wechat-to-anything 0.6.5 → 0.6.7

package/.claude/settings.local.json

@@ -0,0 +1,11 @@
+{
+  "permissions": {
+    "allow": [
+      "mcp__filesystem__directory_tree",
+      "Bash(find /Users/zxw/AITOOL/wechat-to-anything/examples -name *.mjs ! -path */node_modules/* -exec head -40 {})",
+      "Bash(2)",
+      "Bash(ls -la /Users/zxw/AITOOL/wechat-to-anything/examples/*/package.json)",
+      "mcp__filesystem__read_text_file"
+    ]
+  }
+}

package/README.en.md

@@ -12,7 +12,7 @@
 </p>
 
 <p align="center">
-  <a href="#quick-start">Quick Start</a> · <a href="#full-multimodal-matrix">Multimodal</a> · <a href="#media-protocol">Media Protocol</a> · <a href="#multi-agent-mode">Multi-Agent</a> · <a href="#bring-your-own-agent">Custom Agent</a>
+  <a href="#quick-start">Quick Start</a> · <a href="#full-multimodal-matrix">Multimodal</a> · <a href="#media-protocol">Media Protocol</a> · <a href="#multi-agent-mode">Multi-Agent</a> · <a href="#proactive-send-api">Send API</a> · <a href="#bring-your-own-agent">Custom Agent</a>
 </p>
 
 <p align="center">
@@ -38,6 +38,7 @@
 - 📡 **Full multimodal** — Text, images, voice, video, files — bidirectional
 - 🤖 **Multi-Agent** — Connect multiple Agents simultaneously, route with `@` prefix
 - ⌨️ **Typing indicator** — Shows "typing..." while Agent is thinking
+- 📤 **Proactive Send API** — Agent can push multiple messages to simulate human typing rhythm
 
 ### Full Multimodal Matrix
 
@@ -152,6 +153,34 @@ npx wechat-to-anything \
 | `@list` | List all Agents |
 | `@switch gemini` | Switch default |
 
+## Proactive Send API
+
+Bridge starts an HTTP API on `localhost:9099`. Agents can proactively push multiple messages (simulating human typing rhythm):
+
+```bash
+curl -X POST http://localhost:9099/api/send \
+  -H "Content-Type: application/json" \
+  -d '{"to": "user_id", "content": "Hmm..."}'
+```
+
+- `to` — WeChat user ID (bridge passes this via the `user` field when calling agents)
+- `content` — Same formats as agent responses (plain text, `![](url)`, `[audio:path]`, etc.)
+- Use `--port PORT` to customize the port
+
+**Use case**: Agent splits one reply into multiple segments with controlled timing:
+
+```python
+import requests, time
+def send(to, text):
+    requests.post("http://localhost:9099/api/send", json={"to": to, "content": text})
+
+send(user_id, "Hmm...")
+time.sleep(1.5)
+send(user_id, "Let me think")
+time.sleep(2)
+# Final segment returned as normal response
+```
+
 ## Credentials
 
 Login credentials are saved in `~/.wechat-to-anything/credentials.json`. Delete to re-login.

package/README.md

@@ -12,7 +12,7 @@
 </p>
 
 <p align="center">
-  <a href="#快速开始">快速开始</a> · <a href="#全模态支持矩阵">全模态</a> · <a href="#多媒体协议">多媒体协议</a> · <a href="#多-agent-模式">多 Agent</a> · <a href="#接入自己的-agent">自定义 Agent</a>
+  <a href="#快速开始">快速开始</a> · <a href="#全模态支持矩阵">全模态</a> · <a href="#多媒体协议">多媒体协议</a> · <a href="#多-agent-模式">多 Agent</a> · <a href="#主动发送-api">主动发送</a> · <a href="#接入自己的-agent">自定义 Agent</a>
 </p>
 
 <p align="center">
@@ -38,6 +38,7 @@
 - 📡 **全模态** — 文本、图片、语音、视频、文件，双向全覆盖
 - 🤖 **多 Agent** — 同时接入多个 Agent，`@` 路由切换
 - ⌨️ **打字指示器** — Agent 思考时显示"对方正在输入"
+- 📤 **主动发送 API** — Agent 可推送多条消息，模拟真人打字节奏
 
 ### 全模态支持矩阵
 
@@ -152,6 +153,34 @@ npx wechat-to-anything \
 | `@list` | 查看所有 Agent |
 | `@切换 gemini` | 切换默认 |
 
+## 主动发送 API
+
+Bridge 启动时会在 `localhost:9099` 暴露 HTTP API，Agent 可主动推送多条消息（模拟真人打字节奏）：
+
+```bash
+curl -X POST http://localhost:9099/api/send \
+  -H "Content-Type: application/json" \
+  -d '{"to": "user_id", "content": "嗯……"}'
+```
+
+- `to` — 微信用户 ID（bridge 调 agent 时通过 `user` 字段传入）
+- `content` — 支持和 Agent 回复相同的格式（纯文本、`![](url)`、`[audio:path]` 等）
+- 用 `--port PORT` 自定义端口
+
+**用途**：Agent 对一条消息可分多段回复，控制发送间隔：
+
+```python
+import requests, time
+def send(to, text):
+    requests.post("http://localhost:9099/api/send", json={"to": to, "content": text})
+
+send(user_id, "嗯……")
+time.sleep(1.5)
+send(user_id, "让我想想")
+time.sleep(2)
+# 最后一段作为正常 response 返回
+```
+
 ## 凭证
 
 登录凭证保存在 `~/.wechat-to-anything/credentials.json`，删除即可重新登录。

package/bin/cli.mjs

@@ -32,6 +32,12 @@ ${pc.bold("参数:")}
   --openclaw            ${pc.dim("内置 OpenClaw（需先 npm i -g openclaw）")}
   --agent ${pc.dim("name=url")}    ${pc.dim("注册自定义 Agent")}
   --default ${pc.dim("name")}      ${pc.dim("设置默认 Agent")}
+  --port ${pc.dim("PORT")}        ${pc.dim("API 端口（默认 9099），暴露 POST /api/send")}
+
+${pc.bold("API:")}
+  POST http://localhost:PORT/api/send
+  ${pc.dim('{ "to": "user_id", "content": "消息内容" }')}
+  ${pc.dim("Agent 可主动推送多条消息，模拟真人节奏")}
 
 ${pc.dim("Docs: https://github.com/kellyvv/wechat-to-anything")}
 `);
@@ -40,6 +46,7 @@ ${pc.dim("Docs: https://github.com/kellyvv/wechat-to-anything")}
 
 // 解析参数
 let i = 0;
+let port = 9099;
 while (i < args.length) {
   if (args[i] === "--codex") {
     agents.set("codex", "cli://codex");
@@ -71,6 +78,13 @@ while (i < args.length) {
   } else if (args[i] === "--default" && args[i + 1]) {
     defaultAgent = args[i + 1].toLowerCase();
     i += 2;
+  } else if (args[i] === "--port" && args[i + 1]) {
+    port = parseInt(args[i + 1], 10);
+    if (isNaN(port)) {
+      console.error(pc.red(`无效的端口号: ${args[i + 1]}`));
+      process.exit(1);
+    }
+    i += 2;
   } else if (!args[i].startsWith("--")) {
     if (!args[i].startsWith("acp://")) {
       try { new URL(args[i]); } catch {
@@ -114,7 +128,7 @@ if (agents.size === 1 && agents.has("default")) {
 }
 console.log();
 
-import("../cli/bridge.mjs").then((mod) => mod.start(agents, defaultAgent)).catch((err) => {
+import("../cli/bridge.mjs").then((mod) => mod.start(agents, defaultAgent, { port })).catch((err) => {
   console.error(pc.red(err.message));
   process.exit(1);
 });

package/cli/bridge.mjs

@@ -1,4 +1,5 @@
 import pc from "picocolors";
+import { createServer } from "node:http";
 import {
   loadCredentials, loginWithQR, getUpdates,
   sendMessage, sendImageByUrl, sendVideoByUrl,
@@ -14,7 +15,7 @@ import { stripMarkdown } from "./markdown.mjs";
  * 启动桥：WeChat ilinkai API ←→ Agent HTTP
  * 支持文本 + 图片 + 语音 + 文件，双向
  */
-export async function start(agents, defaultAgent) {
+export async function start(agents, defaultAgent, { port = 9099 } = {}) {
   // 兼容旧的单 URL 调用
   if (typeof agents === "string") {
     const url = agents;
@@ -103,6 +104,186 @@ export async function start(agents, defaultAgent) {
   const pendingImages = new Map(); // userId → { base64, timestamp }
   const IMAGE_BUFFER_TTL = 5 * 60_000; // 5 min 过期
 
+  // per-user contextToken 缓存（供 /api/send 使用）
+  const userContextTokens = new Map(); // userId → contextToken
+
+  /**
+   * 统一发送内容（纯文本 / 图片 / 语音 / 视频 / 文件）
+   * 复用 Agent 回复的多媒体协议格式
+   */
+  async function sendContent(to, content, tag = "") {
+    const ct = userContextTokens.get(to) || "";
+
+    // 检查回复是否包含 [audio:path/url]
+    const audioMatch = content.match(/\[audio:(.*?)\]/);
+    // 检查回复是否包含图片（markdown 格式，支持 URL 和 data URI）
+    const imageMatch = content.match(/!\[.*?\]\(((?:https?:\/\/|data:image\/|\/).+?)\)/);
+    // 检查回复是否包含 [video:path/url]
+    const videoMatch = content.match(/\[video:(.*?)\]/);
+    // 检查回复是否包含 [file:path/url]
+    const fileMatch = content.match(/\[file:(.*?)\]/);
+
+    if (audioMatch) {
+      const audioSrc = audioMatch[1];
+      const textPart = content.replace(/\[audio:.*?\]/g, "").trim();
+      console.log(pc.green(`→ [send] [语音] ${audioSrc.slice(0, 60)}`));
+      try {
+        const { execFileSync } = await import("node:child_process");
+        const { statSync, writeFileSync } = await import("node:fs");
+        const { uploadToCdn } = await import("./cdn.mjs");
+        const { buildHeaders, BASE_URL: baseUrl } = await import("./weixin.mjs");
+
+        let audioFile = audioSrc;
+        if (audioSrc.startsWith("http://") || audioSrc.startsWith("https://")) {
+          const resp = await fetch(audioSrc);
+          if (!resp.ok) throw new Error(`下载失败: ${resp.status}`);
+          writeFileSync("/tmp/wxta_audio_in.mp3", Buffer.from(await resp.arrayBuffer()));
+          audioFile = "/tmp/wxta_audio_in.mp3";
+        }
+
+        execFileSync("ffmpeg", ["-y", "-i", audioFile, "-ar", "16000", "-ac", "1", "-f", "s16le", "/tmp/wxta_audio.pcm"], { stdio: "ignore" });
+        execFileSync("python3", ["-c", "import pilk; pilk.encode('/tmp/wxta_audio.pcm', '/tmp/wxta_audio.silk', pcm_rate=16000, tencent=True)"]);
+        const pcmSize = statSync("/tmp/wxta_audio.pcm").size;
+        const durationMs = Math.round((pcmSize / 32000) * 1000);
+
+        const cdn = await uploadToCdn("/tmp/wxta_audio.silk", to, creds.token, 4);
+        const aesKeyB64 = Buffer.from(cdn.aeskey).toString("base64");
+        const crypto = await import("node:crypto");
+
+        const body = JSON.stringify({
+          msg: {
+            from_user_id: "", to_user_id: to,
+            client_id: crypto.randomUUID(),
+            message_type: 2, message_state: 2,
+            item_list: [{
+              type: 3,
+              voice_item: {
+                media: { encrypt_query_param: cdn.downloadParam, aes_key: aesKeyB64 },
+                encode_type: 4, bits_per_sample: 16, sample_rate: 16000, playtime: durationMs,
+              },
+            }],
+            context_token: ct,
+          },
+          base_info: {},
+        });
+        await fetch(`${baseUrl}/ilink/bot/sendmessage`, {
+          method: "POST", headers: buildHeaders(creds.token, body), body,
+        });
+        console.log(pc.green(`→ [语音] 已发送 (${durationMs}ms)`));
+        if (textPart) await sendMessage(creds.token, to, tag + stripMarkdown(textPart), ct);
+      } catch (err) {
+        console.error(pc.red(`   语音发送失败: ${err.message}`));
+        await sendMessage(creds.token, to, tag + content.replace(/\[audio:.*?\]/g, "").trim() || content, ct);
+      }
+    } else if (imageMatch) {
+      const imageUrl = imageMatch[1];
+      const textPart = content.replace(/!\[.*?\]\(((?:https?:\/\/|data:image\/|\/).+?)\)/g, "").trim();
+      console.log(pc.green(`→ [send] [图片] ${imageUrl.slice(0, 60)}`));
+      try {
+        if (textPart) await sendMessage(creds.token, to, tag + stripMarkdown(textPart), ct);
+        await sendImageByUrl(creds.token, to, ct, imageUrl);
+      } catch (err) {
+        console.error(pc.red(`   图片发送失败: ${err.message}`));
+        await sendMessage(creds.token, to, tag + content, ct);
+      }
+    } else if (videoMatch) {
+      const videoSrc = videoMatch[1];
+      const textPart = content.replace(/\[video:.*?\]/g, "").trim();
+      console.log(pc.green(`→ [send] [视频] ${videoSrc.slice(0, 60)}`));
+      try {
+        if (textPart) await sendMessage(creds.token, to, tag + stripMarkdown(textPart), ct);
+        await sendVideoByUrl(creds.token, to, ct, videoSrc);
+      } catch (err) {
+        console.error(pc.red(`   视频发送失败: ${err.message}`));
+        await sendMessage(creds.token, to, tag + stripMarkdown(content), ct);
+      }
+    } else if (fileMatch) {
+      const fileSrc = fileMatch[1];
+      const textPart = content.replace(/\[file:.*?\]/g, "").trim();
+      const fileName = fileSrc.split("/").pop() || "file";
+      console.log(pc.green(`→ [send] [文件] ${fileSrc.slice(0, 60)}`));
+      try {
+        const { writeFileSync, unlinkSync } = await import("node:fs");
+        const { tmpdir } = await import("node:os");
+        const { join } = await import("node:path");
+        const resp = await fetch(fileSrc);
+        if (!resp.ok) throw new Error(`file download failed: ${resp.status}`);
+        const buf = Buffer.from(await resp.arrayBuffer());
+        const tmpPath = join(tmpdir(), `wx-file-${Date.now()}-${fileName}`);
+        writeFileSync(tmpPath, buf);
+        try {
+          const uploaded = await uploadToCdn(tmpPath, to, creds.token, 3);
+          const { sendFileMessage } = await import("./weixin.mjs");
+          await sendFileMessage(creds.token, to, ct, uploaded, fileName);
+          if (textPart) await sendMessage(creds.token, to, tag + stripMarkdown(textPart), ct);
+        } finally {
+          try { unlinkSync(tmpPath); } catch {}
+        }
+      } catch (err) {
+        console.error(pc.red(`   文件发送失败: ${err.message}`));
+        await sendMessage(creds.token, to, tag + stripMarkdown(content), ct);
+      }
+    } else {
+      // 纯文本
+      console.log(pc.green(`→ [send] ${content.slice(0, 80)}${content.length > 80 ? "..." : ""}`));
+      await sendMessage(creds.token, to, tag + stripMarkdown(content), ct);
+    }
+  }
+
+  // ─── HTTP API Server (/api/send) ────────────────────────────────
+  const httpServer = createServer(async (req, res) => {
+    // CORS
+    res.setHeader("Access-Control-Allow-Origin", "*");
+    res.setHeader("Access-Control-Allow-Methods", "POST, OPTIONS");
+    res.setHeader("Access-Control-Allow-Headers", "Content-Type");
+    if (req.method === "OPTIONS") { res.writeHead(204); res.end(); return; }
+
+    if (req.method === "POST" && req.url === "/api/send") {
+      let body = "";
+      let bodySize = 0;
+      for await (const chunk of req) {
+        bodySize += chunk.length;
+        if (bodySize > 1_048_576) { // 1MB limit
+          res.writeHead(413, { "Content-Type": "application/json" });
+          res.end(JSON.stringify({ error: "body too large (max 1MB)" }));
+          return;
+        }
+        body += chunk;
+      }
+      try {
+        const { to, content } = JSON.parse(body);
+        if (!to || !content) {
+          res.writeHead(400, { "Content-Type": "application/json" });
+          res.end(JSON.stringify({ error: "missing 'to' or 'content'" }));
+          return;
+        }
+        console.log(pc.cyan(`← [API] → ${to.slice(0, 12)}...: ${content.slice(0, 60)}`));
+        await sendContent(to, content);
+        res.writeHead(200, { "Content-Type": "application/json" });
+        res.end("{}");
+      } catch (err) {
+        console.error(pc.red(`   /api/send 错误: ${err.message}`));
+        res.writeHead(500, { "Content-Type": "application/json" });
+        res.end(JSON.stringify({ error: err.message }));
+      }
+      return;
+    }
+
+    res.writeHead(404, { "Content-Type": "application/json" });
+    res.end(JSON.stringify({ error: "not found" }));
+  });
+
+  httpServer.on("error", (err) => {
+    if (err.code === "EADDRINUSE") {
+      console.warn(pc.yellow(`⚠️ 端口 ${port} 已被占用，API 未启动（bridge 继续运行）`));
+    } else {
+      console.error(pc.red(`API 服务器错误: ${err.message}`));
+    }
+  });
+  httpServer.listen(port, () => {
+    console.log(pc.green(`📡 API 已启动: http://localhost:${port}/api/send`));
+  });
+
   const loop = async () => {
     while (true) {
       try {
@@ -117,6 +298,9 @@ export async function start(agents, defaultAgent) {
           const contextToken = msg.context_token || "";
           if (!from) continue;
 
+          // 缓存 contextToken 供 /api/send 使用
+          if (contextToken) userContextTokens.set(from, contextToken);
+
           const text = extractText(msg);
           const media = extractMedia(msg);
 
@@ -184,12 +368,6 @@ export async function start(agents, defaultAgent) {
             }
 
           } else if (media?.type === "voice") {
-            // 打印完整 voice_item 结构，用于对照发送格式
-            const voiceItem = (msg.item_list || []).find(i => i.type === 3)?.voice_item;
-            if (voiceItem) {
-              console.log(pc.yellow("📋 收到的 voice_item 完整结构:"));
-              console.log(JSON.stringify(voiceItem, null, 2));
-            }
             const voiceText = media.voiceText || text;
             if (voiceText) {
               console.log(pc.cyan(`← [微信] ${from}: [语音] ${voiceText.slice(0, 80)}`));
@@ -224,65 +402,6 @@ export async function start(agents, defaultAgent) {
             continue;
           }
 
-          // === 语音测试触发器 ===
-          if (text === "语音测试") {
-            console.log(pc.yellow("🎤 语音测试..."));
-            try {
-              const { execSync } = await import("node:child_process");
-              const { statSync } = await import("node:fs");
-              const crypto = await import("node:crypto");
-              const { buildHeaders, BASE_URL: baseUrl } = await import("./weixin.mjs");
-              const { uploadToCdn } = await import("./cdn.mjs");
-
-              // TTS → MP3 → PCM(16kHz) → SILK
-              execSync(`python3 -m edge_tts --text "你好，这是一条AI语音消息测试" --voice zh-CN-XiaoxiaoNeural --write-media /tmp/tts_bridge.mp3`);
-              execSync(`ffmpeg -y -i /tmp/tts_bridge.mp3 -ar 16000 -ac 1 -f s16le /tmp/tts_bridge.pcm 2>/dev/null`);
-              execSync(`python3 -c "import pilk; pilk.encode('/tmp/tts_bridge.pcm', '/tmp/tts_bridge.silk', pcm_rate=16000, tencent=True)"`);
-              const pcmSize = statSync("/tmp/tts_bridge.pcm").size;
-              const durationMs = Math.round((pcmSize / 32000) * 1000);
-              console.log(pc.dim(`   TTS+SILK 完成 (duration=${durationMs}ms)`));
-
-              // CDN 上传 (mediaType=4 = 语音)
-              const cdn = await uploadToCdn("/tmp/tts_bridge.silk", from, creds.token, 4);
-              const aesKeyB64 = Buffer.from(cdn.aeskey).toString("base64");
-              console.log(pc.dim(`   CDN 上传成功 (mediaType=4)`));
-
-              // 发送语音消息
-              const body = JSON.stringify({
-                msg: {
-                  from_user_id: "", to_user_id: from,
-                  client_id: crypto.randomUUID(),
-                  message_type: 2, message_state: 2,
-                  item_list: [{
-                    type: 3,
-                    voice_item: {
-                      media: {
-                        encrypt_query_param: cdn.downloadParam,
-                        aes_key: aesKeyB64,
-                      },
-                      encode_type: 4,
-                      bits_per_sample: 16,
-                      sample_rate: 16000,
-                      playtime: durationMs,
-                    },
-                  }],
-                  context_token: contextToken,
-                },
-                base_info: {},
-              });
-              const res = await fetch(`${baseUrl}/ilink/bot/sendmessage`, {
-                method: "POST",
-                headers: buildHeaders(creds.token, body),
-                body,
-              });
-              console.log(pc.green(`→ [语音] status: ${res.status}`));
-              await sendMessage(creds.token, from, `🎤 语音已发送 (${durationMs}ms)`, contextToken);
-            } catch (err) {
-              console.error(pc.red(`   语音测试失败: ${err.message}`));
-              await sendMessage(creds.token, from, `⚠️ 语音测试失败: ${err.message}`, contextToken);
-            }
-            continue;
-          }
 
           // 解析 @agentName 路由
           let targetAgent = userDefaults.get(from) || defaultAgent;
@@ -328,135 +447,7 @@ export async function start(agents, defaultAgent) {
             const reply = await callAgentAuto(agentUrl, agentMessages, from);
             if (typing) typing.onIdle();
             const agentTag = multiMode ? `[${targetAgent}] ` : "";
-
-            // 检查回复是否包含 [audio:path/url]
-            const audioMatch = reply.match(/\[audio:(.*?)\]/);
-            // 检查回复是否包含图片（markdown 格式，支持 URL 和 data URI）
-            const imageMatch = reply.match(/!\[.*?\]\(((?:https?:\/\/|data:image\/)[^\s)]+)\)/);
-            // 检查回复是否包含 [video:path/url]
-            const videoMatch = reply.match(/\[video:(.*?)\]/);
-            // 检查回复是否包含 [file:path/url]
-            const fileMatch = reply.match(/\[file:(.*?)\]/);
-
-            if (audioMatch) {
-              const audioSrc = audioMatch[1];
-              const textPart = reply.replace(/\[audio:.*?\]/g, "").trim();
-              console.log(pc.green(`→ [${targetAgent}] [语音] ${audioSrc.slice(0, 60)}`));
-              try {
-                const { execSync } = await import("node:child_process");
-                const { statSync, writeFileSync } = await import("node:fs");
-                const { uploadToCdn } = await import("./cdn.mjs");
-                const { buildHeaders, BASE_URL: baseUrl } = await import("./weixin.mjs");
-
-                // 下载或使用本地文件
-                let audioFile = audioSrc;
-                if (audioSrc.startsWith("http://") || audioSrc.startsWith("https://")) {
-                  const resp = await fetch(audioSrc);
-                  if (!resp.ok) throw new Error(`下载失败: ${resp.status}`);
-                  writeFileSync("/tmp/wxta_audio_in.mp3", Buffer.from(await resp.arrayBuffer()));
-                  audioFile = "/tmp/wxta_audio_in.mp3";
-                }
-
-                // 转码: audio → PCM(16kHz) → SILK
-                execSync(`ffmpeg -y -i "${audioFile}" -ar 16000 -ac 1 -f s16le /tmp/wxta_audio.pcm 2>/dev/null`);
-                execSync(`python3 -c "import pilk; pilk.encode('/tmp/wxta_audio.pcm', '/tmp/wxta_audio.silk', pcm_rate=16000, tencent=True)"`);
-                const pcmSize = statSync("/tmp/wxta_audio.pcm").size;
-                const durationMs = Math.round((pcmSize / 32000) * 1000);
-
-                // CDN 上传 + 发送语音（与"语音测试"相同格式）
-                const cdn = await uploadToCdn("/tmp/wxta_audio.silk", from, creds.token, 4);
-                const aesKeyB64 = Buffer.from(cdn.aeskey).toString("base64");
-                const crypto = await import("node:crypto");
-
-                const body = JSON.stringify({
-                  msg: {
-                    from_user_id: "", to_user_id: from,
-                    client_id: crypto.randomUUID(),
-                    message_type: 2, message_state: 2,
-                    item_list: [{
-                      type: 3,
-                      voice_item: {
-                        media: {
-                          encrypt_query_param: cdn.downloadParam,
-                          aes_key: aesKeyB64,
-                        },
-                        encode_type: 4,
-                        bits_per_sample: 16,
-                        sample_rate: 16000,
-                        playtime: durationMs,
-                      },
-                    }],
-                    context_token: contextToken,
-                  },
-                  base_info: {},
-                });
-                await fetch(`${baseUrl}/ilink/bot/sendmessage`, {
-                  method: "POST",
-                  headers: buildHeaders(creds.token, body),
-                  body,
-                });
-                console.log(pc.green(`→ [语音] 已发送 (${durationMs}ms)`));
-                if (textPart) await sendMessage(creds.token, from, agentTag + stripMarkdown(textPart), contextToken);
-              } catch (err) {
-                console.error(pc.red(`   语音发送失败: ${err.message}`));
-                await sendMessage(creds.token, from, agentTag + reply.replace(/\[audio:.*?\]/g, "").trim() || reply, contextToken);
-              }
-            } else if (imageMatch) {
-              // Agent 回复了图片 URL → 直接发到微信
-              const imageUrl = imageMatch[1];
-              const textPart = reply.replace(/!\[.*?\]\(https?:\/\/[^\s)]+\)/g, "").trim();
-              console.log(pc.green(`→ [${targetAgent}] [图片] ${imageUrl.slice(0, 60)}`));
-              try {
-                if (textPart) await sendMessage(creds.token, from, agentTag + stripMarkdown(textPart), contextToken);
-                await sendImageByUrl(creds.token, from, contextToken, imageUrl);
-              } catch (err) {
-                console.error(pc.red(`   图片发送失败: ${err.message}`));
-                await sendMessage(creds.token, from, agentTag + reply, contextToken);
-              }
-            } else if (videoMatch) {
-              // Agent 回复了视频 → CDN 上传发到微信
-              const videoSrc = videoMatch[1];
-              const textPart = reply.replace(/\[video:.*?\]/g, "").trim();
-              console.log(pc.green(`→ [${targetAgent}] [视频] ${videoSrc.slice(0, 60)}`));
-              try {
-                if (textPart) await sendMessage(creds.token, from, agentTag + stripMarkdown(textPart), contextToken);
-                await sendVideoByUrl(creds.token, from, contextToken, videoSrc);
-              } catch (err) {
-                console.error(pc.red(`   视频发送失败: ${err.message}`));
-                await sendMessage(creds.token, from, agentTag + stripMarkdown(reply), contextToken);
-              }
-            } else if (fileMatch) {
-              // Agent 回复了文件 → CDN 上传发到微信
-              const fileSrc = fileMatch[1];
-              const textPart = reply.replace(/\[file:.*?\]/g, "").trim();
-              const fileName = fileSrc.split("/").pop() || "file";
-              console.log(pc.green(`→ [${targetAgent}] [文件] ${fileSrc.slice(0, 60)}`));
-              try {
-                const { writeFileSync, unlinkSync } = await import("node:fs");
-                const { tmpdir } = await import("node:os");
-                const { join } = await import("node:path");
-                const resp = await fetch(fileSrc);
-                if (!resp.ok) throw new Error(`file download failed: ${resp.status}`);
-                const buf = Buffer.from(await resp.arrayBuffer());
-                const tmpPath = join(tmpdir(), `wx-file-${Date.now()}-${fileName}`);
-                writeFileSync(tmpPath, buf);
-                try {
-                  const uploaded = await uploadToCdn(tmpPath, from, creds.token, 3);
-                  const { sendFileMessage } = await import("./weixin.mjs");
-                  await sendFileMessage(creds.token, from, contextToken, uploaded, fileName);
-                  if (textPart) await sendMessage(creds.token, from, agentTag + stripMarkdown(textPart), contextToken);
-                } finally {
-                  try { unlinkSync(tmpPath); } catch {}
-                }
-              } catch (err) {
-                console.error(pc.red(`   文件发送失败: ${err.message}`));
-                await sendMessage(creds.token, from, agentTag + stripMarkdown(reply), contextToken);
-              }
-            } else {
-              // 纯文本回复
-              console.log(pc.green(`→ [${targetAgent}] ${reply.slice(0, 80)}${reply.length > 80 ? "..." : ""}`));
-              await sendMessage(creds.token, from, agentTag + stripMarkdown(reply), contextToken);
-            }
+            await sendContent(from, reply, agentTag);
           } catch (err) {
             if (typing) typing.onCleanup();
             console.error(pc.red(`   ${targetAgent} 错误: ${err.message}`));
@@ -471,6 +462,7 @@ export async function start(agents, defaultAgent) {
   };
 
   process.on("SIGINT", () => {
+    httpServer.close();
     console.log(pc.dim("\n桥已停止"));
     process.exit(0);
   });

package/cli/cdn.mjs

@@ -80,7 +80,7 @@ export async function downloadMediaToFile(encryptQueryParam, aesKeyBase64, ext =
  */
 export async function uploadImageWithThumb(filePath, toUserId, token) {
   const { buildHeaders, BASE_URL } = await import("./weixin.mjs");
-  const { execSync } = await import("child_process");
+  const { execFileSync } = await import("child_process");
 
   const plaintext = await readFile(filePath);
   const rawsize = plaintext.length;
@@ -92,7 +92,7 @@ export async function uploadImageWithThumb(filePath, toUserId, token) {
   // 生成缩略图
   const thumbPath = `/tmp/wxta_thumb_${Date.now()}.jpg`;
   try {
-    execSync(`sips --resampleWidth 120 "${filePath}" --out "${thumbPath}" 2>/dev/null`);
+    execFileSync("sips", ["--resampleWidth", "120", filePath, "--out", thumbPath], { stdio: "ignore" });
   } catch {
     // sips 失败时用原图当缩略图
     await writeFile(thumbPath, plaintext);
@@ -102,7 +102,7 @@ export async function uploadImageWithThumb(filePath, toUserId, token) {
   // 获取缩略图尺寸
   let thumbWidth = 120, thumbHeight = 120;
   try {
-    const sipsOut = execSync(`sips -g pixelWidth -g pixelHeight "${thumbPath}" 2>/dev/null`).toString();
+    const sipsOut = execFileSync("sips", ["-g", "pixelWidth", "-g", "pixelHeight", thumbPath], { encoding: "utf-8", stdio: ["pipe", "pipe", "ignore"] });
     const wm = sipsOut.match(/pixelWidth:\s*(\d+)/);
     const hm = sipsOut.match(/pixelHeight:\s*(\d+)/);
     if (wm) thumbWidth = parseInt(wm[1]);
@@ -183,7 +183,7 @@ export async function uploadImageWithThumb(filePath, toUserId, token) {
  */
 export async function uploadVideoWithThumb(filePath, toUserId, token) {
   const { buildHeaders, BASE_URL } = await import("./weixin.mjs");
-  const { execSync } = await import("child_process");
+  const { execFileSync } = await import("child_process");
 
   const plaintext = await readFile(filePath);
   const rawsize = plaintext.length;
@@ -195,14 +195,14 @@ export async function uploadVideoWithThumb(filePath, toUserId, token) {
   // 获取视频时长
   let playLength = 10;
   try {
-    const dur = execSync(`ffprobe -v error -show_entries format=duration -of csv=p=0 "${filePath}"`, { encoding: "utf-8" }).trim();
+    const dur = execFileSync("ffprobe", ["-v", "error", "-show_entries", "format=duration", "-of", "csv=p=0", filePath], { encoding: "utf-8" }).trim();
     playLength = Math.round(parseFloat(dur));
   } catch {}
 
   // 生成缩略图（第一帧）
   const thumbPath = `/tmp/wxta_video_thumb_${Date.now()}.jpg`;
   try {
-    execSync(`ffmpeg -y -i "${filePath}" -vframes 1 -vf "scale=224:-1" -q:v 5 "${thumbPath}" 2>/dev/null`);
+    execFileSync("ffmpeg", ["-y", "-i", filePath, "-vframes", "1", "-vf", "scale=224:-1", "-q:v", "5", thumbPath], { stdio: "ignore" });
   } catch {
     // ffmpeg 失败时创建空白缩略图（sendmessage 仍需 thumb 注册）
     const { writeFileSync } = await import("node:fs");
@@ -213,7 +213,7 @@ export async function uploadVideoWithThumb(filePath, toUserId, token) {
   // 缩略图尺寸
   let thumbWidth = 224, thumbHeight = 224;
   try {
-    const sipsOut = execSync(`sips -g pixelWidth -g pixelHeight "${thumbPath}" 2>/dev/null`).toString();
+    const sipsOut = execFileSync("sips", ["-g", "pixelWidth", "-g", "pixelHeight", thumbPath], { encoding: "utf-8", stdio: ["pipe", "pipe", "ignore"] });
     const wm = sipsOut.match(/pixelWidth:\s*(\d+)/);
     const hm = sipsOut.match(/pixelHeight:\s*(\d+)/);
     if (wm) thumbWidth = parseInt(wm[1]);

package/cli/weixin.mjs

@@ -254,34 +254,6 @@ export async function sendImageByUrl(token, to, contextToken, imageUrl) {
   );
 }
 
-/**
- * 发送语音消息（base64 音频数据）
- */
-export async function sendVoiceMessage(token, to, contextToken, audioBase64, durationSec) {
-  await apiPost(
-    "ilink/bot/sendmessage",
-    {
-      msg: {
-        from_user_id: "",
-        to_user_id: to,
-        client_id: crypto.randomUUID(),
-        message_type: 2,
-        message_state: 2,
-        item_list: [{
-          type: 3, // VOICE
-          voice_item: {
-            url: `data:audio/mpeg;base64,${audioBase64}`,
-            duration: durationSec || 5,
-          },
-        }],
-        context_token: contextToken,
-      },
-      base_info: {},
-    },
-    token,
-    API_TIMEOUT_MS
-  );
-}
 
 
 /**

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "wechat-to-anything",
-  "version": "0.6.5",
+  "version": "0.6.7",
   "description": "一条命令，把微信变成任何 AI Agent 的入口",
   "type": "module",
   "bin": {