fix: add timeout to MCP server disconnect to prevent hung removals

stack.aclose() on a stuck streamable-http transport hangs indefinitely,
causing 50% CPU on all nodes when removing a broken remote server via
reconcile_sync. Wrap with asyncio.wait_for(timeout=10s) so cleanup
proceeds even if the transport refuses to close cleanly.

turnstone/core/mcp_client.py

@@ -977,8 +977,10 @@ class MCPClientManager:
                 self._sessions.pop(name, None)
                 stack = self._per_server_stacks.pop(name, None)
                 if stack is not None:
-                    with contextlib.suppress(Exception):
-                        await stack.aclose()
+                    try:
+                        await asyncio.wait_for(stack.aclose(), timeout=10)
+                    except (TimeoutError, Exception):
+                        log.warning("Timed out closing MCP server '%s', forcing cleanup", name)
                 # Clean up per-server state (on the event loop thread)
                 self._per_server_tools.pop(name, None)
                 self._per_server_resources.pop(name, None)