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
# 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
# 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 rolerole://tones- Available tone profiles
Tools
The server provides the following tools:
process-with-role- Process a query using a specific rolecreate-role- Create a new roleupdate-role- Update an existing roledelete-role- Delete a custom rolechange-role-tone- Change the tone of a rolestore-memory- Store a memory for a specific roleclear-role-memories- Clear all memories for a role
Prompts
The server provides the following prompts:
role-{roleId}- Use a specific role to process a requestcreate-custom-role- Create a new custom role
Example Usage
Processing a Query with a Role (MCP)
// 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)
// 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
// 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