OpenHands Dynamic Agent Factory

A powerful extension for OpenHands that provides dynamic agent generation capabilities based on technology keywords. This module automatically creates specialized micro-agents for analyzing different technologies like Python, React, Node.js, and SQL using LLM-powered code generation.

Features

• 🤖 Dynamic Agent Generation: Automatically create specialized micro-agents based on technology keywords
• 🔄 OpenHands Integration: Seamless integration with OpenHands' LLM configuration system
• 🛡️ Security First: Built-in code validation and security checks
• 🎯 Technology-Specific: Pre-configured for Python, React, Node.js, and SQL analysis
• 🔌 Extensible: Easy to add new technology triggers and customizations
• ⚡ Production Ready: Comprehensive error handling and validation

Installation

Prerequisites

• Python 3.7 or higher
• OpenHands framework
• Access to an LLM provider (e.g., OpenAI)

Basic Installation

pip install openhands-dynamic-agent-factory

Technology-Specific Dependencies

Choose the dependencies based on your needs:

# Python analysis support
pip install "openhands-dynamic-agent-factory[python]"

# React analysis support
pip install "openhands-dynamic-agent-factory[react]"

# Node.js analysis support
pip install "openhands-dynamic-agent-factory[node]"

# SQL analysis support
pip install "openhands-dynamic-agent-factory[sql]"

# All technologies
pip install "openhands-dynamic-agent-factory[all]"

Quick Start

Basic Usage

from openhands_dynamic_agent_factory import DynamicAgentFactoryLLM

# Initialize the factory
factory = DynamicAgentFactoryLLM()

# Generate a Python code analyzer
result = factory.run({
    "technology_keyword": "python",
    "options": {
        "analysis_type": "security"
    }
})

# Use the generated agent
if result["agent_class"]:
    agent = result["agent_class"]()
    analysis = agent.run({
        "code_snippet": """
        def process_data(user_input):
            return eval(user_input)  # Security risk!
        """,
        "analysis_type": "security"
    })
    print(analysis)

Configuration

Configure your OpenHands LLM settings in your project:

# config.py
from openhands.config import Config

config = Config({
    "llm": {
        "provider": "openai",
        "model": "gpt-4",
        "api_key": "your-api-key",
        "temperature": 0.7
    }
})

The factory will automatically use these settings from your OpenHands configuration.

Supported Technologies

Python Analysis

• Input: Python code snippets
• Analysis Types:
  • Style (PEP 8 compliance)
  • Security (vulnerability detection)
  • Performance (optimization suggestions)
• Outputs: Detailed analysis report with:
  • Code quality metrics
  • Security vulnerabilities
  • Performance recommendations

React Analysis

• Input: React/JSX components
• Features:
  • Component structure analysis
  • Hooks usage patterns
  • Performance optimization
• Outputs:
  • Best practices compliance
  • Performance bottlenecks
  • Accessibility issues

Node.js Analysis

• Input: Node.js/JavaScript code
• Focus Areas:
  • Security vulnerabilities
  • Async/await patterns
  • Scalability issues
• Outputs:
  • Security audit report
  • Performance metrics
  • Best practices suggestions

SQL Analysis

• Input: SQL queries
• Features:
  • Query optimization
  • Injection prevention
  • Performance analysis
• Outputs:
  • Optimization suggestions
  • Security recommendations
  • Execution plan analysis

Advanced Usage

Custom Technology Triggers

Add your own technology-specific analyzers:

from openhands_dynamic_agent_factory import TRIGGER_MAP, TriggerInfo

# Define a new Java analyzer
TRIGGER_MAP["java"] = TriggerInfo(
    class_name="JavaAnalyzer",
    description="Advanced Java code analyzer",
    inputs=["code_snippet", "analysis_type"],
    outputs=["analysis_report", "suggestions"],
    required_imports=["javalang"],
    validation_rules={
        "max_code_length": 10000,
        "required_fields": ["code_snippet"]
    },
    llm_prompt_template="""
    Create a Java code analyzer that:
    1. Parses the input using javalang
    2. Analyzes for {analysis_type}
    3. Returns detailed suggestions
    """
)

Error Handling

The factory provides comprehensive error handling:

try:
    result = factory.run({
        "technology_keyword": "python",
        "options": {"analysis_type": "security"}
    })
    if result["agent_class"] is None:
        error_info = result["generation_info"]
        print(f"Agent generation failed: {error_info['error']}")
        print(f"Details: {error_info.get('details', 'No additional details')}")
except Exception as e:
    print(f"Unexpected error: {str(e)}")

Project Structure

openhands-dynamic-agent-factory/
├── openhands_dynamic_agent_factory/
│   ├── core/
│   │   ├── factory.py         # Main factory implementation
│   │   ├── triggers.py        # Technology trigger definitions
│   │   └── dynamic_agent_factory_llm.py
│   └── __init__.py
├── examples/
│   └── basic_usage.py         # Usage examples
├── README.md
├── LICENSE
└── pyproject.toml

Contributing

1. Fork the repository
2. Create your feature branch:

   git checkout -b feature/amazing-feature
   

3. Make your changes and commit:

   git commit -m 'Add amazing feature'
   

4. Push to your branch:

   git push origin feature/amazing-feature
   

5. Open a Pull Request

Development Setup

1. Clone the repository
2. Install development dependencies:

   pip install -e ".[dev]"
   

3. Run tests:

   pytest tests/
   

License

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

Acknowledgments

• Built on top of the OpenHands framework
• Powered by state-of-the-art LLM capabilities
• Inspired by the need for intelligent, technology-specific code analysis