arXiv subjects

Xiaofei Xie

Publications and source records attributed to Xiaofei Xie.

At least 37 records · Page 2Linked to original sources

arXiv2026

Where Did It Go Wrong? Capability-Oriented Failure Attribution for Vision-and-Language Navigation Agents

Embodied agents in safety-critical applications such as Vision-Language Navigation (VLN) rely on multiple interdependent capabilities (e.g., perception, memory, planning, decision), making failures difficult to localize and attribute. Existing testing methods are largely system-level and provide limited insight into which capability deficiencies cause task failures. We propose a capability-oriented testing approach that enables failure detection and attribution by combining (1) adaptive test case generation via seed selection and mutation, (2) capability oracles for identifying capability-specific errors, and (3) a feedback mechanism that attributes failures to capabilities and guides further test generation. Experiments show that our method discovers more failure cases and more accurately pinpoints capability-level deficiencies than state-of-the-art baselines, providing more interpretable and actionable guidance for improving embodied agents.

cs.MA↗

arXiv2026

Foundation Models for Autonomous Driving System: An Initial Roadmap

Recent advances in foundation models (FMs), including large language models (LLMs), vision-language models (VLMs), and world models, have opened new opportunities for autonomous driving systems (ADSs) in perception, reasoning, decision-making, and interaction. However, ADSs are safety-critical cyber-physical systems, and integrating FMs into them raises substantial software engineering challenges in data curation, system design, deployment, evaluation, and assurance. To clarify this rapidly evolving landscape, we present an initial roadmap, grounded in a structured literature review, for integrating FMs into autonomous driving across three dimensions: FM infrastructure, in-vehicle integration, and practical deployment. For each dimension, we summarize the state of the art, identify key challenges, and highlight open research opportunities. Based on this analysis, we outline research directions for building reliable, safe, and trustworthy FM-enabled ADSs.

cs.SE↗

arXiv2026

AutoEG: Exploiting Known Third-Party Vulnerabilities in Black-Box Web Applications

Large-scale web applications are widely deployed with complex third-party components, inheriting security risks arising from component vulnerabilities. Security assessment is therefore required to determine whether such known vulnerabilities remain practically exploitable in real applications. Penetration testing is a widely adopted approach that validates exploitability by launching concrete attacks against known vulnerabilities in real-world black-box systems. However, existing approaches often fail to automatically generate reliable exploits, limiting their effectiveness in practical security assessment. This limitation mainly stems from two issues: (1) precisely triggering vulnerabilities with correct technical details, and (2) adapting exploits to diverse real-world deployment settings. In this paper, we propose AutoEG, a fully automated multi-agent framework for exploit generation targeting black-box web applications. AutoEG has two phases: First, AutoEG extracts precise vulnerability trigger logic from unstructured vulnerability information and encapsulates it into reusable trigger functions. Second, AutoEG uses trigger functions for concrete attack objectives and iteratively refines exploits through feedback-driven interaction with the target application. We evaluate AutoEG on 104 real-world vulnerabilities with 29 attack objectives, resulting in 660 exploitation tasks and 55,440 exploit attempts. AutoEG achieves an average success rate of 82.41%, substantially outperforming state-of-the-art baselines, whose best performance reaches only 32.88%.

cs.CR↗

arXiv2026

Search-Induced Issues in Web-Augmented LLM Code Generation: Detecting and Repairing Error-Inducing Pages

Web-augmented large language models (LLMs) offer promising capabilities for automatic code generation. However, integrating live web search exposes models to unreliable or malicious content, leading to Search-Induced Issues (SII), a novel failure mode in which external pages mislead LLMs into producing incorrect code. This paper presents a comprehensive empirical study of the prevalence and impact of SII across three commercial search APIs and six advanced LLMs. Our analysis reveals that all evaluated web-augmented LLMs are vulnerable to SII, with root causes arising from either misaligned specifications or flawed code implementations in the searched Error-Inducing Pages (EIPs). To address this challenge, we propose Sherlock, an automated framework that enables LLM service providers to proactively safeguard web-augmented generation systems at scale. Sherlock operates as a continuous pipeline that first detects potential SII instances, then debugs them to identify the responsible EIPs and pinpoint their root causes, and finally repairs them by either annotating misaligned content or replacing erroneous code snippets with evaluated solutions from trusted sources. Experiments show that Sherlock identifies EIPs with an F1 score of up to 95% and repairs 71% to 100% of affected generations across the evaluated models, with modest computational overhead. Our findings and framework provide practical guidance for improving the reliability of web-augmented LLM-based code generation systems in real-world software engineering scenarios.

cs.SE↗

arXiv2026

Human in the Loop for Fuzz Testing: Literature Review and the Road Ahead

Fuzz testing is one of the most effective techniques for detecting bugs and vulnerabilities in software. However, as the basis of fuzz testing, automated heuristics often fail to uncover deep or complex vulnerabilities. As a result, the performance of fuzz testing remains limited. One promising way to address this limitation is to integrate human expert guidance into the paradigm of fuzz testing. Even though some works have been proposed in this direction, there is still a lack of a systematic research roadmap for combining Human-in-the-Loop (HITL) and fuzz testing, hindering the potential for further enhancing fuzzing effectiveness. To bridge this gap, this paper outlines a forward-looking research roadmap for HITL for fuzz testing. Specifically, we highlight the promise of visualization techniques for interpretable fuzzing processes, as well as on-the-fly interventions that enable experts to guide fuzzing toward hard-to-reach program behaviors. Moreover, the rise of Large Language Models (LLMs) introduces new opportunities and challenges, raising questions about how humans can efficiently provide actionable knowledge, how expert meta-knowledge can be leveraged, and what roles humans should play in the intelligent fuzzing loop with LLMs. To address these questions, we survey existing work on HITL fuzz testing and propose a research agenda emphasizing future opportunities in (1) human monitoring, (2) human steering, and (3) human-LLM collaboration. We call for a paradigm shift toward interactive, human-guided fuzzing systems that integrate expert insight with AI-powered automation in the next-generation fuzzing ecosystem.

cs.SE↗

arXiv2026

Demystifying the Lifecycle of Failures in Platform-Orchestrated Agentic Workflows

Agentic workflows built on low-code orchestration platforms enable rapid development of multi-agent systems, but they also introduce new and poorly understood failure modes that hinder reliability and maintainability. Unlike traditional software systems, failures in agentic workflows often propagate across heterogeneous nodes through natural-language interactions, tool invocations, and dynamic control logic, making failure attribution and repair particularly challenging. In this paper, we present an empirical study of platform-orchestrated agentic workflows from a failure lifecycle perspective, with the goal of characterizing failure manifestations, identifying underlying root causes, and examining corresponding repair strategies. We present AgentFail, a dataset of 307 real-world failure cases collected from two representative agentic workflow platforms. Based on this dataset, we analyze failure patterns, root causes, and repair difficulty for various failure root causes and nodes in the workflow. Our findings reveal key failure mechanisms in agentic workflows and provide actionable guidelines for reliable failure repair, and real-world agentic workflow design.

cs.AI↗

arXiv2026

What Makes a Good LLM Agent for Real-world Penetration Testing?

LLM-based agents show promise for automating penetration testing, yet reported performance varies widely across systems and benchmarks. We analyze 28 LLM-based penetration testing systems and evaluate five representative implementations across three benchmarks of increasing complexity. Our analysis reveals two distinct failure modes: Type A failures stem from capability gaps (missing tools, inadequate prompts) that engineering readily addresses, while Type B failures persist regardless of tooling due to planning and state management limitations. We show that Type B failures share a root cause that is largely invariant to the underlying LLM: agents lack real-time task difficulty estimation. As a result, agents misallocate effort, over-commit to low-value branches, and exhaust context before completing attack chains. Based on this insight, we present Excalibur, a penetration testing agent that couples strong tooling with difficulty-aware planning. A Tool and Skill Layer eliminates Type A failures through typed interfaces and retrieval-augmented knowledge. A Task Difficulty Assessment (TDA) mechanism addresses Type B failures by estimating tractability through four measurable dimensions (horizon estimation, evidence confidence, context load, and historical success) and uses these estimates to guide exploration-exploitation decisions within an Evidence-Guided Attack Tree Search (EGATS) framework. Excalibur achieves up to 91% task completion on CTF benchmarks with frontier models (39 to 49% relative improvement over baselines) and compromises 4 of 5 hosts on the GOAD Active Directory environment versus 2 by prior systems. These results show that difficulty-aware planning yields consistent end-to-end gains across models and addresses a limitation that model scaling alone does not eliminate.

cs.CR↗

arXiv2026

Understanding Codebase like a Professional! Human-AI Collaboration for Code Comprehension

Understanding an unfamiliar codebase is an essential task for developers in various scenarios, such as during the onboarding process. Especially when the codebase is large and time is limited, achieving a decent level of comprehension remains challenging for both experienced and novice developers, even with the assistance of LLMs. Existing studies have shown that LLMs often fail to support users in understanding code structures or to provide user-centered, adaptive, and dynamic assistance in real-world settings. To address this, we propose learning from the perspective of a unique role, code auditors, whose work often requires them to quickly familiarize themselves with new code projects on a weekly or even daily basis. To achieve this, we recruited and interviewed 8 code auditing practitioners to understand how they master codebase understanding. We identified four design opportunities for an LLM-based codebase understanding system: supporting cognitive alignment through automated codebase information extraction, decomposition, and representation, as well as reducing manual effort and conversational distraction through interaction design. To validate these four design opportunities, we designed a system prototype, CodeMap, that provides dynamic information extraction and representation aligned with the human cognitive flow and enables interactive switching among hierarchical codebase visualizations. We then conducted a user study with nine experienced developers and six novice developers. Our results demonstrate that CodeMap significantly improved users' perceived intuitiveness, ease of use, and usefulness in supporting code comprehension, while reducing their reliance on reading and interpreting LLM responses by 79% and increasing map usage time by 90% compared to the static visualization analysis tool.

cs.HC↗

arXiv2026

AtPatch: Debugging Transformers via Hot-Fixing Over-Attention

Transformer-based deep neural networks (DNNs) affected by backdoor attacks and unfairness typically exhibit anomalous attention patterns, leading to over-attend to backdoor triggers or protected attributes. Existing neuron-editing mitigation strategies often struggle to handle such situation and most of them lack flexibility and tend to distort feature representations. Motivated by such over-attention phenomenon and software engineering paradigms such as delta debugging and hot patching, we propose AtPatch, a hot-fix method that dynamically redistributes attention maps during model inference. Specifically, for a given input, AtPatch first extracts the attention map from the model's inference process. Then, it uses a pre-trained detector to identify anomalous columns and replace them with unified benign attention. Then, AtPatch rescales other columns to mitigate the impact of over-attention. Finally, AtPatch returns the redistributed attention map to the model for continued inference. Notably, if the detector does not report any anomalous columns, AtPatch directly returns the original attention map to the model. Unlike existing techniques, AtPatch selectively redistributes the attention map, making it better at preserving the model's original functionality. Furthermore, AtPatch's on-the-fly nature allows it to work without modifying model parameters or retraining, making it better suited for deployed models. We conducted extensive experiments to validate AtPatch. Experimental results show that, compared to existing methods, AtPatch can more effectively mitigate backdoor attacks and unfairness while better preserving the model's original functionality.

cs.SE↗

arXiv2026

GenCode: A Generic Data Augmentation Framework for Boosting Deep Learning-Based Code Understanding

Pre-trained code models lead the era of code intelligence, with multiple models designed with impressive performance. However, one important problem, data augmentation for code data that automatically helps developers prepare training data lacks study in this field. In this paper, we introduce a generic data augmentation framework, GenCode, to enhance the training of code understanding models. Simply speaking, GenCode follows a generation-and-selection paradigm to prepare useful training code data. Specifically, it employs code augmentation techniques to generate new code candidates first and then identifies important ones as the training data by influence scores. To evaluate the effectiveness of GenCode, we conduct experiments on four code understanding tasks (e.g., code clone detection) and three pre-trained code models (e.g., CodeT5) and two recent released code-specific Large Language Models (LLMs) (e.g., Qwen2.5-Coder). Compared to the state-of-the-art (SOTA) code augmentation method MixCode, GenCode produces pre-trained code models with 2.92% higher accuracy and 4.90% adversarial robustness on average. For code-specific LLMs, GenCode achieves an average improvement of 0.93% in accuracy and 0.98% in natural robustness.

cs.SE↗

arXiv2026

PTCBENCH: Benchmarking Contextual Stability of Personality Traits in LLM Systems

With the increasing deployment of large language models (LLMs) in affective agents and AI systems, maintaining a consistent and authentic LLM personality becomes critical for user trust and engagement. However, existing work overlooks a fundamental psychological consensus that personality traits are dynamic and context-dependent. To bridge this gap, we introduce PTCBENCH, a systematic benchmark designed to quantify the consistency of LLM personalities under controlled situational contexts. PTCBENCH subjects models to 12 distinct external conditions spanning diverse location contexts and life events, and rigorously assesses the personality using the NEO Five-Factor Inventory. Our study on 39,240 personality trait records reveals that certain external scenarios (e.g., "Unemployment") can trigger significant personality changes of LLMs, and even alter their reasoning capabilities. Overall, PTCBENCH establishes an extensible framework for evaluating personality consistency in realistic, evolving environments, offering actionable insights for developing robust and psychologically aligned AI systems.

cs.CL↗

arXiv2026

Chimera: Harnessing Multi-Agent LLMs for Automatic Insider Threat Simulation

Insider threats pose a persistent and critical security risk, yet are notoriously difficult to detect in complex enterprise environments, where malicious actions are often hidden within seemingly benign user behaviors. Although machine-learning-based insider threat detection (ITD) methods have shown promise, their effectiveness is fundamentally limited by the scarcity of high-quality and realistic training data. Enterprise internal data is highly sensitive and rarely accessible, while existing public and synthetic datasets are either small-scale or lack sufficient realism, semantic richness, and behavioral diversity. To address this challenge, we propose Chimera, an LLM-based multi-agent framework that automatically simulates both benign and malicious insider activities and generates comprehensive system logs across diverse enterprise environments. Chimera models each agent as an individual employee with fine-grained roles and supports group meetings, pairwise interactions, and self-organized scheduling to capture realistic organizational dynamics. Based on 15 insider attacks abstracted from real-world incidents, we deploy Chimera in three representative data-sensitive organizational scenarios and construct ChimeraLog, a new dataset for developing and evaluating ITD methods. We evaluate ChimeraLog through human studies and quantitative analyses, demonstrating its diversity and realism. Experiments with existing ITD methods show substantially lower detection performance on ChimeraLog compared to prior datasets, indicating a more challenging and realistic benchmark. Moreover, despite distribution shifts, models trained on ChimeraLog exhibit strong generalization, highlighting the practical value of LLM-based multi-agent simulation for advancing insider threat detection.

cs.CR↗

arXiv2026

Towards Understanding and Characterizing Vulnerabilities in Intelligent Connected Vehicles through Real-World Exploits

Intelligent Connected Vehicles (ICVs) are a core component of modern transportation systems, and their security is crucial as it directly relates to user safety. Despite prior research, most existing studies focus only on specific sub-components of ICVs due to their inherent complexity. As a result, there is a lack of systematic understanding of ICV vulnerabilities. Moreover, much of the current literature relies on human subjective analysis, such as surveys and interviews, which tends to be high-level and unvalidated, leaving a significant gap between theoretical findings and real-world attacks. To address this issue, we conducted the first large-scale empirical study on ICV vulnerabilities. We began by analyzing existing ICV security literature and summarizing the prevailing taxonomies in terms of vulnerability locations and types. To evaluate their real-world relevance, we collected a total of 649 exploitable vulnerabilities, including 592 from eight ICV vulnerability discovery competitions, Anonymous Cup, between January 2023 and April 2024, covering 48 different vehicles. The remaining 57 vulnerabilities were submitted daily by researchers. Based on this dataset, we assessed the coverage of existing taxonomies and identified several gaps, discovering one new vulnerability location and 13 new vulnerability types. We further categorized these vulnerabilities into 6 threat types (e.g., privacy data breach) and 4 risk levels (ranging from low to critical) and analyzed participants' skills and the types of ICVs involved in the competitions. This study provides a comprehensive and data-driven analysis of ICV vulnerabilities, offering actionable insights for researchers, industry practitioners, and policymakers. To support future research, we have made our vulnerability dataset publicly available.

cs.CR↗

arXiv2025

PentestEval: Benchmarking LLM-based Penetration Testing with Modular and Stage-Level Design

Penetration testing is essential for assessing and strengthening system security against real-world threats, yet traditional workflows remain highly manual, expertise-intensive, and difficult to scale. Although recent advances in Large Language Models (LLMs) offer promising opportunities for automation, existing applications rely on simplistic prompting without task decomposition or domain adaptation, resulting in unreliable black-box behavior and limited insight into model capabilities across penetration testing stages. To address this gap, we introduce PentestEval, the first comprehensive benchmark for evaluating LLMs across six decomposed penetration testing stages: Information Collection, Weakness Gathering and Filtering, Attack Decision-Making, Exploit Generation and Revision. PentestEval integrates expert-annotated ground truth with a fully automated evaluation pipeline across 346 tasks covering all stages in 12 realistic vulnerable scenarios. Our stage-level evaluation of 9 widely used LLMs reveals generally weak performance and distinct limitations across the stages of penetration-testing workflow. End-to-end pipelines reach only 31% success rate, and existing LLM-powered systems such as PentestGPT, PentestAgent, and VulnBot exhibit similar limitations, with autonomous agents failing almost entirely. These findings highlight that autonomous penetration testing demands stronger structured reasoning, where modularization enhances each individual stage and improves overall performance. PentestEval provides the foundational benchmark needed for future research on fine-grained, stage-level evaluation, paving the way toward more reliable LLM-based automation.

cs.SE↗

arXiv2025

Defects4C: Benchmarking Large Language Model Repair Capability with C/C++ Bugs

Automated Program Repair (APR) plays a critical role in enhancing the quality and reliability of software systems. While substantial progress has been made in Java-based APR, largely facilitated by benchmarks like Defects4J, there remains a significant gap in research on C/C++ program repair, despite the widespread use of C/C++ and the prevalence of associated vulnerabilities. This gap is primarily due to the lack of high-quality, open-source benchmarks tailored for C/C++. To address this issue, we introduce Defects4C, a comprehensive and executable benchmark specifically designed for C/C++ program repair. Our dataset is constructed from real-world C/C++ repositories and includes a large collection of bug-relevant commits (9M in total), 248 high-quality buggy functions, and 102 vulnerable functions, all paired with test cases for reproduction. These resources enable rigorous evaluation of repair techniques and support the retraining of learning-based approaches for enhanced performance. Using Defects4C, we conduct a comprehensive empirical study evaluating the effectiveness of 24 state-of-the-art large language models (LLMs) in repairing C/C++ faults. Our findings offer valuable insights into the strengths and limitations of current LLM-based APR techniques in this domain, highlighting both the need for more robust methods and the critical role of Defects4C in advancing future research

cs.SE↗

arXiv2025

Enhancing Automated Program Repair via Faulty Token Localization and Quality-Aware Patch Refinement

Large language models (LLMs) have recently demonstrated strong potential for automated program repair (APR). However, existing LLM-based techniques primarily rely on coarse-grained external feedback (e.g.,test results) to guide iterative patch generation, while lacking fine-grained internal signals that reveal why a patch fails or which parts of the generated code are likely incorrect. This limitation often leads to inefficient refinement, error propagation, and suboptimal repair performance. In this work, we propose TokenRepair, a novel two-level refinement framework that enhances APR by integrating internal reflection for localizing potentially faulty tokens with external feedback for quality-aware patch refinement. Specifically, TokenRepair first performs internal reflection by analyzing context-aware token-level uncertainty fluctuations to identify suspicious or low-confidence tokens within a patch. It then applies Chain-of-Thought guided rewriting to refine only these localized tokens, enabling targeted and fine-grained correction. To further stabilize the iterative repair loop, TokenRepair incorporates a quality-aware external feedback mechanism that evaluates patch quality and filters out low-quality candidates before refinement. Experimental results show that TokenRepair achieves new state-of-the-art repair performance, correctly fixing 88 bugs on Defects4J 1.2 and 139 bugs on HumanEval-Java, demonstrating substantial improvements ranging from 8.2% to 34.9% across all models on Defects4J 1.2 and from 3.3% to 16.1% on HumanEval-Java.

cs.SE↗

arXiv2025

Adversarial Attack on Black-Box Multi-Agent by Adaptive Perturbation

Evaluating security and reliability for multi-agent systems (MAS) is urgent as they become increasingly prevalent in various applications. As an evaluation technique, existing adversarial attack frameworks face certain limitations, e.g., impracticality due to the requirement of white-box information or high control authority, and a lack of stealthiness or effectiveness as they often target all agents or specific fixed agents. To address these issues, we propose AdapAM, a novel framework for adversarial attacks on black-box MAS. AdapAM incorporates two key components: (1) Adaptive Selection Policy simultaneously selects the victim and determines the anticipated malicious action (the action would lead to the worst impact on MAS), balancing effectiveness and stealthiness. (2) Proxy-based Perturbation to Induce Malicious Action utilizes generative adversarial imitation learning to approximate the target MAS, allowing AdapAM to generate perturbed observations using white-box information and thus induce victims to execute malicious action in black-box settings. We evaluate AdapAM across eight multi-agent environments and compare it with four state-of-the-art and commonly-used baselines. Results demonstrate that AdapAM achieves the best attack performance in different perturbation rates. Besides, AdapAM-generated perturbations are the least noisy and hardest to detect, emphasizing the stealthiness.

cs.MA↗

arXiv2025

AutoEmpirical: LLM-Based Automated Research for Empirical Software Fault Analysis

Understanding software faults is essential for empirical research in software development and maintenance. However, traditional fault analysis, while valuable, typically involves multiple expert-driven steps such as collecting potential faults, filtering, and manual investigation. These processes are both labor-intensive and time-consuming, creating bottlenecks that hinder large-scale fault studies in complex yet critical software systems and slow the pace of iterative empirical research. In this paper, we decompose the process of empirical software fault study into three key phases: (1) research objective definition, (2) data preparation, and (3) fault analysis, and we conduct an initial exploration study of applying Large Language Models (LLMs) for fault analysis of open-source software. Specifically, we perform the evaluation on 3,829 software faults drawn from a high-quality empirical study. Our results show that LLMs can substantially improve efficiency in fault analysis, with an average processing time of about two hours, compared to the weeks of manual effort typically required. We conclude by outlining a detailed research plan that highlights both the potential of LLMs for advancing empirical fault studies and the open challenges that required be addressed to achieve fully automated, end-to-end software fault analysis.

cs.SE↗