ClaudeCode-Custom-Skills/skills/claw-setup

🦞 Claw Setup

Professional AI Agent Deployment Made Simple

End-to-end setup of OpenClaw, NanoBot, PicoClaw, ZeroClaw, or NanoClaw with security hardening and personal customization

---

✨ Autonomously developed by GLM 5 Advanced Coding Model

⚠️ Disclaimer: Test in a test environment prior to using on any live system

---

Overview

Claw Setup handles the complete deployment of AI Agent platforms from the Claw family - from selection to production - with security best practices and personalized configuration through interactive brainstorming.

┌─────────────────────────────────────────────────────────────────┐
│                    CLAW SETUP WORKFLOW                          │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│  Phase 1          Phase 2          Phase 3          Phase 4    │
│  ────────         ────────         ────────         ────────   │
│                                                                 │
│  ┌─────────┐     ┌─────────┐     ┌─────────┐     ┌─────────┐  │
│  │ SELECT  │────►│ INSTALL │────►│CUSTOMIZE│────►│ DEPLOY  │  │
│  └─────────┘     └─────────┘     └─────────┘     └─────────┘  │
│      │               │               │               │         │
│      ▼               ▼               ▼               ▼         │
│  Compare         Clone &          Brainstorm       Systemd    │
│  platforms       harden           your use case    & monitor  │
│                  security                                     │
│                                                                 │
│  ┌─────────────────────────────────────────────────────────────┐│
│  │                   SUPPORTED PLATFORMS                       ││
│  │                                                             ││
│  │  🦞 OpenClaw   Full-featured, 1700+ plugins, 215K stars   ││
│  │  🤖 NanoBot    Python, 4K lines, research-ready            ││
│  │  🦐 PicoClaw   Go, <10MB, $10 hardware                     ││
│  │  ⚡ ZeroClaw   Rust, <5MB, 10ms startup                    ││
│  │  💬 NanoClaw   TypeScript, WhatsApp focused                ││
│  │                                                             ││
│  └─────────────────────────────────────────────────────────────┘│
│                                                                 │
└─────────────────────────────────────────────────────────────────┘

Platform Comparison

┌─────────────────────────────────────────────────────────────────┐
│                    PLATFORM COMPARISON                          │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│  Metric        OpenClaw  NanoBot  PicoClaw  ZeroClaw  NanoClaw │
│  ─────────────────────────────────────────────────────────────  │
│  Language      TS        Python   Go        Rust      TS       │
│  Memory        >1GB      ~100MB   <10MB     <5MB      ~50MB    │
│  Startup       ~500s     ~30s     ~1s       <10ms     ~5s      │
│  Binary Size   ~28MB     N/A      ~8MB      3.4MB     ~15MB    │
│  GitHub Stars  215K+     22K      15K       10K       5K       │
│  Plugins       1700+     ~50      ~20       ~15       ~10      │
│  Learning      Medium     Easy     Easy     Medium    Easy     │
│                                                                 │
│  BEST FOR:                                                      │
│  ─────────                                                      │
│  OpenClaw  → Full desktop AI, extensive integrations           │
│  NanoBot   → Research, customization, Python developers        │
│  PicoClaw  → Embedded, low-resource, $10 hardware              │
│  ZeroClaw  → Maximum performance, security-critical            │
│  NanoClaw  → WhatsApp automation, messaging bots               │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘

Decision Flowchart

                    ┌─────────────────┐
                    │ Need AI Agent?  │
                    └────────┬────────┘
                             │
                             ▼
                 ┌───────────────────────┐
                 │ Memory constrained?   │
                 │ (<1GB RAM available)  │
                 └───────────┬───────────┘
                       ┌─────┴─────┐
                       │           │
                      YES         NO
                       │           │
                       ▼           ▼
              ┌──────────────┐  ┌──────────────────┐
              │ Need <10MB?  │  │ Want plugins?    │
              └──────┬───────┘  └────────┬─────────┘
               ┌─────┴─────┐       ┌─────┴─────┐
               │           │       │           │
              YES         NO      YES         NO
               │           │       │           │
               ▼           ▼       ▼           ▼
          ┌────────┐  ┌────────┐ ┌────────┐ ┌────────┐
          │ZeroClaw│  │PicoClaw│ │OpenClaw│ │NanoBot │
          │ (Rust) │  │  (Go)  │ │ (Full) │ │(Python)│
          └────────┘  └────────┘ └────────┘ └────────┘

Quick Start

Option 1: Interactive Setup (Recommended)

"Setup Claw AI assistant on my server"
"Help me choose and install an AI agent platform"

Option 2: Direct Platform Selection

"Setup OpenClaw with all security features"
"Install ZeroClaw on my VPS"
"Deploy NanoBot for research use"

Installation Guides

OpenClaw (Full Featured)

# Prerequisites
sudo apt update && sudo apt install -y nodejs npm git

# Clone official repo
git clone https://github.com/openclaw/openclaw.git
cd openclaw

# Install dependencies
npm install

# Run setup wizard
npm run setup

# Configure environment
cp .env.example .env
nano .env  # Add your API keys

# Start
npm run start

NanoBot (Python Lightweight)

# Quick install via pip
pip install nanobot-ai

# Initialize
nanobot onboard

# Configure (~/.nanobot/config.json)
{
  "providers": {
    "openrouter": { "apiKey": "sk-or-v1-xxx" }
  },
  "agents": {
    "defaults": { "model": "anthropic/claude-opus-4-5" }
  }
}

# Start gateway
nanobot gateway

PicoClaw (Go Ultra-Light)

# Download latest release
wget https://github.com/sipeed/picoclaw/releases/latest/download/picoclaw-linux-amd64
chmod +x picoclaw-linux-amd64
sudo mv picoclaw-linux-amd64 /usr/local/bin/picoclaw

# Create config
mkdir -p ~/.config/picoclaw
picoclaw config init

# Start
picoclaw gateway

ZeroClaw (Rust Minimal)

# Download latest release
wget https://github.com/zeroclaw-labs/zeroclaw/releases/latest/download/zeroclaw-linux-amd64
chmod +x zeroclaw-linux-amd64
sudo mv zeroclaw-linux-amd64 /usr/local/bin/zeroclaw

# Initialize config
zeroclaw init

# Migrate from OpenClaw (optional)
zeroclaw migrate openclaw --dry-run

# Start
zeroclaw gateway

Security Hardening

1. Secrets Management

# Never hardcode API keys - use environment variables
export ANTHROPIC_API_KEY="your-key"
export OPENROUTER_API_KEY="your-key"

# Add to shell profile for persistence
echo 'export ANTHROPIC_API_KEY="your-key"' >> ~/.bashrc

# Use encrypted config files
mkdir -p ~/.config/claw
chmod 700 ~/.config/claw

2. Network Security

# Bind to localhost only
# config.json:
{
  "server": {
    "host": "127.0.0.1",
    "port": 3000
  }
}

# Use nginx reverse proxy for external access
sudo certbot --nginx -d claw.yourdomain.com

3. Systemd Hardened Service

# /etc/systemd/system/claw.service
[Unit]
Description=Claw AI Assistant
After=network.target

[Service]
Type=simple
User=claw
Group=claw
WorkingDirectory=/opt/claw
ExecStart=/usr/local/bin/claw gateway
Restart=on-failure

# Security hardening
NoNewPrivileges=true
PrivateTmp=true
ProtectSystem=strict
ProtectHome=true
ReadWritePaths=/opt/claw/data
Environment="ANTHROPIC_API_KEY=%i"

[Install]
WantedBy=multi-user.target

# Enable service
sudo systemctl daemon-reload
sudo systemctl enable --now claw

Brainstorm Session

After installation, we'll explore your needs:

🎯 Use Case Discovery

Q: What tasks should your AI handle?
   □ Code assistance & development
   □ Research & information gathering
   □ Personal productivity (calendar, reminders)
   □ Content creation & writing
   □ Data analysis & visualization
   □ Home automation
   □ Customer support / chatbot
   □ Other: _______________

🤖 Model Selection

Q: Which AI model(s) to use?
   
   □ Claude (Anthropic)      - Best reasoning
   □ GPT-4 (OpenAI)          - General purpose
   □ Gemini (Google)         - Multimodal
   □ Local models (Ollama)   - Privacy-first
   □ OpenRouter              - Multi-model access

🔌 Integration Planning

Q: Which platforms to connect?
   
   Messaging:
   □ Telegram   □ Discord   □ WhatsApp   □ Slack
   
   Calendar:
   □ Google     □ Outlook   □ Apple      □ None
   
   Storage:
   □ Local      □ Google Drive   □ Dropbox   □ S3
   
   APIs:
   □ Custom REST APIs
   □ Webhooks
   □ Database connections

🎨 Agent Personality

Q: How should your agent behave?
   
   Tone: Professional □ Casual □ Formal □ Playful □
   
   Proactivity:
   □ Reactive (responds only when asked)
   □ Proactive (suggests, reminds, initiates)
   
   Memory:
   □ Session only (fresh each chat)
   □ Persistent (remembers everything)
   □ Selective (configurable retention)

Architecture

┌─────────────────────────────────────────────────────────────────┐
│                    DEPLOYED ARCHITECTURE                        │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│                      ┌─────────────┐                           │
│                      │   Internet  │                           │
│                      └──────┬──────┘                           │
│                             │                                   │
│                     ┌───────▼───────┐                          │
│                     │  nginx/HTTPS  │                          │
│                     │   (Reverse    │                          │
│                     │    Proxy)     │                          │
│                     └───────┬───────┘                          │
│                             │                                   │
│  ┌──────────────────────────┼──────────────────────────────┐  │
│  │                     localhost                            │  │
│  │  ┌─────────┐    ┌─────────▼────────┐    ┌────────────┐  │  │
│  │  │ Config  │    │   CLAW ENGINE    │    │   Data     │  │  │
│  │  │ ~/.config│    │   (Gateway)      │    │  Storage   │  │  │
│  │  │  /claw   │    │   Port: 3000     │    │  ~/claw/   │  │  │
│  │  └─────────┘    └─────────┬────────┘    └────────────┘  │  │
│  │                           │                              │  │
│  │         ┌─────────────────┼─────────────────┐           │  │
│  │         │                 │                 │           │  │
│  │    ┌────▼────┐      ┌─────▼─────┐    ┌─────▼─────┐     │  │
│  │    │  LLM    │      │   Tools   │    │  Memory   │     │  │
│  │    │  APIs   │      │  Plugins  │    │  Context  │     │  │
│  │    │Claude/GPT│     │  Skills   │    │  Store    │     │  │
│  │    └─────────┘      └───────────┘    └───────────┘     │  │
│  │                                                          │  │
│  └──────────────────────────────────────────────────────────┘  │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘

Post-Setup Checklist

□ API keys configured securely
□ Network binding verified (localhost)
□ Firewall configured
□ SSL certificate installed (if external)
□ Systemd service enabled
□ Logs configured and rotating
□ Backup strategy in place
□ Test conversation successful
□ Custom agents created
□ Integrations connected

---

Learn more about GLM 5 Advanced Coding Model

AI Provider Configuration

Supported Providers

┌─────────────────────────────────────────────────────────────────┐
│                    AI PROVIDER OPTIONS                          │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│  Direct Providers              │  Gateways & Aggregators        │
│  ─────────────────             │  ──────────────────────        │
│  • Anthropic (Claude)          │  • OpenRouter (200+ models)   │
│  • OpenAI (GPT-4, o1, o3)      │  • Replicate                  │
│  • Google (Gemini 2.0)         │                               │
│  • Mistral                     │  Fast Inference               │
│  • DeepSeek                    │  ───────────────              │
│  • xAI (Grok)                  │  • Groq (ultra-fast)          │
│                                │  • Cerebras (fastest)         │
│  Local/Self-Hosted             │  • Together AI                │
│  ──────────────────            │                               │
│  • Ollama                      │                               │
│  • LM Studio                   │                               │
│  • vLLM                        │                               │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘

Model Selection Options

Option A: Fetch from Provider

# Automatically fetch available models
"Fetch available models from OpenRouter"
"Show me Groq models"
"What models are available via OpenAI?"

Option B: Custom Model Input

"Add custom model: my-org/fine-tuned-llama"
"Configure local Ollama model: llama3.2:70b"
"Use fine-tuned GPT: ft:gpt-4o:org:custom"

Multi-Provider Setup

{
  "providers": {
    "anthropic": { "apiKey": "${ANTHROPIC_API_KEY}" },
    "openai": { "apiKey": "${OPENAI_API_KEY}" },
    "google": { "apiKey": "${GOOGLE_API_KEY}" },
    "openrouter": { "apiKey": "${OPENROUTER_API_KEY}" },
    "groq": { "apiKey": "${GROQ_API_KEY}" },
    "ollama": { "baseURL": "http://localhost:11434" }
  },
  "models": {
    "default": "anthropic/claude-sonnet-4-5",
    "fast": "groq/llama-3.3-70b-versatile",
    "local": "ollama/llama3.2:70b"
  }
}

Provider Comparison

Provider	Best For	Speed	Cost
Claude	Reasoning, coding	Medium	$$$
GPT-4o	General purpose	Fast	$$$
Gemini	Multimodal	Fast	$$
Groq	Fastest inference	Ultra-fast	$
OpenRouter	Model variety	Varies	-$$
Ollama	Privacy, free	Depends on HW	Free

---

Learn more about GLM 5 Advanced Coding Model