Papers

Evaluating software engineering agents on realistic benchmarks is costly, since each task may require multi-step code exploration, modification, and test execution. Existing efficient evaluation methods select representative subsets to estimate full-benchmark performance, but are largely result-only: they fit historical pass/fail response matrices or static task semantics, discarding how agents solve problems. We propose PTA-IRT, a Privileged Trajectory-Aware Item Response Theory framework that fuses process and outcome signals. Historical execution trajectories supply process-level evidence beyond pass/fail, such as explored context, attempted edits, and solving paths, which PTA-IRT uses as privileged information for calibration subset selection and ability estimation. Under low calibration budgets, PTA-IRT consistently outperforms prior IRT baselines on score and ranking recovery across four SWE benchmarks. Code and data are publicly available at https://github.com/DeepSoftwareAnalytics/PTA-IRT.

The repository-level code generation task requires synthesizing code that satisfies task requirements while remaining consistent with the target repository context. Since real-world repositories often exceed the input length limits of LLMs, existing approaches commonly adopt retrieval-augmented generation (RAG) to provide repository-specific context. Despite improving repository-context retrieval, existing methods typically provide context as task-level support, without explicitly identifying the critical tokens that require fine-grained repository context during generation. During the autoregressive generation process of LLMs, errors often concentrate at a small number of decisive positions: once such tokens are generated incorrectly, subsequent code may follow an incorrect semantic path and eventually lead to functional failure. We refer to these positions as "critical tokens". In this paper, we propose ACToR, an adaptive critical token-aware retrieval framework for repository-level code generation. ACToR identifies critical tokens during generation and triggers targeted retrieval on demand to provide repository context at these decisive positions. In addition, we design a position-aware weighting method for dense retrievers to prioritize context that is more informative for generation. We evaluate ACToR on two representative repository-level benchmarks, RepoExec and CoderEval. Experimental results show that ACToR consistently outperforms state-of-the-art methods, achieving relative improvements of 8.4% on RepoExec and 15.4% on CoderEval. Beyond performance gains, we systematically quantify the impact of critical tokens, revealing their central role in major generation failures and highlighting the necessity of targeted retrieval strategies. We provide the code and data at https://github.com/DeepSoftwareAnalytics/ACToR.

Learning- and LLM-based software vulnerability analysis is only as trustworthy as the data it is trained and evaluated on, yet that data is rarely examined as a first-class object. We investigate the data behind vulnerability analysis through a dataset-centric taxonomy that separates what an artifact is (code, metadata, patches, tests/PoCs, reasoning, traces), how good it is (realism, label evidence, scale, diversity, leakage, availability), and what it is used for. From a systematically assembled corpus of 1522 papers covering 2016-2026 plus foundational earlier work we deep-code a tiered set of 111 anchor papers, backing every affirmative rubric-graded value with a verbatim span, and we report, per attribute, both how much it has been studied and how well datasets achieve it. The results trace an evidence ladder: executable artifacts are the only major type where 15 of the 24 datasets are both graded real-world and carry labels that received an independent check, while code-sample datasets-the largest category in both the auto-tagged corpus and the anchor set-are the least realistic: 20 of the 41 draw their vulnerabilities from authentic projects or CVEs, but only 3 keep the sample at the unit the code is deployed in, and only 2 do both-though these are coarse component tests, and just one code-sample dataset meets the codebook's stricter full-context real-world grade. Among these, leakage goes unaddressed by 49 of the 90 datasets where it applies, more than a quarter say nothing about availability, reasoning data has arrived only recently and is mostly model-generated, and primary trace corpora remain limited to three datasets, the total after a corpus-wide screen and a full-text check of every candidate it surfaced, with further datasets releasing traces secondarily behind benchmarks and harnesses.

As agents move from research prototypes to deployed tools, their capability increasingly depends on model-external execution infrastructure, commonly termed the agent harness. Changing this harness while holding model weights fixed can substantially alter task performance. Current agent evaluations typically report downstream performance under a chosen harness, leaving a model's ability to develop the harness itself comparatively underexplored. We introduce HarnessDev, a benchmark that shifts the unit of evaluation from task outputs to runnable infrastructure. HarnessDev covers two stages. In Creation, the agent starts from a minimal seed and a small number of cases, then builds a complete execution system. In Evolution, it starts from its own created harness and iteratively revises it using downstream execution feedback, with the goal of improving benchmark performance. We then evaluate each constructed harness on capability (task success on held-out benchmarks) and efficiency (execution-token cost). The reported Creation results cover six creator LLMs, four domains, and five downstream benchmarks totaling 2,207 unique downstream instances, with hidden evaluation tasks withheld from development. We find that generated harnesses remain substantially behind mature human-engineered references on code and on search and research, while matching or exceeding the selected references on writing and machine-learning experimentation, with large variation in execution cost. Evolution produces some performance gains, but they are unstable and transfer only partially to held-out tasks. Experiments with a fixed runtime model further show that the gains depend strongly on the model executing the harness, indicating limited transfer across models.

Agentic software engineering benchmarks are typically summarized by nominal category labels such as "bug fix" or "feature implementation," yet benchmarks carrying the same label are built through very different curation pipelines. A label thus reveals little about the engineering work a benchmark demands. We introduce the Spread–Novelty–Centrality (SNC) profile, a three-axis characterization of the demands of repository-level coding tasks, grounded in empirical software engineering research. We apply the profile to five widely used benchmarks and 14,922 trajectories of two model families at three scales, and report three findings. (1) A label is an unreliable proxy for task demands, as every pair of benchmarks is statistically separated on at least two SNC axes, and the separations trace back to specific curation decisions. (2) Agent behaviour reveals demands that the human-written gold solution cannot. Agents produce larger solutions than the gold where problem statements withhold hints and smaller ones where curation inflates the gold. How a task is phrased shapes what an agent produces. (3) Task demands correlate with success uniformly, with resolved runs concentrating in the low-SNC region for every family and scale, whereas the behavioural signatures of success are family-specific. Claude succeeds by matching the scope of the gold solution, and its parity share on files rises from $0.17$ at the smallest scale to $0.54$ at the largest. Qwen succeeds by exceeding the gold scope at every scale, and editing too little marks failure for both families.

Large language models have shown promising capabilities in code refactoring, but existing approaches remain limited to method-level tasks. In this paper, we envision LLM-based refactoring as a continuous component of software maintenance rather than a tool invoked only for occasional manual refactoring. Under this vision, AI agents continuously monitor, evaluate, and improve codebases against explicit and evolving notions of software quality. We present a roadmap organized around five dimensions: the multi-objective optimization problem, quality definition and evaluation, multi-timescale integration of heterogeneous signals, architecture and design pattern, and trust in autonomous refactoring. We further identify integration into continuous delivery pipelines and cost considerations as cross-cutting concerns. For each dimension, we analyze the underlying challenges and pose open research questions. These dimensions define a research agenda for advancing autonomous refactoring from isolated code improvements to system-level quality maintenance.

A single vulnerability in a widely used library can cascade through millions of dependent applications, yet more than half of vulnerability database entries contain missing or incorrect affected-library information. Existing automated approaches neglect the relational structure of vulnerability databases, treating identification as an isolated text retrieval problem. In this paper, we propose Athena, the first graph-based approach for vulnerability affected library identification. Athena models vulnerability databases as a knowledge graph and reformulates the identification problem as knowledge graph completion (KGC). It comprises three key modules: a Modeling module that constructs a security knowledge graph integrating CVEs, libraries, CWE weakness types, CPE products, and software ecosystems; a Completion module that applies a modular KGC backbone to predict missing affected libraries for a given CVE via link prediction; and a Re-ranking module that retrieves KGC candidates and rescores them using a fine-tuned LLM augmented with knowledge graph embeddings, jointly leveraging structural and textual information. Our experiments on VulLib demonstrate that Athena significantly outperforms four state-of-the-art baselines, achieving a 32% improvement in Avg. F1 over the best baseline (i.e., VulLibGen). Notably, our KGC backbone with only 110M parameters already surpasses VulLibGen's best configuration at 7B parameters, demonstrating the effectiveness of graph-based modeling; the re-ranking module then provides substantial further gains, consistently outperforming the best baseline across all evaluated LLM backbones.

The Common Vulnerabilities and Exposures (CVE) database catalogs vulnerability claims in hard- and software, among them those pertaining to blockchain programs a.k.a. smart contracts. We present CVE-Smart-Contracts, a curated dataset of CVE records up to July 2026 referring to Ethereum smart contracts. The dataset contains the vulnerable artifacts (source code and runtime bytecode), labels according to three taxonomies, and function-level locations. The retrieval of CVE records, collection of additional evidence, validation of the correspondence between records and artifacts, label assignment, and vulnerability localization are automated, leaving 15% to manual analysis. The dataset does not validate the original vulnerability claims, but marks a few records obviously wrong as `refuted'. For the sake of reproducibility, all external inputs are retained, so that rerunning the pipelines results in the same outputs. The dataset comprises 491 records linked to deployed contracts, 26 referring to projects (mostly libraries), 45 without validated artifacts, and six records with refuted claims. The dataset supports empirical security research, in particular the evaluation of code analysis and repair techniques.

When a hint turns a failing generated program into a passing one, does it provide missing information or merely steer the model toward a solution it could already produce? We test these hypotheses on HumanEval+ and MBPP+ using executable evaluation. For Qwen2.5-3B-Instruct, adaptive relevant hints rescue 36 of 79 selected failures; an unrelated hint rescues 19, while eight unhinted samples solve 46 and recover 31 of the 36 relevant-hint rescues. Phi-3.5-mini shows the same pattern: relevant hints rescue 42 of 101 failures, an unrelated hint rescues 17, and unhinted sampling solves 57, including 36 of the 42 relevant-hint rescues. Because the hint conditions use different attempt budgets, these comparisons do not isolate a purely semantic effect. Mechanistic tests on Qwen identify a stable activation direction shared by relevant and unrelated hints. Persistently adding this direction yields 14 rescues and 18 regressions, with no detectable net accuracy gain; learned low-rank interventions have a positive but imprecise estimated effect. Full textual specifications solve 22 of 24 context-defined problems, versus 5-11 for tested virtual-KV prefixes. Post-generation hidden-state probes transfer across benchmarks, with pooled AUROC 0.806 and 0.780, but their top-one selection advantage over token confidence is statistically unresolved. Overall, relevant hints can rescue failures, but most rescued solutions are already reachable through ordinary sampling, and the internal interventions tested here do not establish task-general capability transfer.

Extracting physics information from modern particle-physics experiments requires multistage analyses implemented on top of large and highly interconnected software ecosystems. General-purpose large language models (LLMs) often produce unreliable programs for such tasks because a user request alone rarely specifies the required APIs, dependencies, and usage conventions. We organize these software relations before generation and retrieve task-relevant knowledge at inference time. Using the open-source ROOT framework as a representative and reproducible testbed, we evaluate a complete grounding system that combines hybrid retrieval over a heterogeneous software knowledge graph, skill-selected workflow examples, and execution-guided repair. On a benchmark of 275 ROOT tasks, grounding improves first-attempt execution from 58.5% to 76.0% under Claude Code orchestration and from 51.3% to 64.0% under standalone orchestration. Final success increases from 90.5% to 96.0% and from 78.9% to 90.9%, respectively, while the average generation cost per successful task increases by only 1.3% and 3.2%. The gains persist under a strong coding agent, indicating that explicit software knowledge remains valuable even when agentic scaffolding is already in place. Because the method captures software relations common to large codebases rather than facts specific to ROOT or a particular model, it should transfer to other experiment frameworks and proprietary software, especially where documentation is sparse or internal dependencies are complex.

Context: Accurate alignment between pull requests (PRs) and corresponding issues is crucial for efficient software development and maintaining code quality, as misalignments can reduce traceability, hinder defect localization, and decrease maintainability. Objective: This study aims to improve automated PR-issue alignment classification by leveraging fine-tuned large language models (LLMs) across multiple alignment categories, and conducts interpretability analysis to investigate the effects of PR-issue fields on the predictions of fine-tuned LLMs. Method: Our methodology consists of dataset preparation, LLM fine-tuning, and interpretability analysis. We first extended an existing dataset and applied data augmentation to address class imbalance. GPT-4o was then fine-tuned via instruction tuning, and open-source LLMs including CodeLlama-7B, CodeQwen1.5-7B, StableCode-3B, CodeGemma-7B, and Deepseek-Coder-6.7B were fine-tuned using classification-specific heads. Interpretability analysis using Shapley Additive Explanations (SHAP) was conducted to examine the influence of PR-issue fields on predictions for the best-performing open-source LLM. Results: Fine-tuned LLMs outperformed baseline models, achieving average improvements of 6.15% in accuracy and F1-micro, 14.69% in F1-macro, and 6.15% in recall. CodeLlama-7B emerged as the best-performing fine-tuned LLM overall, while interpretability analysis revealed that code diffs together with issue body and PR body contents exert the greatest influence on predictions. Conclusions: Fine-tuning substantially enhances PR-issue alignment classification, improving both accuracy and efficiency. Interpretability analysis provides actionable insights into the dataset features driving alignment decisions, deepening understanding of how LLMs reason over software artifacts.

Fault localization can focus a code model's repair on the statements a failing test implicates, but a targeted edit may succeed merely because it is small, and a second model call may succeed without using the failure at all. We separate these explanations with three arms applied to the same failed candidate: blind whole-solution resampling, spectrum-based localization followed by suspect-span infilling, and same-length infilling at a disjoint random code span. Across three frozen 26-32B models, three benchmarks and 488 failing candidates, plus a separately declared 24B fourth model from a third family, three results follow. First, localization is rarely available: only 9.0% of failing candidates expose a failing public test with a usable spectrum. Second, among the 177 candidates localizable from a strong suite, localized infilling loses decisively to blind resampling at a matched attempt count (3:40, p = 3.0 x 10^-9), opposite to our hypothesis; the loss replicates in a third family at -11.3 points (95% CI [-16.6, -6.8]), and widening the edit does not rescue it. Third, against the random-span placebo localized infilling leads pooled (11:1, Holm-adjusted p = .019), but that lead resolves in no individual model under the analysis our shipped plan designates primary (best Holm p = .087), so we report the location effect as suggestive rather than established. Re-pricing attempts as tokens narrows but does not overturn this: a span attempt spends 21.7 generated tokens against 371.1, yet 16 localized attempts reach 6.8% while one blind attempt already reaches 10.1%. Infilling reproduces the removed span verbatim in 48.9% of attempts, which is why more budget does not help. We restrict every localization conclusion to the 24-32B models tested.

Test-set accuracy is silent on two issues that matter when deploying autoregressive neural sequence models: how much probability mass the system under test (SUT) places on constraint-violating alternatives that are reachable under sampling and what fraction of the input population satisfies a domain requirement. We answer both with probabilistic model checking. The pipeline extracts a discrete-time Markov chain (DTMC) from the SUT's token-by-token generation, verifies formal PCTL specifications with the PRISM model checker, and aggregates the per-input verdicts into a coverage curve over the input space. A soundness theorem establishes the DTMC as an under-approximation, so every verdict yields a certified interval on the SUT's true reachability probability. The coverage built from those verdicts is, therefore, conservative by construction. A counterexample-guided abstraction refinement (CEGAR) loop adaptively tightens the interval, and a maximum-likelihood algorithm extracts the most probable falsifying trace. Two case studies exercise the pipeline. On a GPT-2 computer-aided process-planning (CAPP) model with 100% test accuracy, the pipeline quantifies the probability mass greedy decoding hides, but that is reachable with sampling; and identifies the smallest training fraction at which an ordering requirement holds population-wide, neither of which test accuracy can report. We then verify the SMILES molecular generator with a 50x larger vocabulary. The only change is an external chemical-validity oracle, and the pipeline identifies the gap between structural completeness and chemical validity.

Modern Text-to-SQL systems often follow generate-execute-select pipelines, generating multiple candidate queries then selecting the best one. Listwise selection, by jointly comparing multiple candidates, has been widely adopted, but fine-tuning listwise selectors is costly. We thus propose a fine-tuning-free listwise selector. We replace two major fine-tuning objectives with inference-time strategies: (1) learning selection criteria as ordering and (2) mitigating positional bias. First, we build reusable structured memories instead of learning selection behavior as model parameters. Given a question, MaP-SQL retrieves memories distilled from training data that encode how natural language maps to schema elements, SQL operations, and expected outputs. These memories serve as explicit decision criteria for evaluating candidates in a listwise manner. Second, to mitigate ordering bias of listwise selectors, we aggregate rankings across multiple input permutations, with inference cost optimized by execution results and pointwise scoring. Our approach improves selection accuracy while maintaining efficiency and compatibility with existing large language models. Across Text-to-SQL benchmarks, it produces more stable selection without fine-tuning and fewer unnecessary comparisons than existing methods. On BIRD-dev, it outperforms the previous state-of-the-art selector-based method R^3-SQL by 2.02 execution accuracy points on average using the same candidate sets, with 2.92x fewer tokens.

LLM-based software systems increasingly require effective "evals" as quality gates in the development lifecycle. However, existing work typically addresses individual aspects of eval reliability rather than the full set of practical requirements. We present an end-to-end eval pipeline that combines eval checklist creation, with learned aggregation for checklist responses, to improve agreement across LLM judges and accuracy against human judgments. The framework additionally pro- vides self-consistency, explanations, and prediction uncertainty, and we empirically demonstrate its effectiveness.

In the contemporary landscape of software devel-opment, securing sensitive data is paramount to safeguarding organizational reputation, preventing financial losses, and pro-tecting individuals from identity theft. This paper addresses the pervasive challenge of identifying and rectifying security vulnerabilities early in the development process, emphasizing the role of Static Application Security Testing (SAST) tools. While SAST tools play a crucial role in detecting vulnerabilities, widespread adoption has been hindered by usability issues, including high false positive rates and a lack of native pipeline support. This paper proposes a novel, generalized, and automated process for aggregating SAST tool outputs and integrating them into developers' familiar issue-tracking software. The process streamlines the identification and communication of security vulnerabilities during the development lifecycle, facilitating more efficient remediation efforts. We demonstrate the successful implementation of the proposed process with the SonarQube SAST tool in a GitLab-based development environment. Developers were positive about the structured implementation, real-time feedback, and proactive vulnerability management. However, despite some challenges such as a potential learning curve and tradeoffs between secure coding and workflow disruption, the overall positive impact on security awareness and responsiveness suggests that the proposed process holds promise in enhancing the security posture of software development practices

Large language models (LLMs) have shown proficiency in various software engineering tasks, such as code generation and translation. However, a key limitation in their performance may be their (lack of) understanding of programming-language semantics. Even when explicit semantics are given, it remains unclear whether LLMs apply those rules or lean on priors learned during pre-training instead. We study if LLMs lean on priors or given semantics with a novel task–Program Executability Prediction (PrEx)–that asks models to predict whether a program is semantically valid or invalid (and, if invalid, which formal rule it violates) given the program's syntax and operational semantics. Because PrEx requires both valid and invalid programs, we build a dataset with systematically generated invalid transformations derived from valid programs. We evaluate open-source coding LLMs under two semantic formalisms and two semantic shifts across Human-Written, LLM-Translated, and Fuzzer-Generated program splits. Our findings show that LLMs lean on pre-training priors rather than systematically applying the given rules, performing especially poorly on modified semantics and degrading further as program complexity increases. PrEx is available at https://github.com/EngineeringSoftware/prex.

Code generation aims to automatically generate source code from task requirements and has attracted significant attention with the rapid advancement of large language models (LLMs). Despite remarkable progress, LLMs often struggle to generate correct code for complex software engineering tasks because task descriptions are frequently incomplete, ambiguous, or lack critical contextual information. Existing approaches primarily improve the capabilities of coding agents through more sophisticated tools, skills, and workflows, while largely overlooking the quality of the task requirements themselves. To address this limitation, we draw inspiration from software requirements engineering and propose WiseSpec, a novel requirements-driven agent framework for repository-level code generation. WiseSpec automatically constructs structured and information-rich requirements, assesses their quality through execution-based evaluation, and iteratively refines them to better guide code generation. Experimental results show that WiseSpec consistently outperforms all baselines, achieving an average improvement of 13.17% in %Resolved.

Long-lived AI agents may replace their models, orchestration harnesses, interaction surfaces and hosts while retaining identity, memory and executable body lineage. We present a runtime-independent architecture that separates a continuity-bearing substrate $P_t=(I_t,M_t,B_t)$ from replaceable execution and interaction bindings. Six continuity invariants and a quiesce–checkpoint–validate–bind–rehydrate–resume protocol govern state preservation, capability changes and continuation authority. Enoch implements the design through a reusable software body, private installed state, provider contracts and fenced migration. We report complementary evidence from frozen implementation snapshots and three kinds of live study. An operational case retains an established instance's identity, nonempty memory, body revision and historical task IDs across host, joint Codex-to-Claude model/harness, and chat-surface substitutions. A separate cross-host task resumes from a verified artifact checkpoint. A supervised Codex–Muse–Codex study completes five ordinary round trips and five matched same-runtime controls on their first attempts after a separate qualification run. A planned worker interruption recovers through the native task API with the checkpoint preserved, a stale reply rejected and one task completion recorded. These results support mechanical continuity for the tested deployments and bounded workflows. They do not establish behavioral equivalence, arbitrary-task portability, general exactly-once external effects or unattended reliability.

Verification-aware languages, such as Dafny, integrate logical constructs into code and enable automatic verification of program correctness. However, tests remain helpful in scenarios that verification alone does not address (e.g., to support test-driven development). Existing Dafny test generation tools are implementation-based, limiting their applicability in this context. We present DSpec2Test, a specification-driven test generation tool for Dafny that automatically derives tests from formal specifications, without considering implementation details. Our tool extends Dafny's generate-tests command with a new blackbox mode based on Disjunctive Normal Form (DNF) equivalence class partitioning and optional Boundary Value Analysis (BVA). DSpec2Test relies on the Z3 SMT solver to synthesize inputs and expected outputs that meet the specification-derived constraints. We evaluate DSpec2Test on programs from DafnyBench mutated using MutDafny and compare it against Dafny's existing implementation-driven Block mode. DSpec2Test achieves a 93.9% mutation kill rate on a dataset of 131 mutants, outperforming Block's 82.4%, and uniquely killing 17 mutants. These results suggest that specification-driven testing is an effective and complementary approach for testing Dafny programs.

As quantum computing evolves, the demand for high-quality quantum software is increasing. Classical software quality models, such as McCall's model, provide a useful foundation, but they do not explicitly capture distinctive properties of quantum software, including probabilistic measurement outcomes, backend-dependent behavior, and hybrid quantum-classical workflows. This paper proposes a base quality model for quantum software using McCall's factor-criterion-metric structure as the organizing backbone. It broadens the set of quality factors by drawing on representative classical software quality models and standards and introduces quantum-specific extensions only where needed. We define the resulting factor set and quality criteria, provide a many-to-many mapping of core factor-to-criterion relationships, discuss representative candidate metrics together with the execution, reference, and interpretation conditions needed to use them, and illustrate how the resulting model can be applied across the quantum software lifecycle. The model is intended as a common starting point for structuring, assessing, and communicating quality concerns in quantum software while supporting context-specific adaptation. Further expert feedback, case studies, and empirical studies can be used to evaluate and refine the model and its application across different quantum software contexts.

LLM coding agents are crossing from task-scale to project-scale: single assisted efforts that run for months, across repeated context compactions, on codebases far larger than any context window. We report operational experience from one such effort (a research project under a single continuous Claude Code session line since January 2026, driving a 633,000-line codebase, with its memory subsystem continuously instrumented since July 2026) and describe SIx Harness, the open-source memory and continuity infrastructure that emerged from keeping it coherent. The harness combines per-project long-term memory (hybrid lexical-vector retrieval over SQLite, fully local), precision-gated context injection, anti-recurrence stores for decisions and dead-ends, conventions engineered to survive compaction, and, the part we argue is missing from operational practice, reliability engineering for the memory subsystem itself: a session-start health gate with discriminated failure modes, heartbeat telemetry designed so that no enumerated failure mode can pass unrecorded, and alert-fatigue budgeting borrowed from SRE practice. We present what we believe is the first months-scale instrumented operational record of a persistent agent-memory subsystem in production development use: 78,933 hook invocations; 85 recorded failures, none silent: 84 in the subsystem's first three weeks, one since, none in the final 20 days; an injection layer whose ten-day precision instrument shows zero false fires against an intact denominator; and three production incidents traced from instrument reading to structural fix. From the record we distill seven design principles, state our limitations plainly (N=1, no control arm, self-reported), and publish a tagged pre-registered ablation protocol that any team can run with the released MIT-licensed kit.

The software engineering research community has enthusiastically embraced the integration of Large Language Models (LLMs) into complex techniques to solve a wide variety of tasks. However, the extent to which this investment is strategic remains unclear, as the native capabilities of successive frontier model generations can rapidly render existing techniques obsolete. To assess this research investment, we analyze 35 LLM-based technique papers from ICSE 2026. We evaluate whether their complex tools can be outperformed by the simplest possible alternative: a single, automatically generated prompt executed on a newer generation model, without any iterative refinement. We find that for between 37% and 63% papers, a newer model with a single prompt natively outperforms the heavily engineered tooling proposed just a year prior. We identify that constructive techniques like code generation or repair are more amenable to substitution by a single-prompt. We also identify a surviving set of papers relying on strategies that provide additional insights to the model where newer LLMs will amplify the proposed technique. Our findings raise questions about the cost-benefit proposition of techniques designed as workarounds to temporary model deficits and the need to focus on enduring challenges that scale synergistically with future model generations. Our source codes and results are made publicly available at https://github.com/less-lab-uva/What-Survives-the-Next-Model.

Since the advent of Large Language Models (LLMs), practitioners have increasingly leveraged them to support their software engineering tasks, including automated code repair, showing promising results. Yet, concerns regarding reproducibility and generalizability remain largely unexplored. To further evaluate these concerns and associated impacts, we partially reproduce and conduct an exploratory Java extension of the FeedbackEval benchmark [1], which evaluates how LLMs leverage different feedback types for Python code repair. First, we partially replicate the original study on 394 repair tasks using GPT-4o and Claude 3.5 Sonnet, reproducing and observing the main qualitative trends reported in the original work. Second, we conduct an exploratory Java extension by constructing 100 erroneous repair instances from 50 Java tasks and evaluating feedback effectiveness. Our results show that previous conclusions from Python may be sensitive to benchmark construction, feedback representation, and tooling ecosystem, motivating more controlled multilingual benchmarks. Specifically, while test feedback remains the strongest feedback type in our Python replication, the same ranking is not observed in our Java extension, as simple and JUnit-based test feedback do not differ significantly. We hypothesize that differences in feedback informativeness and tooling ecosystems, such as the verbosity of test frameworks, may partly explain such a difference. Finally, lighter prompts reduce cost without significant differences in repair effectiveness. Overall, our findings confirm key trends under a partially controlled replication and highlight the need for more rigorous multilingual evaluation and careful feedback design in LLMbased repair systems.

Context: Software engineering is moving from AI-assisted practices like vibe coding, in which assistants accelerate individual developers, towards Agentic Software Engineering (ASE), in which autonomous agents are delegated goal-level tasks. However, industry reports a productivity paradox: as individual productivity increases, team throughput, review capacity, and stability degrade because team-scale software engineering discipline is neglected. Objective: This paper aims to establish the conceptual and methodological foundations of Spec-Driven Development (SDD) as an enabling discipline for ASE at team scale and characterize the harness, i.e., the technical and methodological mechanisms through which teams govern agent behavior. Method: We conducted a conceptual analysis drawing predominantly on gray literature, including ASE vision and roadmap papers, practitioner reports, talks, and tooling, because peer-reviewed evidence and a shared academic-industrial vocabulary are not yet established. Results: Using a comparative characterization of the paradigm progression as conceptual framing, the article presents (i) a socio-technical model of SDD in which specifications act as the contract substrate between humans and agents; (ii) an operational characterization of the harness, distinguishing the technical harness around the agent from the methodological harness around the team, with worked examples; and (iii) a typology of five human–agent interaction patterns through which the human role is redefined. Conclusion: We conclude that SDD reconstitutes, in specification-centric form, the contracts that vibe coding dissolves: accountability, verifiability, and transferability. Given the immaturity of the evidence base, this work is presented as a first step toward academic-industrial consensus rather than a validated theory, and outlines a research agenda for future empirical validation.

While it is fairly well known how empirical software engineering (ESE) is used in the academic world, we have limited knowledge of how ESE is practiced in industry. As part of our regular column on empirical software engineering (ACM SIGSOFT SEN-ESE), we want to dedicate a series of articles to interviewing ESE practitioners from various companies. Among other things, we want to understand how ESE processes are implemented in industry, e.g., different research methods, how practitioners decide on what to study, how research results are used within companies and beyond, and if they face recurrent impediments to using ESE methods in industrial contexts. In the first edition of "ESE in Practice", we are joined by Ciera Jaspan and Collin Green from the Developer Intelligence team at Google. This article is a faithful account of our conversation from August 13, 2026, which we edited for our column.

Maintaining an abstraction of the execution history of threads can improve the precision of data race detection in static analysis. Here, we extend the digest framework to handle concurrency constructs and synchronization mechanisms that have been ignored in static race detection. We introduce mechanisms for the commonly used thread barriers, as well as pthread_once, which allows to ensure that an action is executed only once. We also instantiate the framework with an abstraction of locksets held by ancestor threads. We propose a suite of litmus tests to evaluate analyses for these features and compare our implementation to state-of-the-art tools, finding that they lack support.

LLM agents increasingly diagnose incidents and propose remediations. In a GitOps workflow, applying a fix means editing a version-controlled config file, and the obvious implementation, having the model author the edited file or a diff, is what practitioners reach for first. Evaluating that choice on real Kubernetes manifests, we find no text-generation strategy is safe for unattended automation. Unified diffs are unsafe: under strict patching almost none apply, but that is an artifact, since a tolerant tool (GNU patch) applies 96%, yet silently misapplies about 1 in 7 (14-20%) with no error signal. Full-file rewrite is capability-dependent: a small model corrupts the file, while a frontier model is usually correct but non-deterministic (it silently drops a field or edits a neighbor on some runs) and must regenerate the whole file, costing O(file size) per edit. We present an alternative that separates the semantic decision (which resource, field, and value) from the syntactic act of editing the file. The agent emits only a structured field-change intent; a deterministic pipeline indexes manifests by (kind, name), locates the target scalar's exact character span via the YAML parser's node position marks, and replaces only that span in the raw text. Because the file is never re-serialized, the diff is minimal by construction, formatting and comments are preserved, and the edit is correct and deterministic independent of the model, at O(1) generation cost. The contribution is the pairing of an LLM-proposed intent with a deterministic, fail-closed application contract for GitOps. We implement it in KubeAstra (Apache-2.0) and release the benchmark. Our claim is scoped to faithful application of a known change; whether the change is right is left to human PR review.

LLM judges, models that score another system's output, can be gamed by the systems they score. Recent work identifies one defence that works: the judge solves the task itself first and commits to that answer, then accepts a candidate only if the two match. We call this commit-first judging, and ask whether shipped software implements it, and what it costs. We audit the default judge configurations of eight widely used evaluation frameworks. Of the 24 configurations in scope, none implement it. Nine implement a variant the literature measures as ineffective, and share one ancestor prompt, traceable through a copied typographical error. In a controlled experiment, an ordinary best-of-N search with no access to correct answers optimises code against one of these configurations, used exactly as documented. On an interval merging task the judge accepted 90 of 96 candidates in one seed and 93 of 96 in the other; every accepted candidate passed every test the search could see and failed a held-out suite it could not. The judge identified the defective line and cited it as grounds for a perfect score. Commit-first judging removed the effect: 0 of 96 in both seeds. On a second task it made matters worse in both seeds: the judge's committed answer was wrong, and in one seed the population converged on it. This is our main finding. Commit-first judging does not remove the anchor that gets gamed, it moves it from the candidate to the judge's own answer, so evaluation is only as good as the judge is at the task. That precondition is cheap to measure in advance, and is task local rather than scale dependent: a smaller judge solved a task the frontier judge failed and resisted gaming where it did not. We also validate our own instruments: five of fifteen claims in our criteria were wrong against verbatim sources, and two held-out checks were unjustified by their specifications.

End-to-end GUI testing is essential for verifying web applications, yet existing evaluations rely on predefined checklists and are confined to the data and frameworks of web generation benchmarks, leaving the bug-discovery ability of vision-language models (VLMs) systematically untested. We introduce Code-driven Agentic Testing (CAT), a paradigm in which the agent writes Playwright code to drive the browser, gathers feedback, and autonomously explores web applications to uncover bugs. We instantiate CAT with CATJudge, an agentic framework that unifies Browser-Use and Computer-Use tools within a single environment and CATTest, a benchmark of 102 AI-generated web applications with carefully annotated bugs, built through close human-AI collaboration to feature complex interactions and subtle defects. Experiments with mainstream VLMs show that all evaluated models perform poorly, revealing a clear gap between current VLM capabilities and the demands of real-world testing in AI web development. We release our code and data at https://github.com/SleepyWithoutCoffee/CATJudge.