openclaw-control

OpenClaw Control Skill

This skill lets you control and manage the OpenClaw AI assistant.

Core Capabilities

1. Send messages to OpenClaw

Send tasks to OpenClaw via Naga's Agent Server API:

POST /openclaw/send
{
  "message": "你的消息内容",
  "wake_mode": "now"  // 立即执行
}

2. Check OpenClaw status

• GET /openclaw/health - Health check
• GET /openclaw/status - Running status
• GET /openclaw/session - Session information

3. Manage Gateway

• POST /openclaw/gateway/start - Start
• POST /openclaw/gateway/stop - Stop
• POST /openclaw/gateway/restart - Restart
• GET /openclaw/gateway/status - Status

4. Configuration management

• GET /openclaw/config - Get configuration
• POST /openclaw/config/model - Set model
• POST /openclaw/config/hooks - Configure Hooks

Usage Flow

1. Check connection: First call the health check to confirm OpenClaw is available
2. Send task: Use /openclaw/send to send a message
3. Monitor results: Via /openclaw/history to view execution results

Notes

• OpenClaw runs by default on http://127.0.0.1:18789
• A Hooks Token must be configured to send messages
• It is recommended to use the free GLM-4.7 model for testing

Examples

Have OpenClaw execute a task

{
  "message": "帮我整理桌面上的文件，按类型分类",
  "wake_mode": "now"
}

Query the current status

Call GET /openclaw/status to get what OpenClaw is currently doing.