ClaudeCode-Custom-Skills/skills/claw-setup/SKILL.md

name, description, version

name	description	version
claw-setup	Use this skill when the user asks to "setup openclaw", "install nanobot", "deploy zeroclaw", "configure picoclaw", "setup qwen code", "import qwen oauth", "use free qwen tier", "configure AI providers", "add openai provider", "AI agent setup", or mentions setting up AI platforms.	2.0.0

Claw Setup Skill

End-to-end professional setup of AI Agent platforms with 25+ OpenCode-compatible providers and FREE Qwen OAuth cross-platform import.

Two Key Features

┌─────────────────────────────────────────────────────────────────┐
│                    CLAW SETUP FEATURES                          │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│  FEATURE 1: Qwen OAuth Cross-Platform Import (FREE)            │
│  ───────────────────────────────────────────────────            │
│  • FREE: 2,000 requests/day, 60 req/min                        │
│  • Works with: OpenClaw, NanoBot, PicoClaw, ZeroClaw           │
│  • Model: coder-model (qwen3-coder-plus)                       │
│  • Auth: Browser OAuth via qwen.ai                             │
│  • Token refresh: Automatic (ALL platforms)                    │
│  • DEFAULT: Recommended as primary provider                    │
│                                                                 │
│  FEATURE 2: 25+ OpenCode-Compatible AI Providers                │
│  ─────────────────────────────────────────────────              │
│  • All major AI labs: Anthropic, OpenAI, Google, xAI, Mistral  │
│  • Cloud platforms: Azure, AWS Bedrock, Google Vertex          │
│  • Fast inference: Groq, Cerebras                              │
│  • Gateways: OpenRouter (100+ models), Together AI             │
│  • Local: Ollama, LM Studio, vLLM                              │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘

Supported Platforms

Platform	Language	Memory	Qwen OAuth	All Providers	Best For
Qwen Code	TypeScript	~200MB	✅ Native	✅	FREE coding
ZeroClaw	Rust	<5MB	✅ Native	✅	Max performance
OpenClaw	TypeScript	>1GB	✅ Full	✅	Full-featured
NanoBot	Python	~100MB	✅ Full	✅	Research
PicoClaw	Go	<10MB	✅ Full	✅	Embedded
NanoClaw	TypeScript	~50MB	✅ Full	✅	WhatsApp

All platforms have IDENTICAL Qwen OAuth experience:

• FREE tier: 2,000 requests/day
• Auto token refresh
• Same credentials: ~/.qwen/oauth_creds.json
• Same API: dashscope.aliyuncs.com

---

FEATURE 1: Qwen OAuth Cross-Platform Import (FREE)

Get FREE Qwen OAuth

# Install Qwen Code CLI
npm install -g @qwen-code/qwen-code@latest

# Authenticate (FREE - opens browser)
qwen --auth-type qwen-oauth -p "test"

# FREE: 2,000 requests/day, 60 req/min
# Credentials saved to: ~/.qwen/oauth_creds.json

Credentials File Structure

{
  "access_token": "your-access-token",
  "refresh_token": "your-refresh-token",
  "token_type": "Bearer",
  "resource_url": "portal.qwen.ai",
  "expiry_date": 1771774796531
}

API Endpoints

Endpoint	URL
Auth (Browser)	https://portal.qwen.ai
Token Refresh	https://chat.qwen.ai/api/v1/oauth2/token
API Base	https://dashscope.aliyuncs.com/compatible-mode/v1
Chat Completions	/chat/completions

Available Models (FREE Tier)

Model	Best For
coder-model (qwen3-coder-plus)	Coding (DEFAULT)
qwen3-coder-flash	Fast coding
qwen-max	Complex tasks

Set as Default Provider

# ZeroClaw - Native qwen-oauth as default
cat > ~/.zeroclaw/config.toml << EOF
default_provider = "qwen-oauth"
default_model = "coder-model"  # qwen3-coder-plus
EOF

# Other platforms - Set environment variables
export OPENAI_API_KEY=$(cat ~/.qwen/oauth_creds.json | jq -r '.access_token')
export OPENAI_BASE_URL="https://dashscope.aliyuncs.com/compatible-mode/v1"
export OPENAI_MODEL="coder-model"  # qwen3-coder-plus

Import Methods

ZeroClaw (Native Provider - Auto Token Refresh)

ZeroClaw has built-in qwen-oauth provider that handles token refresh automatically:

# Configure ZeroClaw to use native qwen-oauth provider
cat > ~/.zeroclaw/config.toml << EOF
default_provider = "qwen-oauth"
default_model = "qwen3-coder-plus"
default_temperature = 0.7
EOF

# ZeroClaw automatically:
# • Reads ~/.qwen/oauth_creds.json
# • Refreshes expired tokens using refresh_token
# • Uses correct DashScope API endpoint

zeroclaw agent -m "Hello!"

Other Platforms (OpenAI-Compatible Import)

For OpenClaw, NanoBot, PicoClaw, NanoClaw - use OpenAI-compatible endpoint:

# Extract token
export OPENAI_API_KEY=$(cat ~/.qwen/oauth_creds.json | jq -r '.access_token')
export OPENAI_BASE_URL="https://dashscope.aliyuncs.com/compatible-mode/v1"

# Use with any platform
openclaw      # OpenClaw with FREE Qwen
nanobot       # NanoBot with FREE Qwen
picoclaw      # PicoClaw with FREE Qwen
nanoclaw      # NanoClaw with FREE Qwen

Auto Token Refresh (ALL Platforms)

Use the included refresh script to automatically refresh expired tokens:

# Check token status
./scripts/qwen-token-refresh.sh --status

# Refresh if expired
./scripts/qwen-token-refresh.sh

# Run as background daemon (checks every 5 min)
./scripts/qwen-token-refresh.sh --daemon

# Install as systemd service (auto-start on boot)
./scripts/qwen-token-refresh.sh --install
systemctl --user enable --now qwen-token-refresh

How Auto-Refresh Works

┌─────────────────────────────────────────────────────────────────────────────┐
│                    AUTO TOKEN REFRESH FLOW                                   │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                             │
│  1. Check expiry_date in ~/.qwen/oauth_creds.json                          │
│  2. If expired (< 5 min buffer):                                            │
│     POST https://chat.qwen.ai/api/v1/oauth2/token                          │
│     Body: grant_type=refresh_token&refresh_token=xxx                       │
│  3. Response: { access_token, refresh_token, expires_in }                  │
│  4. Update ~/.qwen/oauth_creds.json with new tokens                        │
│  5. Update ~/.qwen/.env with new OPENAI_API_KEY                            │
│  6. Platforms using source ~/.qwen/.env get fresh token                    │
│                                                                             │
│  Systemd service: Runs every 5 minutes, refreshes when needed              │
│                                                                             │
└─────────────────────────────────────────────────────────────────────────────┘

---

FEATURE 2: 25+ OpenCode-Compatible AI Providers

Tier 1: FREE Tier

Provider	Free Tier	Model	Setup
Qwen OAuth	2,000/day	Qwen3-Coder	qwen --auth-type qwen-oauth

Tier 2: Major AI Labs

Provider	SDK Package	Key Models	Features
Anthropic	@ai-sdk/anthropic	Claude 3.5/4/Opus	Extended thinking, PDF support
OpenAI	@ai-sdk/openai	GPT-4o, o1, o3, GPT-5	Function calling, structured output
Google AI	@ai-sdk/google	Gemini 2.5, Gemini 3 Pro	Multimodal, long context
xAI	@ai-sdk/xai	Grok models	Real-time data integration
Mistral	@ai-sdk/mistral	Mistral Large, Codestral	Code-focused models

Tier 3: Cloud Platforms

Provider	SDK Package	Models	Features
Azure OpenAI	@ai-sdk/azure	GPT-5 Enterprise	Azure integration, custom endpoints
Google Vertex	@ai-sdk/google-vertex	Claude, Gemini on GCP	Anthropic on Google infrastructure
Amazon Bedrock	@ai-sdk/amazon-bedrock	Nova, Claude, Llama 3	AWS credentials, regional prefixes

Tier 4: Aggregators & Gateways

Provider	SDK Package	Models	Features
OpenRouter	@openrouter/ai-sdk-provider	100+ models	Multi-provider gateway
Vercel AI	@ai-sdk/vercel	Multi-provider	Edge hosting, rate limiting
Together AI	@ai-sdk/togetherai	Open source models	Fine-tuning, hosting
DeepInfra	@ai-sdk/deepinfra	Open source	Cost-effective hosting

Tier 5: Fast Inference

Provider	SDK Package	Speed	Models
Groq	@ai-sdk/groq	Ultra-fast	Llama 3, Mixtral
Cerebras	@ai-sdk/cerebras	Fastest	Llama 3 variants

Tier 6: Specialized

Provider	SDK Package	Use Case
Perplexity	@ai-sdk/perplexity	Web search integration
Cohere	@ai-sdk/cohere	Enterprise RAG
GitLab Duo	@gitlab/gitlab-ai-provider	CI/CD AI integration
GitHub Copilot	Custom	IDE integration

Tier 7: Local/Self-Hosted

Provider	Base URL	Use Case
Ollama	localhost:11434	Local model hosting
LM Studio	localhost:1234	GUI local models
vLLM	localhost:8000	High-performance serving
LocalAI	localhost:8080	OpenAI-compatible local

---

Quick Import Script

#!/bin/bash
# Usage: source import-qwen-oauth.sh [platform]

CREDS_FILE="$HOME/.qwen/oauth_creds.json"
PLATFORM="${1:-zeroclaw}"

QWEN_TOKEN=$(cat "$CREDS_FILE" | jq -r '.access_token')

case "$PLATFORM" in
    zeroclaw)
        # Native provider - just set config
        sed -i 's/^default_provider = .*/default_provider = "qwen-oauth"/' ~/.zeroclaw/config.toml
        sed -i 's/^default_model = .*/default_model = "qwen3-coder-plus"/' ~/.zeroclaw/config.toml
        ;;
    *)
        # OpenAI-compatible
        export OPENAI_API_KEY="$QWEN_TOKEN"
        export OPENAI_BASE_URL="https://dashscope.aliyuncs.com/compatible-mode/v1"
        ;;
esac

---

Platform Installation

Qwen Code (Native FREE OAuth)

npm install -g @qwen-code/qwen-code@latest
qwen --auth-type qwen-oauth -p "test"

OpenClaw

git clone https://github.com/openclaw/openclaw.git
cd openclaw && npm install && npm run setup

NanoBot

pip install nanobot-ai && nanobot onboard

PicoClaw

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

ZeroClaw

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

---

Usage Examples

"Setup OpenClaw with FREE Qwen OAuth"
"Configure NanoBot with all AI providers"
"Import Qwen OAuth to ZeroClaw"
"Fetch available models from OpenRouter"
"Setup Claw with Anthropic and OpenAI providers"
"Add custom model to my Claw setup"

---

Troubleshooting

Token Expired

# ZeroClaw: Automatic refresh (no action needed)

# Other platforms: Re-authenticate and re-export
qwen --auth-type qwen-oauth -p "test"
export OPENAI_API_KEY=$(cat ~/.qwen/oauth_creds.json | jq -r '.access_token')

API Test

QWEN_TOKEN=$(cat ~/.qwen/oauth_creds.json | jq -r '.access_token')
curl -X POST "https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions" \
  -H "Authorization: Bearer $QWEN_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"model": "qwen3-coder-plus", "messages": [{"role": "user", "content": "Hello"}]}'