EvoGit: Decentralized Code Evolution via Git-Based Multi-Agent Collaboration

2025-06-01

AI Summary

This paper addresses the challenge of decentralized collaborative software development by proposing a multi-agent coordination framework that requires no centralized coordination, explicit communication, or shared memory. Methodologically, it pioneers the use of Git as a distributed collaboration infrastructure, modeling code evolution as a lineage graph—enabling LLM-driven coding agents to asynchronously commit changes while automatically inferring historical dependencies and branching relationships. Humans specify high-level goals and perform only lightweight oversight. Key contributions include: (1) elevating version control systems to first-class primitives for multi-agent collaboration; (2) enabling implicit concurrency and fully traceable, auditable code evolution; and (3) supporting autonomous modular growth. Experiments on zero-shot web application construction and self-evolving solver development demonstrate fully automated generation of functionally complete, well-structured software systems—validating the feasibility of sustained, decentralized software development.

Technology Category

• Multiagent Systems: Coordination and Collaboration
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Systems and Infrastructure for Web, Mobile and WoT: Decentralized Web and Fediverse systems

Abstract

We introduce EvoGit, a decentralized multi-agent framework for collaborative software development driven by autonomous code evolution. EvoGit deploys a population of independent coding agents, each proposing edits to a shared codebase without centralized coordination, explicit message passing, or shared memory. Instead, all coordination emerges through a Git-based phylogenetic graph that tracks the full version lineage and enables agents to asynchronously read from and write to the evolving code repository. This graph-based structure supports fine-grained branching, implicit concurrency, and scalable agent interaction while preserving a consistent historical record. Human involvement is minimal but strategic: users define high-level goals, periodically review the graph, and provide lightweight feedback to promote promising directions or prune unproductive ones. Experiments demonstrate EvoGit's ability to autonomously produce functional and modular software artifacts across two real-world tasks: (1) building a web application from scratch using modern frameworks, and (2) constructing a meta-level system that evolves its own language-model-guided solver for the bin-packing optimization problem. Our results underscore EvoGit's potential to establish a new paradigm for decentralized, automated, and continual software development. EvoGit is open-sourced at https://github.com/BillHuang2001/evogit.

Problem

• Decentralized multi-agent framework for autonomous code evolution
• Git-based coordination without centralized control or shared memory
• Autonomous production of functional software with minimal human input

Innovation

• Git-based phylogenetic graph for decentralized coordination
• Autonomous multi-agent code editing without central control
• Human-guided goal setting and feedback integration