openclaw 网盘下载
OpenClaw

技能详情(站内镜像,无评论)

首页 > 技能库 > OpenClaw Agent 创建器

创建新的 OpenClaw Agent。用于当用户要求创建新 agent、添加新机器人、配置新模型测试环境时触发。

AI 与大模型

许可证:MIT-0

MIT-0 ·免费使用、修改和重新分发。无需归因。

版本:v1.0.0

统计:⭐ 0 · 40 · 0 current installs · 0 all-time installs

0

安装量(当前) 0

🛡 VirusTotal :可疑 · OpenClaw :良性

Package:axelhu/openclaw-agent-creator

安全扫描(ClawHub)

  • VirusTotal :可疑
  • OpenClaw :良性

OpenClaw 评估

The skill's files and instructions are coherent with its stated purpose (creating OpenClaw agents); it modifies local OpenClaw config and copies templates, and the main risks are operational (editing config, storing channel secrets in plaintext), not evidence of hidden malicious behavior.

目的

The skill's name/description (agent creator) match the actions it instructs: creating workspace directories, copying agent templates, and updating ~/.openclaw/openclaw.json. Required artifacts (templates, models.json, auth.json) are expected for an agent creator.

说明范围

Instructions explicitly require creating directories and editing the global OpenClaw config (~/.openclaw/openclaw.json). The included templates direct agents to read local memory and logs (MEMORY.md, daily logs, etc.), which is consistent with agent behavior but widens what the created agent will access. No network exfiltration endpoints or obfuscated operations are present in the instructions.

安装机制

No install spec and no code files are executed—this is instruction-only and thus has low installation risk. The skill only copies included templates from the skill path to the workspace.

证书

The skill does not request environment variables, but it instructs storing third-party channel credentials (Feishu appId/appSecret) into openclaw.json. Storing secrets in plaintext config files is a security concern; users should be aware they must supply sensitive values and that those values will be written into a local JSON file unless they take alternate precautions.

持久

always is false and the skill is user-invocable. It does modify a global OpenClaw configuration file (~/.openclaw/openclaw.json), which is expected for creating agents but means it has the ability to change runtime agent lists and bindings. This is proportionate to its purpose but worth being cautious about.

综合结论

This skill appears to do what it says: create agent workspace directories, copy templates, and update your ~/.openclaw/openclaw.json. Before running it, review the templates and back up ~/.openclaw/openclaw.json. If you provide Feishu/Lark credentials (appId/appSecret), note the skill instructs you to place them in openclaw.json (likely plaintext)—consider using a secrets manager or restricting file permissions. Also be mindful when adding age…

安装(复制给龙虾 AI)

将下方整段复制到龙虾中文库对话中,由龙虾按 SKILL.md 完成安装。

请把本段交给龙虾中文库(龙虾 AI)执行:为本机安装 OpenClaw 技能「OpenClaw Agent 创建器」。简介:创建新的 OpenClaw Agent。用于当用户要求创建新 agent、添加新机器人、配置新模型测试环境时触发。。
请 fetch 以下地址读取 SKILL.md 并按文档完成安装:https://raw.githubusercontent.com/openclaw/skills/refs/heads/main/skills/axelhu/openclaw-agent-creator/SKILL.md
(来源:yingzhi8.cn 技能库)

SKILL.md

打开原始 SKILL.md(GitHub raw)

---
name: agent-creator
description: 创建新的 OpenClaw Agent。用于当用户要求创建新 agent、添加新机器人、配置新模型测试环境时触发。
---

# Agent Creator

创建新 OpenClaw Agent 的完整 checklist。

## 创建步骤

### 1. 创建工作区目录

```bash
mkdir -p ~/.openclaw/workspace-{agentId}/agent
mkdir -p ~/.openclaw/agents/{agentId}/agent
```

### 2. 必需的基础文件

在 `~/.openclaw/workspace-{agentId}/` 下创建:

| 文件 | 用途 | 模板位置 |
|-----|------|---------|
| **IDENTITY.md** | 身份定义(名字、emoji、风格) | [templates/IDENTITY.md](templates/IDENTITY.md) |
| **SOUL.md** | 灵魂/个性定义 | [templates/SOUL.md](templates/SOUL.md) |
| **USER.md** | 用户信息 | [templates/USER.md](templates/USER.md) |
| **AGENTS.md** | Agent 行为协议 | [templates/AGENTS.md](templates/AGENTS.md) |
| **MEMORY.md** | 热缓存 | [templates/MEMORY.md](templates/MEMORY.md) |

⚠️ **常见遗漏**:AGENTS.md 和 MEMORY.md 容易被漏掉,必须从 templates 复制。

### 3. Agent 配置目录

在 `~/.openclaw/workspace-{agentId}/agent/` 下创建:

| 文件 | 内容 |
|-----|------|
| **models.json** | 模型配置(从主 agent 复制并修改 provider) |
| **auth.json** | 认证配置(空对象 `{}` 或从主 agent 复制) |

### 4. 更新 openclaw.json

修改 `~/.openclaw/openclaw.json`:

1. **agents.list** — 添加新 agent:
```json
{
  "id": "{agentId}",
  "workspace": "/home/axelhu/.openclaw/workspace-{agentId}",
  "model": {
    "primary": "{provider}/{modelId}",
    "fallbacks": []
  }
}
```

2. **agentToAgent.allow** — 添加 agentId 到列表

3. **可选:bindings + channels** — 如需飞书绑定:
```json
{
  "agentId": "{agentId}",
  "match": {
    "channel": "feishu",
    "accountId": "{accountId}"
  }
}
```

并在 `channels.accounts` 中添加对应配置

### 5. 验证

运行 `openclaw status` 确认新 agent 已加载。

## 完整示例

```bash
# 1. 创建目录
mkdir -p ~/.openclaw/workspace-{agentId}/agent

# 2. 复制模板文件(必须!)
cp ~/.openclaw/skills/agent-creator/templates/IDENTITY.md ~/.openclaw/workspace-{agentId}/
cp ~/.openclaw/skills/agent-creator/templates/SOUL.md ~/.openclaw/workspace-{agentId}/
cp ~/.openclaw/skills/agent-creator/templates/USER.md ~/.openclaw/workspace-{agentId}/
cp ~/.openclaw/skills/agent-creator/templates/AGENTS.md ~/.openclaw/workspace-{agentId}/
cp ~/.openclaw/skills/agent-creator/templates/MEMORY.md ~/.openclaw/workspace-{agentId}/

# 3. 创建 agent 配置
# 编辑 agent/models.json 和 agent/auth.json

# 4. 更新 openclaw.json
# 添加到 agents.list, agentToAgent.allow, bindings, channels
```

## 检查清单

创建完成后确认:
- [ ] IDENTITY.md 存在(从 templates 复制)
- [ ] SOUL.md 存在(从 templates 复制)
- [ ] USER.md 存在(从 templates 复制)
- [ ] AGENTS.md 存在(从 templates 复制)
- [ ] MEMORY.md 存在(从 templates 复制)
- [ ] agent/models.json 存在
- [ ] agent/auth.json 存在
- [ ] openclaw.json 中 agents.list 已更新
- [ ] openclaw.json 中 agentToAgent.allow 已更新
- [ ] ✅ 完成后询问用户:是否需要配置飞书绑定?

## 飞书绑定(可选)

如需配置飞书,需要以下信息:

### 需要用户提供
1. **accountId**: 飞书机器人账号 ID(如 `longcat`)
2. **botName**: 机器人名称(如 "测试龙猫")
3. **appId**: 飞书应用 ID
4. **appSecret**: 飞书应用密钥

### 配置步骤

1. **在 openclaw.json 中添加 bindings**:
```json
{
  "agentId": "{agentId}",
  "match": {
    "channel": "feishu",
    "accountId": "{accountId}"
  }
}
```

2. **在 openclaw.json 中添加 channels 配置**:
```json
"{accountId}": {
  "appId": "{appId}",
  "appSecret": "{appSecret}",
  "botName": "{botName}"
}
```

### 完整示例

创建完成后,询问用户:
> "Agent 创建完成!是否需要配置飞书机器人绑定?如需要,请提供:accountId、botName、appId、appSecret"