How to send custom stream events from an MCP Client to an app connected to an agent, similar to `get_stream_writer` (langgraph) with `custom` mode?

**Evironment** langchain-mcp-adapters: 0.2.1 ## ❓ Question: Forwarding tool progress events from MCP Client to an agent-connected application Currently, I am able to receive tool execution progress via `on_progress` callbacks on the MCP client side. However, I am not sure how to **forward these progress events back to the application that is connected to the agent**, so that the UI can react dynamically (e.g. display real-time tool progress, loading states, or intermediate steps). --- ``` async def on_progress( progress: float, total: float | None, message: str | None, context: CallbackContext, ): """Handle progress updates from MCP servers.""" percent = (progress / total * 100) if total else progress tool_info = f" ({context.tool_name})" if context.tool_name else "" print(f"[{context.server_name}{tool_info}] Progress: {percent:.1f}% - {message}") async def on_logging_message( params: LoggingMessageNotificationParams, context: CallbackContext, ): """Handle log messages from MCP servers.""" print(f"[{context.server_name}] {params.level}: {params.data}") async def runtime_interceptor( request: MCPToolCallRequest, handler ) -> CallToolResult: request.headers = {"Authorizatahaksion": f"Bearer 123456789"} return await handler(request) async def make_graph(): client = MultiServerMCPClient( { "math": { "transport": "streamable_http", # HTTP-based remote server # Ensure you start your weather server on port 8000 "url": "http://localhost:8000/mcp", }, }, tool_interceptors=[runtime_interceptor], callbacks=Callbacks(on_logging_message=on_logging_message, on_progress=on_progress), ) tools = await client.get_tools() tools.extend([get_current_timestamp, get_all_jobs, get_ownerships, get_activity_logs, create_activity_log]) agent = create_agent( "gpt-4.1", …