Evidence Passport
Agentic refactoring: an empirical study of AI coding agents
Top story 8 in the issue of 24 October 2026, in 38 AI Coding & Developer Tools.
Original source
Agentic refactoring: an empirical study of AI coding agents
Analysing 51,412 refactorings from agent pull requests in open-source Java projects, the authors find agents set out to refactor in about a quarter of commits.
Evidence
Evidence: Peer-reviewed (evidence step 5 of 6)
Published after review by independent experts.
The story in three lines
- Key finding
- Refactoring is a routine, deliberate task for coding agents, appearing as an explicit goal in 26.1% of their commits.
- Significance
- Technical: gives real-world evidence on how agents change code quality, not just whether they pass tests.
- Takeaway
- Teams should review agent refactors as carefully as new features.
Later coverage in the journal
No later coverage in the journal yet.
Corrections and updates
No corrections or updates recorded for this item.
This page is generated from the issue and stays at the same address. It adds nothing that is not in the issue, the source link or the corrections log.