MCP Prompts Examples

This repository contains three examples demonstrating different approaches to building MCP (Model Context Protocol) servers that manage and serve prompts.

Examples

1. Simple (/simple)

A basic MCP server that demonstrates the fundamentals:

• Simple prompt storage and retrieval
• Basic MCP server setup with stdio communication
• Template variable substitution
• Good starting point for understanding MCP concepts

2. Prompt Management (/prompt-management)

A full-featured prompt management system with CRUD operations:

• Add, update, delete, and list prompts
• Template support with variable validation
• Tag-based filtering and organization
• JSON-based prompt definitions with metadata
• Complete prompt lifecycle management

3. Dynamic Prompts (/dynamic-prompts)

Advanced example with dynamic content generation:

• Resource embedding: Prompts that read and include external files/logs
• Multi-step workflows: Pre-defined conversation sequences
• Context-aware prompts: Different behaviors based on input parameters
• File system integration: Safe file reading with sandboxing
• Includes special prompts like analyzer_001 (log/code analysis) and debug_workflow_001 (error debugging workflow)

Getting Started

Each example folder contains:

• server.py - The MCP server implementation
• client.py - A demo client that connects to the server
• readme.md - Detailed documentation for that specific example

To run any example:

cd <example-folder>
python client.py server.py

Choose the example that best fits your use case:

• Simple: Learning MCP basics
• Prompt Management: Building a prompt library/database
• Dynamic Prompts: Creating intelligent, context-aware prompt systems