Peer Revenue Publish or profit

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Computer Science Software Engineering

cs.SE Watch Software Engineering

Monday 7 September

Software Engineering

Trust the Spec, Not the Code - A Specification-First, AI-Assisted Case Study in Online Banking

Eitan Farchi

Formal specification promises early error detection, explicit invariants, and correctness by design, yet its notational cost has kept it out of mainstream practice. We argue that AI removes much of that cost: natural language enriched with lightweight mathematics, written in \LaTeX, can serve as an intermediate specification language that is precise enough to reason over and prove, while a large language model (LLM) reviews it for ambiguity, drafts proofs, and generates the implementation. The specification becomes the artifact one authors, reviews, proves, and refines; the code becomes regenerable output. This paper is a follow-on to a prior study that established the discipline on an organizational-knowledge-growth simulation~\cite{predecessor2026}. Here we replicate the discipline in a different domain—an online-banking fund-transfer service—and extend it. The two domains share one spine: a conservation invariant (knowledge in the prior study, money here), which suggests the approach generalizes across domains. We contribute: (i) a second, independent case study of the method; (ii) a stress-test of the method on a richer problem—scheduled/recurring transfers—whose generated code grows substantially while the invariant and its proof do not; (iii) an AI-proposed runtime coverage model for invariants (“never violated $\neq$ covered”); and (iv) a Z formalization, including paired success/failure operation schemas and an invariant proved over the inductive set of all reachable configurations, together with an experiment in which the AI proposes the Z interfaces itself. We are explicit about the method's limits: the proofs and runtime checks live at the specification level and do not establish that the generated code refines the specification—that step is delegated to the AI. This is a case study, not a controlled experiment.
Read the full abstract

Software Engineering

Scanning the Harness: Configuration Exposures in AI Coding-Agent Supply Chains

Benjamin Kapner, Carmel Soceanu, Alicia Petrunin et al. (+1 more)

AI coding agents rely on repository instructions, skills, hooks, tool-server declarations, and subagent definitions. These artifacts distribute both behavior and access to executable dependencies, making configuration review part of the agent software supply chain. We study 3,171 public GitHub repositories: 2,660 assembled setups and 511 skill collections. Deterministic analysis, mechanical re-derivation, model-assisted adjudication, human review, and platform documentation checks identify six categories of configuration exposure and conformance issues. Unpinned MCP package declarations occur in 9.8% of setups, broad execution grants in 2.5%, and broad skill tool preapproval in 3.8%. Their union covers 409 setups (15.4%); among setups with MCP configuration, 24.5% contain an unpinned declaration. Including required-field and skill-format issues brings the setup rate to 17.9% and the collection rate to 6.8%. Holding the six categories fixed, contextual review changes the setup rate from 18.3% to 17.9%; rule selection explains most of the reduction from the broader candidate set. The findings identify concrete opportunities to pin dependencies, review execution pre approval, and check component conformance before distribution or use. A documented permission exception also exposes a shared error in the scanner and its mechanical audit, motivating version-specific semantic checks. The study provides reproducible evidence about repository declarations; agreement with human judgments informs label validation, while runtime consequences and recall remain unmeasured.
Read the full abstract

All authors Benjamin Kapner, Carmel Soceanu, Alicia Petrunin, Hofni Gartner

Software Engineering

EnvPilot: Systematic Design and Evaluation of an Experience-Augmented Agent for Software Environment Setup

Hanwu Chen, Hanyu Lin, Zhanjiang Yang et al. (+6 more)

Environment Setup is a critical yet complex task in software engineering that relies heavily on expert knowledge. Existing automated environment setup methods lack the ability to accumulate experience from past execution trajectories and to evolve over time. As a result, their performance is limited because they often perform redundant exploration, ignore useful past solutions, and fail to generalize across diverse software ecosystems. We present the systematic design and empirical validation of EnvPilot, an experience-augmented agent that operationalizes trajectory-derived experience reuse for software environment setup. EnvPilot maintains an expandable Trajectory-Derived Memory (TDM), initialized with 667 high-quality experiences. It systematically transforms implicit knowledge from historical execution trajectories into structured experience and retrieves the most relevant guidance during task execution through the Context-aware Retrieval mechanism. This enables EnvPilot to combine multiple validated setup strategies, providing more precise and detailed guidance than methods that rely solely on static project files or web retrieval. To evaluate EnvPilot, we construct AES-Bench, a multilingual benchmark of 112 real-world GitHub instances across 9 programming languages. Experiments show that EnvPilot achieves a new state-of-the-art (SOTA) with a 75.00% Pass@1 success rate while reducing reasoning costs. Our empirical study shows that both the structured experience representation and the Context-aware Retrieval mechanism are essential.
Read the full abstract

All authors Hanwu Chen, Hanyu Lin, Zhanjiang Yang, Linhao Zhang, Aoyan Li, Jinxi Li, Meng Li, Yin Chen, Daoguang Zan

Software Engineering

When Stakeholder-centric Requirements Engineering is Not Enough: An Action Research Study on Legacy System Modernisation

Ruward S. Karper, Damian A. Tamburri, Alessio Ferrari et al. (+1 more)

Legacy system modernisation is a major challenge in digital transformation, especially when organisations depend on long-lived, business-critical systems that are only partly understood. In such context, organisations must define future needs while determining what current systems actually do and which functions to retain, adapt, or replace. Modernisation is therefore not only a technical challenge but also a requirements engineering (RE) problem, shaped by stakeholder perspectives. This study examines how far stakeholder-centric RE can support gap analysis between the system-as-is and the system-to-be in a legacy modernisation context. We conducted an action research study in a multinational energy company engaged in system modernisation. In the study, we applied stakeholder-centric RE practices, including stakeholder identification, semi-structured elicitation interviews, agreement-building through the Delphi method, and prioritisation with the MoSCoW method. The results show that this process was effective in producing requirements stakeholders generally viewed as understandable and correct, but less effective in achieving agreement on how elicited requirements mapped to legacy system functionality. The findings suggest that stakeholder-centric RE is necessary but not sufficient, pointing to the need for uncertainty-aware, iterative, and evidence-based modernisation practices that combine stakeholder perspectives with manual and tool-assisted analysis of legacy systems.
Read the full abstract

All authors Ruward S. Karper, Damian A. Tamburri, Alessio Ferrari, Willem-Jan van den Heuvel

Software Engineering

Quality Metrics for LLM-Generated Asset Administration Shells: A Perturbation-Based Evaluation Approach

Janek Groß, Elena Zentgraf and Jens Heidrich

The rapid digital transformation of manufacturing, often referred to as Industry 4.0, relies on seamless interoperability between physical and software assets. A central enabler is the Asset Administration Shell (AAS), a standardized digital representation of such assets. Recent advances in large language models (LLMs) enable the generation of AAS submodels from unstructured sources such as product datasheets but raise challenges for quality assurance. In particular, unexpected errors, the lack of ground truth references, and the absence of standardized quality metrics hinder reliable adoption. In this work, we evaluate quality metrics for AI-generated AAS using a perturbation-based evaluation framework. By systematically degrading AAS generation along multiple dimensions, we assess how well different metrics reflect quality changes. Based on a dataset of 200 products from multiple manufacturers, we generate 6,400 AAS instances using GPT-4o-mini, Qwen3, and DeepSeek-R1. Our results show that metrics based on exact matching of property names and similarity-based soft matching of property values, in particular value-based recall and name-based F1 score, provide the most reliable indicators of quality degradation. Furthermore, we quantify the impact of different perturbation types and analyze differences across model families and product segments. These findings support the selection of suitable metrics, the tuning of LLM-based pipelines, and the integration of AI-generated AAS into industrial applications.
Read the full abstract

Software Engineering

One Is Not Enough: The Untold Story of Multiple Security Patches for One Vulnerability

Fangyuan Zhang, Lyuye Zhang, Lingling Fan et al. (+6 more)

Security patches (SPs) are the main mechanism for fixing software vulnerabilities, yet a single vulnerability is not always resolved by a single patch: fixes may be completed incrementally, propagated across maintained branches, or replicated across related repositories. When patch records are incomplete, downstream users may observe only part of the required fix set and therefore apply only partial patching. However, comprehensive patch discovery remains difficult because the prevalence and causes of the multi-SP phenomenon are still poorly understood. In this paper, we present the first large-scale empirical study of multi-SP vulnerabilities. By merging four major vulnerability databases, we construct a dataset of 6,053 multi-SP CVEs with 16,260 SPs, showing that 20.6% of CVEs with patches involve multiple SPs and that merging databases increases recognized multi-SP CVE counts by 36-55% over any single source. We further analyze why a vulnerability is associated with multiple SPs and derive a two-level taxonomy with 6 categories and 16 sub-categories. Based on these findings, we develop SPectre, a taxonomy-driven prototype for comprehensive patch discovery. On 300 multi-SP CVEs, after manually verifying ground-truth SPs, SPectre improves multi-SP patch coverage over representative patch localization baselines, achieving 0.927 recall on same-repository cases and 0.873 recall on cross-repository cases after manual ground truth verification. On 100 recent CVEs recorded as single-patch by all public databases, SPectre further discovers 28 previously unreported SPs across 20 CVEs. Our results show that multi-SP vulnerabilities are both prevalent and systematically underreported, motivating stronger patch-completeness awareness, improved vulnerability database curation, and relation-aware security tooling.
Read the full abstract

All authors Fangyuan Zhang, Lyuye Zhang, Lingling Fan, Chengwei Liu, Yinan Li, Liang Huang, Yang Liu, Zheli Liu, Sen Chen

Software Engineering

Recompilation Is Not Enough: Test-Guided Decompiled-C Repair

Yuhan Huang, Puzhuo Liu and Jianlei Chi

Decompiled C often becomes recompilable only after repair, but recompilation alone does not establish test-observed behavior. A recompiled command-line binary can still parse options incorrectly, print different bytes, or return a different exit status. We present a few-step workflow for repairing decompiled C using compiler feedback and related official tests. Compiler and linker diagnostics first guide build repair. Once the repaired C recompiles into a binary, smoke checks and related official tests expose behavioral discrepancies for semantic repair. In a preliminary static-enriched evaluation on 104 Coreutils 9.5 binaries with available decompiler exports and deterministic exact-output smoke comparisons, 91 binaries (87.5%) recompile and pass the test gate; 9 do not recompile within the repair budget, and 4 recompile but still fail the test gate. The result suggests that test-gate feedback can make LLM-assisted repair of decompiled C more auditable than compile-only recovery.
Read the full abstract

Software Engineering

LatentMD: Benchmarking Markdown Boundary Failures in LLM-Generated Text

Sungjune Lee and Myungjoo Kang

Large language models (LLMs) increasingly generate Markdown that is consumed by renderers, agents, code extractors, and structured downstream pipelines. Yet existing evaluations often conflate content quality with format adherence, leaving Markdown boundary failures under-measured. We introduce LatentMD, a benchmark and evaluation protocol for diagnosing CommonMark-level fence-boundary failures in LLM-generated Markdown. LatentMD separates content correctness from boundary correctness, enabling detection of outputs that are content-correct but boundary-broken. The benchmark contains 4,179 prompts and a CLI for scoring arbitrary model outputs. Across 9 LLMs and roughly 37,600 generations, we find that Markdown boundary failures are widespread: 38.0% of valid main-grid outputs are content-correct but boundary-broken, with substantial boundary breakage under unspecified prompts and in a small human-authored validation set. Ablations show that failures are driven primarily by same-family symmetric-delimiter collisions rather than nesting alone, are only partially mitigated by prompt hints, and generalize to Python triple-quote docstrings while JSON remains robust as an asymmetric-delimiter control. LatentMD provides a reproducible diagnostic target for parser-sensitive LLM evaluation.
Read the full abstract

Sunday 6 September

Software Engineering

ReqEvolve: User-Oriented Software Self-Evolution through Automatic Requirement Interpretation

Md Asif Iqbal Fahim and Alessio Ferrari

The paradigm of software self-evolution enables systems to autonomously extend and reconfigure their own capabilities during execution in response to technical specifications. Yet requests for new functionality often originate from end users and are rarely expressed in technical terms. As a result, developers must translate user needs into technical specifications before the system can evolve, delaying early validation of the requested functionality by preventing users from immediately observing the resulting behaviour. To address this gap, we present ReqEvolve, a runtime code generation system that enables user-driven self-evolution by accepting high-level user requests. The system integrates automatic requirements engineering (RE) and test-driven development (TDD) to transform these requests into executable functionality through clarification, specification decomposition, test generation, and runtime integration. We evaluate ReqEvolve on 72 software evolution cases across 18 projects against two baselines: SpecFix, an RE-focused code generation approach, and an ablation variant of our system. ReqEvolve achieves 89.2% Pass@1, outperforming SpecFix by 18.8% (p < 0.01, r = 0.79, large effect) and the ablation baseline by 32.6% (p < 0.001, r = 0.88, large effect). These results provide initial evidence that user-driven self-evolution is a viable paradigm for autonomously extending software capabilities from user requests, thereby accelerating requirements validation prior to developer verification.
Read the full abstract

Software Engineering

The Double Measurement Confound in Agent Benchmarks: De-Scaffolding, Ground-Truth Scoring, and Reliability Beyond the Mean

Yonghong Zhang, Shadi Motaali, Vu Phong Dinh et al. (+6 more)

Agent benchmarks are increasingly used to compare large language models (LLMs) and guide deployment decisions, yet benchmark scores are meaningful only if they measure model capability rather than properties of the evaluation pipeline. We identify a double measurement confound: execution-critical decisions are performed by a fixed scaffold instead of the model, while the scorer evaluates outputs using criteria that may not reflect task correctness. We unify these issues within a measurement-theoretic framework that characterizes when benchmark scores can be interpreted as evidence of model capability, and instantiate it with an audit-and-repair protocol that (i) transfers execution-critical decisions from the scaffold to the model, (ii) replaces shape-based evaluation with seeded ground-truth scoring, and (iii) reports reliability beyond the mean through worst-case and tail-risk metrics. Experiments on ComtradeBench show that the joint intervention transforms a nearly flat leaderboard into a reliability spectrum that distinguishes both average performance and robustness across seeds. Applying the audit to existing benchmarks further shows that scorer validity is benchmark-specific, whereas scaffold ownership is an uncontrolled axis wherever we probed it. Our results suggest that benchmark scores should be interpreted together with their scaffolding level, scoring criterion, and reliability profile, providing a practical framework for more valid evaluation of LLM agents.
Read the full abstract

All authors Yonghong Zhang, Shadi Motaali, Vu Phong Dinh, Avin Piroutiniya, Jorge E. López de Vergara, Luis de Pedro, Ricardo Correia, Isabel M. Parra, Yong Xie

Software Engineering · Cross-listed from cs.CR

AURA-Eval: Evaluation Framework for Acting Under Risk Awareness in LLM Agent Trajectories

Ruoxi Shang, Christina-Maria Androna, Orfeas Menis Mastromichalakis et al. (+6 more)

LLM agents operate in workflows where unsafe actions can have real consequences. Existing safety evaluations often reduce behavior to a single score, obscuring risk recognition, pre-action detection, and safe task completion when a safe solution exists. We introduce AURA-Eval, a framework combining controlled augmentation with granular diagnosis of behavior in tool-use trajectories. Its pipeline identifies safety-critical decision points, generates controlled variations, and constructs counterparts differing in whether a request has a safe fulfillment path. Using 157 sourced trajectories, we generate 1,249 evaluation items and evaluate 20 frontier and open-weight models. We developed rubrics to classify risk detection, action strategy, and scenario-specific action safety. Our results show that LLM agents engage in unsafe behavior more often when no safe fulfillment path exists. In these cases, frontier proprietary models more often recognize risk and exhibit safer behavior by proposing alternatives, while evaluated open-weight models more often directly execute unsafe requests. Increasing impact or reducing opportunities for oversight before execution also exposes greater vulnerability across models.
Read the full abstract

All authors Ruoxi Shang, Christina-Maria Androna, Orfeas Menis Mastromichalakis, Yu Feng, Aniruddhan Ramesh, Rico Angell, Shang Hong Sim, Chrysoula Zerva, Emmanouil Koukoumidis

Software Engineering

Shortcutting the Fix: Identifying and Categorizing Agentic Exploits in Software Engineering Benchmarks

Nikolai Ludwig, Wasi Uddin Ahmad, Somshubra Majumdar et al. (+1 more)

High software engineering benchmark scores often mask exploitative shortcuts, such as retrieving upstream solutions, exploiting local Git histories, or recalling memorized code. We formalize this phenomenon as agentic shortcutting and audit three state-of-the-art open LLMs on SWE-bench Multilingual and DeepSWE using trajectory-level LLM judges. Standard evaluations trigger shortcutting attempts in 60.7%-90.1% and 55.5%-89.1% of trajectories, respectively, with up to 72.1% of tasks resolved with the aid of a successful exploit; disabling network access fails to curb these high attempt rates. Prompting agents to construct solutions independently suppresses attempts to 9.4%-14.2% and 0.9%-10.3%, respectively, while virtually eliminating shortcut-dependent resolutions and preserving core task performance. Our findings demonstrate the necessity of exploit-aware evaluation frameworks to measure repository-level problem solving.
Read the full abstract

All authors Nikolai Ludwig, Wasi Uddin Ahmad, Somshubra Majumdar, Boris Ginsburg

Software Engineering

CAROL: Context-Aware Online Learning for Fuzzer Scheduling

Zirui Liu, Mengfan Xu, Juan Zhai et al. (+2 more)

Ensemble fuzzing runs multiple fuzzers on a target while a scheduler allocates CPU time among them. Existing schedulers base these decisions on compact summaries of past performance and rules fixed before a campaign. Our measurements reveal two limitations. First, past-reward summaries do not reliably capture performance evolution: after accounting for estimation noise, agreement between consecutive-window rankings is statistically indistinguishable from within-window self-agreement. Second, predictive signals vary across targets: on eight of nine targets, a weighting learned from the other eight predicts reward worse than one learned on the current target. We introduce CAROL, an online scheduler that uses each fuzzer's current context. Already available to the dispatch loop, this context describes reward trends, waiting and plateau time, reached code, and estimation uncertainty. CAROL uses context in two ways: a domain-guided method detects whether a fuzzer is rising or rotting and applies a phase-specific learning rule, while a learned method predicts reward from 15 context signals and uses predictive uncertainty for online selection. Across nine Magma targets, CAROL triggers more unique bugs than each of three ensemble-scheduling baselines whenever their results differ, and fewer on none. Compared with the strongest baseline for each target, CAROL gains 11.8% and surpasses an oracle that retrospectively selects the best single fuzzer per target. Removing context eliminates the gain, and the additional bugs are concentrated among those the baselines trigger rarely or never. Run unchanged on five widely used C++ programs, CAROL finds 120 previously unknown crashing defects, deduplicated by site, fault, and entry point; all were reported to maintainers through the projects' stated disclosure channels.
Read the full abstract

All authors Zirui Liu, Mengfan Xu, Juan Zhai, Shenglong Yao, Shiqing Ma

Software Engineering

An Autonomy Aware Metamodel for Human AI Collaboration in Software Engineering

Md Mahade Hasan, Md Toufique Hasan, Zheying Zhang et al. (+5 more)

Artificial Intelligence (AI) is shifting software engineering from tool-supported processes towards AI-first collaboration, where authority is dynamically distributed across human and artificial actors. However, existing method engineering approaches assume static, human-centric control and provide limited support explicitly capturing evolving autonomy. This paper presents a vision for autonomy-aware method engineering by proposing a metamodel that treats autonomy not as a fixed property of an actor, but as a derived, situation-dependent authority assignment determined by task, context, and collaboration pattern. The metamodel formalizes autonomy through four authority dimensions: task execution, task decomposition, task initiation, and collaboration reconfiguration. Through an analytical instantiation with a multi-agent requirements analysis tool, we illustrate how the metamodel supports dynamic authority assignment. This work provides a conceptual foundation for governance-aware, adaptable, and AI-first software engineering methods.
Read the full abstract

All authors Md Mahade Hasan, Md Toufique Hasan, Zheying Zhang, Kalle Kulonen, Konsta Kalliokoski, François Christophe, Tommi Mikkonen, Pekka Abrahamsson

Software Engineering · Cross-listed from cs.AI

A Unified Policy Architecture (UPA): The Governance Kernel for Enterprise AI Operating Systems

Prabhu Raghav, Balamurugan Pandi, Arul Vivek et al. (+2 more)

Enterprise AI is evolving into an Enterprise Operating System where autonomous AI agents can plan, reason, use memory, invoke tools, execute workflows, and collaborate with other agents. This shift creates a new governance challenge: existing authorization, security, guardrails, and compliance mechanisms are fragmented and are not designed to govern autonomous AI as a unified system. This paper introduces the Unified Policy Architecture (UPA), a governance architecture for Enterprise AI Operating Systems. UPA provides a unified policy model for governing AI and agents, tools, workflows, memory, enterprise resources, and agent-to-agent interactions and enterprise business rules. It extends policy control beyond authorisation to include runtime obligations, human approvals, compliance, audit evidence, and governance evaluation. We present UPA's governance model, declarative policy language foundations, policy evaluation semantics, extensible plugins, industry policy packs, and an evaluation framework for enterprise governance. We also identify extensions for multi-agent coordination, provenance-aware policies, and stateful runtime governance. UPA provides a foundation for building secure, accountable, and governable Enterprise Operating Systems for autonomous AI.
Read the full abstract

All authors Prabhu Raghav, Balamurugan Pandi, Arul Vivek, Shek Mohammed, Sridhar S

Software Engineering

From Reading Code to Reading Spec: A Verified Layer for LLM-Driven Codebase Maintenance

Xinhao Zhang, Jingjie Lu, Kunpeng Liu et al. (+1 more)

The rapid growth of LLM-generated code increases software complexity and the maintenance burden on engineers. While LLMs offer a potential automated alternative, this structural complexity hinders their ability to manage codebases directly. We introduce the Provable Representation Of Original Functionality (PROOF), which manages codebases indirectly via structured specifications. To enable full-lifecycle codebase management strictly through these specifications, PROOF abstracts codebase topology into a hierarchical natural-language representation. To establish absolute trust, the system proves semantic equivalence by reconstructing source code exclusively from this specification. This verified foundation drives maintenance requests, executing code modifications while synchronously updating itself to prevent semantic drift. Experiments on real-world repositories confirm the effectiveness of these specifications.
Read the full abstract

All authors Xinhao Zhang, Jingjie Lu, Kunpeng Liu, Fei Xie

Saturday 5 September

Software Engineering

KG-Commit: A Dynamic Knowledge Graph for Online Just-in-Time Software Defect Prediction

Mohsen Hesamolhokama, Mohammad Sina Beyrami Aghbash, Behnam Rohani et al. (+2 more)

Just-in-time software defect prediction (JIT-SDP) aims to identify risky commits as they arrive and provide developers with timely feedback. This need for low latency has led most approaches to rely on commit-level information and overlook the broader project context in which a change occurs. Incorporating this context is challenging because it requires both efficient retrieval for incoming commits and continual maintenance as the repository evolves. We introduce KG-Commit, a dynamic knowledge graph that incrementally maintains repository history, within-file code structure, and commit semantics as the project evolves. It also uses an AST-delta mechanism to track structural changes between file edits and relies on lightweight graph inference running entirely on CPU. Our evaluation on 11 Apache software projects against six baselines shows that KG-Commit achieves the highest aggregate Macro-F1 (0.704), G-Mean (0.706), and AUC (0.809) using our selected inference pipeline. Under a realistic online protocol, it outperforms LR, HGB, RF, and DeepJIT on all 11 projects, LApredict on 10, and JITLine on 9 projects in Macro-F1, with the aggregate paired difference significant in every case. KG-Commit processes each commit in approximately 1.33~s, with a cost that remains stable as the graph grows and is compatible with commit rates observed in real-world projects. These findings show that rich project context can be efficiently maintained and exploited for online JIT-SDP.
Read the full abstract

All authors Mohsen Hesamolhokama, Mohammad Sina Beyrami Aghbash, Behnam Rohani, Mohammadamin Fazli, Jafar Habibi

Software Engineering

SWE-Test: Benchmarking LLM Vulnerability Discovery via Input Prediction

Yuanxiang Shi, Jiayi Lin, Xuanyong Lin et al. (+8 more)

Vulnerability discovery is becoming an important ability of large language model (LLM) agents: agents that silently miss real defects leave critical software exposed. Rigorously measuring this ability is therefore urgent, but existing benchmarks are gameable through data contamination, score recall against an unknowable vulnerability set, often rely on synthetic bugs, and report a single end-to-end verdict that cannot localize where an agent fails. Vulnerability discovery is a composite ability: an agent must comprehend source code, infer input constraints, construct inputs, execute them, and iteratively correct from feedback. We recast its measurement as an input-prediction task with a closed, deterministic ground truth: using coverage-guided fuzzing, we mine deep target branches in real-world C/C++ programs and ask an agent to predict an input that drives execution to a given branch. This decomposes discovery into three task modes over 22 real-world C/C++ programs spanning 15 domains. Open-loop and Feedback-enabled share 60 fixed-target task instances across 16 of these codebases (13 domains), testing input construction without and with a distance oracle to isolate code comprehension from feedback-driven correction. Online Arena instead removes the predefined target and scores path exploration by coverage gain on a separate, partially overlapping pool of 11 programs; agents collectively confirmed 13 distinct bugs across six programs. Evaluating 15 default-effort model-scaffold configurations, the best reaches only 55.0% pass rate in the Feedback-enabled mode, and the mean across seven paired Claude Code configurations is 36.4% with feedback versus 19.3% without. Decomposing failures, we find constraint inference, not navigation, is the dominant bottleneck. We release SWE-Test with a turnkey evaluation environment.
Read the full abstract

All authors Yuanxiang Shi, Jiayi Lin, Xuanyong Lin, Liangcai Su, Yeheng Duan, Wei Wang, Qi Han, Bing Zhao, Wei Hu, Xander Xu, Chenxiong Qian

Software Engineering

Beyond Lexical Metrics: Sentence-Embedding Detection of Reviewer Habituation in AI Code Review

Haoran Yu, Lifei Liu and Danping Zhang

Code review is a key quality checkpoint between AI-generated code and production. As AI coding agents submit pull requests at scale, it is unclear whether reviewers reduce scrutiny with repeated exposure and whether review comments reveal this change. We study 11,429 reviews from 400 repeat reviewers over 207 days, paired with 10,104 human-authored inline comments from AIDev. Approval rates rise from 30.5% in reviewers' early periods to 36.6% in late periods (Wilcoxon p = 8.6 x 10^-8; Cohen's d = 0.25). However, four hand-crafted linguistic features - lexical diversity, Shannon entropy, technical specificity, and constructive actionability - show no monotonic decline across exposure deciles (all Spearman absolute rho <= 0.53, p >= 0.11; Bonferroni-corrected Mann-Whitney p >= 0.36). A logistic-regression classifier based on these features reaches F1 = 0.485, below the majority-class baseline. Sentence-embedding structure does carry signal: reviewers' late-period comment centroids shift farther from their early-period centroids than under within-reviewer random permutations (Wilcoxon p < 0.001), and a small MLP using three embedding statistics reaches F1 = 0.74 under 5-fold reviewer-stratified cross-validation. Granger analysis shows that approval-rate changes predict later shifts in technical specificity at all tested lags (p < 0.001), while the reverse direction is significant at only one of four lags. Reviewer adaptation is therefore detectable in latent distributional structure rather than in classical lexical metrics, and language shifts follow rather than precede changes in approval behavior.
Read the full abstract

Software Engineering

SkillSpec: Intent-Masked Specification Reasoning for Agent Skill Correctness

Yizhuo Zhang, Bo Kang, Yi Yang et al. (+3 more)

Autonomous agent systems increasingly depend on reusable skill abstractions for consolidating experiential knowledge and domain expertise. These artifacts typically bundle free-form instructions with heterogeneous resources. However, ensuring their correctness remains challenging. Their failure modes transcend conventional code defects to subtle semantic inconsistencies such as intent conflicts, which manifest as silent failures masked by the underlying model. Moreover, skill correctness must be grounded in intended task boundaries and generalizability. We propose SkillSpec, a Hoare-style framework that formulates skill correctness as a specification reasoning problem. It transforms a heterogeneous skill repository into a unified graph representation that aligns descriptions, instructions and code artifacts. For each node, SkillSpec derives an ExpectSpec from the surrounding declared intent, and infers FactSpecs from encoded behavior under partially disclosed intent. An intent mask regulates access to holistic, lineage, neighborhood, and local views to balance the bias introduced by excessive context against unsupported inference caused by insufficient context. SkillSpec jointly reasons over these views to flag candidate defects, and automatically validates them in an isolated sandbox. On 515 real-world skills from SkillsBench and widely downloaded repositories, SkillSpec identified 763 manually confirmed defects across 239 skills, achieving 61.2% precision. The node-level analysis across multiple model families shows that specification reasoning is consistently reliable for code nodes, whereas plain-text nodes remain a major bottleneck. Most defects arise at the boundaries between declared intent and implementation, demonstrating that explicit specifications provide a practical foundation for skill quality assurance in real-world agent ecosystems.
Read the full abstract

All authors Yizhuo Zhang, Bo Kang, Yi Yang, Zhiyu Duan, Zhouteng Ye, Shunkun Yang

Software Engineering

SpecCoder: Specification-Aware Code Generation with Curriculum Dual-Task Reinforcement Learning

Yixuan Li, Mingxuan Huang, Jiajing Wang et al. (+4 more)

Large language models (LLMs) have made substantial progress in code generation but still struggle with challenging programming tasks that require understanding rich natural language requirements. These requirements often specify problem goals, input/output formats, constraints, examples, and edge cases. Overlooking even one may produce executable but functionally incorrect code. Existing training-free methods mainly rely on prompting or agent-based workflows, while training-based methods typically optimize final code outputs. However, existing approaches provide limited supervision for learning the intermediate mapping from raw requirements to structured specifications and for grounding them in concrete implementation behavior. Consequently, models may omit critical constraints, and even when an explicit specification is produced, the implementation may fail to reflect it consistently. Motivated by this gap, we propose SpecCoder, a specification-aware two-stage training framework for code generation. SpecCoder first employs specification-guided SFT to train LLMs to derive structured specification analyses and generate code conditioned on them. It then introduces curriculum dual-task GRPO, which jointly optimizes specification-guided generation and discrimination to encourage stronger correspondence between specifications and code behavior. Experiments on APPS, CodeContests, and xCodeEval demonstrate the effectiveness of specification-aware training, with SpecCoder consistently improving both standalone code generation and agent-based workflows. Additional evaluations on BigCodeBench-Hard and ClassEval, alongside human evaluation and perturbation studies, further validate the role of structured specifications in guiding code generation and discrimination.
Read the full abstract

All authors Yixuan Li, Mingxuan Huang, Jiajing Wang, Weidong Yang, Xinyi Liu, Ben Fei, Lipeng Ma

Software Engineering

SemVul: Semantic-Enhanced Graph Neural Networks for Code Property Graph-based Vulnerability Detection

Younas Affan, Leonardo Regano and Giorgio Giacinto

Vulnerabilities in source code are often the root cause of cyberattacks worldwide, as attackers exploit weaknesses in software to gain unauthorized access, steal data, or disrupt services. In this study, we evaluated existing research approaches and propose SemVul, a vulnerability detection pipeline that demonstrates better generalization and higher accuracy in learning vulnerable code patterns. We propose a Code Property Graph-based vulnerability-detection approach combined with semantic-level enhancement, enabling the model to capture both the program's structural flow and the semantic meaning of the code. Our approach integrates both node-level and edge-level semantic embeddings using pre-trained code embedding techniques. We systematically evaluate multiple GNN architectures on publicly available benchmark datasets. SemVul is generic with respect to the programming language and supports multiple architectures. By integrating structural and semantic information, the proposed approach improves vulnerability detection performance. Our results show that SemVul outperforms existing approaches and provides better generalization.
Read the full abstract

Software Engineering

Replications, Revisions, and Reanalyses: Managing Empirical Evidence in Software Engineering

Julian Frattini, Jannik Fischbach, Davide Fucci et al. (+4 more)

One aspired outcome of empirical research on quantitative data is a variance theory, i.e., a quantification of the effect of an independent on a dependent variables. The validity of variance theories stems from the synthesis of multiple pieces of evidence, which increases its validity beyond the findings of a single study. However, research synthesis in SE is rare and if done mostly limited to purely narrative syntheses. At best, researchers perform meta-analyses to synthesize variance theories from several quantitative results. But even meta-analyses only produce reliable results when synthesizing exact replications yet fail to generalize from variations. We aim to extend the frontier of research synthesis beyond the state-of-the-art to systematically manage empirical evidence and its evolution. We apply method engineering to construct a framework for research synthesis from proven, individual method fragments. The framework allows researchers to put new evidence in a clear relation to an existing body of evidence and systematically expand knowledge about a studied phenomenon. We demonstrate the application of this framework to two fields of research by explicitly modeling the relationship between existing pieces of evidence. The framework puts three types of evolution of evidence into relation: (1) replications investigate the same hypothesis in a new context to improve external validity, (2) revisions challenge an existing hypothesis to improve internal validity, and (3) reanalyses replace analysis methods to improve conclusion validity. Through a systematic evolution of evidence and clear assessment criteria for each dimension of validity, the proposed framework can determine the frontier of a field of research. The framework provides a perspective to systematically evolve empirical evidence in SE, supporting more constructive and productive advances in our field.
Read the full abstract

All authors Julian Frattini, Jannik Fischbach, Davide Fucci, Michael Unterkalmsteiner, Daniel Mendez, Robert Feldt, Richard Torkar

Software Engineering

VibeCheck: Assessing the Quality of LLM-Generated Unit Tests: A Multi-agent Empirical Study across Heterogeneous Repositories

Anika Tabassum, Mushahid Intesum, Md. Fahim Arefin et al. (+1 more)

LLM-based IDE agents are increasingly used to generate repository-grounded unit tests, yet common evaluations often rely on execution success or coverage. These metrics can miss deeper quality issues such as weak assertions, missing edge cases, poor isolation, and limited maintainability. This paper presents VibeCheck, an empirical study of unit test generation across 15 student-developed Python and JavaScript/TypeScript repositories. We evaluate Kiro, Antigravity, and Cursor with Claude Sonnet 4.5 as the underlying agent, under repository-only, zero-shot conditions using a five-dimensional rubric covering runnability, assertion strength, logic and edge-case coverage, isolation/determinism, and maintainability. We also apply leave-one-out cross-agent peer evaluation to compare tools and identify failure patterns. Results show a clear execution-adequacy gap: generated tests are often runnable but frequently lack strong assertions and meaningful behavioral coverage. Weak assertions and missing edge cases occur more often than blocking failures, showing that runnable tests can still be shallow. VibeCheck provides a reliability-oriented framework for evaluating LLM-generated tests beyond pass/fail outcomes.
Read the full abstract

All authors Anika Tabassum, Mushahid Intesum, Md. Fahim Arefin, Tarannum Shaila Zaman

Software Engineering

Broken on Arrival: Silently Defective LLM Artifacts in Public Model Registries and How to Catch Them

Aditi Patodiya

Developers increasingly run large language models locally by pulling quantized GGUF artifacts from public registries, yet nothing in the distribution pipeline functionally tests these conversions before they reach users. We executed 327 quantized code-capable model artifacts: 305 from the official Ollama library, spanning 15 model lines at every eligible quantization level at or under 8 GB, and 22 from the most-downloaded community repositories on HuggingFace. Each ran a 15-task smoke suite calibrated so that healthy artifacts pass while a known-broken one fails; suspects then faced full 164-task evaluation, a second inference backend, an independent distributor's conversion of the same model and quantization as referee, and, for community files, re-testing under the artifact's own template. The official library carries five silently defective artifacts, a batch of four Qwen2.5-Coder-3B conversions and one phi3.5-mini conversion, that solve zero of 164 tasks and zero of the smoke suite on both backends while independent conversions of the same models work: 1.6% of official artifacts, 2 of 29 model-and-size conversion groups. The adjudication chain cleared small-model artifacts that a naive threshold would condemn as broken when they are merely collapsed by extreme quantization, and it exposed two older community conversions that degrade badly on CUDA yet pass on Metal: not defective files but backend-dependent failures, a third phenomenon no registry currently tests for. Two confirmed defects produce output whose surface statistics sit inside the healthy range, invisible to any low-noise heuristic short of execution. We release the audit dataset, the quantcheck acceptance-testing tool, and disclosure reports for every confirmed defect (https://github.com/aditi-p31/quantcheck), and argue that model registries need the acceptance gate that package registries already run.
Read the full abstract

Software Engineering

Correct Tests Are Not Enough: Measuring and Training Oracle Conversion in Specification-Based Test Generation

Yunhao Liang, Chengguang Gan, Ruixuan Ying et al. (+3 more)

Generating tests from a natural-language specification requires both an input that exposes faulty behavior and a correct expected output. These requirements need not improve together: a model can increase test correctness by choosing easier inputs, or discover useful inputs whose expected outputs it cannot predict. We study this interaction through executable reward decomposition and suite-level oracle-conversion measurement. Our generator jointly emits five input–output tests in one response. During training, audited reference programs provide correctness feedback, while a fixed bank of faulty programs provides two utility signals: potential input kill and effective kill after checking the generated output. An additive GRPO objective preserves both signals without requiring execution at inference time. On an audited TC-Bench split with 506 training and 142 evaluation tasks, three independently trained Qwen3.5-9B runs at step 75 increase full-test correctness from 28.59% to 42.54%, input kill from 24.06% to 25.27%, and effective full kill from 12.23% to 14.15%. Matched 50-step ablations reveal a trade-off: removing kill rewards yields higher correctness and slightly higher full kill, but lowers input kill to 21.60%. A fixed-input source–oracle crossover on 64 training-pool tasks attributes the principal NoKill-to-FullKill difference to harder input selection rather than worse output prediction on identical inputs. These results identify oracle conversion as a measurable bottleneck and show the benefits and limits of preserving input-utility feedback in joint test generation.
Read the full abstract

All authors Yunhao Liang, Chengguang Gan, Ruixuan Ying, Hanjun Wei, Zhe Cui, Shiwen Ni

Software Engineering

Diffs vs. Whole Files: An Empirical Comparison of Iterative Edit-Based and Direct Generation for Flutter/Dart Code Models

Andrej Andrejev

Large language models used for code editing can be trained and deployed in at least two output regimes: direct generation, where the model emits the entire modified file in one shot, and iterative diff-based generation ("steps"), where the model emits a sequence of localized search/replace edits applied one at a time until it signals completion or a step budget is exhausted. The diff-based regime is attractive because it mirrors how developers edit code and should require far fewer generated tokens per turn. We train two code models - a 100M-parameter model trained from scratch (Rainbow-Pony-100M) and a fine-tuned Qwen2.5-Coder-0.5B - in both regimes on a shared Flutter/Dart dataset, and evaluate all four resulting models on a held-out set of approx 1,790 tasks per model. Direct generation substantially outperforms diff-based generation on every metric we measure - compilation/static-analysis pass rate, bits-per-byte, character-level similarity to the reference, and blinded LLM-judge ratings of goal fulfillment, correctness, and code quality - and the gap persists after controlling for task difficulty via a matched-ID comparison and when restricting to code that compiles on both sides. We then identify a single, architecture-independent mechanism behind the conditions where diff-based generation does win: it is competitive on short, spatially localized edits, and its category-level wins concentrate in exactly the two task categories - refactoring and error-handling/edge-case fixes - with the lowest mean edit-step count in our dataset. We term this task locality and discuss its implications for when an edit-based training regime is and is not the right choice for a code-editing model.
Read the full abstract

Friday 4 September

Software Engineering

Regret Dominates Surprise: Design-Time Requirements Engineering for Agentic-AI Safety

Nuwayyir Almohammadi, Rami Bahsoon and Tao Chen

Requirements engineers for agentic-AI domains face challenges in evaluating, specifying, and operationalizing safe autonomy. Mainstream frameworks, such as Goal-Oriented Requirements Engineering (GORE), lack mechanisms to systematically address these challenges under epistemic uncertainty. We contribute an approach that builds on GORE to model and simulate safe autonomy in agentic-AI systems. We introduce a novel Regret-Dominance Mechanism (MS-RGR) to operationalize safe autonomy. MS-RGR uses two signals: epistemic surprise (novelty detection) and cognitive regret (evaluative risk) to address the trilemma problem: should the agent operate in routine autonomy, undergo reflective reasoning, or escalate to human? We instantiate MS-RGR in elderly care monitoring and autonomous driving. A 100-seed stochastic simulation shows MS-RGR reduces silent failures to near-zero and detects risk approximately 17.5 times faster than a sensor-only baseline, remaining formally traceable via LTL safety properties. A retrospective proxy instantiation applying the DRI gate post-hoc over execution traces from 208 AGENTHARM scenarios across seven LLMs shows the gate improves harmful-task refusal only for models with strong baseline safety (over 80% pre-gate refusal, e.g., 84.1% to 90.9%), indicating MS-RGR amplifies rather than substitutes for model-level safety training. We discuss threats to validity, positioning MS-RGR as initial feasibility evidence for design-time safety constraints in agentic-AI requirements engineering.
Read the full abstract

Software Engineering · Cross-listed from cs.CL

Who Maintains Agent Skills? A Longitudinal Study of Human-Governed, AI-Assisted Skill Maintenance

Chen Shen and Estevam Hruschka

Lifelong LLM agents increasingly rely on external skill artifacts as one element for preserving and reusing capabilities over time. These skills (usually portable Markdown files such as SKILL.md) describe when and how to apply a capability and must be corrected, expanded, and consolidated as tools and usage patterns shift over deployment. Recent work seeks to automate skill curation, but it largely evaluates against automated baselines and treats human maintenance as an unmeasured bottleneck. We study that missing process directly. We mine the full commit histories of five public AI-skill repositories, a purposive sample of AI-tooling organizations, covering 873 commits, 143 skill files, and 254 substantive post-creation edits from October 2025 to June 2026. We code each edit with pre-registered governance, operation, and trigger-evidence codebooks. Three findings emerge. First, every substantive edit is authored or merged through a named human account, while 62% carry an AI co-author trailer, with large repository-level variation. Second, these edits are genuine curation: an audited sample shows that most change skill content, and the coded operations are dominated by additions and corrections. Third, a pre-registered rule-likeness axis fails its reliability gate; reliably coding rule-likeness from commit artifacts remains an open measurement problem. We release the corpus, codebooks, mining scripts, and a replay protocol for automated skill curators. For self-evolving agents, public skill maintenance currently looks less like an autonomous pipeline than a human-governed, AI-assisted loop that future curators must measure against and operate within.
Read the full abstract

Software Engineering

The Impact of GenAI on the Future of Requirements Engineering

Travis Breaux and Anmol Singhal

Recent advances in artificial intelligence (AI), particularly large language models (LLMs), are transforming how we design and build systems by increasing access to domain knowledge and by providing automation support to software engineering (SE). As implementation becomes less expensive through generalist SE agents, engineering effort shifts away from writing correct code and toward expressing, curating, verifying, and evaluating requirements. In this paper, we survey the state of the art in AI for requirements engineering (RE) research leading up to the transformation, before reviewing advances in LLMs. We survey two subsequent research areas: prompt programming, which treats LLM instructions as a program in SE vernacular, and generalist SE agents, which combine multiple LLM advances to yield semi-autonomous processes that complete SE tasks. Finally, we explore the future of requirements engineering along two axes: matters changing how we interact with requirements through the SE process, and matters changing how requirements are experienced by software developers and stakeholders more broadly, including end-users. This article aims to inform how RE researchers can navigate this transformation in the selection of future research priorities.
Read the full abstract