MCP-Server

# Role-Specific Context MCP Server A Model Context Protocol (MCP) server that defines and governs contextual boundaries based on agent roles in your system. ## Overview This MCP server enables role-based context management for AI agents, allowing you to: - Establish clear instructions, objectives, and domain knowledge for each AI agent (Marketing Expert, Songwriter, Executive Assistant, etc.) - Keep role-relevant memory partitioned and scoped, preventing cross-contamination between different agent roles - Adapt tone and style dynamically (serious, witty, sarcastic) per role, with tone profiles baked into the prompt ## Features ### Role Management - Create, update, and delete custom roles - Pre-defined roles with specific expertise domains - Role-specific system prompts and instructions - Customizable tone profiles ### Memory Management - Role-specific memory storage - Memory retrieval based on relevance to current query - Time-to-live (TTL) for memories - Memory limits per role ### MCP Integration - Exposes roles as MCP resources - Provides tools for role management and query processing - Offers prompts for role-based interactions ## Getting Started ### Prerequisites - Node.js 18+ - OpenAI API key ### Installation ```bash # Clone the repository git clone https://github.com/yourusername/role-context-mcp.git cd role-context-mcp # Install dependencies npm install # Set up environment variables echo "OPENAI_API_KEY=your_api_key_here" > .env # Build the project npm run build ``` ### Running the Server ```bash # Run the MCP server npm start # Run the HTTP server for testing npm run start:http ``` ### Configuration The server can be configured by modifying `src/config.ts`. Key configuration options include: - Default roles and their properties - Available tone profiles - Memory management settings - OpenAI model selection ## MCP Integration ### Resources The server exposes the following resources: - `role://{roleId}` - Information about a specific role - `role://tones` - Available tone profiles ### Tools The server provides the following tools: - `process-with-role` - Process a query using a specific role - `create-role` - Create a new role - `update-role` - Update an existing role - `delete-role` - Delete a custom role - `change-role-tone` - Change the tone of a role - `store-memory` - Store a memory for a specific role - `clear-role-memories` - Clear all memories for a role ### Prompts The server provides the following prompts: - `role-{roleId}` - Use a specific role to process a request - `create-custom-role` - Create a new custom role ## Example Usage ### Processing a Query with a Role (MCP) ```typescript // Example of using the process-with-role tool const result = await client.executeToolRequest({ name: 'process-with-role', parameters: { roleId: 'marketing-expert', query: 'How can I improve my social media engagement?', customInstructions: 'Focus on B2B strategies' } }); ``` ### Processing a Query with a Role (HTTP API) ```typescript // Example of using the HTTP API const response = await axios.post('http://localhost:3000/process', { roleId: 'marketing-expert', query: 'How can I improve my social media engagement?', customInstructions: 'Focus on B2B strategies' }); console.log(response.data.response); ``` ### Creating a Custom Role ```typescript // Example of using the create-role tool const result = await client.executeToolRequest({ name: 'create-role', parameters: { id: 'tech-writer', name: 'Technical Writer', description: 'Specializes in clear, concise technical documentation', instructions: 'Create documentation that is accessible to both technical and non-technical audiences', domains: ['technical-writing', 'documentation', 'tutorials'], tone: 'technical', systemPrompt: 'You are an experienced technical writer with expertise in creating clear, concise documentation for complex systems.' } }); ``` ## License MIT