v0.0.6

Release v0.0.6

🚀 New Features

Model Configuration Support

• Added comprehensive model configuration options through the @model decorator
• Introduced fine-grained control over model behavior:
  • Response length control (maxTokens)
  • Response randomness adjustment (temperature)
  • Request retry handling (maxRetries)
  • Conversation step limits (maxSteps)
  • Tool usage control (toolChoice)

🔧 Improvements

Agent System

• Enhanced model configuration handling in agent class
• Improved type safety and default value management
• Better integration with existing agent functionality

Documentation

• Added detailed documentation for @model decorator configuration
• Updated package dependencies
• Included configuration examples and best practices

🧪 Testing

• Added additional generator tests for schema validation
• Improved test coverage for schema caching mechanisms
• Added comprehensive tests for model configuration features

Example Usage

@model('openai:gpt-4', {
maxTokens: 100,
temperature: 0.5,
maxRetries: 3,
maxSteps: 3,
toolChoice: 'auto'
})
class MyAgent extends Agent<string, string> {}

For detailed documentation, visit our GitBook documentation.

📦 Dependencies

• Updated package dependencies to latest compatible versions

---

Full list of changes: Compare with previous version