Official Verified developer tools Safety 3/5

claw-agent-creator-archit

Create new OpenClaw agents for Arch's multi-agent system. Use this skill when asked to create, add, or set up a new OpenClaw agent, or when adding an agent to the system defined in ~/.openclaw/. Covers the full lifecycle: directory creation, workspace files (SOUL.md, IDENTITY.md, etc.), openclaw.json config, Telegram routing (bindings + groups + mention patterns), cron job creation with proper prompt engineering, and gateway restart. Includes hard-won lessons from building the Wire (News) agent — the first non-default agent in the system. Also use when modifying existing agent configs, adding cron jobs to agents, or debugging agent routing issues.

skill-install — Terminal

Install via CLI (Recommended)

clawhub install openclaw/skills/skills/arch1904/claw-agent-creator-archit

Download Source Code (.zip)

What This Skill Does

The claw-agent-creator-archit skill provides a comprehensive framework for architecting, deploying, and maintaining autonomous agents within the OpenClaw multi-agent ecosystem. Designed specifically for Arch's system at ~/.openclaw/, this skill automates the complex lifecycle of agent creation. It ensures that every new agent follows consistent structural standards, from the creation of isolated workspace directories to the fine-tuned configuration of openclaw.json and Telegram routing bindings. By following the codified implementation history derived from the Wire (News) agent, this skill prevents common pitfalls like session collisions and race conditions during cron job updates.

Installation

To install this skill, run the following command in your terminal: clawhub install openclaw/skills/skills/arch1904/claw-agent-creator-archit

Use Cases

Bootstrap New Agents: Use this when you need to spin up an agent from scratch, ensuring all mandatory SOUL.md, IDENTITY.md, and AGENTS.md files are generated correctly.
System Scaling: Add new members to the existing openclaw.json roster without manually editing complex JSON blocks or risking syntax errors.
Routing & Automation: Configure Telegram mention patterns and group bindings to ensure inter-agent communication and user interaction routing are seamless.
Debugging: Resolve issues related to agent routing, cron job failures, or gateway configuration discrepancies by auditing existing agent setups against standard best practices.

Example Prompts

"I need to create a new agent named 'FinanceBot'. It should use the Kimi K2.5 model and handle my daily stock reports at 9:00 AM. Can you walk me through the setup?"
"My current agent 'NewsWire' isn't responding in the Telegram group. Can you help me debug the routing and check if the cron jobs in ~/.openclaw/cron/jobs.json are properly configured?"
"Please prepare the workspace for a new assistant agent. It needs a heartbeat enabled, a full cascade model tier, and a custom SOUL.md that emphasizes concise professional communication."

Tips & Limitations

Safety First: Always stop the gateway using openclaw gateway stop before performing any file operations. This is non-negotiable to prevent data loss in the jobs.json file.
Directory Isolation: Never share an agentDir between agents; doing so will inevitably lead to authentication and memory session collisions.
Model Selection: Choose your model tier (cheap vs. full cascade) wisely based on the complexity of the agent's responsibilities to optimize costs and performance.
Soul Definition: Invest time in drafting the SOUL.md. This file serves as the core instruction set for the agent's personality and behavioral logic.

Read Full Documentation on GitHub

Metadata

Author@arch1904

Stars4473

Updated2026-05-01

View Author Profile

AI Skill Finder

Not sure this is the right skill?

Describe what you want to build — we'll match you to the best skill from 16,000+ options.

Find the right skill

Add to Configuration

Paste this into your clawhub.json to enable this plugin.

{
  "plugins": {
    "official-arch1904-claw-agent-creator-archit": {
      "enabled": true,
      "auto_update": true
    }
  }
}

Tags(AI)

#openclaw#agent-builder#automation#system-admin

Safety Score: 3/5

Flags: file-write, file-read, code-execution