fix(runners): close shared toolsets only from the last runner

[claxman]

Please ensure you have read the contribution guide before creating a pull request.

Link to Issue or Description of Change

1. Link to an existing issue (if applicable):

• Closes: [BUG] When executing Runners(of which agent contains stateful toolset) in parallel, toolset is being released when only one of them has finished #4045

Problem:
Runner.close() always calls toolset.close() for every toolset on the agent. Two Runners sharing one LlmAgent with a CountingToolset printed after_first_close 1 then after_second_close 2. The first close already released the toolset while the second runner was still open. Parallel AgentTool runs on agents that share a toolset hit the same path (await runner.close() after each nested run).

Solution:
Each Runner is a weak holder of the toolsets it collected. close() removes that holder and only calls toolset.close() when no live runner remains. A Runner dropped without close() falls out of the set, so it cannot pin the toolset open. close() also re-collects toolsets so one appended after construct is still released. AgentTool.run_async closes its nested runner in a finally so a raising nested run still unwinds on the same task.

Testing Plan

Unit Tests:

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

PYTHONPATH=src .venv/bin/python -m pytest -q tests/unittests/test_runners.py::TestRunnerWithPlugins::test_shared_toolset_stays_open_until_last_runner_closes tests/unittests/test_runners.py::TestRunnerWithPlugins::test_single_runner_still_closes_its_toolset tests/unittests/test_runners.py::TestRunnerWithPlugins::test_dropped_runner_does_not_pin_shared_toolset tests/unittests/test_runners.py::TestRunnerWithPlugins::test_late_added_toolset_is_closed tests/unittests/tools/test_agent_tool.py::test_parallel_agent_tools_do_not_close_shared_toolset_early tests/unittests/tools/test_agent_tool.py::test_failed_agent_tool_still_closes_shared_toolset

6 passed.

With runners.py and agent_tool.py checked out from b0180620, four of those six fail (test_shared_toolset_stays_open_until_last_runner_closes, test_single_runner_still_closes_its_toolset, test_parallel_agent_tools_do_not_close_shared_toolset_early, test_failed_agent_tool_still_closes_shared_toolset). The dropped-runner and late-added tests pass on main (every Runner.close() already closed the toolset) and guard the WeakSet / re-collect paths.

.venv/bin/python -m mypy src/google/adk/runners.py: Success, no issues.

Manual End-to-End (E2E) Tests:

Not run. The CountingToolset runners and the parallel AgentTool test cover the two cases in the issue.

Additional context

Claim: #4045 (comment)

#4046 added an AgentToolManager singleton and was closed on 2026-05-12 after requested changes stayed open.

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.