fix(tools): run async tools that are wrapped by sync decorators

[techi101]

Link to Issue or Description of Change

2. Or, if no issue exists, describe the change:

Problem:

FunctionTool._invoke_callable decides whether to await a tool with inspect.iscoroutinefunction(). That check does not see through a plain sync decorator, which is a common pattern for logging, retries or rate limiting:

def logged(fn):
  @functools.wraps(fn)
  def wrapper(*args, **kwargs):
    return fn(*args, **kwargs)
  return wrapper

@logged
async def get_price(item: str) -> dict:
  return {"item": item, "price": 3}

The declaration is still built correctly (inspect.signature follows __wrapped__), so the model calls the tool normally. But the sync branch returns the coroutine without awaiting it, so the tool body never runs and the model receives:

{'result': <coroutine object get_price at 0x...>}

Solution:

On the sync path (both the direct call and the bound sync-callable runner), await the result if it is awaitable. This matches how ADK already treats user callables elsewhere, e.g. before/after tool callbacks in utils/_callback_pipeline.py. Regular sync and async tools are unaffected.

Related: #7012 handles an awaitable returned by a require_confirmation predicate by failing closed. With this change _invoke_callable awaits such a result, so a sync wrapper around an async predicate returns its real bool.

Testing Plan

Unit Tests:

• I have added or updated unit tests for my change.
• All unit tests pass locally.

Added to tests/unittests/tools/test_function_tool.py:

• test_run_async_awaits_async_function_behind_sync_wrapper
• test_run_async_awaits_async_function_behind_sync_wrapper_with_runner (same case with a thread-pool sync callable runner bound)

Both fail on main with assert <coroutine object ...> == {'item': 'apple', 'price': 3} and pass with this change.

pytest tests/unittests/tools/test_function_tool.py
61 passed

pytest tests/unittests/tools tests/unittests/flows tests/unittests/agents tests/unittests/workflow tests/unittests/test_runners.py
5625 passed, 2 skipped, 7 xfailed, 3 failed

The 3 failures are tests/unittests/tools/test_skill_toolset.py::test_integration_shell_*, which also fail on main on my machine (Windows, no /bin/bash), so they are unrelated.

Manual End-to-End (E2E) Tests:

The agent above, run through InMemoryRunner with a mock model that calls get_price(item="apple"):

mock = testing_utils.MockModel.create(responses=[
    [types.Part.from_function_call(name="get_price", args={"item": "apple"})],
    "done",
])
runner = testing_utils.InMemoryRunner(LlmAgent(name="root", model=mock, tools=[get_price]))
for event in runner.run("price of an apple?"):
  ...  # print function_response parts

Before (main):

function_response: {'result': <coroutine object get_price at 0x00000200178D96C0>}

After:

function_response: {'item': 'apple', 'price': 3}

Checklist

• I have read the CONTRIBUTING.md document.
• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas.
• I have added tests that prove my fix is effective or that my feature works.
• New and existing unit tests pass locally with my changes.
• I have manually tested my changes end-to-end.
• Any dependent changes have been merged and published in downstream modules.