SWE-Debate: Competitive Multi-Agent Debate for Software Issue Resolution

2025-07-31

AI Summary

Existing agent-based software repair approaches often converge to local optima and struggle to identify cross-module fault patterns. To address this, we propose SWE-Debate: a competitive multi-agent debate framework that constructs multiple fault propagation paths via code dependency graphs and orchestrates specialized agents—each endowed with distinct reasoning perspectives—to engage in three rounds of structured debate. Its core innovations include (1) an MCTS-guided code modification strategy and (2) a phased debate coordination mechanism that jointly foster diverse reasoning through competition and converge toward robust repairs through collaboration. Evaluated on the SWE-bench benchmark, SWE-Debate achieves state-of-the-art performance among open-source agent frameworks, significantly outperforming mainstream baselines. This demonstrates that structured multi-agent debate substantially enhances cross-module fault localization and repair efficacy.

Technology Category

• Multiagent Systems: Agreement, Argumentation & Negotiation
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Issue resolution has made remarkable progress thanks to the advanced reasoning capabilities of large language models (LLMs). Recently, agent-based frameworks such as SWE-agent have further advanced this progress by enabling autonomous, tool-using agents to tackle complex software engineering tasks. While existing agent-based issue resolution approaches are primarily based on agents' independent explorations, they often get stuck in local solutions and fail to identify issue patterns that span across different parts of the codebase. To address this limitation, we propose SWE-Debate, a competitive multi-agent debate framework that encourages diverse reasoning paths and achieves more consolidated issue localization. SWE-Debate first creates multiple fault propagation traces as localization proposals by traversing a code dependency graph. Then, it organizes a three-round debate among specialized agents, each embodying distinct reasoning perspectives along the fault propagation trace. This structured competition enables agents to collaboratively converge on a consolidated fix plan. Finally, this consolidated fix plan is integrated into an MCTS-based code modification agent for patch generation. Experiments on the SWE-bench benchmark show that SWE-Debate achieves new state-of-the-art results in open-source agent frameworks and outperforms baselines by a large margin.

Problem

• Improves issue resolution by addressing local solution stagnation
• Enhances cross-codebase issue pattern identification
• Proposes competitive multi-agent debate for diverse reasoning

Innovation

• Competitive multi-agent debate framework
• Fault propagation traces via code dependency
• MCTS-based code modification agent