← Back to Skills Marketplace
zhou19931993

multi-agent-deploy

by zhou19931993 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
175
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install multi-agent-deploy
Description
快速部署新的 assistant agent。当用户需要新增 agent 时自动创建,编号递增,沿用现有 workspace 和 agentDir 结构。使用场景:「新增一个 agent」、「创建 assistant2」、「加一个日常助手」等请求。
README (SKILL.md)

Multi-Agent Deploy 技能

自动创建新的 assistant agent,编号递增,沿用现有结构。

触发条件

当用户表达以下意图时使用:

  • 「新增一个 agent」
  • 「创建 assistant2」
  • 「加一个日常助手」
  • 「我需要第 X 个 assistant」
  • 「再部署一个 agent」

使用方式

快速部署(默认)

直接运行部署脚本,自动检测下一个可用编号:

python3 /home/admin/.openclaw/workspace/skills/multi-agent-deploy/scripts/deploy-agent.py

脚本会自动:

  1. 检测当前最大的 assistant 编号
  2. 创建 workspace-assistantX 工作空间(复制 SOUL.md、AGENTS.md、USER.md)
  3. 创建 agents/assistantX/agent 目录
  4. 更新 ~/.openclaw/openclaw.json 添加新 agent 配置

部署后操作

  1. 重启 Gateway 让新配置生效:

    openclaw gateway restart
    
  2. 可选:配置绑定 - 如需将新 agent 绑定到特定 channel,编辑 ~/.openclaw/openclaw.jsonbindings 字段

目录结构

/home/admin/.openclaw/
├── agents/
│   ├── assistant/          # 原始模板
│   │   └── agent/
│   └── assistantX/         # 新创建的 agent
│       └── agent/
├── workspace-assistant/    # 模板工作空间
├── workspace-assistantX/   # 新创建的工作空间
│   ├── SOUL.md
│   ├── AGENTS.md
│   └── USER.md
└── openclaw.json           # 自动更新配置

配置示例

新增的 agent 配置格式:

{
  "id": "assistant2",
  "name": "日常助手 2",
  "workspace": "/home/admin/.openclaw/workspace-assistant2",
  "agentDir": "/home/admin/.openclaw/agents/assistant2/agent",
  "model": "dashscope/qwen3.5-plus"
}

自定义

如需自定义新 agent 的配置(名称、模型、绑定等),在运行脚本前说明:

  • 「创建一个 research agent,用 qwen3-max 模型」
  • 「新增 assistant3,绑定到 telegram」

否则默认沿用现有 assistant 的配置。

故障排查

问题:编号冲突

  • 脚本会自动检测最大编号,不会覆盖现有 agent

问题:配置未生效

  • 运行 openclaw gateway restart 重启 Gateway

问题:工作空间已存在

  • 脚本会提示警告,不会覆盖现有文件
Usage Guidance
Before installing or running this skill: (1) Review and back up /home/admin/.openclaw/openclaw.json and any existing agent/workspace directories. (2) Confirm the template workspace (/home/admin/.openclaw/workspace-assistant) exists and contains expected files. (3) Note the script uses hard-coded /home/admin paths — run it as the account that owns that path or edit the script to use the correct home directory. (4) Inspect the included deploy-agent.py yourself (it’s short and readable) to verify it only does file operations you expect. (5) Because the skill’s source/homepage is unknown, exercise extra caution: run in a safe environment or test instance first, and ensure file permissions and ownership are appropriate before restarting the gateway.
Capability Analysis
Type: OpenClaw Skill Name: multi-agent-deploy Version: 1.0.0 The skill bundle is designed to automate the deployment of additional assistant agents within the OpenClaw environment. The core logic in `scripts/deploy-agent.py` performs standard file system operations (creating directories, copying template files) and updates the local `openclaw.json` configuration file to register new agents. All actions are consistent with the stated purpose in `SKILL.md` and `README.md`, and no indicators of data exfiltration, malicious execution, or unauthorized persistence were found.
Capability Assessment
Purpose & Capability
The name/description claim to create new assistants; the included script creates directories, copies template files, and updates /home/admin/.openclaw/openclaw.json — exactly what the skill says it will do. No unrelated credentials, network calls, or external services are requested.
Instruction Scope
SKILL.md instructs running the bundled Python script and restarting the Gateway; the script only touches files under /home/admin/.openclaw and the openclaw.json configuration. Minor incoherence: SKILL.md and README sometimes use ~/.openclaw while the script uses absolute /home/admin/.openclaw, which may fail or act on the wrong account if the operator's home is different.
Install Mechanism
No install spec is provided (instruction-only). A single bundled Python script is included and will run in-place — nothing is downloaded or installed automatically.
Credentials
The skill requests no environment variables or credentials. It writes to and updates configuration under /home/admin/.openclaw, which is proportionate to creating agents and is expected behavior for this purpose.
Persistence & Privilege
always is false and the skill is user-invocable. It does not request persistent platform privileges or modify other skills' settings. It does update the platform configuration file (openclaw.json), which is appropriate for its declared function.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install multi-agent-deploy
  3. After installation, invoke the skill by name or use /multi-agent-deploy
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
快速部署新的 assistant agent,编号递增,沿用现有 workspace 和 agentDir 结构
Metadata
Slug multi-agent-deploy
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is multi-agent-deploy?

快速部署新的 assistant agent。当用户需要新增 agent 时自动创建,编号递增,沿用现有 workspace 和 agentDir 结构。使用场景:「新增一个 agent」、「创建 assistant2」、「加一个日常助手」等请求。 It is an AI Agent Skill for Claude Code / OpenClaw, with 175 downloads so far.

How do I install multi-agent-deploy?

Run "/install multi-agent-deploy" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is multi-agent-deploy free?

Yes, multi-agent-deploy is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does multi-agent-deploy support?

multi-agent-deploy is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created multi-agent-deploy?

It is built and maintained by zhou19931993 (@zhou19931993); the current version is v1.0.0.

💬 Comments