Automated repair agents fix easy bugs by over-engineering them. an issta 2026 study of five repair…
automated repair agents fix easy bugs by over-engineering them. an issta 2026 study of five repair agents across 500 real-world tasks found they excel at simple fixes but stumble on logic-intensive bugs, often with verbose patches.
the harder the bug, the more confident the rewrite.
Context
The paper Understanding Automated Program Repair Agents through the Lens of Traceability: An Empirical Study (arXiv 2506.08311, ISSTA 2026, authors from Columbia University and IBM Research) analyses five repair agents: AutoCodeRover, Agentless, SWE-Agent, OpenHands-CodeAct and MASAI. It reports that agents excel at simple fixes but struggle with logic-intensive bugs, often generating verbose, overfitted patches that pass existing test suites without solving the root cause, and that test generation and regression test selection are bottlenecks. It reports SWE-Agent patches were unnecessarily complex and longer than developer solutions, while Agentless was typically shorter.
The benchmark is SWE-bench Verified with 500 tasks, but the analyses use subsets, a manual review of 50 tasks, an edit-pattern analysis of over 400 tasks and a moderate-tier analysis of 261 issues, so 500 real-world tasks is the benchmark size and not shown as the evaluation size for each agent. Results use GPT-4o and Claude-3.5-Sonnet backbones, so they are bounded to those versions and models. Verbosity differs by agent. No measurement of confidence was found, so the harder the bug, the more confident the rewrite is the author's gloss.
Watch next
- Reruns with newer models and the final published version.
Sources
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 20 September 2026 at 19:05 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →