Metraipolis › archive › dc_54247285dfb4
CuraLoc: Repository-Level Fault Localization via Planned Code Inspection — arXiv:2610.31809
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.31809
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-26T20:11:22.857Z
- SHA-256 of the archived text
408306f10a3cba9f04ca3d3938d47c16549374021cc239f25a797e612c789dd9- Stable URL of this page
https://metraipolis.com/documento/dc_54247285dfb4
How to cite
CuraLoc: Repository-Level Fault Localization via Planned Code Inspection — arXiv:2610.31809 — arXiv (en), https://arxiv.org/abs/2610.31809 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_54247285dfb4 (Metraipolis archive dc_54247285dfb4, sha256 408306f10a3cba9f04ca3d3938d47c16549374021cc239f25a797e612c789dd9)
BibTeX
@misc{dc54247285dfb4,
title = {CuraLoc: Repository-Level Fault Localization via Planned Code Inspection — arXiv:2610.31809},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.31809},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_54247285dfb4; sha256 408306f10a3cba9f04ca3d3938d47c16549374021cc239f25a797e612c789dd9},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.31809. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
A bug report describes what went wrong, but rarely where the code should change. Existing repository-level localization methods often let language models iteratively decide what code to inspect next, requiring repeated model calls and substantial token usage. We investigate whether these inspection decisions can instead be planned before source retrieval. In this research, we introduce CuraLoc, which first selects candidate files and represents their symbols using compact behavioral and structural descriptions. In a single planning decision, the model jointly selects starting symbols, chooses code relationships to inspect from each, and forms hypotheses about likely edits. Deterministic retrieval collects the requested source for a final ranking informed by those hypotheses, concentrating model reasoning on choosing and comparing code rather than directing each retrieval step. We demonstrate the efficacy of our proposed approach on the SWE-Bench Lite and Verified datasets. On SWE Bench Verified with GPT-5.2, CuraLoc raises Func@1 from LocAgent's 53.8% to 59.3% while using 90.3% fewer recorded tokens, and also outperforms RepoSearcher (56.2%) and Agentless (39.6%) on Func@1. On SWE-bench Lite, CuraLoc achieves a high 83.8% Func@5 accuracy improving on Agentless by more than 15% in relative accuracy. Additionally, CuraLoc supports different model backbones, with strong localization accuracy demonstrated using both GPT-5.2 and GPT-6 Astra. These results provide evidence that repository-level code inspection can be planned effectively without iterative model-controlled navigation. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Software Engineering (cs.SE). Pagina del lavoro: https://arxiv.org/abs/2610.31809 — 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 2026-10-26T19:49:03.360Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).