[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-82404-en":3,"doc-seo-82404-105":29,"detail-sidebar-cat-0-en-105":90},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},82404,1099514068365,"Aurelia","https://ap-avatar.wpscdn.com/avatar/10000253d8d9f28188e?_k=1776742907772140068",8,"Research & Report","Writing Bug Reports for Software Repair Agents: What Information Matters Most","Agentic-first software repair workflows assign AI agents the initial task of generating fixes from issue reports, turning those reports into the agent’s primary specification rather than background documentation for humans. This work studies which bug-report information types make an issue “agent-ready” for accurate patch generation. Using SWE-bench Verified, the study classifies bug reports, annotates each sentence by information type, runs mini-swe-agent across multiple LLM backbones, and fits a binomial regression to quantify impact on success probability.","Writing Bug Reports for Software Repair Agents:  \nWhat Information Matters Most?  \nVincenzo Luigi Bruno∗†, Alessandro Giagnorio∗†, Daniele Bifolco∗‡, Leon Wienges†, Massimiliano Di Penta‡, and Gabriele Bavota†  \n† USI Universita` della Svizzera italiana, Lugano, Switzerland  \nEmail: {vincenzo.luigi.bruno, alessandro.giagnorio, leon.wienges, [gabriele.bavota](gabriele.bavota}@usi.ch)[}](gabriele.bavota}@usi.ch)[@usi.ch](gabriele.bavota}@usi.ch)  \n‡ University of Sannio, Benevento, Italy  \nEmail: [d.bifolco@studenti.unisannio.it](d.bifolco@studenti.unisannio.it), [dipenta@unisannio.it](dipenta@unisannio.it)  \n∗ These authors contributed equally to this research.  \narXiv :2607 .09553v 1 [ cs . SE] 10 Jul 2026  \nAbstract—Software development is increasingly moving toward agentic-first workflows. This includes AI agents responsible for generating initial fixes for submitted issue reports. In this setting, issue reports are no longer merely documentation for human maintainers; they become the primary task specification for the agent. However, little is known about how such reports should be written to maximize the agent’s chances of producing a correct fix. We study what makes a bug report agent-ready. Starting from the SWE-bench Verified benchmark (i.e., a collection of 500 real repository issues with human-written gold patches and test suites for evaluating generated fixes) we manually classify each issue by change type (e.g., bug fix vs refactoring) and annotate each sentence with its information type, such as observed behavior, expected behavior, reproduction steps, localization cues, and suggested fixes. We focus on the 441 issues representing bug reports, and we run on them mini-swe-agent using three LLM backbones (i.e., GPT-5-mini, MiniMax M2.5, and Gemini 3 Flash). We then fit a binomial regression model to estimate the incremental association between each information type and agent success, controlling for confounding factors. Our results suggest that agentic-first reports benefit most from information that narrows the agent’s search and repair space. Localization cues, such as references to affected code areas, are positively associated with successful repairs, while suggested fixes, expressed either in code or natural language, show some of the strongest positive associations with pass probability. An ablation study removing selected information types confirms that agents benefit less from information traditionally useful to humans, such as reproduction steps, and more from sentences that expose a repair direction, either through bug localization or a suggested fix.  \nIndex Terms—Agentic-first issues, Empirical Study  \nI. INTRODUCTION  \nSoftware development workflows are increasingly incorporating LLM-based agents that can inspect repositories, reason over issues, modify code, and propose patches [1]–[3] . In these workflows, the human developer is no longer always the first actor. For example, in the context of bug fixing, a common interaction pattern is an agentic-first approach: an issue is assigned to an AI agent, the agent attempts to generate a patch, and the human developer reviews the result before accepting, modifying, or rejecting it. This shift changes the role of issue reports. So far, bug reports have been written primarily for human maintainers. A useful report helps a developer understand what went wrong, reproduce the problem, locate the relevant code, and reason about a possible fix [4] . In an  \nagentic-first workflow, however, the issue report also becomes the task specification for an automated repair process.  \nThe agent receives the issue as its initial description of the problem and must decide where to look, what behavior to reproduce, and how to construct a patch. Therefore, the practical question is no longer only what makes a bug report useful for humans [4], but also what makes a bug report effective for agents.  \nThe distinction between human-oriented and agent-oriented issue quality is releva","cbCaibcaLcknRzKV","https://ap.wps.com/l/cbCaibcaLcknRzKV","pdf",238190,1,13,"English","en",105,"# Introduction\n## Human- vs agent-oriented issue quality\n## Research objective and evaluation setting\n## Annotation and information-type taxonomy\n## Modeling approach and experimental results","[{\"question\":\"What problem does the paper address in agentic-first software repair workflows?\",\"answer\":\"It examines how to write bug reports so AI repair agents can use them as effective task specifications, not just documentation for human maintainers.\"},{\"question\":\"Which benchmark and dataset are used to evaluate bug-report information types?\",\"answer\":\"The study uses SWE-bench Verified, selecting 441 issues that correspond to defects and excluding 59 tasks related to feature implementations or refactorings.\"},{\"question\":\"What information types are most strongly associated with successful repair by the agents?\",\"answer\":\"Localization cues that narrow the agent’s search space are positively associated with successful repairs, and suggested fixes—whether in code or natural language—show some of the strongest positive associations with pass probability.\"}]",1784180157,33,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":27},"writing-bug-reports-for-software-repair-agents-what-information-matters-most","",{"@graph":35,"@context":84},[36,53,67],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/writing-bug-reports-for-software-repair-agents-what-information-matters-most/82404/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"What problem does the paper address in agentic-first software repair workflows?","Question",{"text":74,"@type":75},"It examines how to write bug reports so AI repair agents can use them as effective task specifications, not just documentation for human maintainers.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"Which benchmark and dataset are used to evaluate bug-report information types?",{"text":79,"@type":75},"The study uses SWE-bench Verified, selecting 441 issues that correspond to defects and excluding 59 tasks related to feature implementations or refactorings.",{"name":81,"@type":72,"acceptedAnswer":82},"What information types are most strongly associated with successful repair by the agents?",{"text":83,"@type":75},"Localization cues that narrow the agent’s search space are positively associated with successful repairs, and suggested fixes—whether in code or natural language—show some of the strongest positive associations with pass probability.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,119,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":105,"slug":137},19,"General","general"]