Skip to main content

openclaw-china-integration

Expert in integrating OpenClaw AI with Chinese IM platforms (DingTalk, Feishu, QQ, WeChat Work, WeChat Official Account)

インストールへ移動

ソース情報

リポジトリ
reason-machines/hermes-skills
ソースの最終更新活動
2026年5月17日 07:26
検出された SKILL.md の言語
英語
スター
5
フォーク
0

インストール方法

デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。

ソースファイルを確認

インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。

SKILL.md を表示中

SKILL.md
ソースの指示 · 読み取り専用プレビュー
name
openclaw-china-integration
description
Expert in integrating OpenClaw AI with Chinese IM platforms (DingTalk, Feishu, QQ, WeChat Work, WeChat Official Account)
triggers
["how do I connect OpenClaw to DingTalk","integrate OpenClaw with WeChat Work","set up OpenClaw China plugin","configure QQ bot with OpenClaw","add Feishu channel to OpenClaw","deploy OpenClaw to Chinese IM platforms","troubleshoot OpenClaw China connection","what Chinese messaging platforms does OpenClaw support"]
# OpenClaw China Integration Skill > Skill by [ara.so](https://ara.so) — Hermes Skills collection. OpenClaw China is a comprehensive plugin collection that extends OpenClaw to support major Chinese instant messaging platforms including DingTalk (钉钉), Feishu (飞书), QQ, WeChat Work (企业微信), WeChat Official Account (微信公众号), and WeChat Customer Service (微信客服). ## What OpenClaw China Does - **Multi-Platform Support**: Connects OpenClaw AI agents to 6+ Chinese IM platforms - **Unified Interface**: Provides consistent message handling across different platforms - **Rich Features**: Supports text, Markdown, images, files, voice messages, and streaming responses - **Enterprise Ready**: Handles both internal (WeChat Work) and external (WeChat Customer Service) users - **Low Configuration**: Simplified setup process with guided installation ## Installation ### Quick Setup (Recommended) ```bash npx @openclaw-china/setup ``` This interactive installer will: 1. Check your OpenClaw installation 2. Guide you through platform selection 3. Generate configuration files 4. Install required dependencies ### Manual Installation ```bash # Install OpenClaw China packages npm install @openclaw-china/dingtalk @openclaw-china/qqbot @openclaw-china/wecom @openclaw-china/wecom-app @openclaw-china/wecom-kf @openclaw-china/wechat-mp # Or install specific channels npm install @openclaw-china/dingtalk # DingTalk only npm install @openclaw-china/qqbot # QQ Bot only npm install @openclaw-china/wecom # WeChat Work Robot only ``` ## Platform Selection Guide | Platform | Complexity | Public IP Required | Use Case | |----------|------------|-------------------|----------| | **DingTalk** | Simple | No | Enterprise internal | | **QQ Bot** | Simple | No | Public bot service | | **WeChat Work Robot** | Simple | No | Enterprise internal (recommended) | | **WeChat Work App** | Medium | Yes | Connect to personal WeChat | | **WeChat Customer Service** | Medium | Yes | External customer support | | **WeChat Official Account** | Medium | Yes | Public followers | ## Configuration ### DingTalk Configuration ```typescript // config/channels.dingtalk.ts export default { channels: { dingtalk: { enabled: true, clientId: process.env.DINGTALK_CLIENT_ID, clientSecret: process.env.DINGTALK_CLIENT_SECRET, // Optional: Enable multiple accounts accounts: [ { clientId: process.env.DINGTALK_CLIENT_ID_1, clientSecret: process.env.DINGTALK_CLIENT_SECRET_1 } ] } } } ``` Environment variables: ```bash DINGTALK_CLIENT_ID=your_client_id DINGTALK_CLIENT_SECRET=your_client_secret ``` ### QQ Bot Configuration ```typescript // config/channels.qqbot-china.ts export default { channels: { 'qqbot-china': { enabled: true, appId: process.env.QQ_APP_ID, token: process.env.QQ_BOT_TOKEN, // Enable streaming for C2C (private chat) streaming: true, // Optional: Enable multiple bots accounts: [ { appId: process.env.QQ_APP_ID_1, token: process.env.QQ_BOT_TOKEN_1, secret: process.env.QQ_BOT_SECRET_1 } ] } } } ``` ### WeChat Work Robot (Long Connection) ```typescript // config/channels.wecom.ts export default { channels: { wecom: { enabled: true, corpId: process.env.WECOM_CORP_ID, corpSecret: process.env.WECOM_CORP_SECRET, // Connection mode: 'ws' (WebSocket) or 'webhook' mode: 'ws', // Optional: Robot token for webhook mode token: process.env.WECOM_ROBOT_TOKEN, encodingAESKey: process.env.WECOM_ENCODING_AES_KEY } } } ``` ### WeChat Work Self-Built App (Can Connect to Personal WeChat) ```typescript // config/channels.wecom-app.ts export default { channels: { 'wecom-app': { enabled: true, corpId: process.env.WECOM_APP_CORP_ID, agentId: process.env.WECOM_APP_AGENT_ID, secret: process.env.WECOM_APP_SECRET, token: process.env.WECOM_APP_TOKEN, encodingAESKey: process.env.WECOM_APP_ENCODING_AES_KEY, // Public URL for callbacks callbackUrl: process.env.WECOM_APP_CALLBACK_URL } } } ``` ### WeChat Customer Service (External Users) ```typescript // config/channels.wecom-kf.ts export default { channels: { 'wecom-kf': { enabled: true, corpId: process.env.WECOM_KF_CORP_ID, secret: process.env.WECOM_KF_SECRET, token: process.env.WECOM_KF_TOKEN, encodingAESKey: process.env.WECOM_KF_ENCODING_AES_KEY, // Welcome message when user enters session welcomeMessage: '您好!我是AI助手,有什么可以帮您?' } } } ``` ### WeChat Official Account ```typescript // config/channels.wechat-mp.ts export default { channels: { 'wechat-mp': { enabled: true, appId: process.env.WECHAT_MP_APP_ID, appSecret: process.env.WECHAT_MP_APP_SECRET, token: process.env.WECHAT_MP_TOKEN, encodingAESKey: process.env.WECHAT_MP_ENCODING_AES_KEY, // Render Markdown to plain text for WeChat compatibility renderMarkdown: true, // Active delivery mode: 'split' (multiple messages) or 'merged' (single message) activeDeliveryMode: 'split', // ASR (Automatic Speech Recognition) for voice messages asrConfig: { enabled: true, secretId: process.env.TENCENT_CLOUD_SECRET_ID, secretKey: process.env.TENCENT_CLOUD_SECRET_KEY, engineType: '16k_zh' // 16kHz Chinese } } } } ``` ## Real Code Examples ### Handling Incoming Messages ```typescript // Custom message handler for DingTalk import { DingTalkChannel } from '@openclaw-china/dingtalk'; const channel = new DingTalkChannel({ clientId: process.env.DINGTALK_CLIENT_ID, clientSecret: process.env.DINGTALK_CLIENT_SECRET }); // Listen for messages channel.on('message', async (message) => { console.log('Received:', message.content); // Process with OpenClaw const response = await openclawAgent.chat(message.content); // Send response await channel.sendMessage({ conversationId: message.conversationId, content: response.text, messageType: 'markdown' // or 'text' }); }); ``` ### Sending Proactive Messages (Scheduled Tasks) ```typescript // Send scheduled message via QQ Bot import { QQBotChannel } from '@openclaw-china/qqbot'; const qqBot = new QQBotChannel({ appId: process.env.QQ_APP_ID, token: process.env.QQ_BOT_TOKEN }); async function sendDailyReport() { const report = await generateReport(); await qqBot.sendMessage({ userId: 'user_openid', content: report, messageType: 'markdown' }); } // Schedule daily at 9 AM cron.schedule('0 9 * * *', sendDailyReport); ``` ### Handling File Uploads ```typescript // Process file from WeChat Work import { WeComAppChannel } from '@openclaw-china/wecom-app'; const wecomApp = new WeComAppChannel({ corpId: process.env.WECOM_APP_CORP_ID, agentId: process.env.WECOM_APP_AGENT_ID, secret: process.env.WECOM_APP_SECRET }); wecomApp.on('file', async (fileMessage) => { // Download file const fileBuffer = await wecomApp.downloadMedia(fileMessage.mediaId); // Process with OpenClaw const analysis = await openclawAgent.analyzeFile(fileBuffer); // Send response await wecomApp.sendMessage({ userId: fileMessage.fromUserId, content: `文件分析结果:\n${analysis}` }); }); ``` ### Streaming Response (QQ Bot C2C) ```typescript // Enable streaming for real-time chat experience import { QQBotChannel } from '@openclaw-china/qqbot'; const qqBot = new QQBotChannel({ appId: process.env.QQ_APP_ID, token: process.env.QQ_BOT_TOKEN, streaming: true // Enable streaming for private chat }); qqBot.on('message', async (message) => { // Streaming only works for C2C (private chat) if (message.messageType === 'C2C') { const stream = await openclawAgent.chatStream(message.content); for await (const chunk of stream) { await qqBot.updateStreamMessage({ messageId: message.id, content: chunk.text }); } } else { // Group chat falls back to regular message const response = await openclawAgent.chat(message.content); await qqBot.sendMessage({ groupId: message.groupId, content: response.text }); } }); ``` ### Voice Message with ASR (WeChat Official Account) ```typescript // Automatic voice-to-text conversion import { WeChatMPChannel } from '@openclaw-china/wechat-mp'; const wechatMP = new WeChatMPChannel({ appId: process.env.WECHAT_MP_APP_ID, appSecret: process.env.WECHAT_MP_APP_SECRET, asrConfig: { enabled: true, secretId: process.env.TENCENT_CLOUD_SECRET_ID, secretKey: process.env.TENCENT_CLOUD_SECRET_KEY, engineType: '16k_zh' } }); wechatMP.on('voice', async (voiceMessage) => { // ASR automatically converts voice to text const text = voiceMessage.recognizedText; console.log('Voice recognized:', text); // Process with OpenClaw const response = await openclawAgent.chat(text); await wechatMP.sendMessage({ userId: voiceMessage.fromUserId, content: response.text }); }); ``` ## Common Patterns ### Multi-Account Setup ```typescript // config/channels.dingtalk.ts export default { channels: { dingtalk: { enabled: true, accounts: [ { name: 'team-a', clientId: process.env.DINGTALK_TEAM_A_CLIENT_ID, clientSecret: process.env.DINGTALK_TEAM_A_CLIENT_SECRET }, { name: 'team-b', clientId: process.env.DINGTALK_TEAM_B_CLIENT_ID, clientSecret: process.env.DINGTALK_TEAM_B_CLIENT_SECRET } ] } } } ``` ### Group Chat Management ```typescript // Handle @mentions in WeChat Work group import { WeComChannel } from '@openclaw-china/wecom'; const wecom = new WeComChannel({ corpId: process.env.WECOM_CORP_ID, corpSecret: process.env.WECOM_CORP_SECRET, mode: 'ws' }); wecom.on('message', async (message) => { // Check if bot is mentioned in group if (message.isGroupChat && message.isMentioned) { const cleanContent = message.content.replace(/@\w+/g, '').trim(); const response = await openclawAgent.chat(cleanContent, { context: { groupId: message.groupId, userId: message.fromUserId } }); await wecom.sendMessage({ groupId: message.groupId, content: response.text, mentionedUsers: [message.fromUserId] // @user in response }); } }); ``` ### Message Formatting ```typescript // Convert Markdown to platform-specific format function formatMessage(text: string, platform: string): string { switch (platform) { case 'dingtalk': case 'wecom': case 'qqbot': // These support Markdown return text; case 'wechat-mp': // Convert Markdown to plain text return text .replace(/\*\*(.+?)\*\*/g, '$1') // Bold .replace(/\*(.+?)\*/g, '$1') // Italic .replace(/`(.+?)`/g, '$1') // Code .replace(/^#+\s+(.+)$/gm, '$1'); // Headers default: return text; } } ``` ## Troubleshooting ### Connection Issues **Problem**: WebSocket connection fails for WeChat Work Robot ```typescript // Solution: Check network and enable debug logging const wecom = new WeComChannel({ corpId: process.env.WECOM_CORP_ID, corpSecret: process.env.WECOM_CORP_SECRET, mode: 'ws', debug: true // Enable debug logs });
GitHubで見る
この SKILL.md は非常に大きいため、SkillsMP では最初のセクションだけを表示しています。 GitHubで見る