Failing job: Tests: Python 3.10 on windows-latest (Job ID: 64080726796)
All other jobs: PASS (ubuntu-latest 3.10, ubuntu-latest 3.13, lowest-deps, integration)

Log excerpt showing the timeout:

tests\cli\test_run.py .......+++++++++++++++++++++++++++++++++ Timeout +++++++++++++++++++++++++++++++++++

8 tests passed, then test_run_mcp_config (9th test) timed out.

Captured stderr at timeout:

DEBUG  Inferred transport: <MCPConfigTransport(config='mcpServers={'test_server': StdioMCPServer(...)}')>
DEBUG  Inferred transport: <FastMCPTransport(server='FastMCP Proxy-115d')>
DEBUG  [Client-aa99] called list_tools
DEBUG  Stdio transport connected

The stdio transport had just connected when the 5-second timeout fired. The event loop was blocked in select.select() waiting on Windows pipe I/O — a known source of non-deterministic latency.

Stack trace at timeout:

Stack of MainThread (3420):
  asyncio\base_events.py:1871, in _run_once
    event_list = self._selector.select(timeout)
  selectors.py:324, in select
    r, w, _ = self._select(self._readers, self._writers, [], timeout)
  selectors.py:315, in _select
    r, w, x = select.select(r, w, w, timeout)

PR diff: Only README.md, docs/changelog.mdx, docs/getting-started/welcome.mdx, and docs/updates.mdx were modified — all prose/documentation changes.

• tests/cli/test_run.py — Contains TestMCPConfig::test_run_mcp_config, the failing test. It spawns a real subprocess via stdio and uses a 5-second pytest timeout.
• README.md, docs/changelog.mdx, docs/getting-started/welcome.mdx, docs/updates.mdx — The only files changed in this PR; none are Python files.