LLAMOSC: LLM Agent-Based Modeling for Open Source Communities

A framework that pairs Large Language Models with agent-based modeling to simulate and sustain open-source communities.

active 2024 –present
Artificial Intelligence

LLAMOSC (LLM-Powered Agent-Based Model for Open Source Communities) is a framework for simulating open-source community dynamics using LLM-driven agents. It integrates AutoCodeRover for code-aware behavior, models the full pull-request lifecycle between contributor and maintainer agents, and supports both authoritarian (benevolent dictator) and decentralized (meritocratic) governance algorithms. The framework was developed during GSoC ‘24 under OREL/INCF and builds on prior GSoC ‘22 and ‘23 work.