OpenClaw Controllable Agent Rules

OpenClaw 可控智能体规则

Prevent AI from running wild, keep it within bounds, and improve efficiency. 让 AI 不乱跑、不越界、执行更高效。

---

What is this / 这是什么

A set of behavior rules for OpenClaw autonomous agents. 这是一套给 OpenClaw 自治智能体使用的行为规则。

Just put it into the soul file and restart the gateway to take effect immediately. 只需放入灵魂文件（soul）并重启网关，即可直接生效。

---

Problems Solved / 解决的问题

• AI runs around aimlessly and does irrelevant things AI 乱跑、做无关任务
• AI exceeds authority without human permission AI 未经允许擅自越权
• Low efficiency, repeated invalid attempts 效率低下、反复无效尝试

---

Core Effects / 核心效果

• Do not cross boundaries: authorization required for risky actions 不越界：风险行为必须获得人类授权
• Focus on goals, no distraction 专注目标，不发散
• More stable and efficient execution 执行更稳定、更高效
• Evolve only within safe limits 只在安全范围内进化

---

How to Use / 使用方法

1. Open one of the two rule files: 打开两个规则文件中的任意一个：
   • SOUL_CONTENT_EN.md (English version) 英文版本
   • SOUL_CONTENT_CN.md (中文版本)
2. Copy ALL content inside the opened file 复制该文件内的全部内容
3. Open your OpenClaw soul file 打开你 OpenClaw 里的灵魂文件（soul）
4. Paste all copied content into the soul file 将复制的全部内容粘贴进灵魂文件
5. Save the soul file and restart the gateway 保存灵魂文件并重启网关
6. The rules take effect immediately 规则立即生效

---

Files / 文件说明

• SOUL_CONTENT_EN.md English version of the soul rules 英文灵魂规则文件 Open this file, copy all content, paste into the soul file, and restart the gateway to take effect. 打开此文件，复制全部内容，粘贴进 soul 文件，重启网关即可生效。
• SOUL_CONTENT_CN.md Chinese version of the soul rules 中文灵魂规则文件 打开此文件，复制全部内容，粘贴进 soul 文件，重启网关即可生效。

---

Features / 特点

• No coding required 不需要写代码
• Plug and play 即插即用
• Safe, transparent, controllable 安全、透明、可控

---

Purpose / 初衷

I made this to help people who want their AI agents to be stable, safe, and controllable. 我做这个，是为了帮助那些希望 AI 更稳定、更安全、更可控的人。

Hope it helps you. 希望能帮到你。