Resolving Java Code Repository Issues with iSWE Agent

arXiv cs.AI / 3/13/2026

📰 NewsIdeas & Deep AnalysisTools & Practical UsageModels & Research

共有:

Key Points

iSWE Agent introduces a Java-focused automated issue resolver with two sub-agents (localization and editing) and uses novel rule-based Java static analysis and transformation tools.
It targets the gap that most prior AI-based issue resolution work is Python-centric, demonstrating state-of-the-art performance on Java benchmarks (Multi-SWE-bench and SWE-PolyBench).
The approach combines rule-based static analysis with model-based techniques to improve enterprise software development and extend automated issue resolution to Java.
The work highlights potential productivity gains for software teams by accelerating issue localization and automated code edits in Java-enabled projects.

Abstract

Resolving issues on code repositories is an important part of software engineering. Various recent systems automatically resolve issues using large language models and agents, often with impressive performance. Unfortunately, most of these models and agents focus primarily on Python, and their performance on other programming languages is lower. In particular, a lot of enterprise software is written in Java, yet automated issue resolution for Java is under-explored. This paper introduces iSWE Agent, an automated issue resolver with an emphasis on Java. It consists of two sub-agents, one for localization and the other for editing. Both have access to novel tools based on rule-based Java static analysis and transformation. Using this approach, iSWE achieves state-of-the-art issue resolution rates across the Java splits of both Multi-SWE-bench and SWE-PolyBench. More generally, we hope that by combining the best of rule-based and model-based techniques, this paper contributes towards improving enterprise software development.