test(mcp): assert stored tool_filter in McpToolset init test

[anxkhn]

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):

• N/A (no existing issue; this is a small test-fidelity fix, described below).

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

Problem:

test_init_with_tool_filter_list in
tests/unittests/tools/mcp_tool/test_mcp_toolset.py is meant to check that a
list tool_filter passed to McpToolset(...) is retained, but its only
assertion is:

assert toolset._is_tool_selected is not None

_is_tool_selected is a regular method inherited from BaseToolset
(src/google/adk/tools/base_toolset.py), so on any instance it is a bound method
that is always non-None, independent of the tool_filter value. The assertion
can never fail and never checks what the test name promises: that the list filter
was stored. A regression that dropped or mis-stored tool_filter in __init__
would still pass green. The test's own comment even noted the intended
verification ("checking the filtering behavior in get_tools") was never written
here.

Solution:

Assert the stored value directly:

assert toolset.tool_filter == tool_filter

tool_filter is stored as a public attribute in BaseToolset.__init__
(self.tool_filter = tool_filter) and forwarded by McpToolset.__init__ via
super().__init__(...), so this is a valid, non-tautological check. It mirrors
the sibling init tests in the same file, which assert
toolset._connection_params == <params>. The actual get_tools list-filtering
behavior is already covered separately by test_get_tools_with_list_filter, so
no behavioral coverage is lost. The stale two-line comment is replaced with a
single accurate one. Production code is untouched.

Testing Plan

Unit Tests:

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

Effectiveness (red -> green) was verified by temporarily setting
self.tool_filter = None in BaseToolset.__init__ to simulate a dropped filter:
the old assertion still passed (confirming it was vacuous), while the new
assertion failed with AssertionError: assert None == ['tool1', 'tool2']. The
temporary production change was reverted; this PR contains only the test change.

pytest summary (local):

# target test
tests/unittests/tools/mcp_tool/test_mcp_toolset.py::TestMcpToolset::test_init_with_tool_filter_list
-> 1 passed

# full file
tests/unittests/tools/mcp_tool/test_mcp_toolset.py
-> 35 passed

Manual End-to-End (E2E) Tests:

Not applicable. This is a unit-test-only change with no runtime or user-facing
behavior change.

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. (N/A: unit-test-only change.)
• Any dependent changes have been merged and published in downstream modules. (None.)

Additional context

Diff is +2/-3 in a single file (tests/unittests/tools/mcp_tool/test_mcp_toolset.py);
no production code is changed. pre-commit (isort, pyink, addlicense, ADK
Compliance Checks) passes on the changed file.