LLM Integration

About

Streamlines the integration of large language models within the XSky framework, supporting a wide range of providers including OpenAI, Anthropic, and Google. It provides robust mechanisms for model configuration, real-time streaming response handling, and Human-in-the-Loop interactions. By offering strategies for token optimization and conversation compression, this skill ensures that agentic workflows remain cost-effective and performant while facilitating the easy addition of new AI SDK adapters.

Key Features

• Multi-provider support for OpenAI, Anthropic, Google, and AWS Bedrock
• 0 GitHub stars
• Human-in-the-Loop patterns for confirmation, input, and selection dialogs
• Advanced streaming callback management for text, tool usage, and agent results
• Extensible provider factory system for adding custom AI SDK adapters
• Automated token optimization through conversation compression and model tiering

Use Cases

• Optimizing application performance and cost by managing token usage and model selection
• Implementing interactive AI workflows that require human validation or feedback
• Configuring specialized models for different agent tasks like planning or memory compression