warden-agent-builder

⚠Review·Scanned 2/18/2026

This skill scaffolds, tests, and deploys LangGraph agents for Warden Protocol, providing templates, scripts, and deployment guidance. It instructs running shell commands (e.g., python scripts/init-agent.py, npm install), using secrets in OPENAI_API_KEY/LANGSMITH_API_KEY, and calling network endpoints such as https://smith.langchain.com and arbitrary agent URLs like https://api.example.com.

from clawhub.ai·va734bac·90.5 KB·0 installs
Scanned from 1.0.0 at a734bac · Transparency log ↗
$ vett add clawhub.ai/kulotzkih/warden-agent-builderReview findings below

Warden Agent Builder Skill for OpenClaw

This skill enables OpenClaw bots to build and deploy LangGraph agents for the Warden Protocol's Agentic Wallet ecosystem.

Overview

Warden Protocol is an "Agentic Wallet for the Do-It-For-Me economy" with an active Agent Builder Incentive Programme offering up to $10,000 for top agents. This skill provides everything needed to:

  • Build LangGraph-based agents (TypeScript or Python)
  • Follow Warden's requirements and best practices
  • Deploy to LangSmith Deployments or custom infrastructure
  • Participate in the incentive programme
  • Register with Warden Studio

Skill Structure

warden-agent-builder/
├── SKILL.md                          # Main skill file (always read first)
├── references/
│   ├── langgraph-patterns.md         # Detailed LangGraph patterns & examples
│   └── deployment-guide.md           # API integration & deployment guide
├── scripts/
│   ├── init-agent.py                 # Quick agent initialization script
│   └── test-agent.py                 # Agent testing script
└── assets/
    └── example-configs.md            # Example configurations & templates

How OpenClaw Should Use This Skill

1. When to Trigger This Skill

The skill should be used when users mention:

  • "Build a Warden agent"
  • "Create an agent for Warden Protocol"
  • "Warden incentive programme"
  • "LangGraph agent"
  • "Deploy to Warden"
  • Specific Warden templates (Weather Agent, CoinGecko Agent, Portfolio Agent)

2. Step-by-Step Usage

Step 1: Read SKILL.md Always start by reading the main SKILL.md file which contains:

  • Requirements checklist
  • Template options (quick start, weather, CoinGecko, portfolio)
  • Building instructions
  • Deployment options
  • Best practices

Step 2: Use Reference Files When Needed

Read references/langgraph-patterns.md when:

  • User needs detailed code examples
  • Implementing Schema-Guided Reasoning (SGR)
  • Setting up complex workflows
  • Handling errors and edge cases

Read references/deployment-guide.md when:

  • Setting up API integrations
  • Deploying to production
  • Configuring Docker/Kubernetes
  • Implementing monitoring

Step 3: Use Scripts

The skill includes two helper scripts:

scripts/init-agent.py - Initialize new agent projects

python scripts/init-agent.py my-agent --template typescript --description "My Warden agent"

scripts/test-agent.py - Test deployed agents

python scripts/test-agent.py https://api.example.com --api-key [YOUR-API-KEY]

Step 4: Reference Assets

Use assets/example-configs.md for:

  • Sample agent configurations
  • Environment variable templates
  • Docker Compose examples
  • Warden Studio registration format

Example OpenClaw Workflows

Example 1: Building a New Agent

User: "Build me a Warden agent that tracks Bitcoin prices"

OpenClaw Actions:
1. Read SKILL.md to understand requirements
2. Identify: User wants ORIGINAL agent (not CoinGecko clone)
3. Suggest studying CoinGecko agent pattern, then building unique features
4. Use init-agent.py script to create NEW project structure
5. Read langgraph-patterns.md for SGR implementation pattern
6. Implement CUSTOM price tracking logic with unique features
7. Read deployment-guide.md for API setup
8. Generate code files for NEW agent
9. Provide deployment instructions
10. Emphasize: "This should be different from CoinGecko Agent"

Example 2: Deploy Existing Agent

User: "Deploy my agent to Warden"

OpenClaw Actions:
1. Read SKILL.md deployment section
2. Check if LangSmith Deployments or self-hosted
3. Read deployment-guide.md for specific instructions
4. Verify requirements checklist
5. Provide deployment commands
6. Explain Warden Studio registration

Example 3: Debug Agent

User: "My Warden agent has errors"

OpenClaw Actions:
1. Use test-agent.py script to diagnose
2. Read langgraph-patterns.md for error handling patterns
3. Check deployment-guide.md for common issues
4. Suggest fixes based on error type

Key Warden Requirements

Agents MUST meet these requirements to be eligible for the incentive programme:

✓ Framework: LangGraph only (TypeScript or Python) ✓ Access: API-accessible (no UI needed) ✓ Deployment: LangSmith Deployments OR custom infrastructure ✓ API Keys Required:

  • OpenAI API key (for LLM calls)
  • LangSmith API key (for LangSmith Deployments deployment)
  • Additional keys based on your agent's functionality ✓ Isolation: One agent per LangGraph instance ✓ Phase 1 Restrictions:
  • Cannot access user wallets
  • Cannot store data on Warden infrastructure
  • (Confirm current limitations in the latest Warden docs)

Available Templates

The Warden repository contains example agents to study, not templates to recreate:

  1. LangGraph Quick Start - Minimal chatbot for learning basics
  2. Weather Agent ⚠️ - Study for simple data fetching patterns (don't rebuild)
  3. CoinGecko Agent ⚠️ - Study for intermediate SGR patterns (don't rebuild)
  4. Portfolio Analysis Agent ⚠️ - Study for advanced architecture (don't rebuild)

These exist to teach patterns. Build something NEW for the incentive programme!

Example Agent Ideas

Web3 Use Cases:

  • Gas price optimizer
  • NFT rarity analyzer
  • DeFi yield comparator
  • Wallet health checker
  • Transaction explainer
  • Token price alerts

General Use Cases:

  • News aggregator
  • Research assistant
  • Data visualizer
  • API orchestrator
  • Workflow automator

Installation for OpenClaw

To use this skill with OpenClaw:

  1. Place the warden-agent-builder folder in your OpenClaw skills/ directory
  2. OpenClaw will automatically detect the skill from SKILL.md frontmatter
  3. The skill triggers when users mention Warden or LangGraph agents

Resources

Quick Start Commands

# Clone Warden community agents
git clone https://github.com/warden-protocol/community-agents.git

# Initialize new agent (TypeScript)
python scripts/init-agent.py my-agent --template typescript

# Initialize new agent (Python)
python scripts/init-agent.py my-agent --template python

# Test deployed agent
python scripts/test-agent.py https://your-agent.com --api-key [YOUR-API-KEY]

# Deploy to LangSmith Deployments
# Use the LangSmith Deployments UI after pushing to GitHub

Support

For questions or issues with this skill:

  • Review the SKILL.md file first
  • Check references/ for detailed guides
  • Consult assets/example-configs.md for templates
  • Join Warden Discord #developers channel

Next Step: Publish on Warden Studio

If the user needs guided UI steps to publish in Warden Studio, switch to the OpenClaw skill: https://www.clawhub.ai/Kryptopaid/warden-studio-deploy

License

This skill is provided for use with OpenClaw bots to build Warden Protocol agents. See individual file headers for specific licensing.

Built for OpenClaw by Claude Compatible with Warden Protocol Agent Builder Incentive Programme