arXiv subjects

Chunqiu Steven Xia

Publications and source records attributed to Chunqiu Steven Xia.

17 recordsLinked to original sources

arXiv2026

SEC-bench Pro: Can Language Models Solve Long-Horizon Software Security Tasks?

Finding a real vulnerability in complicated systems is a challenging, long-horizon task that demands reasoning across an entire codebase to produce a working proof-of-concept (PoC). However, such critical security problems remain understudied. We present SEC-bench Pro, a benchmark that measures how well frontier models hunt real vulnerabilities by reproducing working PoC inputs from disclosed reports, where each task pairs a concrete bug with the instructions for triggering it. We also demonstrate the limitations of existing rule-based judges for grading generated PoCs, and propose a novel LLM-based judge for more precise grading. We instantiate SEC-bench Pro with 344 validated vulnerabilities across three targets, the V8 and SpiderMonkey browser engines and the Linux kernel, covering critical vulnerability families including memory-safety, sandbox, JIT, race-condition, and kernel-subsystem bugs. Across six frontier commercial and open-weight models and three coding agents, the strongest, Codex with GPT-5.5, solves 58% of instances overall. We also observe that Claude Code with Opus 4.6 tends to time out but solves most instances it completes. In contrast, open-weight models struggle; for example, GLM-5 solves only 13 of the 344 instances. During construction and evaluation, SEC-bench Pro also surfaced three vulnerabilities in V8 and SpiderMonkey, including a sandbox escape that was fixed and earned a $20,000 Google Vulnerability Reward Program bounty. More recently, SEC-bench Pro has been adopted by OpenAI to evaluate the long-horizon security capabilities of its newest models. Overall, SEC-bench Pro exposes where long-horizon vulnerability discovery succeeds, where it fails, and how different grading choices change the evaluation landscape, offering insights for security-centric model evaluation and training. Our artifact is available at https://github.com/SEC-bench/SEC-bench-Pro.

cs.CR↗

arXiv2026

Do These Violent Delights Have Violent Ends? Measuring the Post-Merge Fate of Agentic Code

Agentic coding tools are increasingly used to make autonomous repository-level changes to real-world projects. Prior work has largely evaluated these contributions at the pre-merge stage, through outcomes such as pull request acceptance and review effort. Far less is known about what happens to agentic code post-merge. Yet merge success alone does not reveal whether a contribution will remain stable or require bug fixes and other corrective maintenance downstream. We conduct a longitudinal empirical analysis of agentic and human contributions across 182 repositories, tracking their post-merge fate over time, characterizing the intent of subsequent modifications, and analyzing the defects and vulnerabilities they introduce. While the overall maintenance rates are similar, agentic contributions require significantly higher rates of corrective maintenance and introduce more security weaknesses and dependency vulnerabilities. We also find statistically significant evidence that agentic maintenance burden is associated with repository characteristics. In particular, each 10 percentage-point increase in a project's no-review rate is associated with roughly a 6% increase in agentic maintenance burden on average. As coding agents become pervasive in software development, our findings highlight the need to evaluate and design agentic tools not only to produce mergeable changes, but to produce contributions that remain secure and maintainable.

cs.SE↗

arXiv2026

SwarmResearch: Orchestrating Coding Agents for Open-Ended Discovery

Long-running coding agents such as autoresearch can persistently discover optimizations for open-ended problems. However, they tend to converge onto a single high-level approach, then proceed with low-level edits while missing other superior approaches to the problem. We hypothesize two harness-level design choices contribute to this behavior: accumulating context in a single long-running agent and only exposing a single program state to edit. We introduce SwarmResearch, an orchestrator-subagent harness in which a Shepherd Agent uses global context to steer a population of Search Agents, each operating with local context in their respective git branch. On open-ended optimization tasks, SwarmResearch discovers better or comparable solutions to state-of-the-art LLM-guided evolution and multi-agent techniques on 13/15 tasks, driven by higher-level exploration. Compared with fixed scaling of serial and parallel agents, SwarmResearch's orchestrator-guided scaling discovers better-performing solutions by adapting parallelism at different search depths.

cs.AI↗

arXiv2025

Live-SWE-agent: Can Software Engineering Agents Self-Evolve on the Fly?

Large Language Models (LLMs) are reshaping almost all industries, including software engineering. In recent years, a number of LLM agents have been proposed to solve real-world software problems. Such software agents are typically equipped with a suite of coding tools and can autonomously decide the next actions to form complete trajectories to solve end-to-end software tasks. While promising, they typically require dedicated design and may still be suboptimal, since it can be extremely challenging and costly to exhaust the entire agent scaffold design space. Recognizing that software agents are inherently software themselves that can be further refined/modified, researchers have proposed a number of self-improving software agents recently, including the Darwin-Gödel Machine (DGM). Meanwhile, such self-improving agents require costly offline training on specific benchmarks and may not generalize well across different LLMs or benchmarks. In this paper, we propose Live-SWE-agent, the first live software agent that can autonomously and continuously evolve itself on-the-fly during runtime when solving real-world software problems. More specifically, Live-SWE-agent starts with the most basic agent scaffold with only access to bash tools (e.g., mini-SWE-agent), and autonomously evolves its own scaffold implementation while solving real-world software problems. Our evaluation on the widely studied SWE-bench Verified benchmark shows that LIVE-SWE-AGENT can achieve an impressive solve rate of 77.4% without test-time scaling, outperforming all existing software agents, including the best proprietary solution. Moreover, Live-SWE-agent outperforms state-of-the-art manually crafted software agents on the recent SWE-Bench Pro benchmark, achieving the best-known solve rate of 45.8%.

cs.SE↗

arXiv2025

UniDebugger: Hierarchical Multi-Agent Framework for Unified Software Debugging

Software debugging is a time-consuming endeavor involving a series of steps, such as fault localization and patch generation, each requiring thorough analysis and a deep understanding of the underlying logic. While large language models (LLMs) demonstrate promising potential in coding tasks, their performance in debugging remains limited. Current LLM-based methods often focus on isolated steps and struggle with complex bugs. In this paper, we propose the first end-to-end framework, FixAgent, for unified debugging through multi-agent synergy. It mimics the entire cognitive processes of developers, with each agent specialized as a particular component of this process rather than mirroring the actions of an independent expert as in previous multi-agent systems. Agents are coordinated through a three-level design, following a cognitive model of debugging, allowing adaptive handling of bugs with varying complexities. Experiments on extensive benchmarks demonstrate that FixAgent significantly outperforms state-of-the-art repair methods, fixing 1.25$\times$ to 2.56$\times$ bugs on the repo-level benchmark, Defects4J. This performance is achieved without requiring ground-truth root-cause code statements, unlike the baselines. Our source code is available on https://github.com/AcceptePapier/UniDebugger.

cs.SE↗

arXiv2024

Revisiting the Plastic Surgery Hypothesis via Large Language Models

Automated Program Repair (APR) aspires to automatically generate patches for an input buggy program. Traditional APR tools typically focus on specific bug types and fixes through the use of templates, heuristics, and formal specifications. However, these techniques are limited in terms of the bug types and patch variety they can produce. As such, researchers have designed various learning-based APR tools with recent work focused on directly using Large Language Models (LLMs) for APR. While LLM-based APR tools are able to achieve state-of-the-art performance on many repair datasets, the LLMs used for direct repair are not fully aware of the project-specific information such as unique variable or method names. The plastic surgery hypothesis is a well-known insight for APR, which states that the code ingredients to fix the bug usually already exist within the same project. Traditional APR tools have largely leveraged the plastic surgery hypothesis by designing manual or heuristic-based approaches to exploit such existing code ingredients. However, as recent APR research starts focusing on LLM-based approaches, the plastic surgery hypothesis has been largely ignored. In this paper, we ask the following question: How useful is the plastic surgery hypothesis in the era of LLMs? Interestingly, LLM-based APR presents a unique opportunity to fully automate the plastic surgery hypothesis via fine-tuning and prompting. To this end, we propose FitRepair, which combines the direct usage of LLMs with two domain-specific fine-tuning strategies and one prompting strategy for more powerful APR. Our experiments on the widely studied Defects4j 1.2 and 2.0 datasets show that FitRepair fixes 89 and 44 bugs (substantially outperforming the best-performing baseline by 15 and 8), respectively, demonstrating a promising future of the plastic surgery hypothesis in the era of LLMs.

cs.SE↗

arXiv2024

Practical Program Repair in the Era of Large Pre-trained Language Models

Automated Program Repair (APR) aims to help developers automatically patch software bugs. However, current state-of-the-art traditional and learning-based APR techniques face the problem of limited patch variety, failing to fix complicated bugs. This is mainly due to the reliance on bug-fixing datasets to craft fix templates or directly predict potential patches. Large Pre-Trained Language Models (PLMs), trained using billions of text/code tokens, can potentially help avoid this issue. Very recently, researchers have directly leveraged PLMs for APR without relying on any bug-fixing datasets. Meanwhile, such existing work either failed to include state-of-the-art PLMs or was not evaluated on realistic datasets. In this work, we perform the first extensive study on directly applying PLMs for APR. We select 9 recent state-of-the-art PLMs, including both generative and infilling models, ranging from 125M to 20B in size. We designed 3 different repair settings to evaluate the different ways we can use PLMs to generate patches. We apply the PLMs under these repair settings on 5 datasets across 3 different languages and compare different PLMs in the number of bugs fixed, generation speed and compilation rate. Our study demonstrates that directly applying state-of-the-art PLMs can already substantially outperform all existing APR techniques on all our datasets. Among the studied PLMs, the scaling effect exists for APR where larger models tend to achieve better performance. Also, we show for the first time that suffix code after the buggy line (adopted in infilling-style APR) is important in not only generating more fixes but more patches with higher compilation rate. Besides patch generation, the PLMs consider correct patches to be more natural than other ones, and can even be leveraged for effective patch ranking or patch correctness checking.

cs.SE↗

arXiv2024

Fuzz4All: Universal Fuzzing with Large Language Models

Fuzzing has achieved tremendous success in discovering bugs and vulnerabilities in various software systems. Systems under test (SUTs) that take in programming or formal language as inputs, e.g., compilers, runtime engines, constraint solvers, and software libraries with accessible APIs, are especially important as they are fundamental building blocks of software development. However, existing fuzzers for such systems often target a specific language, and thus cannot be easily applied to other languages or even other versions of the same language. Moreover, the inputs generated by existing fuzzers are often limited to specific features of the input language, and thus can hardly reveal bugs related to other or new features. This paper presents Fuzz4All, the first fuzzer that is universal in the sense that it can target many different input languages and many different features of these languages. The key idea behind Fuzz4All is to leverage large language models (LLMs) as an input generation and mutation engine, which enables the approach to produce diverse and realistic inputs for any practically relevant language. To realize this potential, we present a novel autoprompting technique, which creates LLM prompts that are wellsuited for fuzzing, and a novel LLM-powered fuzzing loop, which iteratively updates the prompt to create new fuzzing inputs. We evaluate Fuzz4All on nine systems under test that take in six different languages (C, C++, Go, SMT2, Java and Python) as inputs. The evaluation shows, across all six languages, that universal fuzzing achieves higher coverage than existing, language-specific fuzzers. Furthermore, Fuzz4All has identified 98 bugs in widely used systems, such as GCC, Clang, Z3, CVC5, OpenJDK, and the Qiskit quantum computing platform, with 64 bugs already confirmed by developers as previously unknown.

cs.SE↗

arXiv2024

Keep the Conversation Going: Fixing 162 out of 337 bugs for $0.42 each using ChatGPT

Automated Program Repair (APR) aims to automatically generate patches for buggy programs. Recent APR work has been focused on leveraging modern Large Language Models (LLMs) to directly generate patches for APR. Such LLM-based APR tools work by first constructing an input prompt built using the original buggy code and then queries the LLM to generate patches. While the LLM-based APR tools are able to achieve state-of-the-art results, it still follows the classic Generate and Validate repair paradigm of first generating lots of patches and then validating each one afterwards. This not only leads to many repeated patches that are incorrect but also miss the crucial information in test failures as well as in plausible patches. To address these limitations, we propose ChatRepair, the first fully automated conversation-driven APR approach that interleaves patch generation with instant feedback to perform APR in a conversational style. ChatRepair first feeds the LLM with relevant test failure information to start with, and then learns from both failures and successes of earlier patching attempts of the same bug for more powerful APR. For earlier patches that failed to pass all tests, we combine the incorrect patches with their corresponding relevant test failure information to construct a new prompt for the LLM to generate the next patch. In this way, we can avoid making the same mistakes. For earlier patches that passed all the tests, we further ask the LLM to generate alternative variations of the original plausible patches. In this way, we can further build on and learn from earlier successes to generate more plausible patches to increase the chance of having correct patches. While our approach is general, we implement ChatRepair using state-of-the-art dialogue-based LLM -- ChatGPT. By calculating the cost of accessing ChatGPT, we can fix 162 out of 337 bugs for \$0.42 each!

cs.SE↗

arXiv2024

Less Training, More Repairing Please: Revisiting Automated Program Repair via Zero-shot Learning

Due to the promising future of Automated Program Repair (APR), researchers have proposed various APR techniques, including heuristic-based, template-based, and constraint-based techniques. Among such classic APR techniques, template-based techniques have been widely recognized as state of the art. However, such template-based techniques require predefined templates to perform repair, and their effectiveness is thus limited. To this end, researchers leveraged the recent advances in Deep Learning to further improve APR. Such learning-based techniques view APR as a Neural Machine Translation problem, using the buggy/fixed code snippets as the source/target languages for translation. In this way, such techniques heavily rely on large numbers of high-quality bug-fixing commits, which can be extremely costly and challenging to construct. Furthermore, the edit variety of these learning-based techniques are limited to the available bug-fixes within their training datasets. Therefore, in this paper, we aim to revisit the learning-based APR problem, and propose AlphaRepair, to leverage zero-shot learning directly using large pre-trained code models for APR. Our main insight is instead of modeling what a repair edit should look like, we can directly predict what the correct code is based on the context information. We have implemented AlphaRepair as a practical multilingual APR tool based on the recent CodeBERT model. Our results on the widely used Defects4J benchmark show that AlphaRepair can substantially outperform state-of-the-art APR tools. We also studied the impact of different design choices and show that AlphaRepair performs even better on a newer version of Defects4J (2.0) with 3.3X more fixes than best performing baseline, indicating that AlphaRepair can potentially avoid the dataset-overfitting issue of existing learning-based techniques.

cs.SE↗

arXiv2024

Agentless: Demystifying LLM-based Software Engineering Agents

Recent advancements in large language models (LLMs) have significantly advanced the automation of software development tasks, including code synthesis, program repair, and test generation. More recently, researchers and industry practitioners have developed various autonomous LLM agents to perform end-to-end software development tasks. These agents are equipped with the ability to use tools, run commands, observe feedback from the environment, and plan for future actions. However, the complexity of these agent-based approaches, together with the limited abilities of current LLMs, raises the following question: Do we really have to employ complex autonomous software agents? To attempt to answer this question, we build Agentless -- an agentless approach to automatically solve software development problems. Compared to the verbose and complex setup of agent-based approaches, Agentless employs a simplistic three-phase process of localization, repair, and patch validation, without letting the LLM decide future actions or operate with complex tools. Our results on the popular SWE-bench Lite benchmark show that surprisingly the simplistic Agentless is able to achieve both the highest performance (32.00%, 96 correct fixes) and low cost ($0.70) compared with all existing open-source software agents! Furthermore, we manually classified the problems in SWE-bench Lite and found problems with exact ground truth patch or insufficient/misleading issue descriptions. As such, we construct SWE-bench Lite-S by excluding such problematic issues to perform more rigorous evaluation and comparison. Our work highlights the current overlooked potential of a simple, interpretable technique in autonomous software development. We hope Agentless will help reset the baseline, starting point, and horizon for autonomous software agents, and inspire future work along this crucial direction.

cs.SE↗

arXiv2024

Top Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM

LLMs have become the go-to choice for code generation tasks, with an exponential increase in the training, development, and usage of LLMs specifically for code generation. To evaluate the ability of LLMs on code, both academic and industry practitioners rely on popular handcrafted benchmarks. However, prior benchmarks contain only a very limited set of problems, both in quantity and variety. Further, due to popularity and age, many benchmarks are prone to data leakage where example solutions can be readily found on the web and thus potentially in training data. Such limitations inevitably lead us to inquire: Is the leaderboard performance on existing benchmarks reliable and comprehensive enough to measure the program synthesis ability of LLMs? To address this, we introduce EvoEval -- a program synthesis benchmark suite created by evolving existing benchmarks into different targeted domains for a comprehensive evaluation of LLM coding abilities. Our study on 51 LLMs shows that compared to the high performance obtained on standard benchmarks like HumanEval, there is a significant drop in performance (on average 39.4%) when using EvoEval. Additionally, the decrease in performance can range from 19.6% to 47.7%, leading to drastic ranking changes amongst LLMs and showing potential overfitting of existing benchmarks. Furthermore, we showcase various insights, including the brittleness of instruction-following models when encountering rewording or subtle changes as well as the importance of learning problem composition and decomposition. EvoEval not only provides comprehensive benchmarks, but can be used to further evolve arbitrary problems to keep up with advances and the ever-changing landscape of LLMs for code. We have open-sourced our benchmarks, tools, and complete LLM generations at https://github.com/evo-eval/evoeval

cs.SE↗

arXiv2023

Copiloting the Copilots: Fusing Large Language Models with Completion Engines for Automated Program Repair

During Automated Program Repair (APR), it can be challenging to synthesize correct patches for real-world systems in general-purpose programming languages. Recent Large Language Models (LLMs) have been shown to be helpful "copilots" in assisting developers with various coding tasks, and have also been directly applied for patch synthesis. However, most LLMs treat programs as sequences of tokens, meaning that they are ignorant of the underlying semantics constraints of the target programming language. This results in plenty of statically invalid generated patches, impeding the practicality of the technique. Therefore, we propose Repilot, a general code generation framework to further copilot the AI "copilots" (i.e., LLMs) by synthesizing more valid patches during the repair process. Our key insight is that many LLMs produce outputs autoregressively (i.e., token by token), resembling human writing programs, which can be significantly boosted and guided through a Completion Engine. Repilot synergistically synthesizes a candidate patch through the interaction between an LLM and a Completion Engine, which 1) prunes away infeasible tokens suggested by the LLM and 2) proactively completes the token based on the suggestions provided by the Completion Engine. Our evaluation on a subset of the widely-used Defects4j 1.2 and 2.0 datasets shows that Repilot outperforms state-of-the-art techniques by fixing 27% and 47% more bugs, respectively. Moreover, Repilot produces more valid and correct patches than the base LLM with the same budget. While we focus on leveraging Repilot for APR in this work, the overall approach is also generalizable to other code generation tasks.

cs.SE↗

arXiv2023

Is Your Code Generated by ChatGPT Really Correct? Rigorous Evaluation of Large Language Models for Code Generation

Program synthesis has been long studied with recent approaches focused on directly using the power of Large Language Models (LLMs) to generate code. Programming benchmarks, with curated synthesis problems and test-cases, are used to measure the performance of various LLMs on code synthesis. However, these test-cases can be limited in both quantity and quality for fully assessing the functional correctness of the generated code. Such limitation in the existing benchmarks begs the following question: In the era of LLMs, is the code generated really correct? To answer this, we propose EvalPlus -- a code synthesis evaluation framework to rigorously benchmark the functional correctness of LLM-synthesized code. EvalPlus augments a given evaluation dataset with large amounts of test-cases newly produced by an automatic test input generator, powered by both LLM- and mutation-based strategies. While EvalPlus is general, we extend the test-cases of the popular HumanEval benchmark by 80x to build HumanEval+. Our extensive evaluation across 26 popular LLMs (e.g., GPT-4 and ChatGPT) demonstrates that HumanEval+ is able to catch significant amounts of previously undetected wrong code synthesized by LLMs, reducing the pass@k by up-to 19.3-28.9%. We also surprisingly found that test insufficiency can lead to mis-ranking. For example, both WizardCoder-CodeLlama and Phind-CodeLlama now outperform ChatGPT on HumanEval+, while none of them could on HumanEval. Our work not only indicates that prior popular code synthesis evaluation results do not accurately reflect the true performance of LLMs for code synthesis, but also opens up a new direction to improve such programming benchmarks through automated testing. We have open-sourced our tools, enhanced datasets as well as all LLM-generated code at https://github.com/evalplus/evalplus to facilitate and accelerate future LLM-for-code research.

cs.SE↗

arXiv2023

Large Language Models are Edge-Case Fuzzers: Testing Deep Learning Libraries via FuzzGPT

Deep Learning (DL) library bugs affect downstream DL applications, emphasizing the need for reliable systems. Generating valid input programs for fuzzing DL libraries is challenging due to the need for satisfying both language syntax/semantics and constraints for constructing valid computational graphs. Recently, the TitanFuzz work demonstrates that modern Large Language Models (LLMs) can be directly leveraged to implicitly learn all the constraints to generate valid DL programs for fuzzing. However, LLMs tend to generate ordinary programs following similar patterns seen in their massive training corpora, while fuzzing favors unusual inputs that cover edge cases or are unlikely to be manually produced. To fill this gap, this paper proposes FuzzGPT, the first technique to prime LLMs to synthesize unusual programs for fuzzing. FuzzGPT is built on the well-known hypothesis that historical bug-triggering programs may include rare/valuable code ingredients important for bug finding. Traditional techniques leveraging such historical information require intensive human efforts to design dedicated generators and ensure the validity of generated programs. FuzzGPT demonstrates that this process can be fully automated via the intrinsic capabilities of LLMs (including fine-tuning and in-context learning), while being generalizable and applicable to challenging domains. While FuzzGPT can be applied with different LLMs, this paper focuses on the powerful GPT-style models: Codex and CodeGen. Moreover, FuzzGPT also shows the potential of directly leveraging the instruct-following capability of the recent ChatGPT for effective fuzzing. Evaluation on two popular DL libraries (PyTorch and TensorFlow) shows that FuzzGPT can substantially outperform TitanFuzz, detecting 76 bugs, with 49 already confirmed as previously unknown bugs, including 11 high-priority bugs or security vulnerabilities.

cs.SE↗

arXiv2023

Large Language Models are Zero-Shot Fuzzers: Fuzzing Deep-Learning Libraries via Large Language Models

Detecting bugs in Deep Learning (DL) libraries (e.g., TensorFlow/PyTorch) is critical for almost all downstream DL systems in ensuring effectiveness/safety for end users. Meanwhile, traditional fuzzing techniques can be hardly effective for such a challenging domain since the input DL programs need to satisfy both the input language (e.g., Python) syntax/semantics and the DL API input/shape constraints for tensor computations. To address these limitations, we propose TitanFuzz - the first approach to directly leveraging Large Language Models (LLMs) to generate input programs for fuzzing DL libraries. LLMs are titanic models trained on billions of code snippets and can auto-regressively generate human-like code snippets. Our key insight is that modern LLMs can also include numerous code snippets invoking DL library APIs in their training corpora, and thus can implicitly learn both language syntax/semantics and intricate DL API constraints for valid DL program generation. More specifically, we use both generative and infilling LLMs (e.g., Codex/InCoder) to generate and mutate valid/diverse input DL programs for fuzzing. Our experimental results demonstrate that TitanFuzz can achieve 30.38%/50.84% higher code coverage than state-of-the-art fuzzers on TensorFlow/PyTorch. Furthermore, TitanFuzz is able to detect 65 bugs, with 41 already confirmed as previously unknown bugs. This paper demonstrates that modern titanic LLMs can be leveraged to directly perform both generation-based and mutation-based fuzzing studied for decades, while being fully automated, generalizable, and applicable to domains challenging for traditional approaches (such as DL systems). We hope TitanFuzz can stimulate more work in this promising direction of LLMs for fuzzing.

cs.SE↗

arXiv2023

Conversational Automated Program Repair

Automated Program Repair (APR) can help developers automatically generate patches for bugs. Due to the impressive performance obtained using Large Pre-Trained Language Models (LLMs) on many code related tasks, researchers have started to directly use LLMs for APR. However, prior approaches simply repeatedly sample the LLM given the same constructed input/prompt created from the original buggy code, which not only leads to generating the same incorrect patches repeatedly but also miss the critical information in testcases. To address these limitations, we propose conversational APR, a new paradigm for program repair that alternates between patch generation and validation in a conversational manner. In conversational APR, we iteratively build the input to the model by combining previously generated patches with validation feedback. As such, we leverage the long-term context window of LLMs to not only avoid generating previously incorrect patches but also incorporate validation feedback to help the model understand the semantic meaning of the program under test. We evaluate 10 different LLM including the newly developed ChatGPT model to demonstrate the improvement of conversational APR over the prior LLM for APR approach.

cs.SE↗