_input(input)
         config_list = get_config_list(config, len(tool_calls))

libs/prebuilt/tests/test_tool_node.py

@@ -9,7 +9,6 @@ from typing import (
     NoReturn,
     TypeVar,
 )
-from unittest.mock import Mock
 
 import pytest
 from langchain_core.messages import (
@@ -51,27 +50,26 @@ from .model import FakeToolCallingModel
 pytestmark = pytest.mark.anyio
 
 
-def _create_mock_runtime(store: BaseStore | None = None) -> Mock:
-    """Create a mock Runtime object for testing ToolNode outside of graph context.
-
-    This helper is needed because ToolNode._func expects a Runtime parameter
-    which is injected by RunnableCallable from config["configurable"]["__pregel_runtime"].
-    When testing ToolNode directly (outside a graph), we need to provide this manually.
-    """
-    mock_runtime = Mock()
-    mock_runtime.store = store
-    mock_runtime.context = None
-    mock_runtime.stream_writer = lambda *args, **kwargs: None
-    return mock_runtime
-
-
 def _create_config_with_runtime(store: BaseStore | None = None) -> RunnableConfig:
-    """Create a RunnableConfig with mock Runtime for testing ToolNode.
+    """Create a RunnableConfig for testing ToolNode.
+
+    Since ToolNode now has a default Runtime, this helper can be simplified.
+    It only needs to inject a store if one is provided, otherwise an empty config works.
+
+    Args:
+        store: Optional store to inject via runtime. If None, no runtime is needed.
 
     Returns:
-        RunnableConfig with __pregel_runtime in configurable dict.
+        RunnableConfig, optionally with __pregel_runtime if store is provided.
     """
-    return {"configurable": {"__pregel_runtime": _create_mock_runtime(store)}}
+    if store is None:
+        # No runtime needed - ToolNode will use DEFAULT_RUNTIME
+        return {}
+
+    # Create a mock runtime only when we need to inject a store
+    from langgraph.runtime import Runtime
+    runtime = Runtime(context=None, store=store, stream_writer=lambda *args, **kwargs: None)
+    return {"configurable": {"__pregel_runtime": runtime}}
 
 
 def tool1(some_val: int, some_other_val: str) -> str: