autodegen 0.1.5

Production-ready crypto trading analysis agent with LLM hardening

AutoDegen 🚀

Production-ready cryptocurrency trading analysis agent with enterprise-grade LLM hardening.

🌟 Features

🛡️ Enterprise-Grade LLM Security

• Zero Meta Commentary: Complete elimination of "I'll analyze..." and "The key will be..." responses
• Structured Output Enforcement: Guaranteed consistent report format with Pydantic validation
• Production Hardening: Circuit breaker, retry logic, timeout handling, and tool call firewall

📊 Comprehensive Market Analysis

• Multi-Expert AI Coordination: Technical, market, and news sentiment analysis
• Real-Time Data: Integration with Binance, DeFiLlama, and crypto news sources
• Advanced Technical Indicators: RSI, MACD, moving averages, Smart Money Concepts
• Multi-Timeframe Analysis: 1h, 4h, daily perspectives with confluence signals

⚡ High-Performance Architecture

• Async/Await: Non-blocking operations for maximum throughput
• Resilience Patterns: Circuit breaker prevents cascade failures
• Comprehensive Testing: 52+ test cases with 100% production readiness validation
• Type Safety: Full type hints with mypy compatibility

🚀 Quick Start

Installation

# Basic installation
pip install autodegen

# With CLI support  
pip install autodegen[cli]

# Development installation
pip install autodegen[dev]

Environment Setup

Create a .env file or set environment variables:

FIREWORKS_API_KEY=your_fireworks_api_key      # Required
OPENAI_API_KEY=your_openai_key               # Optional  
BINANCE_API_KEY=your_binance_key             # Optional
BINANCE_API_SECRET=your_binance_secret       # Optional

Python API Usage

import asyncio
from autodegen import TradingAgent

async def main():
    # Initialize agent
    agent = TradingAgent()
    
    # Get structured analysis
    analysis = await agent.analyze("BTC", timeframe="4h")
    
    print(analysis.title)        # **COMPREHENSIVE TECHNICAL ANALYSIS: BTC**
    print(analysis.market_state) # Current market conditions...
    print(analysis.outlook)      # Bullish bias maintained above...
    
    # Get market overview
    overview = await agent.get_market_overview()
    print(overview)

# Run analysis
asyncio.run(main())

CLI Usage

# Analyze a cryptocurrency
autodegen analyze BTC --timeframe 4h --format rich

# Get market overview  
autodegen market

# Check system health
autodegen health

# List supported symbols
autodegen symbols

📖 Advanced Usage

Structured vs Unstructured Output

# Structured output (recommended)
structured_analysis = await agent.analyze("ETH", structured=True)
print(f"Market State: {structured_analysis.market_state}")
print(f"Trading Levels: {structured_analysis.trading_levels}")

# Text output  
text_analysis = await agent.analyze("ETH", structured=False)
print(text_analysis)  # Raw markdown-formatted text

Error Handling

from autodegen import TradingAgent, APIError, ConfigurationError

try:
    agent = TradingAgent()
    result = await agent.analyze("BTC")
except ConfigurationError as e:
    print(f"Setup error: {e}")
except APIError as e:
    print(f"API error: {e} (status: {e.status_code})")
except Exception as e:
    print(f"Unexpected error: {e}")

Health Monitoring

# Check system health
health = await agent.health_check()

if health["status"] == "healthy":
    print("All systems operational")
else:
    print(f"System status: {health['status']}")
    for component, status in health["components"].items():
        print(f"  {component}: {status}")

🏗️ Architecture

AutoDegen
├── 🧠 LLM Core (Production Hardened)
│   ├── Circuit Breaker Pattern
│   ├── Retry with Exponential Backoff  
│   ├── Tool Call Firewall
│   └── Structured Output Enforcement
│
├── 📊 Market Analysis Engine
│   ├── Technical Expert (Charts & Indicators)
│   ├── Market Expert (Liquidity & Volume)
│   └── News Expert (Sentiment Analysis)
│
├── 🔌 Data Sources
│   ├── Binance API (Price & Volume)
│   ├── DeFiLlama (DeFi Metrics)
│   └── News APIs (Market Sentiment)
│
└── 🛡️ Security & Reliability
    ├── Input Validation & Sanitization
    ├── Rate Limiting & Quotas
    └── Comprehensive Logging

📈 Supported Cryptocurrencies

Major Coins: BTC, ETH, BNB, SOL, ADA, XRP, DOT, AVAX

DeFi Tokens: UNI, COMP, CRV, LINK, AAVE

Layer 1s: ATOM, NEAR, ALGO, FTM, ICP

And more: MATIC, VET, FLOW, SAND, MANA

Use autodegen symbols to see the full list

🔧 Configuration

API Keys Priority

1. Constructor parameters
2. Environment variables
3. .env file (if python-dotenv installed)

Timeout & Retry Settings

agent = TradingAgent()

# Custom timeout and structured output
analysis = await agent.analyze(
    symbol="BTC",
    timeout=60.0,        # 60 second timeout
    structured=True      # Enforce schema validation
)

🧪 Production Hardening

This agent includes enterprise-grade production hardening tested with 52+ comprehensive test cases:

🛡️ LLM Response Security

• Meta Commentary Elimination: Zero tolerance for "I'll analyze...", "The key will be..." responses
• Structured Schema Enforcement: Guaranteed report format with required sections
• Content Validation: Multi-layer filtering of banned tokens and meta statements

🔒 Security Features

• Tool Call Firewall: Blocks dangerous function calls (execute_code, run_shell)
• Input Sanitization: Automatic cleanup of potentially harmful content
• Circuit Breaker: Prevents cascade failures with automatic recovery

⚡ Reliability Patterns

• Retry Logic: Exponential backoff with jitter (3 attempts default)
• Timeout Handling: Graceful degradation on slow responses
• Health Monitoring: Real-time component status checking

🧪 Testing & Quality

# Install development dependencies
pip install autodegen[dev]

# Run the production quality test battery  
python -m pytest tests/run_quality_battery.py

# Test specific components
python -m pytest tests/ -v

Quality Metrics:

• ✅ 52/52 tests passing (100% success rate)
• ✅ Zero meta commentary leakage
• ✅ 100% structured output compliance
• ✅ Enterprise resilience patterns validated

🤝 Contributing

We welcome contributions! Please see our Contributing Guide.

Development Setup

git clone https://github.com/taygundogan/cryptotrading-agent
cd cryptotrading-agent
pip install -e .[dev]

# Run tests
python tests/run_quality_battery.py

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Built with production-grade LLM hardening techniques
• Inspired by enterprise trading system architectures
• Powered by Fireworks AI, OpenAI, and crypto data providers

📞 Support

• Documentation: GitHub README
• Issues: GitHub Issues
• Discussions: GitHub Discussions

---

⚠️ Disclaimer: This tool is for educational and informational purposes only. Cryptocurrency trading involves significant risk. Always do your own research and consider your risk tolerance before making trading decisions.