True Random Outcome Generator MCP Server

This server instance implements the Model Context Protocol (MCP) to expose a mechanism for fetching high-quality, verifiable random numbers using the external service provided by random.org. It is designed to allow clients to request arbitrary-sided random selections.

Core Functionality

Tool: generate_result

This function simulates a generalized random selection process, parameterized by the number of available outcomes.

• Parameter: sides (Integer, optional, defaults to 2).
  • Determines the total count of possible results.
  • Leverages external true random data for selection integrity.
  • Behavioral Mapping:
    • If sides is 2: Output is strictly mapped to "Heads" or "Tails".
    • If sides is 3: Output is mapped to "Heads", "Tails", or an underscore symbol ("_").
    • For any sides greater than 3: The response indicates the specific outcome number, e.g., "It landed on side X".
    • Handles invalid inputs (zero, negative, or non-integer sides) gracefully.

Technical Setup and Deployment

To integrate this server within a local MCP environment, follow these deployment steps:

1. Dependency Acquisition: bash npm install

2. Compilation/Bundling: bash npm run build

3. Continuous Development (Watch Mode): bash npm run watch

Client Integration

When configuring your local client (e.g., Claude Desktop) to recognize this utility, update its configuration file (claude_desktop_config.json on Unix-like systems, or %APPDATA%/Claude/claude_desktop_config.json on Windows) as follows:

{
  "mcpServers": {
    "rng_utility": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-coin-flip"]
    }
  }
}

Interaction Examples

Users can invoke this tool using plain language prompts:

• "Perform a standard coin toss."
• "Simulate rolling a D20 (20-sided die)."
• "Acquire a random integer value within the range [1, 100] using the tool."

The underlying engine guarantees the result fidelity by polling the specified external randomness source.

Diagnostics and Inspection

Debugging asynchronous stdio communication benefits from specialized tools. We suggest using the official utility:

npx @modelcontextprotocol/inspector node build/index.js

Contribution Guidelines

Feedback, feature proposals, and code submissions via Pull Requests are highly valued.

Legal

This software is distributed under the terms of the MIT License.

---

WIKIPEDIA NOTE: The original context included extensive historical and technical documentation regarding XMLHttpRequest (XHR), which has been omitted here to focus purely on the MCP tool's definition and functionality.