candidate-kb
Manage candidate knowledge base. Use when ingesting CVs, adding profile information, or querying candidate details for resume/cover letter generation.
An MCP server that provides on-demand skill loading for AI coding assistants.
---
name: openclaw-wecom-channel
description: "企业微信 (WeCom) Channel 插件 — 让 OpenClaw AI Agent 通过企业微信收发消息。支持消息加解密、Token 自动管理、访问控制策略。"
homepage: https://github.com/darrryZ/openclaw-wecom-channel
metadata: { "openclaw": { "emoji": "💬", "requires": { "bins": ["node"], "network": true } } }
---
# OpenClaw 企业微信 Channel 插件
企业微信 (WeCom/WxWork) 消息通道插件,让 OpenClaw AI Agent 通过企业微信收发消息,与 Telegram、Discord、Signal 等并列为原生 Channel。
## 功能
- **📩 接收消息** — 企业微信用户发送文本,Agent 自动回复
- **📤 主动推送** — Agent 通过企业微信 API 主动发送消息
- **🔐 消息加解密** — 完整实现企业微信 AES-256-CBC 消息加解密(WXBizMsgCrypt 标准)
- **🔑 Token 管理** — access_token 自动缓存 + 提前 5 分钟刷新
- **🛡️ 访问控制** — open / pairing / allowlist 三种策略
- **⚡ 智能回复** — 5 秒内被动回复,超时自动降级为主动推送
## 前置条件
- OpenClaw 已安装并运行
- 企业微信管理员权限(创建自建应用)
- 公网可达的回调 URL(推荐 Cloudflare Tunnel)
## 快速开始
### 1. 安装插件
```bash
# 克隆到 OpenClaw extensions 目录
git clone https://github.com/darrryZ/openclaw-wecom-channel.git ~/.openclaw/extensions/wecom
```
### 2. 配置 OpenClaw
编辑 `~/.openclaw/openclaw.json`:
```json
{
"channels": {
"wecom": {
"enabled": true,
"corpId": "你的企业ID",
"agentId": 1000003,
"secret": "应用Secret",
"token": "回调Token",
"encodingAESKey": "回调EncodingAESKey",
"port": 18800,
"dmPolicy": "open"
}
},
"plugins": {
"entries": {
"wecom": { "enabled": true }
}
}
}
```
### 3. 配置公网回调(Cloudflare Tunnel)
```bash
cloudflared tunnel create wecom-tunnel
cloudflared tunnel route dns wecom-tunnel wecom.yourdomain.com
cloudflared tunnel run --edge-ip-version 4 --url http://localhost:18800 wecom-tunnel
```
企业微信后台回调 URL 设置为:`https://wecom.yourdomain.com/wecom/callback`
### 4. 重启 Gateway
```bash
openclaw gateway restart
```
## 详细文档
完整的配置指南、企业微信后台设置步骤、故障排查请参考 README.md。
## 链接
- **GitHub**: https://github.com/darrryZ/openclaw-wecom-channel
- **OpenClaw**: https://github.com/openclaw/openclaw
- **企业微信开发文档**: https://developer.work.weixin.qq.com/document/
---
name: openclaw-wecom-channel
description: "企业微信 (WeCom) Channel 插件 — 让 OpenClaw AI Agent 通过企业微信收发消息。支持消息加解密、Token 自动管理、访问控制策略。"
homepage: https://github.com/darrryZ/openclaw-wecom-channel
metadata: { "openclaw": { "emoji": "💬", "requires": { "bins": ["node"], "network": true } } }
---
# OpenClaw 企业微信 Channel 插件
企业微信 (WeCom/WxWork) 消息通道插件,让 OpenClaw AI Agent 通过企业微信收发消息,与 Telegram、Discord、Signal 等并列为原生 Channel。
## 功能
- **📩 接收消息** — 企业微信用户发送文本,Agent 自动回复
- **📤 主动推送** — Agent 通过企业微信 API 主动发送消息
- **🔐 消息加解密** — 完整实现企业微信 AES-256-CBC 消息加解密(WXBizMsgCrypt 标准)
- **🔑 Token 管理** — access_token 自动缓存 + 提前 5 分钟刷新
- **🛡️ 访问控制** — open / pairing / allowlist 三种策略
- **⚡ 智能回复** — 5 秒内被动回复,超时自动降级为主动推送
## 前置条件
- OpenClaw 已安装并运行
- 企业微信管理员权限(创建自建应用)
- 公网可达的回调 URL(推荐 Cloudflare Tunnel)
## 快速开始
### 1. 安装插件
```bash
# 克隆到 OpenClaw extensions 目录
git clone https://github.com/darrryZ/openclaw-wecom-channel.git ~/.openclaw/extensions/wecom
```
### 2. 配置 OpenClaw
编辑 `~/.openclaw/openclaw.json`:
```json
{
"channels": {
"wecom": {
"enabled": true,
"corpId": "你的企业ID",
"agentId": 1000003,
"secret": "应用Secret",
"token": "回调Token",
"encodingAESKey": "回调EncodingAESKey",
"port": 18800,
"dmPolicy": "open"
}
},
"plugins": {
"entries": {
"wecom": { "enabled": true }
}
}
}
```
### 3. 配置公网回调(Cloudflare Tunnel)
```bash
cloudflared tunnel create wecom-tunnel
cloudflared tunnel route dns wecom-tunnel wecom.yourdomain.com
cloudflared tunnel run --edge-ip-version 4 --url http://localhost:18800 wecom-tunnel
```
企业微信后台回调 URL 设置为:`https://wecom.yourdomain.com/wecom/callback`
### 4. 重启 Gateway
```bash
openclaw gateway restart
```
## 详细文档
完整的配置指南、企业微信后台设置步骤、故障排查请参考 README.md。
## 链接
- **GitHub**: https://github.com/darrryZ/openclaw-wecom-channel
- **OpenClaw**: https://github.com/openclaw/openclaw
- **企业微信开发文档**: https://developer.work.weixin.qq.com/document/
Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Includes source scripts; review them before use. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.