Skip to content

This project sets up a 6-agent version of the strategy game Avalon using Agentscope. Each agent follows its own strategy, creating an interactive environment for experimenting with decision-making and multi-agent systems.

Notifications You must be signed in to change notification settings

FreYun/Avalon-muti-agent-game

Repository files navigation

Avalon Game Project

This is an implementation of the Avalon board game using AI agents as players. The game features multiple AI agents taking on different roles and interacting through natural language to play the classic social deduction game.

Features

  • Multi-agent gameplay with AI players
  • Natural language interactions between agents
  • Support for all classic Avalon roles:
    • Merlin
    • Percival
    • Morgana
    • Assassin
    • Villagers
  • Configurable game settings
  • Detailed game logs and conversation history

Requirements

  • Python 3.8+

Installation

  1. Clone this repository
  2. Install dependencies:

Modify agent

You can modify agent configurations in modify_agent.py, including roles, prompts, personalities, etc. After making changes, run modify_agent.py to generate the corresponding JSON file that will be used to configure the agent model. Then in avalon.py, change model_configs to point to the corresponding JSON file for the changes to take effect.

Run the game

Run avalon.py to start the game. The game will prompt you to enter the number of players and the number of rounds. After the game starts, the agents will interact with each other through natural language to play the game.

About

This project sets up a 6-agent version of the strategy game Avalon using Agentscope. Each agent follows its own strategy, creating an interactive environment for experimenting with decision-making and multi-agent systems.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages