OpenClaw Starter Template

A ready-to-use OpenClaw configuration template. Clone this and get running in minutes.

What's Included

• openclaw.json.example — Main configuration template (fill in your keys)
• workspace/ — Workspace files (AGENTS.md, SOUL.md, USER.md, etc.)
• skills/ — Pre-installed skills
  • agent-browser — Browser automation for AI agents
  • agentic-coding — Production coding workflows
  • clawhub-cli — Install more skills from clawhub.com
  • healthcheck — Track water & sleep
  • self-improving-agent — Self-improving agent system
• completions/ — Shell completions (bash, zsh, fish, PowerShell)

Prerequisites

1. Node.js (v18+) — https://nodejs.org
2. OpenClaw — npm install -g openclaw
3. An API key from one of:
   • OpenRouter (recommended, access to many models)
   • OpenAI
   • Or run a local model via LM Studio

Quick Start

1. Clone this repo

git clone <your-fork-url> ~/.openclaw-setup
cd ~/.openclaw-setup

2. Install OpenClaw (if not already)

npm install -g openclaw

3. Copy config

# Copy the example config
cp openclaw.json.example ~/.openclaw/openclaw.json

4. Edit config with your keys

Open ~/.openclaw/openclaw.json and replace these placeholders:

5. Copy workspace & skills

# Copy workspace
cp -r workspace/* ~/.openclaw/workspace/

# Copy skills
cp -r skills/* ~/.openclaw/skills/

6. Start OpenClaw

# Run the onboard wizard (first time)
openclaw onboard

# Or start the gateway directly
openclaw gateway start

7. Connect your channel (optional)

Telegram:

1. Message @BotFather on Telegram
2. Create a new bot with /newbot
3. Copy the bot token into openclaw.json → channels.telegram.botToken
4. Restart gateway: openclaw gateway restart

Configuration

Models

By default, this template uses:

To use a local model, install LM Studio, load a model, and start the local server on port 8033.

Customization

• SOUL.md — Edit to change personality
• USER.md — Add info about yourself for better context
• AGENTS.md — Customize agent behavior & rules
• TOOLS.md — Add your local tool configs (cameras, SSH, etc.)

Adding More Skills

# Search for skills
clawhub search <keyword>

# Install a skill
clawhub install <skill-name>

# List installed skills
clawhub list

Browse: https://clawhub.com

Common Commands

openclaw status          # Check system status
openclaw gateway start   # Start the gateway
openclaw gateway stop    # Stop the gateway
openclaw gateway restart # Restart the gateway
openclaw doctor --fix    # Diagnose & fix issues
openclaw onboard         # Run setup wizard
openclaw cron list       # List cron jobs

Telegram Commands (once connected)

• /status — Show current model & stats
• /model qwen — Switch to local model
• /model codex — Switch to Codex
• /model router — Switch to OpenRouter
• /reasoning — Toggle thinking mode

File Structure

~/.openclaw/
├── openclaw.json          # Main config (you create this)
├── workspace/             # Workspace files
│   ├── AGENTS.md          # Agent behavior rules
│   ├── SOUL.md            # Personality
│   ├── USER.md            # About you
│   ├── TOOLS.md           # Local tool configs
│   ├── IDENTITY.md        # Agent identity
│   └── HEARTBEAT.md       # Periodic checks
├── skills/                # Installed skills
├── credentials/           # (auto-generated, gitignored)
└── identity/              # (auto-generated, gitignored)

Security

• Never commit openclaw.json with real API keys
• Never share your credentials/ or identity/ folders
• The gateway runs on localhost by default (safe)

Troubleshooting

Gateway won't start:

openclaw doctor --fix

Model not responding:

• Check your API key is valid
• Check API credits/balance
• Try switching models: /model qwen

Telegram not working:

• Verify bot token in config
• Make sure you started a chat with the bot
• Check gateway is running: openclaw gateway status

Links

• Docs: https://docs.openclaw.ai
• Discord: https://discord.com/invite/clawd
• Skills: https://clawhub.com
• GitHub: https://github.com/openclaw/openclaw

Based on the personal setup of @devta0