home / mcp / elrond mcp - thinking augmentation server

Elrond MCP - Thinking Augmentation Server

Experiment in getting LLMs to critique better.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "dogonthehorizon-elrond-mcp": {
      "command": "python",
      "args": [
        "/path/to/elrond_mcp/main.py"
      ],
      "env": {
        "GEMINI_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

Elrond MCP is a Thinking Augmentation Server that analyzes proposals through multiple critique perspectives and then synthesizes actionable insights to support better decision-making. You can run it locally, deploy it for integration with AI assistants, and connect to tools that help you evaluate ideas from positive, neutral, and negative viewpoints.

How to use

You interact with the Elrond MCP server through an MCP client. Start a session to submit a proposal and receive: executive summaries from three critique perspectives, a structured analysis for feasibility, risks, benefits, and implementation, and a synthesized consensus with next steps. Use the client’s consult or analyze flow to submit your proposal, then review the three critique responses and the final synthesis for a comprehensive view.

How to install

Prerequisites: Python 3.13 or higher and a Google Gemini API key.

Setup steps you should follow in order:

1. Clone the repository and enter its directory.

2. Install dependencies using the development setup.

3. Configure your Gemini API key for authentication.

Additional content

Running in development mode lets you test locally with live feedback from the critique and synthesis agents.

# Development mode using uv
uv run python main.py

# Development mode using MCP CLI (if installed)
mcp dev elrond_mcp/server.py

Claude Desktop integration

You can install the Elrond MCP server for Claude Desktop and configure it within Claude’s MCP settings.

mcp install elrond_mcp/server.py --name "Elrond Thinking Augmentation"

Manual Claude Desktop configuration example:

{
  "elrond_mcp": {
    "command": "python",
    "args": ["/path/to/elrond_mcp/main.py"],
    "env": {
      "GEMINI_API_KEY": "your-api-key-here"
    }
  }
}

Using the Tools

Augment Thinking Tool: Analyze any proposal through multi-perspective critique to surface feasibility, risks, and opportunities.

Check System Status Tool: Monitor the health and configuration of the thinking augmentation system to ensure API keys and models are available.

Available tools

Augment Thinking Tool

Analyzes proposals through three critique perspectives to identify feasibility, risks, benefits, implementation considerations, stakeholder impact, and resources, then summarizes key insights.

Check System Status Tool

Verifies configuration, API key accessibility, model availability, and overall system health to ensure reliable operation.