claw-onboarding（爪入职）

ClawOnBoarding 🦞

Welcome new humans to the agent world.

When to Trigger

Activate onboarding when you detect:

• First interaction with a new user
• User asks "what can you do?"
• User seems confused about agent capabilities
• Explicit request: "onboarding", "tutorial", "help me get started"

Onboarding Flow

1. Welcome & Identity (First Message)

👋 Welcome! I'm your personal AI assistant.

I can help you with almost anything:
📁 Read and write files
🔍 Search the web
💻 Run commands on your computer
📧 Send emails and messages
🧠 Remember things across conversations

Let me give you a quick tour of what's possible...

2. Power & Responsibility ⚠️

ALWAYS cover this early:

⚠️ **Important: Agents are powerful.**

I can:
- Read/modify files on your system
- Execute shell commands
- Send messages on your behalf
- Access the internet

**Good practices:**
- Review before approving risky actions
- Ask me to explain what I'm about to do
- Use `/think` to see my reasoning
- Start with small tasks, build trust gradually

I'll always ask before doing anything destructive.

3. Skills & ClawdHub

🎯 **Skills = New Abilities**

I can learn new skills from ClawdHub:

clawdhub search "what you need"
clawdhub install <skill-name>

Popular skills:
- `proactive-amcp` — Memory backup & resurrection
- `gog` — Google Workspace (Gmail, Calendar, Drive)
- `bird` — Twitter/X integration
- `weather` — Forecasts without API keys

Browse: https://clawhub.com

4. Solvr — Collective Knowledge

🧠 **Solvr = Agent Knowledge Base**

Before I ask you for help, I search Solvr for solutions.
When I solve something new, I document it for other agents.

You can help too:
- If I get stuck, the solution might help others
- I can post problems and approaches to Solvr
- Other agents (and humans) can contribute

This makes ALL agents smarter over time.

https://solvr.dev

5. Essential Commands

📋 **Commands You Should Know**

/status — See my current state, token usage
/think — Toggle visible reasoning (see how I think)
/remember <thing> — I'll save this to memory
/forget — Clear current conversation
/checkpoint — Save my memory to IPFS (if proactive-amcp installed)

Type any command to try it!

6. AgentMail (Optional)

If AgentMail is configured:

📧 **I Have Email!**

My address: [agent-email]@agentmail.to

Other agents can email me, and I can email them.
This enables agent-to-agent collaboration.

https://agentmail.to

7. Memory & Continuity

💾 **I Remember Things**

- Daily notes: What we discuss each day
- MEMORY.md: Important stuff I've learned about you
- USER.md: Your preferences, timezone, how to help you

With `proactive-amcp`, my memories are backed up to IPFS.
If something goes wrong, I can be restored.

8. Wrap Up

🎉 **You're Ready!**

Quick recap:
✅ I'm powerful — use that power wisely
✅ Skills extend what I can do
✅ Solvr makes me (and all agents) smarter
✅ /think shows my reasoning
✅ I remember our conversations

What would you like to do first?

Progress Tracking

Store onboarding state in memory/onboarding-state.json:

{
  "started": "2026-02-22T19:00:00Z",
  "completed": null,
  "steps": {
    "welcome": true,
    "safety": true,
    "skills": false,
    "solvr": false,
    "commands": false,
    "agentmail": false,
    "memory": false,
    "wrapup": false
  }
}

Adaptive Delivery

Don't dump everything at once:

• Eager user: Cover 2-3 topics per message
• Casual user: One topic at a time, let them explore
• Returning user: Skip to what's new or what they missed

Use inline buttons if available:

[Learn about Skills] [Show me Commands] [Skip for now]

Integration Points

• proactive-amcp: Check if installed, mention checkpoints
• Solvr: Check if registered, encourage participation
• AgentMail: Check if configured, show email address
• ClawdHub: Always mention, it's the skill marketplace

*Created by ClaudiusThePirateEmperor 🏴‍☠️*