@scarlet-mesh/mcp-labs

Red Hat Labs Finder MCP Server

Overview

This MCP (Model Context Protocol) server, lab-finder, provides access to Red Hat's comprehensive collection of labs, tools, guides, configurations, and learning resources. It helps users discover relevant Red Hat resources based on their specific needs and queries, making it easier to find appropriate learning materials, troubleshooting guides, and configuration tools.

Features

• Resource Discovery: Search and discover Red Hat labs, tools, and guides based on user queries
• Semantic Matching: Intelligent matching of user queries with available Red Hat resources
• Comprehensive Database: Access to over 500 Red Hat product experience applications and resources
• Multi-language Support: Supports resources in multiple languages (primarily English)
• Real-time Data: Fetches up-to-date information from Red Hat's GraphQL API

Tools

The server provides the following MCP tool:

• find-lab: Get recommendations for Red Hat labs, tools, guides, configurations, or resources based on user needs

Data Source

The server fetches lab and resource data from Red Hat's GraphQL endpoint:

https://graphql.redhat.com

The server queries the product_experience_apps collection to retrieve comprehensive information about available Red Hat resources.

Setup

Prerequisites

• Node.js
• Internet connection (for accessing Red Hat's GraphQL API)
• An MCP client (e.g., a compatible AI tool or application)

Installation

• Clone this repository or navigate to the labs package:

• Install the dependencies:

  npm install
  

Running the Server

• Start the server:

  npm start
  

  The server will listen for MCP requests via standard input/output.

Usage

To use the server, send MCP requests from a compatible client. Here's an example request:

• Find Red Hat resources:

  {
    "tool_name": "find-lab",
    "input": {
      "query": "OpenShift container security best practices"
    }
  }
  

Example queries that work well with this server:

• "Kubernetes troubleshooting guide"
• "Red Hat Enterprise Linux configuration"
• "Ansible automation tutorials"
• "Container security scanning"
• "OpenShift deployment strategies"
• "RHEL system administration"

Refer to the MCP client documentation for instructions on sending requests and handling responses.

Data Structure

The server uses the following data structures:

interface Lab {
  id: string;
  name: string;
  description: string;
  url: string;
  lang: string;
}

Search Algorithm

The server implements an intelligent scoring algorithm that:

• Tokenizes queries: Splits user queries into meaningful words (filtering out words shorter than 3 characters)
• Multi-field matching: Searches across resource names, descriptions, and IDs
• Weighted scoring: Gives higher scores to matches in resource names (2x weight) compared to descriptions (1x weight)
• Ranked results: Returns the best matching resource based on calculated relevance scores

Supported Query Types

The server can help with queries related to:

• Learning Materials: Tutorials, guides, and educational content
• Configuration Guides: Step-by-step configuration instructions
• Troubleshooting Help: Problem-solving and debugging resources
• Security Tools: Security scanning and hardening guides
• Automation Tools: Ansible playbooks and automation resources
• Platform-specific Resources: RHEL, OpenShift, Fedora, and other Red Hat products

Response Format

When a matching resource is found, the server returns:

• Name: The official name of the resource
• Description: Detailed description of what the resource provides
• URL: Direct link to access the resource

If no matching resources are found, the server provides helpful guidance on rephrasing queries or trying different topics.