Reproduce “How do Agents Refactor: An Empirical Study” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Lukas Ottenhof, Daniel Penner, Abram Hindle, Thibaud Lutellier
The authors compare detected refactoring actions and code-smell changes in agent and developer pull requests. They find agent actions are led by annotation edits, while developer actions are more varied and include structural changes. Most agents’ smell-change distributions are not statistically distinguishable from developers’; Cursor differs, but its result rests on seven refactoring commits and does not establish causation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.