feat(streaming): 支持流式文本增量与结束信号功能

- 新增 StreamDelta 和 StreamEnd 类型，支持流式数据增量传输
- 扩展 LLMProvider trait，添加带回调的 chat_with_streaming 接口
- 修改 OpenAI Provider 实现，支持流式聊天回调传输增量数据
- Agent 流处理改为异步消费增量消息并传递给前端
- 保证流式增量和最终消息使用相同消息 ID 以便前端替换
- 修改消息总线和协议层，支持携带和识别流式消息的消息 ID
- 客户端 CLI 通过增量输出实现交互式流式响应显示
- Web 前端接收流式增量消息，追加到对应消息，实现实时显示
- 各通道（飞书、微信）支持转发流式增量和结束消息
- 任务工具运行时添加消息 ID 支持，保持消息一致性
- 统一消息构造函数新增流式增量和结束信号的构建方法

src/agent/agent_loop.rs

@@ -7,7 +7,7 @@ use crate::domain::messages::{ContentBlock, ToolCall};
 use crate::observability::{
     Observer, ObserverEvent, ToolExecutionOutcome, ToolExecutionState, truncate_args,
 };
-use crate::providers::{ChatCompletionRequest, LLMProvider, Message, create_provider};
+use crate::providers::{ChatCompletionRequest, LLMProvider, Message, StreamDelta, StreamCallback, create_provider};
 use crate::text::{char_count, take_prefix_chars, take_suffix_chars};
 use crate::tools::{ToolContext, ToolRegistry};
 use async_trait::async_trait;
@@ -663,6 +663,17 @@ pub trait EmittedMessageHandler: Send + Sync + 'static {
     async fn handle_tool_result(&self, message: ChatMessage, _duration_ms: Option<u64>) {
         self.handle(message).await;
     }
+
+    /// Handle a streaming delta. Default is no-op.
+    async fn handle_stream_delta(&self, _delta: &StreamDelta) {
+        // Non-streaming channels ignore this
+    }
+
+    /// Set the message ID to use for stream deltas (so the final assistant message
+    /// can share the same ID, enabling front-end replacement).
+    async fn set_stream_message_id(&self, _id: &str) {
+        // Default: no-op for handlers that don't stream
+    }
 }
 
 /// 装饰器：在内部 emitter 广播前，先将消息持久化到 DB
@@ -706,6 +717,15 @@ impl<H: EmittedMessageHandler> EmittedMessageHandler for PersistingEmittedMessag
         }
         self.inner.handle_tool_result(message, duration_ms).await;
     }
+
+    async fn set_stream_message_id(&self, id: &str) {
+        self.inner.set_stream_message_id(id).await;
+    }
+
+    async fn handle_stream_delta(&self, delta: &StreamDelta) {
+        // Deltas are transient — do NOT persist, just forward to inner handler
+        self.inner.handle_stream_delta(delta).await;
+    }
 }
 
 pub trait SkillProvider: Send + Sync + 'static {
@@ -958,9 +978,34 @@ impl AgentLoop {
                 tools,
             };
 
-            let response = match (*self.provider).chat(request).await {
+            // Set up streaming delta consumer
+            // Pre-generate the message ID so stream deltas and the final assistant
+            // message share the same ID — this lets the front-end replace the
+            // streamed message with the authoritative response.
+            let streaming_message_id = uuid::Uuid::new_v4().to_string();
+            if let Some(handler) = &self.emitted_message_handler {
+                handler.set_stream_message_id(&streaming_message_id).await;
+            }
+
+            let (delta_tx, mut delta_rx) = tokio::sync::mpsc::channel::<StreamDelta>(256);
+            let consumer_handler = self.emitted_message_handler.clone();
+            let consumer_task = tokio::spawn(async move {
+                while let Some(delta) = delta_rx.recv().await {
+                    if let Some(ref handler) = consumer_handler {
+                        handler.handle_stream_delta(&delta).await;
+                    }
+                }
+            });
+            let stream_callback: StreamCallback = std::sync::Arc::new(move |delta: StreamDelta| {
+                // try_send is non-blocking and safe to call from within a tokio runtime
+                let _ = delta_tx.try_send(delta);
+            });
+
+            let response = match (*self.provider).chat_with_streaming(request, stream_callback).await {
                 Ok(response) => response,
                 Err(e) => {
+                    // delta_tx is dropped with the callback; await consumer to finish
+                    let _ = consumer_task.await;
                     tracing::error!(
                         provider = %self.provider.name(),
                         model = %self.provider.model_id(),
@@ -979,6 +1024,22 @@ impl AgentLoop {
                 }
             };
 
+            // Close delta channel and wait for consumer to finish processing
+            // (delta_tx is dropped when the callback closure is dropped)
+            let _ = consumer_task.await;
+
+            // Signal stream end if handler exists
+            let had_streaming = self.emitted_message_handler.is_some();
+            if had_streaming {
+                let end_delta = StreamDelta {
+                    content: String::new(),
+                    reasoning_content: None,
+                };
+                if let Some(handler) = &self.emitted_message_handler {
+                    handler.handle_stream_delta(&end_delta).await;
+                }
+            }
+
             #[cfg(debug_assertions)]
             tracing::debug!(
                 iteration,
@@ -989,12 +1050,17 @@ impl AgentLoop {
 
             // If no tool calls, this is the final response
             if response.tool_calls.is_empty() {
-                let assistant_message = if let Some(reasoning_content) = response.reasoning_content
+                let mut assistant_message = if let Some(reasoning_content) = response.reasoning_content
                 {
                     ChatMessage::assistant_with_reasoning(response.content, reasoning_content)
                 } else {
                     ChatMessage::assistant(response.content)
                 };
+                // Use the same ID as the stream deltas so the front-end can replace
+                // the streamed message with this authoritative response.
+                if had_streaming {
+                    assistant_message.id = streaming_message_id;
+                }
                 emitted_messages.push(assistant_message.clone());
                 self.emit_live_tool_call_message(assistant_message.clone()).await;
                 return Ok(AgentProcessResult {
@@ -1011,7 +1077,7 @@ impl AgentLoop {
             );
 
             // Add assistant message with tool calls
-            let assistant_message =
+            let mut assistant_message =
                 if let Some(reasoning_content) = response.reasoning_content.clone() {
                     ChatMessage::assistant_with_tool_calls_and_reasoning(
                         response.content.clone(),
@@ -1024,6 +1090,11 @@ impl AgentLoop {
                         response.tool_calls.clone(),
                     )
                 };
+            // Use the same ID as stream deltas so the front-end can replace
+            // the streamed message with this authoritative response.
+            if had_streaming {
+                assistant_message.id = streaming_message_id;
+            }
             messages.push(assistant_message.clone());
             emitted_messages.push(assistant_message);
             self.emit_live_tool_call_message(

src/bus/message.rs

@@ -401,6 +401,10 @@ pub struct OutboundMessage {
     pub tool_name: Option<String>,
     pub tool_arguments: Option<serde_json::Value>,
     pub reasoning_content: Option<String>,
+    /// Carry the originating ChatMessage.id so the WS layer can use it
+    /// instead of generating a random UUID. Critical for stream delta → assistant_response
+    /// ID matching on the front-end.
+    pub message_id: Option<String>,
 }
 
 #[derive(Debug, Clone, PartialEq, Eq)]
@@ -412,11 +416,15 @@ pub enum OutboundEventKind {
     SchedulerNotification,
     ErrorNotification,
     TaskStarted,
+    /// 流式文本增量
+    StreamDelta,
+    /// 流式结束信号
+    StreamEnd,
 }
 
 impl OutboundMessage {
     pub fn is_stream_delta(&self) -> bool {
-        self.metadata.get("_stream_delta").is_some()
+        matches!(self.event_kind, OutboundEventKind::StreamDelta | OutboundEventKind::StreamEnd)
     }
 
     pub fn assistant(
@@ -441,6 +449,7 @@ impl OutboundMessage {
             tool_name: None,
             tool_arguments: None,
             reasoning_content: None,
+            message_id: None,
         }
     }
 
@@ -496,6 +505,7 @@ impl OutboundMessage {
             tool_name: Some(tool_name),
             tool_arguments: Some(tool_arguments),
             reasoning_content: None,
+            message_id: None,
         }
     }
     
@@ -526,6 +536,7 @@ impl OutboundMessage {
             tool_name: Some(tool_name),
             tool_arguments: None,
             reasoning_content: None,
+            message_id: None,
         }
     }
     
@@ -556,6 +567,61 @@ impl OutboundMessage {
             tool_name: Some(tool_name),
             tool_arguments: None,
             reasoning_content: None,
+            message_id: None,
+        }
+    }
+    
+    /// 构造流式文本增量消息
+    pub fn stream_delta(
+        channel: impl Into<String>,
+        chat_id: impl Into<String>,
+        session_id: Option<String>,
+        message_id: impl Into<String>,
+        delta: impl Into<String>,
+        reasoning_delta: Option<String>,
+        metadata: HashMap<String, String>,
+    ) -> Self {
+        Self {
+            channel: channel.into(),
+            chat_id: chat_id.into(),
+            session_id,
+            content: delta.into(),
+            reply_to: None,
+            media: Vec::new(),
+            metadata,
+            event_kind: OutboundEventKind::StreamDelta,
+            role: "assistant".to_string(),
+            tool_call_id: Some(message_id.into()),
+            tool_name: None,
+            tool_arguments: None,
+            reasoning_content: reasoning_delta,
+            message_id: None,
+        }
+    }
+    
+    /// 构造流式结束信号
+    pub fn stream_end(
+        channel: impl Into<String>,
+        chat_id: impl Into<String>,
+        session_id: Option<String>,
+        message_id: impl Into<String>,
+        metadata: HashMap<String, String>,
+    ) -> Self {
+        Self {
+            channel: channel.into(),
+            chat_id: chat_id.into(),
+            session_id,
+            content: String::new(),
+            reply_to: None,
+            media: Vec::new(),
+            metadata,
+            event_kind: OutboundEventKind::StreamEnd,
+            role: "assistant".to_string(),
+            tool_call_id: Some(message_id.into()),
+            tool_name: None,
+            tool_arguments: None,
+            reasoning_content: None,
+            message_id: None,
         }
     }
     
@@ -582,6 +648,7 @@ impl OutboundMessage {
                             metadata.clone(),
                         );
                         resp.reasoning_content = message.reasoning_content.clone();
+                        resp.message_id = Some(message.id.clone());
                         outbound.push(resp);
                     }
 
@@ -613,6 +680,7 @@ impl OutboundMessage {
                         metadata.clone(),
                     );
                     resp.reasoning_content = message.reasoning_content.clone();
+                    resp.message_id = Some(message.id.clone());
                     vec![resp]
                 }
             }

src/channels/feishu.rs

@@ -2406,7 +2406,7 @@ impl Channel for FeishuChannel {
     }
 
     async fn send(&self, msg: OutboundMessage) -> Result<(), ChannelError> {
-        if matches!(msg.event_kind, OutboundEventKind::ToolResult | OutboundEventKind::ToolPending)
+        if matches!(msg.event_kind, OutboundEventKind::ToolResult | OutboundEventKind::ToolPending | OutboundEventKind::StreamDelta | OutboundEventKind::StreamEnd)
             || msg.metadata.get("is_subagent_event").map(|v| v == "true").unwrap_or(false)
         {
             return Ok(());

src/channels/wechat.rs

@@ -313,6 +313,8 @@ impl Channel for WechatChannel {
             OutboundEventKind::ToolResult
                 | OutboundEventKind::ToolPending
                 | OutboundEventKind::ToolCall
+                | OutboundEventKind::StreamDelta
+                | OutboundEventKind::StreamEnd
         ) || msg.metadata.get("is_subagent_event").map(|v| v == "true").unwrap_or(false)
         {
             return Ok(());

src/client/mod.rs

@@ -24,6 +24,9 @@ pub async fn run(gateway_url: &str) -> Result<(), Box<dyn std::error::Error>> {
 
     let mut input = InputHandler::new();
     let mut current_session_id: Option<String> = None;
+    // Track message IDs that were already streamed so we can skip
+    // the duplicate AssistantResponse that arrives afterwards.
+    let mut streamed_message_ids: std::collections::HashSet<String> = std::collections::HashSet::new();
     input.write_output("picobot CLI - Commands: /new [title], /save [filepath], /quit\n").await?;
 
     // Main loop: poll both stdin and WebSocket
@@ -36,8 +39,11 @@ pub async fn run(gateway_url: &str) -> Result<(), Box<dyn std::error::Error>> {
                         let text = text.to_string();
                         if let Ok(outbound) = parse_message(&text) {
                             match outbound {
-                                WsOutbound::AssistantResponse { content, .. } => {
-                                    input.write_response(&content).await?;
+                                WsOutbound::AssistantResponse { id, content, .. } => {
+                                    // Skip if already fully streamed via StreamDelta
+                                    if !streamed_message_ids.remove(&id) {
+                                        input.write_response(&content).await?;
+                                    }
                                 }
                                 WsOutbound::ToolCall { tool_name, arguments, .. } => {
                                     input.write_output(&format!("Tool call: {}\n{}\n", tool_name, format_json(&arguments))).await?;
@@ -64,6 +70,18 @@ pub async fn run(gateway_url: &str) -> Result<(), Box<dyn std::error::Error>> {
                                 WsOutbound::SessionSaved { session_id, filepath } => {
                                     input.write_output(&format!("Saved session {} to: {}\n", session_id, filepath)).await?;
                                 }
+                                WsOutbound::StreamDelta { id, delta, .. } => {
+                                    // Track that this message is being streamed
+                                    streamed_message_ids.insert(id.clone());
+                                    // 在终端直接输出流式增量文本
+                                    if !delta.is_empty() {
+                                        input.write_output(&delta).await?;
+                                    }
+                                }
+                                WsOutbound::StreamEnd { .. } => {
+                                    // 流式结束，输出换行
+                                    input.write_output("\n").await?;
+                                }
                                 _ => {}
                             }
                         }

src/gateway/session.rs

@@ -2,6 +2,7 @@ use crate::agent::{AgentError, AgentLoop, ContextCompressor, EmittedMessageHandl
 #[cfg(test)]
 use crate::bus::SYSTEM_CONTEXT_SCHEDULED_PROMPT;
 use crate::bus::{ChatMessage, MessageBus, OutboundMessage};
+use crate::providers::StreamDelta;
 use crate::config::LLMProviderConfig;
 use crate::protocol::WsOutbound;
 use crate::scheduler::ScheduledAgentTaskOptions;
@@ -56,6 +57,7 @@ pub struct BusToolCallEmitter {
     chat_id: String,
     metadata: HashMap<String, String>,
     store: Arc<SessionStore>,
+    stream_message_id: std::sync::Mutex<Option<String>>,
 }
 
 impl BusToolCallEmitter {
@@ -72,6 +74,7 @@ impl BusToolCallEmitter {
             chat_id: chat_id.into(),
             metadata,
             store,
+            stream_message_id: std::sync::Mutex::new(None),
         }
     }
 }
@@ -116,6 +119,43 @@ impl EmittedMessageHandler for BusToolCallEmitter {
             self.persist_todo_write_result(&message);
         }
     }
+
+    async fn handle_stream_delta(&self, delta: &StreamDelta) {
+        // Get or create the stream message ID
+        let message_id = {
+            let mut guard = self.stream_message_id.lock().unwrap();
+            guard.get_or_insert_with(|| Uuid::new_v4().to_string()).clone()
+        };
+
+        // Empty content + no reasoning = stream end signal
+        let outbound = if delta.content.is_empty() && delta.reasoning_content.is_none() {
+            OutboundMessage::stream_end(
+                &self.channel_name,
+                &self.chat_id,
+                None,
+                &message_id,
+                self.metadata.clone(),
+            )
+        } else {
+            OutboundMessage::stream_delta(
+                &self.channel_name,
+                &self.chat_id,
+                None,
+                &message_id,
+                &delta.content,
+                delta.reasoning_content.clone(),
+                self.metadata.clone(),
+            )
+        };
+
+        if let Err(error) = self.bus.publish_outbound(outbound).await {
+            tracing::error!(error = %error, channel = %self.channel_name, "Failed to publish stream delta");
+        }
+    }
+
+    async fn set_stream_message_id(&self, id: &str) {
+        *self.stream_message_id.lock().unwrap() = Some(id.to_string());
+    }
 }
 
 impl BusToolCallEmitter {

src/protocol/mod.rs

@@ -265,6 +265,25 @@ pub enum WsOutbound {
     },
     #[serde(rename = "execution_cancelled")]
     ExecutionCancelled { message: String },
+    #[serde(rename = "stream_delta")]
+    StreamDelta {
+        id: String,
+        delta: String,
+        #[serde(default, skip_serializing_if = "Option::is_none")]
+        reasoning_delta: Option<String>,
+        #[serde(default, skip_serializing_if = "Option::is_none")]
+        subagent_task_id: Option<String>,
+        #[serde(default, skip_serializing_if = "Option::is_none")]
+        topic_id: Option<String>,
+    },
+    #[serde(rename = "stream_end")]
+    StreamEnd {
+        id: String,
+        #[serde(default, skip_serializing_if = "Option::is_none")]
+        subagent_task_id: Option<String>,
+        #[serde(default, skip_serializing_if = "Option::is_none")]
+        topic_id: Option<String>,
+    },
     #[serde(rename = "todo_list")]
     TodoList {
         todos: Vec<TodoItemSummary>,

src/protocol/ws_adapter.rs

@@ -104,7 +104,7 @@ pub(crate) fn ws_outbound_from_outbound_message(message: &OutboundMessage) -> Ve
                 })
                 .collect();
             vec![WsOutbound::AssistantResponse {
-                id: uuid::Uuid::new_v4().to_string(),
+                id: message.message_id.clone().unwrap_or_else(|| uuid::Uuid::new_v4().to_string()),
                 content: message.content.clone(),
                 role: message.role.clone(),
                 attachments,
@@ -174,6 +174,18 @@ pub(crate) fn ws_outbound_from_outbound_message(message: &OutboundMessage) -> Ve
             subagent_type: message.metadata.get("task_subagent_type").cloned().unwrap_or_default(),
             topic_id: message.metadata.get("topic_id").cloned(),
         }],
+        OutboundEventKind::StreamDelta => vec![WsOutbound::StreamDelta {
+            id: message.tool_call_id.clone().unwrap_or_default(),
+            delta: message.content.clone(),
+            reasoning_delta: message.reasoning_content.clone(),
+            subagent_task_id: message.metadata.get("subagent_task_id").cloned(),
+            topic_id: message.metadata.get("topic_id").cloned(),
+        }],
+        OutboundEventKind::StreamEnd => vec![WsOutbound::StreamEnd {
+            id: message.tool_call_id.clone().unwrap_or_default(),
+            subagent_task_id: message.metadata.get("subagent_task_id").cloned(),
+            topic_id: message.metadata.get("topic_id").cloned(),
+        }],
     }
 }
 

src/providers/mod.rs

@@ -9,7 +9,7 @@ pub use crate::domain::messages::ToolCall;
 pub use crate::domain::tools::{Tool, ToolFunction};
 pub use traits::{
     ChatCompletionRequest, ChatCompletionResponse, LLMProvider, Message, ProviderRuntimeConfig,
-    Usage,
+    StreamCallback, StreamDelta, Usage,
 };
 
 pub fn create_provider(

src/providers/openai.rs

@@ -6,7 +6,7 @@ use serde_json::{Value, json};
 use std::collections::HashMap;
 use std::time::Duration;
 
-use super::traits::Usage;
+use super::traits::{StreamCallback, StreamDelta, Usage};
 use super::{ChatCompletionRequest, ChatCompletionResponse, LLMProvider, ToolCall};
 use crate::domain::messages::ContentBlock;
 
@@ -352,10 +352,11 @@ impl OpenAIProvider {
         })
     }
 
-    /// 内部流式聊天实现
-    async fn chat_streaming(
+    /// 内部流式聊天实现，可选传入流式回调
+    async fn chat_streaming_internal(
         &self,
         request: &ChatCompletionRequest,
+        stream_callback: Option<&StreamCallback>,
     ) -> Result<ChatCompletionResponse, Box<dyn std::error::Error + Send + Sync>> {
         tracing::debug!(provider = %self.name, model = %self.model_id, "Starting streaming chat");
 
@@ -444,11 +445,23 @@ impl OpenAIProvider {
                                         // 提取内容增量
                                         if let Some(content) = delta.get("content").and_then(|c| c.as_str()) {
                                             accumulator.add_content(content);
+                                            if let Some(cb) = &stream_callback {
+                                                cb(StreamDelta {
+                                                    content: content.to_string(),
+                                                    reasoning_content: None,
+                                                });
+                                            }
                                         }
 
                                         // 提取推理内容增量
                                         if let Some(reasoning) = delta.get("reasoning_content").and_then(|r| r.as_str()) {
                                             accumulator.add_reasoning_content(reasoning);
+                                            if let Some(cb) = &stream_callback {
+                                                cb(StreamDelta {
+                                                    content: String::new(),
+                                                    reasoning_content: Some(reasoning.to_string()),
+                                                });
+                                            }
                                         }
 
                                         // 提取工具调用增量
@@ -523,9 +536,21 @@ impl OpenAIProvider {
                             if let Some(delta) = choice.get("delta") {
                                 if let Some(content) = delta.get("content").and_then(|c| c.as_str()) {
                                     accumulator.add_content(content);
+                                    if let Some(cb) = &stream_callback {
+                                        cb(StreamDelta {
+                                            content: content.to_string(),
+                                            reasoning_content: None,
+                                        });
+                                    }
                                 }
                                 if let Some(reasoning) = delta.get("reasoning_content").and_then(|r| r.as_str()) {
                                     accumulator.add_reasoning_content(reasoning);
+                                    if let Some(cb) = &stream_callback {
+                                        cb(StreamDelta {
+                                            content: String::new(),
+                                            reasoning_content: Some(reasoning.to_string()),
+                                        });
+                                    }
                                 }
                                 if let Some(tool_calls) = delta.get("tool_calls").and_then(|t| t.as_array()) {
                                     for tool_call in tool_calls {
@@ -737,8 +762,8 @@ impl LLMProvider for OpenAIProvider {
     ) -> Result<ChatCompletionResponse, Box<dyn std::error::Error + Send + Sync>> {
         // 检查是否启用流式输出
         if self.is_streaming_enabled() {
-            // 优先尝试流式输出
-            match self.chat_streaming(&request).await {
+            // 优先尝试流式输出（无回调）
+            match self.chat_streaming_internal(&request, None).await {
                 Ok(response) => return Ok(response),
                 Err(e) => {
                     tracing::debug!(
@@ -890,6 +915,28 @@ impl LLMProvider for OpenAIProvider {
         })
     }
 
+    async fn chat_with_streaming(
+        &self,
+        request: ChatCompletionRequest,
+        callback: StreamCallback,
+    ) -> Result<ChatCompletionResponse, Box<dyn std::error::Error + Send + Sync>> {
+        if self.is_streaming_enabled() {
+            match self.chat_streaming_internal(&request, Some(&callback)).await {
+                Ok(response) => return Ok(response),
+                Err(e) => {
+                    tracing::debug!(
+                        provider = %self.name,
+                        model = %self.model_id,
+                        error = %e,
+                        "Streaming (with callback) failed, falling back to non-streaming"
+                    );
+                }
+            }
+        }
+        // 回退到非流式
+        self.chat(request).await
+    }
+
     fn ptype(&self) -> &str {
         "openai"
     }

src/providers/traits.rs

@@ -4,6 +4,7 @@ use crate::config::LLMProviderConfig;
 use async_trait::async_trait;
 use serde::{Deserialize, Serialize};
 use std::collections::HashMap;
+use std::sync::Arc;
 
 #[derive(Debug, Clone)]
 pub struct ProviderRuntimeConfig {
@@ -137,6 +138,18 @@ pub struct Usage {
     pub total_tokens: u32,
 }
 
+/// 流式响应中的增量事件
+#[derive(Debug, Clone)]
+pub struct StreamDelta {
+    /// 文本内容增量
+    pub content: String,
+    /// 推理内容增量
+    pub reasoning_content: Option<String>,
+}
+
+/// 流式回调类型：每收到一个 delta 就调用一次
+pub type StreamCallback = Arc<dyn Fn(StreamDelta) + Send + Sync>;
+
 #[async_trait]
 pub trait LLMProvider: Send + Sync {
     async fn chat(
@@ -144,6 +157,17 @@ pub trait LLMProvider: Send + Sync {
         request: ChatCompletionRequest,
     ) -> Result<ChatCompletionResponse, Box<dyn std::error::Error + Send + Sync>>;
 
+    /// 带流式回调的 chat：每收到一个 SSE delta 就调用 callback。
+    /// 返回值与 `chat()` 相同（完整的 ChatCompletionResponse）。
+    /// 默认实现忽略 callback，直接调用 chat()。
+    async fn chat_with_streaming(
+        &self,
+        request: ChatCompletionRequest,
+        _callback: StreamCallback,
+    ) -> Result<ChatCompletionResponse, Box<dyn std::error::Error + Send + Sync>> {
+        self.chat(request).await
+    }
+
     fn ptype(&self) -> &str;
 
     fn name(&self) -> &str;

src/tools/task/runtime.rs

@@ -499,6 +499,7 @@ impl SubAgentRuntime for DefaultSubAgentRuntime {
                 tool_name: None,
                 tool_arguments: None,
                 reasoning_content: None,
+                message_id: None,
             };
 
             if let Err(e) = bus.publish_outbound(event).await {

web/src/hooks/useChat.ts

@@ -27,6 +27,7 @@ import type {
   SchedulerJobSessionLookup,
   Channel,
   ChannelList,
+  StreamDelta,
 } from '../types/protocol'
 
 // 简化后的层级状态
@@ -446,14 +447,55 @@ export function useChat(): UseChatReturn {
         break
       }
 
+      case 'stream_delta': {
+        const msg = message as StreamDelta
+        if (msg.topic_id && msg.topic_id !== selectedTopicRef.current) return
+        setMessages((prev) => {
+          const existingIdx = prev.findIndex(m => m.id === msg.id && m.type === 'message')
+          if (existingIdx >= 0) {
+            // 追加到已有消息
+            const updated = [...prev]
+            const existing = updated[existingIdx]
+            updated[existingIdx] = {
+              ...existing,
+              content: existing.content + msg.delta,
+              reasoningContent: msg.reasoning_delta
+                ? (existing.reasoningContent || '') + msg.reasoning_delta
+                : existing.reasoningContent,
+            }
+            return updated
+          }
+          // 创建新消息
+          return [
+            ...prev,
+            {
+              id: msg.id,
+              role: 'assistant' as const,
+              content: msg.delta,
+              timestamp: Math.floor(Date.now() / 1000),
+              type: 'message' as const,
+              reasoningContent: msg.reasoning_delta,
+            },
+          ]
+        })
+        setIsLoading(false)
+        break
+      }
+
+      case 'stream_end': {
+        // 流式结束，无需额外操作，后续 assistant_response 会替换完整内容
+        break
+      }
+
       case 'assistant_response': {
         const msg = message as AssistantResponse
         // 按 topic_id 隔离：如果消息属于其他话题则丢弃
         if (msg.topic_id && msg.topic_id !== selectedTopicRef.current) return
         const role = msg.role === 'user' || msg.role === 'tool' ? msg.role : 'assistant'
-        setMessages((prev) => [
-          ...prev,
-          {
+        setMessages((prev) => {
+          // 如果流式消息已存在（相同 id），替换它
+          const existingIdx = prev.findIndex(m => m.id === msg.id && m.type === 'message')
+          const newMsg: ChatMessage = {
             id: msg.id,
             role,
             content: msg.content,
@@ -461,8 +503,14 @@ export function useChat(): UseChatReturn {
             type: 'message',
             attachments: msg.attachments,
             reasoningContent: msg.reasoning_content,
-          },
-        ])
+          }
+          if (existingIdx >= 0) {
+            const updated = [...prev]
+            updated[existingIdx] = newMsg
+            return updated
+          }
+          return [...prev, newMsg]
+        })
         setIsLoading(false)
 
         // 当前话题无描述时，可能刚触发了异步生成，标记需要刷新

web/src/types/protocol.ts

@@ -256,6 +256,22 @@ export interface ExecutionCancelled {
   message: string
 }
 
+export interface StreamDelta {
+  type: 'stream_delta'
+  id: string
+  delta: string
+  reasoning_delta?: string
+  subagent_task_id?: string
+  topic_id?: string
+}
+
+export interface StreamEnd {
+  type: 'stream_end'
+  id: string
+  subagent_task_id?: string
+  topic_id?: string
+}
+
 export type WsOutbound =
   | AssistantResponse
   | ToolCall
@@ -263,6 +279,8 @@ export type WsOutbound =
   | ToolPending
   | WsError
   | TaskStarted
+  | StreamDelta
+  | StreamEnd
   | SessionEstablished
   | SessionCreated
   | SessionList