Skip to content

README.md

Latest commit

 

History

History
87 lines (62 loc) · 2.82 KB

README.md

File metadata and controls

87 lines (62 loc) · 2.82 KB

Grid Operations MCP Suite

This project implements a Model Context Protocol (MCP) suite for power grid operations research, enabling Large Language Models (LLMs) to interact with real-time grid data, analytics tools, and visualization resources.

Features

  • MCP Server: Provides grid operations tools for load analysis, outage prediction, and visualization.
  • MCP Client: Connects to the server, manages queries, and handles tool execution.
  • Host Integration: Formats prompts and tool descriptions for LLMs.
  • Model Flexibility: Supports multiple LLM providers via AI Suite integration.
  • Best Practices: Includes guidelines for security, performance, and error handling.

Getting Started

Prerequisites

  • Python 3.8 or higher
  • Required packages:
    • mcp
    • aisuite
    • pandas
    • matplotlib
    • numpy
    • aioconsole
    • python-dotenv
  • Set up environment variables (e.g., API keys) in a .env file.

Running the Server

python grid_ops_server.py

Running the Client

python grid_ops_client.py grid_ops_server.py [initial_model]
  • Replace [initial_model] with your chosen LLM model (e.g., openai:gpt-4o).

Usage

  • Enter operational queries at the prompt.
  • Switch LLM models dynamically using model:provider:model_name.
  • Type context to view the current operational context.
  • Type quit to exit the client.

Example Query

Analyze load patterns in the Northeast region for the last 48 hours.

Best Practices

  • Use descriptive tool names and detailed JSON schemas for clarity.
  • Validate all tool inputs and handle errors gracefully.
  • Enable authentication and HTTPS for production environments.
  • Monitor performance and optimize tool implementations as needed.

Troubleshooting

  • Refer to mcp_troubleshooting.md for solutions to common issues.

Project Structure

  • insecure_demos/mcp/
    • attack-mcp-client.py: Demonstrates a vulnerable MCP client implementation (for demo purposes only).
    • vuln-mcp.py: Contains vulnerable MCP code examples (for demo purposes only).
  • ai_suite_integration.py: Manages integration with LLM providers.
  • grid_ops_client.py: Client for interacting with the MCP server.
  • grid_ops_host.py: Handles LLM host integration.
  • grid_ops_research_example.py: Example script for grid operations research.
  • grid_ops_server.py: Implements the MCP server with grid tools and resources.
  • mcp_best_practices.py: Contains best practices and configuration details.
  • mcp_theory.py: Theoretical background and concepts for MCP.
  • mcp_troubleshooting.md: Troubleshooting guide for common issues.
  • README.md: Project documentation (this file).

License

This project is licensed under the MIT License.

Acknowledgments

  • Inspired by Andrew Ng's AI Suite and the Model Context Protocol (MCP) community.