Metraipolis › archive › dc_80d70f54fee8
Enforcing Code Reasoning in Agentic Scaffold for More Effective Issue Resolution — arXiv:2610.23782
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.23782
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T09:37:12.159Z
- SHA-256 of the archived text
b27047c25b5b5d8c7eff23942057bf0c8d61edfd690c4fa03e351ebe3b086538- Stable URL of this page
https://metraipolis.com/documento/dc_80d70f54fee8
How to cite
Enforcing Code Reasoning in Agentic Scaffold for More Effective Issue Resolution — arXiv:2610.23782 — arXiv (en), https://arxiv.org/abs/2610.23782 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_80d70f54fee8 (Metraipolis archive dc_80d70f54fee8, sha256 b27047c25b5b5d8c7eff23942057bf0c8d61edfd690c4fa03e351ebe3b086538)
BibTeX
@misc{dc80d70f54fee8,
title = {Enforcing Code Reasoning in Agentic Scaffold for More Effective Issue Resolution — arXiv:2610.23782},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.23782},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_80d70f54fee8; sha256 b27047c25b5b5d8c7eff23942057bf0c8d61edfd690c4fa03e351ebe3b086538},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.23782. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Reasoning is central to ReAct-based programming agents, yet whether and how they reason about code remains unclear. We study 7,955 issue-resolution trajectories from six single- and multi-agent scaffolds across four programming languages on SWE-bench-Verified, SWE-bench-Pro, and SWE-bench-Lite. We identify 12 forms of code reasoning in 66.95% of trajectories. Code reasoning is associated with higher resolution rates, and its intensity increases with task complexity. However, it remains largely unguided: agents may focus on surface-level symptoms during localization, fail to refine near-correct patches, overlook issue requirements, and generate tests with limited behavioral diversity. To address these limitations, we propose DeskCheck, an agentic scaffold that systematically guides code reasoning throughout repository-level issue resolution. DeskCheck simulates execution paths to localize root causes, simulates candidate patches to refine near-misses, audits patches against issue requirements, and reasons about corner-case inputs and expected outputs before generating tests. On 731 SWE-bench-Pro instances, DeskCheck improves resolution rates over mini-SWE-Agent and Claude Code by 6.29% to 10.53%, at an additional cost of $0.02-$0.28 per instance (average: $0.14). Per-phase analyses show higher method-level localization recall, more successful repairs of imprecisely localized issues, and better detection of missed corner cases in near-correct patches. Exact McNemar tests show statistically significant improvements across all subject LLMs (p < 10^-10), while repeated experiments support robustness to agent nondeterminism. These findings motivate systematic code reasoning as a means of improving programming agents. Tema degli esploratori: calcolo e infrastrutture (elenco arXiv cs.SE, nuovi lavori del giorno). Categorie arXiv: Software Engineering (cs.SE). Pagina del lavoro: https://arxiv.org/abs/2610.23782 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-19T09:31:59.711Z dalla pagina https://arxiv.org/list/cs.SE/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).