arXiv:2602.20048v1 Announce Type: new Abstract: Modern code intelligence agents operate in contexts exceeding 1 million tokens--far beyond the scale where humans manually locate relevant files. Yet agents consistently fail to discover architecturally critical files when solving real-world coding tasks. We identify the Navigation Paradox: agents perform poorly not due to context limits, but because navigation and retrieval are fundamentally distinct problems. Through 258 automated trials across 30 benchmark tasks on a production FastAPI repository, we demonstrate that graph-based structural navigation via CodeCompass--a Model Context Protocol server exposing dependency graphs--achieves 99.4% task completion on hidden-dependency tasks, a 23.2 percentage-point improvement over vanilla agents (76.2%) and 21.2 points over BM25 retrieval (78.2%).However, we uncover a critical adoption gap: 58% of trials with graph access made zero tool calls, and agents required explicit prompt engineering to adopt the tool consistently. Our findings reveal that the bottleneck is not tool availability but behavioral alignment--agents must be explicitly guided to leverage structural context over lexical heuristics. We contribute: (1) a task taxonomy distinguishing semantic-search, structural, and hidden-dependency scenarios; (2) empirical evidence that graph navigation outperforms retrieval when dependencies lack lexical overlap; and (3) open-source infrastructure for reproducible evaluation of navigation tools.

arXiv:2602.18492v1 Announce Type: cross Abstract: Large Language Models (LLMs) are now good enough at coding that developers can describe intent in plain language and let the tool produce the first code draft, a workflow increasingly built into tools like GitHub Copilot, Cursor, and Replit. What is missing is a reliable way to tell which model written queries are safe to accept without sending everything to a human. We study the application of an LLM jury to run this review step. We first benchmark 15 open models on 82 MySQL text to SQL tasks using an execution grounded protocol to get a clean baseline of which models are strong. From the six best models we build unanimous committees of sizes 1 through 6 that see the prompt, schema, and candidate SQL and accept it only when every member says it is correct. This rule matches safety first deployments where false accepts are more costly than false rejects. We measure true positive rate, false positive rate and Youden J and we also look at committees per generator. Our results show that single model judges are uneven, that small unanimous committees of strong models can cut false accepts while still passing many good queries, and that the exact committee composition matters significantly.

arXiv:2602.18548v1 Announce Type: cross Abstract: Design-to-code translates high-fidelity UI designs into executable front-end implementations, but progress remains hard to compare due to inconsistent datasets, toolchains, and evaluation protocols. We introduce 1D-Bench, a benchmark grounded in real e-commerce workflows, where each instance provides a reference rendering and an exported intermediate representation that may contain extraction errors. 1D is short for one day, representing the efficient completion of design-to-code tasks in less than one day. Models take both as input, using the intermediate representation as structural cues while being evaluated against the reference rendering, which tests robustness to intermediate representation defects rather than literal adherence. 1D-Bench requires generating an executable React codebase under a fixed toolchain with an explicit component hierarchy, and defines a multi-round setting in which models iteratively apply component-level edits using execution feedback. Experiments on commercial and open-weight multimodal models show that iterative editing generally improves final performance by increasing rendering success and often improving visual similarity. We further conduct a pilot study on post-training with synthetic repair trajectories and reinforcement learning based editing, and observe limited and unstable gains that may stem from sparse terminal rewards and high-variance file-level updates.

arXiv:2602.18571v1 Announce Type: cross Abstract: While significant progress has been made in automating various aspects of software development through coding agents, there is still significant room for improvement in their bug fixing capabilities. Debugging and investigation of runtime behavior remains largely a manual, developer-driven process. Popular coding agents typically rely on either static analysis of the code or iterative test-fix cycles, which is akin to trial and error debugging. We posit that there is a wealth of rich runtime information that developers routinely access while debugging code, which agents are currently deprived of due to design limitations. Despite how prevalent debuggers are in modern IDEs and command-line tools, they have surprisingly not made their way into coding agents. In this work, we introduce Debug2Fix, a novel framework that incorporates interactive debugging as a core component of a software engineering agent via a subagent architecture. We incorporate debuggers for Java and Python into our agent framework and evaluate against GitBug-Java and SWE-Bench-Live and achieve >20% improvement in performance compared to the baseline for certain models. Furthermore, using our framework, we're able to make weaker models like GPT-5 and Claude Haiku 4.5 match or exceed the performances of stronger models like Claude Sonnet 4.5, showing that better tool design is often just as important as switching to a more expensive model. Finally, we conduct systematic ablations demonstrating the importance of both the subagent architecture and debugger integration.

arXiv:2602.18800v1 Announce Type: cross Abstract: It is now common practice in software development for large language models (LLMs) to be used to generate program code. It is desirable to evaluate the robustness of LLMs for this usage. This paper is concerned in particular with how sensitive LLMs are to variations in descriptions of the coding tasks. However, existing techniques for evaluating this robustness are unsuitable for code generation because the input data space of natural language descriptions is discrete. To address this problem, we propose a robustness evaluation method called scenario domain analysis, which aims to find the expected minimal change in the natural language descriptions of coding tasks that would cause the LLMs to produce incorrect outputs. We have formally proved the theoretical properties of the method and also conducted extensive experiments to evaluate the robustness of four state-of-the-art art LLMs: Gemini-pro, Codex, Llamma2 and Falcon 7B, and have found that we are able to rank these with confidence from best to worst. Moreover, we have also studied how robustness varies in different scenarios, including the variations with the topic of the coding task and with the complexity of its sample solution, and found that robustness is lower for more complex tasks and also lower for more advanced topics, such as multi-threading and data structures.

arXiv:2602.19441v1 Announce Type: cross Abstract: Autonomous coding agents increasingly contribute to software development by submitting pull requests on GitHub; yet, little is known about how these contributions integrate into human-driven review workflows. We present a large empirical study of agent-authored pull requests using the public AIDev dataset, examining integration outcomes, resolution speed, and review-time collaboration signals. Using logistic regression with repository-clustered standard errors, we find that reviewer engagement has the strongest correlation with successful integration, whereas larger change sizes and coordination-disrupting actions, such as force pushes, are associated with a lower likelihood of merging. In contrast, iteration intensity alone provides limited explanatory power once collaboration signals are considered. A qualitative analysis further shows that successful integration occurs when agents engage in actionable review loops that converge toward reviewer expectations. Overall, our results highlight that the effective integration of agent-authored pull requests depends not only on code quality but also on alignment with established review and coordination practices.

arXiv:2602.19718v1 Announce Type: cross Abstract: The rapid adoption of Generative AI (GenAI) in the software development life cycle (SDLC) increases computational demand, which can raise the carbon footprint of development activities. At the same time, organizations are increasingly embedding governance mechanisms into GenAI-assisted development to support trust, transparency, and accountability. However, these governance mechanisms introduce additional computational workloads, including repeated inference, regeneration cycles, and expanded validation pipelines, increasing energy use and the carbon footprint of GenAI-assisted development. This paper proposes Carbon-Aware Governance Gates (CAGG), an architectural extension that embeds carbon budgets, energy provenance, and sustainability-aware validation orchestration into human-AI governance layers. CAGG comprises three components: (i) an Energy and Carbon Provenance Ledger, (ii) a Carbon Budget Manager, and (iii) a Green Validation Orchestrator, operationalized through governance policies and reusable design patterns.

arXiv:2602.19843v1 Announce Type: cross Abstract: As LLM-based Multi-Agent Systems (MAS) are increasingly deployed for complex tasks, ensuring their reliability has become a pressing challenge. Since MAS coordinate through unstructured natural language rather than rigid protocols, they are prone to semantic failures (e.g., hallucinations, misinterpreted instructions, and reasoning drift) that propagate silently without raising runtime exceptions. Prevailing evaluation approaches, which measure only end-to-end task success, offer limited insight into how these failures arise or how effectively agents recover from them. To bridge this gap, we propose MAS-FIRE, a systematic framework for fault injection and reliability evaluation of MAS. We define a taxonomy of 15 fault types covering intra-agent cognitive errors and inter-agent coordination failures, and inject them via three non-invasive mechanisms: prompt modification, response rewriting, and message routing manipulation. Applying MAS-FIRE to three representative MAS architectures, we uncover a rich set of fault-tolerant behaviors that we organize into four tiers: mechanism, rule, prompt, and reasoning. This tiered view enables fine-grained diagnosis of where and why systems succeed or fail. Our findings reveal that stronger foundation models do not uniformly improve robustness. We further show that architectural topology plays an equally decisive role, with iterative, closed-loop designs neutralizing over 40% of faults that cause catastrophic collapse in linear workflows. MAS-FIRE provides the process-level observability and actionable guidance needed to systematically improve multi-agent systems.

arXiv:2602.19844v1 Announce Type: cross Abstract: LLM-enabled applications are rapidly reshaping the software ecosystem by using large language models as core reasoning components for complex task execution. This paradigm shift, however, introduces fundamentally new reliability challenges and significantly expands the security attack surface, due to the non-deterministic, learning-driven, and difficult-to-verify nature of LLM behavior. In light of these emerging and unavoidable safety challenges, we argue that such risks should be treated as expected operational conditions rather than exceptional events, necessitating a dedicated incident-response perspective. Consequently, the primary barrier to trustworthy deployment is not further improving model capability but establishing system-level threat monitoring mechanisms that can detect and contextualize security-relevant anomalies after deployment -- an aspect largely underexplored beyond testing or guardrail-based defenses. Accordingly, this position paper advocates systematic and comprehensive monitoring of security threats in LLM-enabled applications as a prerequisite for reliable operation and a foundation for dedicated incident-response frameworks.

arXiv:2511.02780v3 Announce Type: replace-cross Abstract: Smart contracts operate in a highly adversarial environment, where vulnerabilities can lead to substantial financial losses. Thus, smart contracts are subject to security audits. In auditing, proof-of-concept (PoC) exploits play a critical role by demonstrating to the stakeholders that the reported vulnerabilities are genuine, reproducible, and actionable. However, manually creating PoCs is time-consuming, error-prone, and often constrained by tight audit schedules. We introduce PoCo, an agentic framework that automatically generates executable PoC exploits from natural-language vulnerability descriptions written by auditors. PoCo autonomously generates PoC exploits in an agentic manner by interacting with a set of codeexecution tools in a Reason-Act-Observe loop. It produces fully executable exploits compatible with the Foundry testing framework, ready for integration into audit reports and other security tools. We evaluate PoCo on a dataset of 23 real-world vulnerability reports. PoCo consistently outperforms the Zero-shot and Workflow baselines, generating well-formed and logically correct PoCs. Our results demonstrate that agentic frameworks can significantly reduce the effort required for high-quality PoCs in smart contract audits. Our contribution provides actionable knowledge for the smart contract security community.

arXiv:2602.13379v1 Announce Type: cross Abstract: LLM-based agents are becoming increasingly capable, yet their safety lags behind. This creates a gap between what agents can do and should do. This gap widens as agents engage in multi-turn interactions and employ diverse tools, introducing new risks overlooked by existing benchmarks. To systematically scale safety testing into multi-turn, tool-realistic settings, we propose a principled taxonomy that transforms single-turn harmful tasks into multi-turn attack sequences. Using this taxonomy, we construct MT-AgentRisk (Multi-Turn Agent Risk Benchmark), the first benchmark to evaluate multi-turn tool-using agent safety. Our experiments reveal substantial safety degradation: the Attack Success Rate (ASR) increases by 16% on average across open and closed models in multi-turn settings. To close this gap, we propose ToolShield, a training-free, tool-agnostic, self-exploration defense: when encountering a new tool, the agent autonomously generates test cases, executes them to observe downstream effects, and distills safety experiences for deployment. Experiments show that ToolShield effectively reduces ASR by 30% on average in multi-turn interactions. Our code is available at https://github.com/CHATS-lab/ToolShield.

arXiv:2602.13611v1 Announce Type: cross Abstract: Recently, large language models (LLMs) are extensively utilized to enhance development efficiency, leading to numerous benchmarks for evaluating their performance. However, these benchmarks predominantly focus on implementation, overlooking the equally critical aspect of software design. This gap raises two pivotal questions: (1) Can LLMs handle software design? (2) Can LLMs write code following the specific designs? To investigate these questions, this paper proposes DesBench, a design-aware benchmark for evaluating LLMs on three software design-related tasks: design-aware code generation, object-oriented modeling, and the design of acceptance test cases. DesBench comprises 30 manually crafted Java projects that include requirement documents, design models, implementations, and acceptance tests, amounting to a total of 30 design models, 194 Java classes, and 737 test cases. We evaluated seven state-of-the-art LLMs, including three DeepSeek R1, two Qwen2.5, and two GPT models, using DesBench. The results reveal that LLMs remain significantly challenged by the intricacies of software design: (1) For code generation, LLMs struggle to produce correct implementations when provided with only high-level or no designs. (2) In object-oriented modeling, while LLMs can accurately identify objects and classes, they face challenges in defining operations and inter-class relationships. (3) Acceptance test cases generated by LLMs from functional requirements achieve code coverage quality comparable to those written by humans. Our research highlights the current limitations of LLMs in managing software design and calls for further investigation into new design methodologies and languages suitable for LLM-based development.

arXiv:2602.13851v1 Announce Type: cross Abstract: Formal specifications are crucial for building verifiable and dependable software systems, yet generating accurate and verifiable specifications for real-world C programs remains challenging. This paper empirically evaluates the extent to which formal-analysis tools can automatically generate and verify ACSL specifications without human or learning-based assistance. We conduct a controlled study on a recently released dataset of 506 C programs, repurposing it from interactive, developer-driven workflows to an automated evaluation setting. Five ACSL generation systems are compared: a rule-based Python script, Frama-C's RTE plugin, and three large language models--DeepSeek-V3.2, GPT-5.2, and OLMo 3.1 32B Instruct. All generated specifications are verified under identical conditions using the Frama-C WP plugin powered by multiple SMT solvers, allowing a direct comparison of annotation quality, solver sensitivity, and proof stability. Our results provide new empirical evidence on the capabilities and limitations of automated ACSL generation, complementing prior survey-based work.

arXiv:2602.13921v1 Announce Type: cross Abstract: Repository-level bug localization-the task of identifying where code must be modified to fix a bug-is a critical software engineering challenge. Standard Large Language Modles (LLMs) are often unsuitable for this task due to context window limitations that prevent them from processing entire code repositories. As a result, various retrieval methods are commonly used, including keyword matching, text similarity, and simple graph-based heuristics such as Breadth-First Search. Graph Neural Networks (GNNs) offer a promising alternative due to their ability to model complex, repository-wide dependencies; however, their application has been hindered by the lack of a dedicated benchmark. To address this gap, we introduce GREPO, the first GNN benchmark for repository-scale bug localization tasks. GREPO comprises 86 Python repositories and 47294 bug-fixing tasks, providing graph-based data structures ready for direct GNN processing. Our evaluation of various GNN architectures shows outstanding performance compared to established information retrieval baselines. This work highlights the potential of GNNs for bug localization and established GREPO as a foundation resource for future research, The code is available at https://github.com/qingpingmo/GREPO.

arXiv:2602.14012v1 Announce Type: cross Abstract: The integration of LLMs into vulnerability detection (VD) has shifted the field toward interpretable and context-aware analysis. While post-training methods have shown promise in general coding tasks, their systematic application to VD remains underexplored. In this paper, we present the first comprehensive investigation into the post-training pipeline for LLM-based VD, spanning from cold-start SFT to off-policy preference optimization and on-policy RL, uncovering how data curation, stage interactions, reward mechanisms, and evaluation protocols collectively dictate the efficacy of model training and assessment. Our study identifies practical guidelines and insights: (1) SFT based on rejection sampling greatly outperforms rationalization-based supervision, which can introduce hallucinations due to ground-truth leakage. (2) While increased SFT epochs constantly benefit preference optimization, excessive SFT inhibits self-exploration during RL, ultimately limiting performance gains. (3) Coarse-grained reward signals often mislead RL, whereas fine-grained root-cause judgments ensure reliable credit assignment. Specification-based rewards offer further benefits but incur significant effort in specification generation. (4) Although filtering extremely hard-to-detect vulnerability samples improves RL training efficiency, the cost of performance loss should be considered in practical applications. (5) Models trained under GRPO significantly outperform those using SFT and preference optimization (i.e., DPO and ORPO), as well as a series of zero-shot SOTA LLMs, underscoring the significant potential of on-policy RL for LLM-based VD. (6) In contrast to binary matching that tends to overestimate performance, LLM-as-a-Judge based on root-cause analysis provides a more robust evaluation protocol, although its accuracy varies across judge models with different levels of security expertise.

arXiv:2602.14591v1 Announce Type: cross Abstract: This paper presents an automated method for classifying source code changes during the software development process based on clustering of change metrics. The method consists of two steps: clustering of metric vectors computed for each code change, followed by expert mapping of the resulting clusters to predefined change classes. The distribution of changes into clusters is performed automatically, while the mapping of clusters to classes is carried out by an expert. Automation of the distribution step substantially reduces the time required for code change review. The k-means algorithm with a cosine similarity measure between metric vectors is used for clustering. Eleven source code metrics are employed, covering lines of code, cyclomatic complexity, file counts, interface changes, and structural changes. The method was validated on five software systems, including two open-source projects (Subversion and NHibernate), and demonstrated classification purity of P_C = 0.75 +/- 0.05 and entropy of E_C = 0.37 +/- 0.06 at a significance level of 0.05.

arXiv:2410.15756v3 Announce Type: replace-cross Abstract: Ensuring correctness is crucial for code generation. Formal verification offers a definitive assurance of correctness, but demands substantial human effort in proof construction and hence raises a pressing need for automation. The primary obstacle lies in the severe lack of data-there is much fewer proofs than code snippets for Large Language Models (LLMs) to train upon. In this paper, we introduce SAFE, a framework that overcomes the lack of human-written proofs to enable automated proof generation of Rust code. SAFE establishes a self-evolving cycle where data synthesis and fine-tuning collaborate to enhance the model capability, leveraging the definitive power of a symbolic verifier in telling correct proofs from incorrect ones. SAFE also re-purposes the large number of synthesized incorrect proofs to train the self-debugging capability of the fine-tuned models, empowering them to fix incorrect proofs based on the verifier's feedback. SAFE demonstrates superior efficiency and precision compared to GPT-4o. Through tens of thousands of synthesized proofs and the self-debugging mechanism, we improve the capability of open-source models, initially unacquainted with formal verification, to automatically write proofs for Rust code. This advancement leads to a significant improvement in performance, achieving a 52.52% accuracy rate in a benchmark crafted by human experts, a significant leap over GPT-4o's performance of 14.39%.

arXiv:2505.13766v3 Announce Type: replace-cross Abstract: Software Quality Assurance (SQA) is critical for delivering reliable, secure, and efficient software products. The Software Quality Assurance Process aims to provide assurance that work products and processes comply with predefined provisions and plans. Recent advancements in Large Language Models (LLMs) present new opportunities to enhance existing SQA processes by automating tasks like requirement analysis, code review, test generation, and compliance checks. Simultaneously, established standards such as ISO/IEC 12207, ISO/IEC 25010, ISO/IEC 5055, ISO 9001/ISO/IEC 90003, CMMI, and TMM provide structured frameworks for ensuring robust quality practices. This paper surveys the intersection of LLM-based SQA methods and these recognized standards, highlighting how AI-driven solutions can augment traditional approaches while maintaining compliance and process maturity. We first review the foundational software quality standards and the technical fundamentals of LLMs in software engineering. Next, we explore various LLM-based SQA applications, including requirement validation, defect detection, test generation, and documentation maintenance. We then map these applications to key software quality frameworks, illustrating how LLMs can address specific requirements and metrics within each standard. Empirical case studies and open-source initiatives demonstrate the practical viability of these methods. At the same time, discussions on challenges (e.g., data privacy, model bias, explainability) underscore the need for deliberate governance and auditing. Finally, we propose future directions encompassing adaptive learning, privacy-focused deployments, multimodal analysis, and evolving standards for AI-driven software quality.

arXiv:2507.19457v2 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly adapted to downstream tasks via reinforcement learning (RL) methods like Group Relative Policy Optimization (GRPO), which often require thousands of rollouts to learn new tasks. We argue that the interpretable nature of language often provides a much richer learning medium for LLMs, compared to policy gradients derived from sparse, scalar rewards. To test this, we introduce GEPA (Genetic-Pareto), a prompt optimizer that thoroughly incorporates natural language reflection to learn high-level rules from trial and error. Given any AI system containing one or more LLM prompts, GEPA samples trajectories (e.g., reasoning, tool calls, and tool outputs) and reflects on them in natural language to diagnose problems, propose and test prompt updates, and combine complementary lessons from the Pareto frontier of its own attempts. As a result of GEPA's design, it can often turn even just a few rollouts into a large quality gain. Across six tasks, GEPA outperforms GRPO by 6% on average and by up to 20%, while using up to 35x fewer rollouts. GEPA also outperforms the leading prompt optimizer, MIPROv2, by over 10% (e.g., +12% accuracy on AIME-2025), and demonstrates promising results as an inference-time search strategy for code optimization. We release our code at https://github.com/gepa-ai/gepa .

arXiv:2601.12522v2 Announce Type: replace-cross Abstract: Software bugs cost technology providers (e.g., AT&T) billions annually and cause developers to spend roughly 50% of their time on bug resolution. Traditional methods for bug localization often analyze the suspiciousness of code components (e.g., methods, documents) in isolation, overlooking their connections with other components in the codebase. Recent advances in Large Language Models (LLMs) and agentic AI techniques have shown strong potential for code understanding, but still lack causal reasoning during code exploration and struggle to manage growing context effectively, limiting their capability. In this paper, we present a novel agentic technique for bug localization -- CogniGent -- that overcomes the limitations above by leveraging multiple AI agents capable of causal reasoning, call-graph-based root cause analysis and context engineering. It emulates developers-inspired debugging practices (a.k.a., dynamic cognitive debugging) and conducts hypothesis testing to support bug localization. We evaluate CogniGent on a curated dataset of 591 bug reports using three widely adopted performance metrics and compare it against six established baselines from the literature. Experimental results show that our technique consistently outperformed existing traditional and LLM-based techniques, achieving MAP improvements of 23.33-38.57% at the document and method levels. Similar gains were observed in MRR, with increases of 25.14-53.74% at both granularity levels. Statistical significance tests also confirm the superiority of our technique. By addressing the reasoning, dependency, and context limitations, CogniGent advances the state of bug localization, bridging human-like cognition with agentic automation for improved performance.