reasoning-agents-framework

by ashishkhar · Python · ★ 0

A modular framework for building multi-agent reasoning systems using the [A2A protocol](https://github.com/google/a2a) and [MCP tools](https://modelcontextprotocol.io/).

Install

pip install git+https://github.com/ashishkhar/reasoning-agents-framework.git

Claude Desktop config

Add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "reasoning-agents-framework": {
      "command": "uvx",
      "args": [
        "git+https://github.com/ashishkhar/reasoning-agents-framework.git"
      ]
    }
  }
}

From the README

A **modular framework** for building multi-agent reasoning systems using the [A2A protocol](https://github.com/google/a2a) and [MCP tools](https://modelcontextprotocol.io/). **Give it your data → It reasons and analyzes → Returns verifiable results** | Component | Role | Analogy | |-----------|------|---------| | **A2A Agents** | Think, plan, reason | 🧠 Brain | | **MCP Tools** | Execute, query, validate | 🔧 Hands | | **Manager** | Orchestrate, delegate, synthesize | 👔 Manager | | **Workers** | Specialize in specific domains | 👷 Specialists | **Result**: An AI system that can *think* abo…

Read full README on GitHub →