Pith. sign in

REVIEW 27 cited by

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

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2507.23348 v1 pith:BLSOYEZT submitted 2025-07-31 cs.SE cs.CLcs.LG

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

classification cs.SE cs.CLcs.LG
keywords issueagentsswe-debateconsolidateddebatereasoningresolutionachieves
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original 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.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 27 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Formal Security Analysis of Agent Protocol Composition

    cs.CR 2026-06 unverdicted novelty 7.0

    AgentThread analyzes five agent protocols with formal TLA+ invariants and SDK tests, reporting 35 specification findings, 80 implementation tests, 30 composition-only failures, and a cross-protocol responsibility gap ...

  2. Dockerless: Environment-Free Program Verifier for Coding Agents

    cs.SE 2026-06 unverdicted novelty 7.0

    Dockerless uses agentic repository exploration to verify patches without execution, enabling SFT and RL training of coding agents that reach 62.0/50.0/35.2% resolve rates on SWE-bench Verified/Multilingual/Pro while m...

  3. LLM Agents Can See Code Repositories

    cs.SE 2026-06 unverdicted novelty 7.0

    Visual graphs of repository structure added to text inputs for multimodal LLM agents reduce token consumption by up to 26% while maintaining or improving issue-resolution accuracy.

  4. SWE-Explore: Benchmarking How Coding Agents Explore Repositories

    cs.SE 2026-06 unverdicted novelty 7.0

    SWE-Explore is a new benchmark evaluating repository exploration by coding agents on 848 issues across 203 repositories, using line-level ground truth from successful agent trajectories and showing agentic methods out...

  5. EquiMem: Calibrating Shared Memory in Multi-Agent Debate via Game-Theoretic Equilibrium

    cs.AI 2026-05 unverdicted novelty 7.0

    EquiMem calibrates shared memory in multi-agent debate by computing a game-theoretic equilibrium from agent queries and paths, outperforming heuristics and LLM validators across benchmarks while remaining robust to ad...

  6. Refute-or-Promote: An Adversarial Stage-Gated Multi-Agent Review Methodology for High-Precision LLM-Assisted Defect Discovery

    cs.CR 2026-04 unverdicted novelty 7.0

    Refute-or-Promote applies adversarial multi-agent review with kill gates and empirical verification to filter LLM defect candidates, killing 79-83% before disclosure and yielding 4 CVEs plus multiple accepted fixes ac...

  7. Dynamic analysis enhances issue resolution

    cs.SE 2026-03 conditional novelty 7.0

    DAIRA integrates dynamic tracing into LLM agents to achieve 79.4% resolution rate on SWE-bench Verified for code defect repair.

  8. CodeOCR: On the Effectiveness of Vision Language Models in Code Understanding

    cs.CL 2026-02 unverdicted novelty 7.0

    Multimodal LLMs process code as images to achieve up to 8x token compression, with visual cues like syntax highlighting aiding tasks and clone detection remaining resilient or even improving under compression.

  9. In Line with Context: Repository-Level Code Generation via Context Inlining

    cs.SE 2026-01 unverdicted novelty 7.0

    InlineCoder reframes repository-level code generation as function-level coding by using a draft anchor to inline the target function into its call graph for upstream usage and downstream dependency context.

  10. SWE-QA: Can Language Models Answer Repository-level Code Questions?

    cs.CL 2025-09 unverdicted novelty 7.0

    SWE-QA creates a new repository-level code QA benchmark with 576 pairs and an agentic LLM framework, showing promise but open challenges for models handling complex codebases.

  11. Repository-Aware Metamorphic Relation Generation for Augmented Reality Applications using Large Language Models

    cs.SE 2026-07 conditional novelty 6.0

    A context-aware LLM pipeline generates metamorphic relations for 142 AR repositories; hierarchical context plus agentic deliberation yields 3,760 refined MRs that human raters judge mostly valid and testable, with 5 c...

  12. PhoenixRepair: Rethinking Repair Strategy Exploration in Software Agents

    cs.AI 2026-07 conditional novelty 6.0

    A multi-agent repair framework that samples multiple edit locations and iteratively reflects on patch attempts reaches 76.0% Pass@1 on SWE-bench-Verified, up to a 7.8% relative gain over SWE-agent.

  13. Know Before Fix: QA-Driven Repository Knowledge Acquisition for Software Issue Resolution

    cs.SE 2026-07 conditional novelty 6.0

    QA-driven pre-repair knowledge acquisition (Questioner + Answerer, then Resolver) raises SWE-bench Verified Pass@1 by up to 4.4 points over Mini-SWE-Agent and other pre-repair methods.

  14. SHERLOC: Structured Diagnostic Localization for Code Repair Agents

    cs.CL 2026-06 unverdicted novelty 6.0

    SHERLOC introduces a training-free structured localization method that achieves SOTA accuracy on SWE-Bench and boosts repair agent resolve rates by 5.95 pp while reducing token use.

  15. LLM Agents Can See Code Repositories

    cs.SE 2026-06 conditional novelty 6.0

    Adding visual dependency-graph images to a text-based coding agent cuts token consumption by up to 26% while keeping issue-resolution accuracy roughly unchanged.

  16. Code Is More Than Text: Uncertainty Estimation for Code Generation

    cs.CL 2026-06 unverdicted novelty 6.0

    Three code-specific uncertainty axes (lexical, algorithmic, functional) yield an ensemble that raises average AUROC from 0.696 to 0.776 across five code LLMs, with one single-pass signal matching multi-pass baselines ...

  17. When Helping Hurts and How to Fix It: Multi-Agent Debate for Data Cleaning

    cs.AI 2026-06 unverdicted novelty 6.0

    Multi-agent debate degrades generation but boosts error detection in data cleaning; a derived benefit condition predicts outcomes across tasks and generalizes to other domains.

  18. From Soliloquy to Agora: Memory-Enhanced LLM Agents with Decentralized Debate for Optimization Modeling

    math.OC 2026-04 unverdicted novelty 6.0

    Agora-Opt uses decentralized debate among LLM agent teams plus a read-write memory bank to produce more accurate optimization models from text than prior LLM methods.

  19. GALA: Multimodal Graph Alignment for Bug Localization in Automated Program Repair

    cs.SE 2026-04 unverdicted novelty 6.0

    GALA uses hierarchical graph alignment between UI screenshots and code structures to achieve state-of-the-art bug localization in multimodal automated program repair on SWE-bench.

  20. Beyond Fixed Tests: Repository-Level Issue Resolution as Coevolution of Code and Behavioral Constraints

    cs.SE 2026-04 unverdicted novelty 6.0

    Agent-CoEvo is a multi-agent LLM framework that coevolves code patches and test patches to resolve repository-level issues, outperforming fixed-test baselines on SWE-bench Lite and SWT-bench Lite.

  21. Dynamic analysis enhances issue resolution

    cs.SE 2026-03 unverdicted novelty 6.0

    Embedding dynamic analysis into an LLM repair agent yields a claimed 79.4% resolution rate on SWE-bench Verified while cutting token use by about 25%.

  22. GlimpRouter: Efficient Collaborative Inference by Glimpsing One Token of Thoughts

    cs.AI 2026-01 unverdicted novelty 6.0

    GlimpRouter uses the entropy of the first token in each reasoning step to decide whether to invoke a large model, yielding 10.7% higher accuracy and 25.9% lower latency than a standalone large model on AIME25.

  23. Free-MAD: Consensus-Free Multi-Agent Debate

    cs.AI 2025-09 conditional novelty 6.0

    Free-MAD picks the winning answer by scoring the full trajectory of agents' answers across debate rounds, beating majority voting with fewer rounds.

  24. SWE-MeM: Learning Adaptive Memory Management for Long-Horizon Coding Agents

    cs.SE 2026-06 unverdicted novelty 5.0

    SWE-MeM introduces adaptive memory management for coding agents via synthesized trajectories and Memory-aware GRPO, reporting 43.4% and 60.2% resolve rates on SWE-Bench Verified for 4B and 30B models while beating bas...

  25. Code as Agent Harness

    cs.CL 2026-05 accept novelty 5.0

    A survey that organizes existing work on LLM-based agents around code as the central harness, structured in three layers of interfaces, mechanisms, and multi-agent scaling, with applications across domains and listed ...

  26. Detection Time Distribution Predicted Using Absorbing Boundary Conditions and Imaginary Potentials

    quant-ph 2026-03 unverdicted novelty 5.0

    Three inequivalent quantum detection-time proposals yield waveguide arrival distributions showing partial reflection, spin-orientation independence (for tested parameters), and width dependence when spin couples to th...

  27. Readability-Robust Code Summarization via Meta Curriculum Learning

    cs.SE 2026-01 conditional novelty 4.0

    A meta-curriculum fine-tuning method improves summary quality on obfuscated Python code while slightly improving quality on clean code.