How do Agents Refactor: An Empirical Study

TL;DR

This study empirically analyzes how software development agents perform Java refactoring, revealing that they mainly focus on annotation changes and may introduce more code smells compared to human developers.

Contribution

First empirical comparison of agentic versus developer refactoring in Java, highlighting differences in refactoring types and impact on code quality.

Findings

Agent refactorings mainly involve annotation changes.

Cursor refactoring increases code smells significantly.

Developers perform more diverse structural refactorings.

Abstract

Software development agents such as Claude Code, GitHub Copilot, Cursor Agent, Devin, and OpenAI Codex are being increasingly integrated into developer workflows. While prior work has evaluated agent capabilities for code completion and task automation, there is little work investigating how these agents perform Java refactoring in practice, the types of changes they make, and their impact on code quality. In this study, we present the first analysis of agentic refactoring pull requests in Java, comparing them to developer refactorings across 86 projects per group. Using RefactoringMiner and DesigniteJava 3.0, we identify refactoring types and detect code smells before and after refactoring commits. Our results show that agent refactorings are dominated by annotation changes (the 5 most common refactoring types done by agents are annotation related), in contrast to the diverse structural improvements typical of developers. Despite these differences in refactoring types, we find Cursor to be the only model to show a statistically significant increase in refactoring smells.