This Python implementation offers a state-based agent orchestration system leveraging the Model Context Protocol (MCP) to streamline LLM interactions. It allows for the creation of resources, tools, and prompts, defining clear agent states like planning, researching, and executing tasks. With this tool, you can customize agent behavior, add new states, define custom tools, and integrate with Claude for Desktop to create sophisticated, context-aware AI workflows.
주요 기능
01State-based agent orchestration
020 GitHub stars
03Integration with Claude for Desktop
04Model Context Protocol (MCP) integration
05Tool creation with the @mcp.tool() decorator
06Customizable agent states and transitions