Design MCP-driven Refactoring Agents

planningChallenge

Prompt Content

Outline a multi-agent system using the Model Context Protocol (MCP) for a long-horizon code refactoring task. Define the core agents (e.g., Planner, Code Editor, Tester, Memory Manager), their roles, the types of messages they will exchange via MCP, and how GPT-5 (or equivalent) will be integrated into their decision-making and action generation. Explain how a vector database will serve as the system's persistent memory to maintain context across multiple refactoring iterations. Use LangGraph or AutoGen to illustrate the workflow.

Try this prompt

Open the workspace to execute this prompt with free credits, or use your own API keys for unlimited usage.

Usage Tips

Copy the prompt and paste it into your preferred AI tool (Claude, ChatGPT, Gemini)

Customize placeholder values with your specific requirements and context

For best results, provide clear examples and test different variations