Showing 1–50 of 341 results for author: Lo, D

CoSA: Context-Aware Severity Assessment via Context Analysis with Large Language Models

Authors: Jinfeng Jiang, Yikun Li, Chengran Yang, Ting Zhang, Wen Bin Leow, Yide Yin, Eng Lieh Ouh, Lwin Khin Shar, David Lo

Abstract: Accurate vulnerability severity assessment is essential for prioritizing remediation, yet manually assessing Common Vulnerability Scoring System (CVSS) base metrics remains labor-intensive. Existing automated approaches often fail to capture the repository-level evidence required for assessing many CVSS base metrics. Such repository-aware assessment is challenging because relevant evidence is scat… ▽ More Accurate vulnerability severity assessment is essential for prioritizing remediation, yet manually assessing Common Vulnerability Scoring System (CVSS) base metrics remains labor-intensive. Existing automated approaches often fail to capture the repository-level evidence required for assessing many CVSS base metrics. Such repository-aware assessment is challenging because relevant evidence is scattered across the entire repository under heavy noise. To address these challenges, we present CoSA, a Context-aware vulnerability Severity Assessment approach that infers CVSS base metrics from repository artifacts. CoSA constructs a code property graph (CPG) and applies a two-stage repository-pruning strategy: lightweight static pruning to preserve structurally proximal context, followed by an agentic large language model (LLM)-guided pruning step to retain CVSS-relevant context while collecting supporting evidence. The LLM then consolidates the retrieved repository context into compact, CVSS metric-wise textual summaries, which are fed into a lightweight transformer predictor. We also construct a higher-quality repository-level dataset comprising 6,816 CVSS labeled instances spanning 90 Common Weakness Enumeration (CWE) types. Experiments on real-world vulnerabilities show that CoSA consistently outperforms function-level and pure-LLM baselines. It improves prediction accuracy by 14.4% and Macro-F1 by 15.3% over the best-performing baseline, suggesting that explicit, metric-oriented repository context retrieval is crucial for practical and reliable automated severity assessment. △ Less

Submitted 13 August, 2026; originally announced August 2026.

Second Thought: Reasoning in Parallel as LLM Agents Act and Observe

Authors: Zhensu Sun, Chengran Yang, Yunbo Lyu, Jieke Shi, David Lo

Abstract: LLM agents in the ReAct paradigm alternate between reasoning, acting, and observing, but deliberate reasoning is confined to the Thought phase: while the agent serializes an action and waits for the environment, its reasoning is frozen. We identify this recurring interval for Action and Observation as a reasoning idle window and ask whether it can host additional reasoning in parallel that serves… ▽ More LLM agents in the ReAct paradigm alternate between reasoning, acting, and observing, but deliberate reasoning is confined to the Thought phase: while the agent serializes an action and waits for the environment, its reasoning is frozen. We identify this recurring interval for Action and Observation as a reasoning idle window and ask whether it can host additional reasoning in parallel that serves future turns. Therefore, we propose Second Thought, a training-free inference framework that forks four auxiliary branches the instant each Thought phase concludes, decodes them concurrently with the main loop, and merges the generated thoughts back when the environment observation arrives. In this way, Second Thought relocates the added reasoning off the main thread's sequential decoding path. Across three agentic benchmarks and three reasoning LLMs, Second Thought lowers the average turn count in all nine (model,benchmark) pairs and reduces main thread decoding in six of them by up to 43% (roughly 20% on average among those settings), while leaving it essentially unchanged in a seventh; Pass@1 shows no significant change in seven of nine pairs and the two significant differences are +12.4 and +10.2 points. Against a compute-matched control that forces an equivalent budget onto the main thread's own reasoning, it attains strictly higher Pass@1 with 1.3 to 3.2 less sequential decoding in all four settings where the control applies. △ Less

Submitted 13 August, 2026; originally announced August 2026.

A Unified Issue Resolution Benchmark for Requirement Clarification, Planning, and Code Generation for Coding Agents

Authors: Xin Zhou, Chun Yong Chong, Kisub Kim, Yun Peng, Rui Shu, Zihan Wu, Xu Han, Guowen Yuan, Zeyang Zhuang, Jounghoon Kim, Jeongjin Ju, Seongmin Ju, Taein Yoon, David Lo

Abstract: Large language model-powered coding agents are increasingly used to modify existing code repositories, for example, by adding features or fixing bugs. Yet existing repository-level benchmarks typically evaluate only whether the final patch passes tests. Satisfying a user request requires a long chain of interdependent reasoning and decisions: an agent must recover explicit and implicit requirement… ▽ More Large language model-powered coding agents are increasingly used to modify existing code repositories, for example, by adding features or fixing bugs. Yet existing repository-level benchmarks typically evaluate only whether the final patch passes tests. Satisfying a user request requires a long chain of interdependent reasoning and decisions: an agent must recover explicit and implicit requirements, formulate a repository-grounded implementation plan, and translate it into correct code. A pass/fail outcome cannot characterize how an unsuccessful trajectory diverges from the requirements and implementation process needed for a correct patch. To address this gap, we introduce SWE-RPG, a repository-level benchmark that combines executable patch evaluation with validated ground-truth references (GTs) for (1) Requirement Clarification and (2) Implementation Planning. These intermediate GTs support retrospective, GT-aligned diagnosis of complete coding-agent trajectories across clarification, planning, code generation, and artifact submission. SWE-RPG comprises 163 tasks from 31 Python and Java repositories, including 113 bug fixes and 50 feature additions. We evaluate 3 coding agents, including Claude Code, Codex, and OpenCode, with 6 large language model backends, including Claude-Sonnet-5 and GPT-5.6-Terra. Results show that the evaluated popular coding agents still struggle to implement user requests in existing repositories, achieving an average resolved rate of only 31.5% on SWE-RPG. Intermediate-GT diagnosis further identifies implicit requirement recovery as the main bottleneck, accounting for 24.5%--46.0% of agent runs. This result suggests implicit-requirement recovery as a key candidate direction for improving coding agents. The benchmark data and evaluation code are available at https://github.com/Xin-Zhou-smu/SWE-RPG-Bench. △ Less

Submitted 9 August, 2026; originally announced August 2026.

Comments: 9 pages

Beyond Text Matching: Towards Reference-Free Evaluation for Human-Oriented Binary Reverse Engineering

Authors: Xiuwei Shang, Li Hu, Xiao Jiang, Jieke Shi, Junda He, Zhou Yang, Shaoyin Cheng, Guoqiang Chen, Weiming Zhang, David Lo

Abstract: Human-Oriented Binary Reverse Engineering (HOBRE) aims to transform decompiled pseudocode into a more human-friendly representation, thereby reducing the cognitive burden of reverse analysis and improving efficiency. However, reliably evaluating HOBRE outputs remains a fundamental challenge: human evaluation is costly, time-consuming, and difficult to scale, while existing automated metrics either… ▽ More Human-Oriented Binary Reverse Engineering (HOBRE) aims to transform decompiled pseudocode into a more human-friendly representation, thereby reducing the cognitive burden of reverse analysis and improving efficiency. However, reliably evaluating HOBRE outputs remains a fundamental challenge: human evaluation is costly, time-consuming, and difficult to scale, while existing automated metrics either require executable test cases and runtime environments that are often unavailable for real-world binaries, or rely on high-quality source code references that are typically inaccessible and fail to capture semantically equivalent but lexically diverse outputs. Although LLM-as-a-Judge paradigm is naturally well-suited to HOBRE evaluation, its effectiveness remains underexplored. This paper presents the first systematic investigation of the LLM-as-a-Judge paradigm for HOBRE across three representative tasks: function name recovery, binary code summarization, and decompilation optimization. We introduce BinJudgeBench, the first expert-annotated, reference-free evaluation benchmark based on multi-dimensional human judgment, where LLM-as-a-Judge achieves an average correlation of 63.20\% with human judgment, outperforming traditional automated metrics at 35.04\%. By analyzing judge configurations across backbone LLMs, prompting strategies, and decoding temperatures, we find that no ``one-size-fits-all'' configuration exists, as the optimal setup varies across tasks and individual samples. To address this, we propose BinJudge, which employs a lightweight routing mechanism to adaptively select the optimal judge configuration for each task and sample. BinJudge improves correlation with human experts by 4.5\%-24.7\% and reduces API cost to 0.06$\times$-0.84$\times$ of that of static best configurations, providing a scalable, cost-effective, and high-fidelity automated evaluation scheme for HOBRE. △ Less

Submitted 7 August, 2026; originally announced August 2026.

Comments: Accepted by the 41st IEEE/ACM International Conference on Automated Software Engineering (ASE 2026)

Understanding and Improving Model Editing for Secure Code Generation

Authors: Weifeng Sun, Quanjun Zhang, Yuchen Chen, Chengran Yang, Gou Tan, David Lo

Abstract: Large language models (LLMs) are widely used for code generation, yet they can reproduce vulnerable implementations learned from insecure training patterns. Prior work has mainly explored inference-time hardening, which reduces insecure generations without modifying the target model but relies on auxiliary components and adds runtime overhead. We conduct the first systematic study of model editing… ▽ More Large language models (LLMs) are widely used for code generation, yet they can reproduce vulnerable implementations learned from insecure training patterns. Prior work has mainly explored inference-time hardening, which reduces insecure generations without modifying the target model but relies on auxiliary components and adds runtime overhead. We conduct the first systematic study of model editing as a model-level hardening mechanism for secure code generation. We evaluate 3 state-of-the-art editing methods across diverse LLM families and compare them with CoSec, a representative inference-time approach, focusing on security, robustness, generalization, and functional correctness. Model editing yields larger security gains than CoSec on seen vulnerability types, improving security ratios by 15%-25% over vanilla models, with gains remaining stable under prompt perturbations. However, these improvements transfer unreliably to unseen vulnerabilities and can reduce functional correctness. To mitigate this trade-off, we propose SafeEdit, a post-edit refinement method combining functional tuning with edit-aware regularization. Across eight target LLMs, SafeEdit improves Pass@1 over UltraEdit by 11.73/13.70/15.50 percentage points at T=0.1/0.4/0.8 while largely preserving security. Compared with CoSec, it achieves relative security-ratio gains of 7.54%-12.04%. Additional evaluation on CodeGuard+ confirms improved joint secure-and-correct generation. SafeEdit and CoSec are also complementary, and their combination can further improve security while maintaining strong functional correctness. Overall, our results provide evidence-backed guidance for applying model editing to secure code generation. △ Less

Submitted 7 August, 2026; originally announced August 2026.

Comments: ISSTA 2026

How Reasoning Shapes Social Bias in LLM-Generated Code?

Authors: Weifeng Sun, Jieke Shi, Zhou Yang, Yuchen Chen, Hongyan Li, Meng Yan, David Lo

Abstract: Large language models (LLMs) are increasingly used for code generation, yet generated programs may exhibit social bias through unfair or differential treatment of sensitive demographic attributes. While prior work mainly studies direct code generation, bias in reasoning-based generation remains underexplored. We conduct the first systematic study of social bias in reasoning-based code generation,… ▽ More Large language models (LLMs) are increasingly used for code generation, yet generated programs may exhibit social bias through unfair or differential treatment of sensitive demographic attributes. While prior work mainly studies direct code generation, bias in reasoning-based generation remains underexplored. We conduct the first systematic study of social bias in reasoning-based code generation, evaluating 9 standard LLMs and large reasoning models (LRMs) on realistic bias-sensitive tasks across three human-centered decision scenarios. We find that reasoning generally reduces bias, lowering the average bias rate from 0.64 to 0.40, but the effect varies substantially across models. Meanwhile, code quality is not consistently preserved, with the average quality dropping from 0.72 to 0.59. Biased reasoning strongly predicts biased code, and adjusting generation configurations alone is insufficient for robust mitigation. Based on these findings, we propose ProbeDebias, a reasoning-aware framework that detects and rewrites biased reasoning traces before code generation. ProbeDebias achieves 87.76% F1 for reasoning-bias detection and reduces code bias by 83.73% on average while largely preserving quality. Compared with SOTA baselines, it further reduces average bias by 52.70%-54.42% and improves quality by 9.79%-36.79%. These results highlight the value of reasoning-stage analysis for trustworthy code generation. △ Less

Submitted 7 August, 2026; originally announced August 2026.

Comments: ASE 2026

AgentChaos: Chaos Engineering for Agent Systems via Programmatic Fault Injection

Authors: Gou Tan, Zhensu Sun, Jieke Shi, Ting Zhang, Zilong He, Qingfu Wu, Shuai Liang, Weifeng Sun, Junda He, Pengfei Chen, Chuanfu Zhang, Lwin Khin Shar, David Lo

Abstract: Agent systems rely on LLM APIs for every response, but these APIs can return server errors, truncated responses, or corrupted content that propagates through downstream agents and causes task failure. Evaluating robustness under these faults is crucial for reliable deployment. Existing fault injection methods are offline, require source code modification, or cannot modify specific response fields.… ▽ More Agent systems rely on LLM APIs for every response, but these APIs can return server errors, truncated responses, or corrupted content that propagates through downstream agents and causes task failure. Evaluating robustness under these faults is crucial for reliable deployment. Existing fault injection methods are offline, require source code modification, or cannot modify specific response fields. A comprehensive evaluation also requires a systematic fault taxonomy because different fault types affect downstream agents differently. We propose AgentChaos, a chaos engineering framework for controlled, runtime, non-intrusive LLM API fault injection. Since all agent systems access LLMs through the same HTTP interface, we inject faults at this shared layer without modifying source code. We define crash, omission, and value faults on content and tool call fields, intercept and modify LLM API responses at runtime, and verify whether each fault is triggered to filter untriggered tasks and avoid underestimating fault impact. Evaluations across agent systems, benchmarks, and backbone LLMs under 65 fault configurations show that all systems degrade under fault injection, with pass@1 dropping by up to 50 percentage points. The ranking is consistent across models, suggesting that robustness depends on system implementation rather than model capability. Existing fault diagnosis methods achieve below 53% accuracy on fault type and below 56% on fault step, leaving room for improvement. We further reveal practical findings for agent system developers. △ Less

Submitted 7 August, 2026; originally announced August 2026.

Comments: Accepted at the 41st IEEE/ACM International Conference on Automated Software Engineering (ASE 2026)

AgentExecutor: Partial Code Execution via Agentic Context Generation

Authors: Junkai Chen, Chengran Yang, Xing Hu, Zhenhao Li, Xin Xia, David Lo

Abstract: Executing code snippets is essential for dynamic program analysis, but it remains challenging to execute an arbitrary code snippet due to issues like missing context and incomplete dependencies. Existing approaches to partial code execution, such as LExecutor and Treefix, leverage the power of language models to infer missing information and enable execution. However, they suffer from (i) limited… ▽ More Executing code snippets is essential for dynamic program analysis, but it remains challenging to execute an arbitrary code snippet due to issues like missing context and incomplete dependencies. Existing approaches to partial code execution, such as LExecutor and Treefix, leverage the power of language models to infer missing information and enable execution. However, they suffer from (i) limited action spaces and feedback, and (ii) rigid optimization strategies, which restrict their effectiveness and efficiency. In this paper, we propose AgentExecutor, a novel multi-agent framework for partial code execution. Our approach introduces a three-phase design: execution environment preparation, dynamic exploration with iterative refinement, and prefix evolution via program synthesis. Supported by the power of LLM agents who can think, act, and get feedback iteratively, AgentExecutor is able to autonomously explore a richer action space, enabling diverse operations such as creating resource files and resolving environment configuration. Furthermore, it adopts adaptive optimization strategies, including coverage-guided context pruning and prefix evolution via program synthesis, to systematically improve the execution quality of partial code. We evaluate AgentExecutor on two widely used datasets comprising Stack Overflow snippets and open-source project code. The results show that AgentExecutor achieves up to 94% and 90% code coverage, outperforming the state-of-the-art approach Treefix by 19.9% and 13.8%, respectively. In addition, AgentExecutor significantly reduces execution time (by up to 80.3%) and cost (by up to 56.6%). These findings demonstrate that AgentExecutor provides an effective and efficient solution for partial code execution. △ Less

Submitted 6 August, 2026; originally announced August 2026.

Comments: ASE 2026

We Must Have Missed This Comment: Detecting and Repairing Stale Function References in Linux Kernel Comments

Authors: Kexin Sun, Yunbo Lyu, Xutong Ma, Hongyu Kuang, Ratnadira Widyasari, He Zhang, Xiaoxing Ma, Julia Lawall, David Lo

Abstract: As the Linux kernel evolves, code comments may become outdated, as the functions they reference can be refactored or removed independently without corresponding updates to the comments. Such stale function references can mislead maintainers and thus hinder code comprehension. Prior work on detecting code-comment inconsistency mainly focused on addressing semantic misalignment between Javadoc comme… ▽ More As the Linux kernel evolves, code comments may become outdated, as the functions they reference can be refactored or removed independently without corresponding updates to the comments. Such stale function references can mislead maintainers and thus hinder code comprehension. Prior work on detecting code-comment inconsistency mainly focused on addressing semantic misalignment between Javadoc comments and their directly annotated functions, making them inapplicable to this type of externally induced staleness in the Linux kernel. Therefore, we propose ReCite, a three-stage approach to identify and repair such stale references: (1) detecting unresolved function-form symbols -- symbols in comments that appear to reference functions but for which no matching function can be found in the current codebase, (2) tracing the evolution history of each unresolved symbol through the Git history, and (3) generating LLM-based repair suggestions grounded in the evolution history and current code context. On Linux kernel v6.18-rc1, ReCite detects 869 stale references with generated repair suggestions. A manual evaluation on 200 sampled repairs shows that 178 (89.0%) provide useful repair guidance, with 85 (42.5%) directly applicable. Of our 75 submitted patches, 50 have been accepted. We also empirically study all unresolved function-form symbols. △ Less

Submitted 4 August, 2026; originally announced August 2026.

Comments: Accepted at ASE 2026

Fail-Fast, Restart-Smart: Early Failure Prediction and Restart for SWE Agentic Tasks

Authors: Chenyu Wang, Yunbo Lyu, Junda He, Zhou Yang, Chenxing Zhong, Yaniv Harel, David Lo

Abstract: Software engineering (SWE) agents resolve repository-level issues through long trajectories that grow increasingly expensive as context accumulates. Failed runs tend to be longer and exhibit redundant exploration or looping, suggesting that some failures may be detectable before completion. Early termination, however, risks interrupting trajectories that would otherwise succeed; conversely, an uns… ▽ More Software engineering (SWE) agents resolve repository-level issues through long trajectories that grow increasingly expensive as context accumulates. Failed runs tend to be longer and exhibit redundant exploration or looping, suggesting that some failures may be detectable before completion. Early termination, however, risks interrupting trajectories that would otherwise succeed; conversely, an unsuccessful trajectory may still contain useful repository edits. We present FailFast-RestartSmart, a two-stage controller for a single active trajectory. FailFast is a lightweight 0.6B monitor trained with terminal and dense fail-to-pass supervision to predict failure from observable prefixes without policy logits or hidden states. Upon an alarm, RestartSmart launches a fresh same-policy rollout without prior prompt history and offers the interrupted repository diff as an optional overlay that the agent may inspect, apply, or discard. On SWE-bench Verified, a monitor trained solely on Qwen3.6-27B trajectories transfers to three other policies, including a closed-API model, and saves 14.6%-20.4% of execution tokens at a target 5% false-positive rate; on Qwen3.6-27B, its 20.4% saving exceeds the 12.5% achieved by our per-step AgentStop adaptation. At a target 25% false-positive rate, RestartSmart raises Qwen3.6-27B resolution from 66.6% to 71.8%, whereas cold restart reaches only 66.8%. Together, these results support early termination with sequential same-policy recovery. △ Less

Submitted 4 August, 2026; originally announced August 2026.

Lossless Tensor Compression as Program Synthesis

Authors: Jieke Shi, Junda He, Wenjia Jiang, Weifeng Sun, Shidong Pan, Zhensu Sun, Chengran Yang, Peixin Zhang, Yifan Jia, Zhou Yang, Thong Hoang, Xiwei Xu, Zhenchang Xing, David Lo

Abstract: Model checkpoints are growing in both number and size, which makes archival, transfer, and deployment increasingly costly. General-purpose compressors can reduce storage requirements but ignore tensor structure, whereas existing tensor-specific compressors rely on fixed and format-specific pipelines. We present Brevis, which formulates lossless tensor compression as program synthesis. We design a… ▽ More Model checkpoints are growing in both number and size, which makes archival, transfer, and deployment increasingly costly. General-purpose compressors can reduce storage requirements but ignore tensor structure, whereas existing tensor-specific compressors rely on fixed and format-specific pipelines. We present Brevis, which formulates lossless tensor compression as program synthesis. We design a typed domain-specific language (DSL) that captures recurring tensor structures, such as repeated regions and floating-point fields, through a set of reversible operators. Given a tensor, Brevis synthesizes a self-contained DSL program that reconstructs it bit-exactly. A checkpoint-specific production prior, learned from a small representative sample of tensors, guides a bounded A* search to synthesize compact programs, which can later be executed directly for bit-exact decompression. On 10 public checkpoints spanning language, audio, and image generation models, Brevis reduces 2.13 TB of checkpoint data to 1.41 TB, a 33.93% storage reduction. It produces archives up to 30.87% smaller than those of four general-purpose compressors, including zstd and gzip, and smaller archives than the tensor-specific compressors ZipNN and DFloat11. Under a practical concurrency configuration, Brevis achieves 3.60 GB/s compression and 6.61 GB/s decompression while preserving every source byte. △ Less

Submitted 3 August, 2026; originally announced August 2026.

Vul4Py: Benchmarking Automated Vulnerability Repair in Python with Paired Exploit and Functional Oracles

Authors: Tan Bui, Ting Zhang, Ferdian Thung, Yunpeng Xiong, Penghao Jiang, Xin Zhou, David Lo

Abstract: Automated Vulnerability Repair (AVR) has advanced rapidly across program analysis, machine learning, and Large Language Models (LLMs), but a verifiable, head-to-head comparison of AVR approaches on Python is still missing. Python underpins critical web, data, and machine-learning infrastructure, yet existing Python benchmarks accept a patch on the strength of a proof-of-concept exploit alone, or a… ▽ More Automated Vulnerability Repair (AVR) has advanced rapidly across program analysis, machine learning, and Large Language Models (LLMs), but a verifiable, head-to-head comparison of AVR approaches on Python is still missing. Python underpins critical web, data, and machine-learning infrastructure, yet existing Python benchmarks accept a patch on the strength of a proof-of-concept exploit alone, or apply a functional test only on the subset of entries whose upstream project happens to ship one. Both therefore miss functional regressions, in which a patch defeats the exploit but breaks unrelated behavior. We present Vul4Py, a Python AVR benchmark in which every entry carries a paired oracle: an exploit oracle that must fail on the vulnerable revision and pass on the fixed one, together with a project-native pytest functional oracle that must pass on both. Vul4Py comprises 100 real vulnerabilities from 60 open-source projects, spanning 60 distinct CWEs and the years 2017 to 2025, each packaged with a pinned, reproducible per-instance environment. Using Vul4Py, we compare six approaches in three categories: a specialized vulnerability repair tool, directly prompted LLMs, and software engineering agents. The agents dominate: OpenHands repairs 41 of 100 vulnerabilities, against 4 for the strongest directly prompted LLM and 2 for the specialized tool, despite all three sharing the same backbone model. The paired oracle is what makes these counts trustworthy: it rejects 15 of the 119 patches that an exploit-only oracle would accept, and 98 of the 104 patches it admits are manually confirmed to be semantically equivalent to the developer's patches △ Less

Submitted 1 August, 2026; originally announced August 2026.

Demystifying Solana Bots: From GitHub Blueprints to On-Chain Fingerprints

Authors: Xiaoye Zheng, Yujing Chen, Minghao Wu, David Lo, Difan Xie, Daoyuan Wu, Xiaohu Yang, Zhiyuan Wan

Abstract: Solana is an emerging blockchain platform designed for high throughput and low transaction fees, making it inexpensive to submit transactions at scale and, consequently, increasing exposure to bot spamming and related financial exploitation. Solana bots are typically off-chain software systems that operate in a competitive on-chain execution environment by constructing and submitting transactions,… ▽ More Solana is an emerging blockchain platform designed for high throughput and low transaction fees, making it inexpensive to submit transactions at scale and, consequently, increasing exposure to bot spamming and related financial exploitation. Solana bots are typically off-chain software systems that operate in a competitive on-chain execution environment by constructing and submitting transactions, and the bot-related transactions on the decentralized exchanges exceed 250 million dollars in daily trading volume in January 2026. Prior studies on Solana have examined system performance, smart-contract security, and specific on-chain phenomena. However, we still lack a systematic understanding of what Solana bots implement in practice and how these implementations manifest as observable on-chain execution fingerprints. To address this gap, we performed a large-scale empirical study of Solana bots from two complementary views: (i) 586 bot repositories collected from GitHub, and (ii) 200 bot addresses on Solana, with over 44 million on-chain transactions. Our study derives an implementation-grounded taxonomy of Solana bots comprising 15 categories grouped into five domains (e.g., Trading Operations, MEV, and On-chain Analytics), identifies a largely shared five-stage operational pipeline manifested in bot implementations, and uncovers systematic variation in on-chain trading behaviors of Solana bots across diverse trading platforms and assets. Based on our findings, we highlight future research directions, and provide recommendations for building and operating bots on the Solana blockchain. △ Less

Submitted 31 July, 2026; v1 submitted 30 July, 2026; originally announced July 2026.

Comments: Accepted at ASE 2026

Graph Is the Verifier: Agentic Reinforcement Learning for Interprocedural Vulnerability Detection

Authors: Yikun Li, Ting Zhang, Jiakun Liu, Jinfeng Jiang, Yuheng Yieh, Yixin Yang, Wen Bin Leow, Yide Yin, Yintong Huo, Eng Lieh Ouh, Lwin Khin Shar, David Lo

Abstract: Real-world vulnerabilities often span multiple functions, yet most learning-based detectors classify each function in isolation: on a sample of real CVEs, we find that 71.7% of vulnerable functions require evidence from outside the function to be classified correctly. Agentic reinforcement learning (RL) could close this gap by enabling a model to gather that evidence itself, but it lacks a reliabl… ▽ More Real-world vulnerabilities often span multiple functions, yet most learning-based detectors classify each function in isolation: on a sample of real CVEs, we find that 71.7% of vulnerable functions require evidence from outside the function to be classified correctly. Agentic reinforcement learning (RL) could close this gap by enabling a model to gather that evidence itself, but it lacks a reliable reward, since a reward defined on the final verdict alone can be obtained without performing any investigation. We propose VulAgentRL, an agentic RL framework for interprocedural vulnerability detection built on a Code Property Graph (CPG). The CPG serves two roles: at inference time the policy queries it for callers, callees, dataflow, and other queries, and at training time the same graph verifies the evidence the policy cites. Because every CPG node carries a persistent integer identifier, this verification is an exact comparison rather than a textual match, so the reward credits verdicts that are supported by evidence. We further initialize the policy by distilling teacher investigations, and show that this warm start is necessary, since RL cannot acquire tool-use behavior it never samples. Under a repository-level split that prevents leakage, VulAgentRL outperforms state-of-the-art baselines, including frontier models, on the strict pair-wise-correct metric while issuing fewer tool calls, and its advantage persists on an out-of-distribution corpus and under class imbalance. △ Less

Submitted 29 July, 2026; originally announced July 2026.

Multi-level Code Optimization via Mixture of Prompts

Authors: Yun Peng, Jun Wan, Jiakun Liu, Shuzheng Gao, David Lo, Xiaoxue Ren

Abstract: Runtime efficiency is a critical factor that impacts both software quality and user satisfaction. There are many approaches proposed for code optimization to improve runtime efficiency. Traditional code optimization methods operate on intermediate representations (IRs) during compilation for static languages. They are effective but struggle to handle dynamic languages that do not require compilati… ▽ More Runtime efficiency is a critical factor that impacts both software quality and user satisfaction. There are many approaches proposed for code optimization to improve runtime efficiency. Traditional code optimization methods operate on intermediate representations (IRs) during compilation for static languages. They are effective but struggle to handle dynamic languages that do not require compilation. Recently, large language models (LLMs) have been leveraged to directly optimize source code in dynamic languages. However, these methods fail to identify suitable optimization targets and usually conduct incomprehensive single-level optimization. To address these challenges, we propose Optimo, a multi-level LLM-based code optimization approach built on a novel Mixture-of-Prompts (MoP) architecture. In the MoP architecture, Optimo identifies time-critical code structures as performance bottlenecks via differential profiling. These structures are then routed to some optimization strategies, akin to expert models in MoE, each tailored to optimize specific code patterns. Unlike traditional approaches that focus only on statement-level optimizations, Optimo operates at four levels of abstraction, ranging from coarse-grained algorithmic improvements to fine-grained optimizations in API usage. We evaluate Optimo on two code efficiency benchmarks, COFFE and Effibench. Our results demonstrate that Optimo achieves an up to 57.48% opt%, i.e., the percentage of optimized programs that are correct and at least 10% faster than the original programs, and an up to 3.97x speedup when optimizing human-written code, and it consistently outperforms the best baseline by up to 96.51% in terms of opt%. Furthermore, Optimo achieves an up to 42.42% opt% and an up to 13.51x speedup when optimizing LLM-generated code. △ Less

Submitted 26 July, 2026; originally announced July 2026.

Comments: This paper has been accepted by ASE 2026

PoCEvolve: Generating Proof-of-Concept Exploits from Security Patches with Vulnerability-Aware Prompt Evolution

Authors: Duc Manh Tran, Ratnadira Widyasari, Ivana Clairine Irsan, Huihui Huang, Ting Zhang, Shar Lwin Khin, Ouh Eng Lieh, Hong Jin Kang, David Lo

Abstract: Ideally, the detailed information about a vulnerability should be made available together with the fixing commit. In practice, however, such details often become available only long after the commit, even when a CVE has already been published. During this window, the patch is already public, so attackers can reverse-engineer it, yet defenders lack the details needed to assess exposure, prioritize,… ▽ More Ideally, the detailed information about a vulnerability should be made available together with the fixing commit. In practice, however, such details often become available only long after the commit, even when a CVE has already been published. During this window, the patch is already public, so attackers can reverse-engineer it, yet defenders lack the details needed to assess exposure, prioritize, and validate the fix. Executable evidence, such as a proof-of-concept (PoC) exploit, could fill this gap. Prior work has automated PoC generation, but the state-of-the-art approach, PoCGen, assumes that a detailed vulnerability report is already available, which is precisely what is missing during this window. In this paper, we first present an empirical study quantifying the long delay between the fixing commit and the availability of a detailed vulnerability report. We then introduce PoCEvolve, a vulnerability-aware prompt-evolution framework that generates PoCs directly from vulnerability-fixing commits. Given a vulnerability-fixing commit, PoCEvolve synthesizes a corresponding PoC exploit. To learn from unsuccessful generation attempts, PoCEvolve assesses the usefulness of different dimensions of vulnerability-related context, including the inferred vulnerable API and code-coverage information. These assessments guide prompt evolution towards more effective exploit-generation prompts. We evaluate PoCEvolve on SecBench.VFC.js, where PoCEvolve achieves a PoC generation success rate of 58.4%, corresponding to relative improvements of 20.7% over PoCGen and 200.0% over the LLM baseline with GPT-4o-mini. With a recent model, Qwen3.7-Plus, PoCEvolve achieves a higher success rate of 85.3%. When detailed vulnerability reports are available, PoCEvolve achieves a success rate of 71.7%, improving over PoCGen by 11.1%. △ Less

Submitted 24 July, 2026; originally announced July 2026.

"Go Home Copilot, You're Drunk": Understanding Developer Responses to Agent-Generated Code Review Comments

Authors: Shamse Tasnim Cynthia, Ratnadira Widyasari, Banani Roy, Ting Zhang, David Lo

Abstract: Code review is a critical quality assurance practice in software engineering development, and AI coding agents are increasingly generating review comments on pull requests. However, little is known about how developers actually respond to such agent-generated feedback. In this paper, we present the first large-scale empirical study on the resolution of agent-generated code review comments. We anal… ▽ More Code review is a critical quality assurance practice in software engineering development, and AI coding agents are increasingly generating review comments on pull requests. However, little is known about how developers actually respond to such agent-generated feedback. In this paper, we present the first large-scale empirical study on the resolution of agent-generated code review comments. We analyze $54{,}791$ comments generated by five widely used coding agents (i.e., Copilot, Cursor, Codex, Devin, and Claude) across $342$ Python repositories on GitHub. We examine (1) resolution rates across agents and comment types, (2) the role of developer experience, and (3) characteristics that influence comment usefulness. Our results show that resolution rate varies considerably across agents, with Copilot accounting for the majority of resolved comments (72.9\%). Core developers resolve the majority of agent-generated feedback, particularly for \textit{design} and \textit{evolvability}-related comments, while peripheral developers are more involved in resolving \textit{functional defect} comments. Through open card sorting of 470 unresolved comment discussions, we identify \textit{ten} discussion patterns explaining why comments remain unresolved, with \textit{incorrect suggestions} and \textit{intentional design decisions} being the most prevalent. Finally, our analysis reveals that the presence of an inline \textit{code suggestion} is the strongest predictor of comment resolution, while lengthy and complex comments are less likely to be acted upon. Our findings provide insights for improving AI-generated code review feedback and its integration into development workflows. △ Less

Submitted 29 July, 2026; v1 submitted 24 July, 2026; originally announced July 2026.

ICAE-Bench: Evaluating Coding Agents as Interactive Project Builders

Authors: Zhongyuan Peng, Dan Huang, Chuyu Zhang, Caijun Xu, Changyi Xiao, Shibo Hong, David Lo, Lin Qiu, Xuezhi Cao, Jiyuan He, Yixin Cao

Abstract: The recent emergence of vibe-coding workflows is changing what coding agents are expected to do. Instead of merely completing code under fully specified instructions, agents are increasingly expected to transform incomplete product intent into working software by combining various abilities including planning, requirement clarification, tool use, debugging, and repository-level construction. Yet e… ▽ More The recent emergence of vibe-coding workflows is changing what coding agents are expected to do. Instead of merely completing code under fully specified instructions, agents are increasingly expected to transform incomplete product intent into working software by combining various abilities including planning, requirement clarification, tool use, debugging, and repository-level construction. Yet existing benchmarks have not fully caught up with this shift, evaluating agents on static, fully specified tasks. In this paper, we introduce ICAE-Bench, a benchmark for evaluating coding agents under interactive project-building settings. The basic idea is to start from a fuzzy product requirement, simulating the dynamic paradigm with an automated User Agent. To make this setting both realistic and evaluable, ICAE-Bench introduces three key designs. First, to avoid the ambiguity of unconstrained fuzzy requirements, each task derives ambiguity from a precise real open-source repository with executable behavior. Second, to ensure high-quality and reproducible user simulation, ICAE-Bench grounds interaction through User Agent Data, allowing the User Agent to reveal hidden constraints without inventing new requirements or leaking implementation artifacts. Third, to evaluate open-ended repositories fairly, ICAE-Bench uses standardized black-box tests together with multi-dimensional diagnostics, including functional correctness, semantic and API similarity, structural fidelity, design quality, and interaction quality. △ Less

Submitted 23 July, 2026; originally announced July 2026.

SciCodePile: A 128GB Corpus and Executable Benchmark for Challenging Scientific Code Generation

Authors: Weifeng Sun, Ye Fan, Yuchen Chen, Gou Tan, Jieke Shi, Yuan Yidi, Swee Liang Wong, Jonathan Pan, David Lo

Abstract: Large language models (LLMs) excel at general-purpose code generation, yet how well they handle scientific code remains an open question. Existing datasets and benchmarks are limited in scale, domain coverage, or executable verification, leaving the true gap between current LLMs and reliable scientific code generators inadequately assessed. To address these limitations, we present SciCodePile, the… ▽ More Large language models (LLMs) excel at general-purpose code generation, yet how well they handle scientific code remains an open question. Existing datasets and benchmarks are limited in scale, domain coverage, or executable verification, leaving the true gap between current LLMs and reliable scientific code generators inadequately assessed. To address these limitations, we present SciCodePile, the largest scientific code corpus to date, constructed from 37,737 public repositories and collectively comprising 128GB of code that spans multiple computational science disciplines. From this corpus, we further curate an executable benchmark of 200 tasks, each equipped with a sandboxed execution environment and an automated test harness for functional verification. We evaluate 15 LLMs from both open-source and closed-source families on three tasks: prefix-to-suffix completion, fill-in-the-middle infilling, and executable code generation. Results show that scientific code generation remains highly challenging: The best CodeBLEU reaches only 38.13 and 38.37 on the two completion tasks, while the strongest model achieves just 12.30\% Pass@1 on the executable benchmark, underscoring how far current models remain from reliable scientific code generation. To demonstrate the training utility of SciCodePile, we further show that continued pretraining on our corpus improves CodeBLEU by $\times$2.84 on scientific code completion, and instruction tuning on our data improves Pass@1 by $\times$4.79 on the executable benchmark. All code and data are available at https://huggingface.co/SciCodePile. △ Less

Submitted 21 July, 2026; originally announced July 2026.

Semantic Drift in Bug Resolution: How Behavioral Signals Propagate from Reports to Tests and Patches

Authors: Wendkûuni C. Ouédraogo, Yinghua Li, Xueqi Dang, Paweł Borsukiewicz, Liang Xiao, Lingfeng Bao, Anil Koyuncu, Jacques Klein, David Lo, Tegawendé F. Bissyandé

Abstract: Desc2Fix is a framework for measuring semantic alignment between bug reports, triggering tests, and developer-written fixes. Alignment is operationalized through structured behavioral anchors (e.g., reproduction steps, API/exception cues, expected vs. actual behavior), deterministic similarity metrics (ROUGE, SBERT, CodeBERT, OpenAI embeddings), and LLM-based judgments grounded in coverage, correc… ▽ More Desc2Fix is a framework for measuring semantic alignment between bug reports, triggering tests, and developer-written fixes. Alignment is operationalized through structured behavioral anchors (e.g., reproduction steps, API/exception cues, expected vs. actual behavior), deterministic similarity metrics (ROUGE, SBERT, CodeBERT, OpenAI embeddings), and LLM-based judgments grounded in coverage, correctness, and specificity. Our analysis covers 2,857 report-test-patch triplets from Defects4J and SWT-Bench using two widely adopted instruction-tuned LLMs from distinct model families. LLMs reliably extract structured signals (up to 90% completeness) and exhibit strong cross-model consistency, yielding a stable semantic input contract for downstream reasoning. However, alignment is highly representation-sensitive: lexical similarity alone is insufficient, full diffs provide the most stable basis for judging report-patch correspondence, and structured summaries trade surface overlap for stronger correspondence at the level of individual actions and entities. Across more than 182,000 LLM-based alignment judgments, both models exhibit systematic optimism relative to humans (1-2 points on 5-point scales) and only modest rank agreement, motivating bias-aware evaluation. Behavioral alignment is measurable but not reducible to similarity, and structured anchors combined with embedding-based proxies provide reproducible signals for ranking and filtering tests and patches. Desc2Fix could support more reliable test generation, fault localization, patch ranking, and bug report authoring. △ Less

Submitted 13 August, 2026; v1 submitted 20 July, 2026; originally announced July 2026.

Insecure Coding Preferences in Long-Term Memory: Security Risks for LLM-based Code Generation

Authors: Yuchen Chen, Wei Cheng, Yuan Xiao, Zhou Yang, Weifeng Sun, Chunrong Fang, Xiang Chen, Baowen Xu, David Lo, Zhenyu Chen

Abstract: LLM-based systems increasingly incorporate long-term memory to improve cross-session continuity. However, once insecure coding preferences are stored, they may silently influence security-critical decisions in subsequent generations. In this study, we conduct the first systematic empirical study on the impact of insecure coding preferences stored in long-term memory on the security of LLM-based co… ▽ More LLM-based systems increasingly incorporate long-term memory to improve cross-session continuity. However, once insecure coding preferences are stored, they may silently influence security-critical decisions in subsequent generations. In this study, we conduct the first systematic empirical study on the impact of insecure coding preferences stored in long-term memory on the security of LLM-based code generation. We evaluate four LLMs (ChatGPT, Gemini, Qwen, and Grok) across five programming languages (Python, C, C++, Go, and JavaScript). Our results show that insecure memories significantly increase the risk of generating vulnerable code by 2.7-50.3 percentage points (pp). Moreover, they create a 5.4-14.0 percentage-point risk-warning gap, where warning-rate increases lag behind vulnerability-rate increases. Further analysis reveals that insecure memories are difficult to overwrite through normal interactions and can broadly influence model outputs even when prompts are phrased differently. Finally, we evaluate three mitigation strategies: security-requirement appending and memory storage reduce vulnerability rates by 19.7-33.6 pp but may degrade functional correctness by up to 15.9 pp; memory-level safety filtering achieves a 100\% detection rate on our evaluated risky memory entries and restores generation behavior to the without-memory baseline. Based on these findings, we provide actionable suggestions to improve the security of long-term memory in LLM-based code generation. △ Less

Submitted 20 July, 2026; originally announced July 2026.

Comments: Accepted to the 35th ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA 2026)

How Do Practitioners Build SE Agents? Insights from a Mixed-Methods Study

Authors: Yunbo Lyu, David Williams, Jieke Shi, Zhensu Sun, Chao Peng, Zhou Yang, Federica Sarro, David Lo

Abstract: The rise of Software Engineering (SE) agents, i.e., LLM-based agents that can understand large codebases and carry out engineering tasks with limited human intervention, has been marked by rapid advances and adoption, but little is known about how developers build these systems in practice: existing studies mine repositories or examine deployment, but few investigate how SE agents are constructed.… ▽ More The rise of Software Engineering (SE) agents, i.e., LLM-based agents that can understand large codebases and carry out engineering tasks with limited human intervention, has been marked by rapid advances and adoption, but little is known about how developers build these systems in practice: existing studies mine repositories or examine deployment, but few investigate how SE agents are constructed. Through semi-structured interviews with 20 practitioners from 12 organizations and an online survey of 80 practitioners, this paper is the first to study how SE processes are changing in the development of SE agents and what challenges developers face. We find that as implementation becomes cheaper, bottlenecks shift rather than disappear: long-standing work in requirements, coordination, and deployment becomes more visible, while reviewing generated code and evaluating agent behavior become new and increasingly central forms of work. We characterize a seven-stage workflow and five process shifts, including a move toward evaluation-driven development, in which evaluation is increasingly defined early and steers iteration, and the emergence of specifications as first-class artifacts that teams test and version alongside code. We further identify six challenges that teams face, together with 12 corresponding practices they use or propose to address them, including unreliable evaluation signals, comprehension debt as code outpaces understanding, and behavioral changes introduced by provider-side model updates. △ Less

Submitted 25 July, 2026; v1 submitted 12 July, 2026; originally announced July 2026.

From Generic to Personalized: Exploring Persona-Aware Code Review Explanations

Authors: Shamse Tasnim Cynthia, Ratnadira Widyasari, Banani Roy, Italo Santos, David Lo

Abstract: Code review is essential for ensuring software quality and supporting collaboration, yet prior work shows that developers can interpret code review comments differently. These differences can hinder effective communication, particularly in collaborative settings. To address this challenge, we explore the potential of personified code review explanations. We report initial findings from an ongoing… ▽ More Code review is essential for ensuring software quality and supporting collaboration, yet prior work shows that developers can interpret code review comments differently. These differences can hinder effective communication, particularly in collaborative settings. To address this challenge, we explore the potential of personified code review explanations. We report initial findings from an ongoing mixed-methods user study in which developers evaluated persona-aligned review comments across multiple code snippets. Our results suggest that preferences for explanation styles vary across problem-solving styles, experience levels, and roles. Across problem-solving style profiles, developers valued explanatory depth, learning support, practical suggestions, and risk awareness over conciseness, highlighting the need to balance personalization with clarity and trust. Based on these findings, we outline a vision for inclusive, human-centered AI-assisted code review systems that adapt feedback to developers' problem-solving preferences. △ Less

Submitted 9 July, 2026; originally announced July 2026.

Comments: Presented at Journal Ahead Workshop (JAWs) 2026

RepoRescue: An Empirical Study of LLM Agents on Whole-Repository Compatibility Rescue

Authors: Zhihao Lin, Mingyi Zhou, Zhensu Sun, Yizhuo Yang, Renyu Yang, David Lo, Li Li

Abstract: Open-source libraries and tools are widely reused, but compatibility maintenance is expensive. Once maintainers leave, useful repositories can stop working as runtimes and dependencies evolve. We study whether LLM agents can adapt old repositories to modern environments, a task we call compatibility rescue. Unlike bug repair, compatibility rescue starts from a repository that worked in its origina… ▽ More Open-source libraries and tools are widely reused, but compatibility maintenance is expensive. Once maintainers leave, useful repositories can stop working as runtimes and dependencies evolve. We study whether LLM agents can adapt old repositories to modern environments, a task we call compatibility rescue. Unlike bug repair, compatibility rescue starts from a repository that worked in its original environment but fails after ecosystem drift. RepoRescue gives agents only the repository and its failing modern environment; the agent must diagnose the failure, locate affected code, and produce a source-code rescue that restores the historical test suite. We build RepoRescue from 193 Python and 122 Java repositories, each verified to pass historically and fail after modernization. We evaluate five deployed agent systems on Python and three on Java. Beyond full-patch pass rate, we rerun patches after removing test-file edits to measure source-only repair, add a runtime-enforced regime that blocks test edits, and validate practical use for repositories whose suites pass after rescue. We find that Claude Code systems sometimes edit failing tests even when prompted not to; with runtime blocking, Kimi still rescues 41.5% of repositories. Systems are complementary: their union reaches 62.7%, exceeding the best single system by 10.9 points. Difficulty concentrates in cross-file coordination: on 14 repositories requiring coordinated whole-codebase changes, GPT-5.2 through Codex passes all 14, while every Claude Code system passes at most two. Finally, a passing suite is only an initial signal: among 34 unmaintained Python candidates whose suites pass after rescue, 22 work in realistic scenarios and 12 pass bug-hunt with patches that address the compatibility failure. RepoRescue benchmarks compatibility rescue with source-only auditing, runtime enforcement, practical validation, and reasoning labels. △ Less

Submitted 1 July, 2026; originally announced July 2026.

Are Performance-Optimization Benchmarks Reliably Measuring Coding Agents?

Authors: Zhi Chen, Zhensu Sun, Yuling Shi, David Lo, Lingxiao Jiang

Abstract: Repository-level performance-optimization benchmarks such as GSO, SWE-Perf and SWE-fficiency evaluate coding agents by applying patches to real repositories and comparing runtime against unoptimized baselines and official reference patches. Their leaderboard scores are increasingly used as evidence of coding-agent progress, but those scores can conflate runtime instability, benchmark-specific scor… ▽ More Repository-level performance-optimization benchmarks such as GSO, SWE-Perf and SWE-fficiency evaluate coding agents by applying patches to real repositories and comparing runtime against unoptimized baselines and official reference patches. Their leaderboard scores are increasingly used as evidence of coding-agent progress, but those scores can conflate runtime instability, benchmark-specific scoring rules, and how many tasks are already solved by at least one public submission. We audit these issues across the three benchmarks. First, we replay the official reference patches for 740 code optimization tasks across four common types of Google Cloud machines. Most benchmark tasks can be replayed, but their reference patches satisfy the original benchmark validity rules in every cross-machine replay for only 39/102 GSO tasks, 11/140 SWE-Perf tasks, and 411/498 SWE-fficiency tasks; SWE-Perf is especially fragile because many reference patches produce close-to-zero runtime changes. Second, we show that public submission rankings depend strongly on the benchmark scoring rule. Among eight public submissions shared by GSO and SWE-fficiency, the official rankings disagree on 9 of 28 pairwise submission comparisons, and SWE-fficiency's leaderboard scoring rule assigns the worst ten tasks overly high score weights of 58.5%-82.8%. Third, looking across 10 public submissions for each task, we find that at least one submission matches or beats the reference patch on 85.3% (384/450) of replay-valid GSO and SWE-fficiency tasks, and beats the unoptimized base code on 99.8% (449/450). Our study complements leaderboard scores by identifying tasks with more reliable performance signals, quantifying per-task score contributions, and exposing the remaining performance gaps that are hidden by aggregate rankings. △ Less

Submitted 16 July, 2026; v1 submitted 1 July, 2026; originally announced July 2026.

Comments: 12 pages, 7 figures. Public data: https://github.com/chenzhi-cz/performance-optimization-benchmark-reliability

Knowledge-Enhanced Agentic Vulnerability Repair

Authors: Sicong Cao, Hao Ma, Le Yu, Kangyi Ding, Xiaolei Liu, Terry Yue Zhuo, Bo Wang, Xingwei Lin, Xiaobing Sun, Linzhang Wang, David Lo

Abstract: Frontier foundation models have changed the math on vulnerability discovery, but the bigger challenge is how the remediation side keeps up. Despite recent progresses in Automated Vulnerability Repair (AVR), current solutions struggle to reliably identify the root causes of vulnerabilities, and insufficiently utilize the prior fix knowledge to guide the patch generation process, thus undermining th… ▽ More Frontier foundation models have changed the math on vulnerability discovery, but the bigger challenge is how the remediation side keeps up. Despite recent progresses in Automated Vulnerability Repair (AVR), current solutions struggle to reliably identify the root causes of vulnerabilities, and insufficiently utilize the prior fix knowledge to guide the patch generation process, thus undermining their effectiveness in practice. To address this gap, we propose KeaRepair, a novel agentic AVR approach that grounds patch generation in verified program facts and high-level vulnerability knowledge. Specifically, KeaRepair first extracts multi-dimensional vulnerability knowledge from historical vulnerability-patch pairs from dual complementary views, and constructs dedicated retrieval knowledge bases. It then employs a tool-augmented agent that performs ReAct-style reasoning to collect verified program facts for vulnerability diagnosis. Finally, based on the diagnostic results, KeaRepair performs knowledge-level retrieval-augmented patch generation and iteratively refines patches through a closed-loop validation process involving compilation, PoC replay, and test-suite execution. Experimental results show that KeaRepair significantly outperforms existing AVR approaches on 55 reproducible C/C++ vulnerabilities. When paired with Gemini-3.1-Pro, KeaRepair successfully repairs 46 vulnerabilities, achieving a repair rate of 83.64%. Moreover, KeaRepair fixes six unique vulnerabilities that none of the baselines can address, and further demonstrates strong cross-language generalizability. △ Less

Submitted 1 July, 2026; originally announced July 2026.

Reproducing FACTER: Fairness via Conformal Thresholding and Prompt Repair

Authors: Oscar Miró López-Feliu, Daimy van Loo, Xanthos Kekkos, Mikel Blom, Clara Rus

Abstract: Fayyazi et al. (2025) recently proposed FACTER, a model-agnostic framework designed to jointly enforce fairness and statistical coverage in LLM-based recommendation through conformal thresholding and iterative prompt repair. In this work, we conduct a reproducibility study of the FACTER framework across diverse architectures and dataset sparsity levels, evaluating both the original open-ended gene… ▽ More Fayyazi et al. (2025) recently proposed FACTER, a model-agnostic framework designed to jointly enforce fairness and statistical coverage in LLM-based recommendation through conformal thresholding and iterative prompt repair. In this work, we conduct a reproducibility study of the FACTER framework across diverse architectures and dataset sparsity levels, evaluating both the original open-ended generation task and a constrained re-ranking extension. Under the strict reproduction, we observe a divergence in recommendation utility, which we trace to underspecified target-set evaluation in the original study. We then use the constrained re-ranking setting to evaluate FACTER when the candidate set is fixed, and introduce a static Fair Zero-Shot baseline to isolate the contribution of the iterative prompt repair loop. Our analysis shows that FACTER consistently reduces adaptive-threshold violation counts, but that these reductions are not consistently reflected under the fixed threshold or in global fairness metrics. In the constrained ranking setting, static fairness instructions achieve comparable semantic-parity outcomes to FACTER's dynamic repair loop, suggesting that the additional online repair mechanism provides limited benefit in this formulation. All code and reproduction artifacts are available at https://github.com/oscar-omlf/facter-repr. △ Less

Submitted 26 June, 2026; originally announced June 2026.

Comments: 29 pages. Accepted by Transactions on Machine Learning Research (TMLR), 2026. OpenReview: https://openreview.net/forum?id=4BPFVex4EM. Code: https://github.com/oscar-omlf/facter-repr

Humanizing Automatically Generated Unit Test Suites with LLM-Based Refactoring

Authors: Wendkûuni C. Ouédraogo, Yinghua Li, Xueqi Dang, Paweł Borsukiewicz, Liang Xiao, Lingfeng Bao, Anil Koyuncu, Jacques Klein, David Lo, Tegawendé F. Bissyandé

Abstract: Search-based test generation tools such as EvoSuite produce compilable and high-coverage unit tests at scale, but their suites are often hard to read and maintain. LLMs can generate more natural tests, yet direct generation remains brittle, with compilation rates of only 51-78% in our study. We introduce TestHumanizer, a hybrid SBST+LLM approach that uses LLMs as controlled refactoring layers over… ▽ More Search-based test generation tools such as EvoSuite produce compilable and high-coverage unit tests at scale, but their suites are often hard to read and maintain. LLMs can generate more natural tests, yet direct generation remains brittle, with compilation rates of only 51-78% in our study. We introduce TestHumanizer, a hybrid SBST+LLM approach that uses LLMs as controlled refactoring layers over compilable SBST suites to improve naming, structure, and developer-oriented clarity while preserving behavior and compilation validity. We evaluate TestHumanizer on 350 classes from Defects4J and SF110. EvoSuite generates 15 suites per class, and each suite is refactored under three context configurations using gpt-4o and mistral-large-2407, yielding 31,500 refactorings. TestHumanizer reaches 88-98% compilation rates, close to EvoSuite's 100% baseline and clearly above direct LLM generation. Structural coverage is largely preserved, typically within 1-2 percentage points, and 86-95% of refactorings satisfy a composite faithful-refactoring threshold. Refactored suites also improve predicted readability, reduce control-flow and cognitive complexity, and mitigate structural smells. The summary-based setting offers the most robust trade-off, while long code-centric prompts are more prone to hallucination-induced failures. A developer study on 30 classes and 444 test methods confirms significant gains in perceived readability and willingness to adopt, with Wilcoxon p less than 0.01 and substantial inter-rater agreement. Overall, LLMs are most effective not as standalone generators but as validation-gated refinement layers over robust SBST outputs. △ Less

Submitted 1 July, 2026; v1 submitted 26 June, 2026; originally announced June 2026.

To Run or Not to Run: Analyzing the Cost-Effectiveness of Code Execution in LLM-Based Program Repair

Authors: Zhihao Lin, Junhua Zhu, Mingyi Zhou, Xin Wang, Zhensu Sun, Renyu Yang, David Lo, Li Li

Abstract: LLM-based agents for program repair are increasingly built on a "generate-run-revise" paradigm, iteratively executing tests to evaluate and refine patches. This execution-based approach has become standard practice in state-of-the-art systems. However, executions can be time-consuming and expensive, yet their impact on these agents remains underexplored. In this paper, we conduct a two-stage empir… ▽ More LLM-based agents for program repair are increasingly built on a "generate-run-revise" paradigm, iteratively executing tests to evaluate and refine patches. This execution-based approach has become standard practice in state-of-the-art systems. However, executions can be time-consuming and expensive, yet their impact on these agents remains underexplored. In this paper, we conduct a two-stage empirical study over execution behavior in LLM-based program repair. To characterize execution behavior at scale, we first analyze 7,745 agent traces from SWE-bench leaderboard submissions. Second, we evaluate 3,000 end-to-end repair attempts across 200 SWE-bench instances and three agents (Claude Code, Codex, and the open-source OpenCode) under four execution paradigms, which allows for a fine-grained comparison of performance and cost. Our analysis reveals three key observations: (1) Code execution is used across all agents and models analyzed, with an average of 8.8 test runs per task. Execution behavior varies substantially across agents and models, with frequency ranging from 2 to 19 per task, and late-stage executions consistently achieve higher success rates than early-stage ones. (2) Execution restrictions have little effect on repair success: on commercial agents with SOTA models the resolve-rate gap between Prohibited and Unrestricted is only 1.25 percentage points and not statistically significant, while Prohibited saves substantial token and wall-clock cost. (3) Execution benefit is concentrated rather than uniform. These patterns suggest that current agents apply execution indiscriminately, paying its cost on instances where it provides little benefit. Execution, therefore, should be treated as a resource with an explicit cost-benefit tradeoff, not a default capability. △ Less

Submitted 25 June, 2026; originally announced June 2026.

Comments: Accepted to the ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA 2026). 23 pages

Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning

Authors: NVIDIA, :, Aaron Blakeman, Aaron Thomas, Aastha Jhunjhunwala, Abhibha Gupta, Abhinav Khattar, Adam Rajfer, Adi Renduchintala, Adil Asif, Aditya Vavre, Adriana Flores Miranda, Ahmad Bilal, Aileen Zaman, Ajay Hotchandani, Akanksha Shukla, Akhiad Bercovich, Aleksander Ficek, Alex Gronskiy, Alex Kondratenko, Alex Steiner, Alex Ye, Alexander Bukharin, Alexandre Milesi, Ali Taghibakhshi , et al. (549 additional authors not shown)

Abstract: We introduce Nemotron 3 Ultra, a 550 billion total and 55 billion active parameter Mixture-of-Experts Hybrid Mamba-Attention language model. We pre-trained Nemotron 3 Ultra on 20 trillion text tokens, then extended the context length to 1M tokens, and post-trained using Supervised Fine Tuning (SFT), Reinforcement Learning (RL), and Multi-teacher On-Policy Distillation (MOPD). Nemotron 3 Ultra is o… ▽ More We introduce Nemotron 3 Ultra, a 550 billion total and 55 billion active parameter Mixture-of-Experts Hybrid Mamba-Attention language model. We pre-trained Nemotron 3 Ultra on 20 trillion text tokens, then extended the context length to 1M tokens, and post-trained using Supervised Fine Tuning (SFT), Reinforcement Learning (RL), and Multi-teacher On-Policy Distillation (MOPD). Nemotron 3 Ultra is our most capable model yet, employing multiple key technologies - LatentMoE, Multi Token Prediction (MTP), NVFP4 pre-training, multi-environment RLVR, MOPD, and reasoning budget control. Nemotron 3 Ultra achieves up to ~6x higher inference throughput as compared to state-of-the-art publicly available LLMs while attaining on-par accuracy. The state-of-the-art accuracy, high inference throughput, and 1M token context length make Nemotron 3 Ultra ideal for long-running autonomous agentic tasks. We open-source the base, post-trained, and quantized checkpoints, along with the training data and recipe on HuggingFace. △ Less

Submitted 12 June, 2026; originally announced June 2026.

Agent System Operations: Categorization, Challenges, and Future Directions

Authors: Zexin Wang, Changhua Pei, Yuanhao Liu, Jingjing Li, Yintong Huo, Quan Zhou, Haotian Si, Hang Cui, Zihan Liu, Gaogang Xie, Fei Sun, Dan Pei, David Lo

Abstract: As the reasoning capabilities of Large Language Models (LLMs) continue to advance, LLM-based agent systems offer advantages in flexibility and interpretability over traditional systems, garnering increasing attention. However, despite the widespread research interest and industrial application of agent systems, these systems, like their traditional counterparts, frequently encounter anomalies. The… ▽ More As the reasoning capabilities of Large Language Models (LLMs) continue to advance, LLM-based agent systems offer advantages in flexibility and interpretability over traditional systems, garnering increasing attention. However, despite the widespread research interest and industrial application of agent systems, these systems, like their traditional counterparts, frequently encounter anomalies. These anomalies lead to instability and insecurity, hindering their further development. Therefore, a comprehensive and systematic approach to the operation and maintenance of agent systems is urgently needed. Unfortunately, current research on the operations of agent systems is sparse. To address this gap, we have undertaken a survey on agent system operations with the aim of establishing a clear framework for the field, defining the challenges, and facilitating further development. Specifically, this paper begins by systematically defining anomalies within agent systems, categorizing them into intra-agent anomalies and inter-agent anomalies. Next, we introduce a novel and comprehensive operational framework for agent systems, dubbed Agent System Operations (AgentOps). We provide detailed definitions and explanations of its four key stages: monitoring, anomaly detection, root cause localization, and resolution. △ Less

Submitted 31 May, 2026; originally announced June 2026.

How Agentic AI Coding Assistants Become the Attacker's Shell

Authors: Yue Liu, Yanjie Zhao, Yunbo Lyu, Ting Zhang, Haoyu Wang, David Lo

Abstract: Agentic AI coding assistants can edit files, run commands, and access the internet on behalf of developers. However, their reliance on unvetted external artifacts introduces a new attack vector. Hidden instructions in external artifacts can hijack these assistants, turning them into an attacker's shell to run unauthorized commands. In this article, we examine how these prompt injection attacks wor… ▽ More Agentic AI coding assistants can edit files, run commands, and access the internet on behalf of developers. However, their reliance on unvetted external artifacts introduces a new attack vector. Hidden instructions in external artifacts can hijack these assistants, turning them into an attacker's shell to run unauthorized commands. In this article, we examine how these prompt injection attacks work, measure their prevalence, discuss the limitations and challenges of current defenses, and suggest future research directions. △ Less

Submitted 25 May, 2026; originally announced May 2026.

Automated Repair of TEE Partitioning Issues via DSL-Guided and LLM-Assisted Patching

Authors: Chengyan Ma, Jieke Shi, Ruidong Han, Ye Liu, Feng Li, Yuqing Niu, David Lo

Abstract: Trusted Execution Environments (TEEs) provide hardware-based isolation to protect sensitive data and computations from potentially compromised operating systems (OS). However, TEE applications inevitably interact with the untrusted OS through SDK interfaces, and improper partitioning can introduce severe vulnerabilities such as data leakage and code injection. While prior work has proposed static… ▽ More Trusted Execution Environments (TEEs) provide hardware-based isolation to protect sensitive data and computations from potentially compromised operating systems (OS). However, TEE applications inevitably interact with the untrusted OS through SDK interfaces, and improper partitioning can introduce severe vulnerabilities such as data leakage and code injection. While prior work has proposed static analysis tools to detect such issues, automated repair remains largely unexplored. This problem is particularly challenging due to three TEE-specific factors: the lack of standardized secure development guidelines, the difficulty of extracting semantic information from low-level C code, and the absence of mature testing and validation methods. In this work, we present TEERepair, a framework for automatically repairing bad partitioning issues in TEE applications. Our approach tackles the above challenges by introducing a domain-specific language (DSL) to encode repair rules that express and capture common TEE security patterns, which are instantiated as patch templates with placeholders for context-specific variables. We then leverage large language models (LLMs) to reason about code semantics and synthesize context-aware patches, and further generate test clients to validate the repairs. We evaluate TEERepair on the TEE Partitioning Errors Benchmark (PartitioningE-Bench), achieving a significantly higher repair success rate of 87.6% compared to baselines. Furthermore, applying TEERepair to real-world TEE projects, we submitted 5 repair pull requests, 2 of which have been confirmed and merged by project maintainers. △ Less

Submitted 21 May, 2026; originally announced May 2026.

Comments: Accepted by the ACM International Conference on the Foundations of Software Engineering (FSE 2026)

Finding Missing Input Validation in TEEs via LLM-Assisted Symbolic Execution

Authors: Chengyan Ma, Jieke Shi, Ruidong Han, Ye Liu, Yuqing Niu, David Lo

Abstract: Trusted Execution Environments (TEEs) provide hardware-enforced isolation that protects sensitive code and data from untrusted software. Despite their strong security guarantees, analyzing TEE applications remains challenging due to the high cost and complexity of configuring complete TEE build and runtime environments, as well as the limited observability imposed by hardware isolation. This paper… ▽ More Trusted Execution Environments (TEEs) provide hardware-enforced isolation that protects sensitive code and data from untrusted software. Despite their strong security guarantees, analyzing TEE applications remains challenging due to the high cost and complexity of configuring complete TEE build and runtime environments, as well as the limited observability imposed by hardware isolation. This paper presents SymTEE, a novel large language model (LLM)-assisted symbolic execution framework for detecting missing input validation issues in TEE applications without requiring real TEE setups. SymTEE begins by leveraging Abstract Syntax Tree (AST) analysis to extract TEE code slices that may lack sufficient input validation, and then employs an LLM (GPT-5 in our case) to automatically convert the extracted slices into KLEE-compatible harness programs containing lightweight mock execution environments for symbolic analysis. Evaluations on 26 vulnerabilities (11 real-world and 15 synthetic) show that SymTEE achieves 100% precision and 92.3% recall in detecting missing input validation vulnerabilities while incurring an average analysis cost of only $0.05. These results demonstrate the effectiveness and practicality of SymTEE's pioneering paradigm of LLM-assisted symbolic execution, where LLMs autonomously generate mock environments to enable automated security analysis without complex setup, providing a more accessible and scalable framework for trusted computing systems. △ Less

Submitted 21 May, 2026; originally announced May 2026.

Comments: Accepted by 2026 IEEE/ACM Third International Conference on AI Foundation Models and Software Engineering (FORGE '26)

Beyond the Tip of the Iceberg: Understanding SATD in Dockerfiles through the Lens of Co-evolution

Authors: Wei Minn, Yan Naing Tun, Biniam Fesseha Demissie, Rui'ang Hu, Jiakun Liu, Mariano Ceccato, Lwin Khin Shar, David Lo

Abstract: Dockerfiles enable the creation of portable container-based execution environments for the application code, and have become an important part of the modern software development process. As Dockerfiles are a form of Infrastructure-as-Code (IaC), they can include temporary workarounds and other suboptimal implementations, leading to the accrual of technical debt that affects their reliability, secu… ▽ More Dockerfiles enable the creation of portable container-based execution environments for the application code, and have become an important part of the modern software development process. As Dockerfiles are a form of Infrastructure-as-Code (IaC), they can include temporary workarounds and other suboptimal implementations, leading to the accrual of technical debt that affects their reliability, security, and maintainability in the future. Prior work characterized self-admitted technical debt (SATD) in Dockerfile comments and the surrounding file chunks. This single-file view is incomplete since source code evolution involves changes across different types of software artifacts such as production, test, build, and other configuration files. Thus, we address this gap by studying SATD events in Dockerfiles alongside the related source code. We find that approximately 27% of admission events and 40% of repayment events are coupled to non-Dockerfile artifacts, and coupling sources are subtype-specific. We also observed that coupled SATD in general are repaid significantly faster overall (p = 0.0201), while coupled SATD regarding missing functionalities persists longer than its isolated counterparts; Lastly, we conducted open and axial coding of coupled SATD events, and we observe that external dependency issues, more particularly regarding unreleased upstream packages and bug fixes, are the most common cause of admission triggers in the source code; we also observe that architectural refactoring is the most common prerequisite for the repayment of SATD in Dockerfiles. These findings indicate that both practitioners (e.g. developers and project managers) and SATD researchers should integrate the source code-side co-evolution, rather than the single-file view, as the primary unit of analysis. △ Less

Submitted 20 May, 2026; originally announced May 2026.

Comments: Submitted to Empirical Software Engineering Journal

ACM Class: D.2.7; D.2.9; K.6.1; K.6.3

An Execution-Verified Multi-Language Benchmark for Code Semantic Reasoning

Authors: Yikun Li, Jinfeng Jiang, Ting Zhang, Chengran Yang, Chenxing Zhong, Yin Yide, Leow Wen Bin, Eng Lieh Ouh, Lwin Khin Shar, David Lo

Abstract: Evaluating whether large language models (LLMs) can recover execution-relevant program structure, rather than only produce code that passes tests, remains an open problem. Existing code benchmarks emphasize test-passing outputs, from standalone programming tasks (HumanEval, MBPP, LiveCodeBench) to repository repair (SWE-Bench); this is useful, but offers limited diagnostic signal about which progr… ▽ More Evaluating whether large language models (LLMs) can recover execution-relevant program structure, rather than only produce code that passes tests, remains an open problem. Existing code benchmarks emphasize test-passing outputs, from standalone programming tasks (HumanEval, MBPP, LiveCodeBench) to repository repair (SWE-Bench); this is useful, but offers limited diagnostic signal about which program semantics a model can recover from source. We introduce TraceEval, to our knowledge the first execution-verified, multi-language benchmark for code semantic reasoning: recovering a program's runtime call structure from source code. Unlike prior call-graph benchmarks that rely on static-tool output or hand-annotated ground truth, every positive edge in TraceEval is mechanically witnessed by validation execution, eliminating annotator disagreement and label noise for observed behavior. TraceEval consists of (i) 10,583 real-world programs (2,129 test, 8,454 train) extracted from 1,600+ open-source repositories across Python, JavaScript, and Java via an LLM-assisted harness-generation pipeline with tracer validation; and (ii) a reproducible pipeline that converts any open-source repository into new verified benchmark instances. We evaluate 10 LLMs at zero-shot on the held-out test split. The strongest model, Claude-Opus-4.6, reaches an average F1 of 72.9% across the three languages. To demonstrate the train split's utility as a supervision substrate, we fine-tune the Qwen2.5-Coder family on it: lifts of up to +55.6 F1 bring tuned Qwen2.5-Coder-32B to 71.2%, within 1.7 F1 of zero-shot Claude-Opus-4.6. We release the benchmark, pipeline, baselines, and a datasheet at https://github.com/yikun-li/TraceEva △ Less

Submitted 10 May, 2026; originally announced May 2026.

Affinity Tailor: Dynamic Locality-Aware Scheduling at Scale

Authors: Jin Xin Ng, Ori Livneh, Richard O'Grady, Josh Don, Peng Ding, Samuel Grossman, Luis Otero, Chris Kennelly, David Lo, Carlos Villavieja

Abstract: Modern large multicore systems often run multiple workloads that share CPUs under schedulers such as Linux CFS. To keep CPUs busy, these schedulers load-balance runnable work, causing each workload to execute on many cores. This weakens locality at the microarchitectural level: workloads lose reuse in caches, branch predictors, and prefetchers, and interfere more with one another - especially on c… ▽ More Modern large multicore systems often run multiple workloads that share CPUs under schedulers such as Linux CFS. To keep CPUs busy, these schedulers load-balance runnable work, causing each workload to execute on many cores. This weakens locality at the microarchitectural level: workloads lose reuse in caches, branch predictors, and prefetchers, and interfere more with one another - especially on chiplet-based systems, where spreading execution across cores also spreads it across LLC boundaries. A natural alternative is strict CPU partitioning, but hard partitions leave capacity idle when workloads do not fully use their reserved CPUs. We present Affinity Tailor, a userspace-guided kernel scheduling system built on a key insight: the kernel can preserve locality for workloads that share CPUs by treating demand-sized, topologically compact CPU sets as affinity hints rather than hard partitions. A userspace controller estimates each workload's CPU demand online and assigns a preferred CPU set sized to that demand, chosen to be as disjoint as possible from other workloads while spanning as few LLC domains as possible. The kernel then uses this set as an affinity hint, steering threads toward those CPUs while still allowing execution elsewhere when needed to preserve utilization. Deployed at Google, Affinity Tailor delivers geometric-mean per-CPU throughput gains of 12% on chiplet-based systems and 3% on non-chiplet systems over Linux CFS. Furthermore, faster execution reduces memory residency, yielding per-GB throughput gains of 3-7%. Our findings suggest that future schedulers should treat spatial locality as a first-class objective, even at the expense of work-conservation. △ Less

Submitted 30 April, 2026; originally announced April 2026.

Tail-aware N-version Machine Learning Models for Reliable API Recommendation

Authors: Aoi Matsuda, Fumio Machida, David Lo

Abstract: Machine learning (ML)-based API recommendation helps developers efficiently identify suitable APIs to complement the application code. However, code datasets used to train ML models often exhibit a long-tail distribution, leading to unreliable API recommendations, especially for infrequently used API methods at the tail of the distribution. To address this issue, we propose N-version API Recommend… ▽ More Machine learning (ML)-based API recommendation helps developers efficiently identify suitable APIs to complement the application code. However, code datasets used to train ML models often exhibit a long-tail distribution, leading to unreliable API recommendations, especially for infrequently used API methods at the tail of the distribution. To address this issue, we propose N-version API Recommendation (NvRec), which leverages N different versions of ML models to enhance the reliability of API sequence recommendations by suppressing unreliable outputs entailing tail APIs. NvRec leverages a set of available ML models and profiles their performance on individual API methods with their tail properties. The generated model profile is used at inference time to filter out unreliable API recommendations and determine the final output. We implement NvRec using five API recommendation models, including CodeBERT, CodeT5, MulaRec, UniXcoder, and CodeT5+, and evaluate it on a public benchmark dataset constructed from compilable Java projects. For the three-version NvRec, we find that the combination of CodeT5, MulaRec, and UniXcoder achieves the highest true accept rate of 83.8%, with a rejection rate of 80.7%, when majority voting is restricted to highly reliable candidates. In contrast, the five-version configuration achieves its highest true accept rate of 83.1% with simple majority voting, while reducing the rejection rate to 69.0%. Overall, the five-version configuration offers a better balance between true accept rate and rejection rate. △ Less

Submitted 30 April, 2026; originally announced April 2026.

Comments: EASE2026

LLM-Assisted Empirical Software Engineering: Systematic Literature Review and Research Agenda

Authors: Victoria Gomes, Delaney Selb, Fabio Palomba, Rodrigo Spinola, David Lo

Abstract: Context: Empirical Software Engineering (ESE) faces increasing challenges due to data scale, methodological complexity, and reproducibility concerns. Large Language Models (LLMs) have emerged as promising tools to support empirical workflows, yet their use remains fragmented, with no comprehensive synthesis to guide responsible adoption. Aims: This study analyzes how LLMs are used in ESE, examin… ▽ More Context: Empirical Software Engineering (ESE) faces increasing challenges due to data scale, methodological complexity, and reproducibility concerns. Large Language Models (LLMs) have emerged as promising tools to support empirical workflows, yet their use remains fragmented, with no comprehensive synthesis to guide responsible adoption. Aims: This study analyzes how LLMs are used in ESE, examining supported tasks, phases of the empirical lifecycle, integration into workflows, reported benefits and limitations, and the extent of reproducibility-related reporting. It also identifies gaps and future research directions. Method: We conducted a systematic literature review of peer-reviewed papers (2020-2025) across 12 leading software engineering venues, resulting in 50 primary studies analyzed through qualitative and quantitative synthesis. Results: We identified 69 LLM-assisted tasks, mainly in mining software repositories and controlled experiments, focusing on classification, filtering, and evaluation. LLMs are used across multiple phases but are concentrated in data processing and analysis. Their integration is largely automation-oriented, with limited decision-support use. Benefits emphasize efficiency and scalability, while limitations include hallucinations, inconsistency, prompt sensitivity, and reproducibility issues. Reporting practices are often incomplete. Conclusion: LLM use in ESE is growing but remains automation-driven, with gaps in human-centered integration and transparency. We outline implications and research agenda for responsible use. △ Less

Submitted 28 April, 2026; originally announced April 2026.

TitanCA: Lessons from Orchestrating LLM Agents to Discover 100+ CVEs

Authors: Ting Zhang, Yikun Li, Chengran Yang, Ratnadira Widyasari, Yue Liu, Ngoc Tan Bui, Phuc Thanh Nguyen, Yan Naing Tun, Ivana Clairine Irsan, Huu Hung Nguyen, Huihui Huang, Jinfeng Jiang, Lwin Khin Shar, Eng Lieh Ouh, David Lo, Hong Jin Kang, Yide Yin, Wen Bin Leow

Abstract: Software vulnerabilities remain one of the most persistent threats to modern digital infrastructure. While static application security testing (SAST) tools have long served as the first line of defense, they suffer from high false-positive rates. This article presents TitanCA, a collaborative project between Singapore Management University and GovTech Singapore that orchestrates multiple large lan… ▽ More Software vulnerabilities remain one of the most persistent threats to modern digital infrastructure. While static application security testing (SAST) tools have long served as the first line of defense, they suffer from high false-positive rates. This article presents TitanCA, a collaborative project between Singapore Management University and GovTech Singapore that orchestrates multiple large language model (LLM)-powered agents into a unified vulnerability discovery pipeline. Applied in open-source software, TitanCA has discovered 203 confirmed zero-day vulnerabilities and yielded 118 CVEs. We describe the four-module architecture, i.e., matching, filtering, inspection, and adaptation, and share key lessons from building and deploying an LLM-based vulnerability discovery solution in practice. △ Less

Submitted 1 June, 2026; v1 submitted 20 April, 2026; originally announced April 2026.

Comments: Accepted by IEEE Security & Privacy Magazine

Nemotron 3 Super: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning

Authors: NVIDIA, :, Aakshita Chandiramani, Aaron Blakeman, Abdullahi Olaoye, Abhibha Gupta, Abhilash Somasamudramath, Abhinav Khattar, Adeola Adesoba, Adi Renduchintala, Adil Asif, Aditya Agrawal, Aditya Vavre, Ahmad Kiswani, Aishwarya Padmakumar, Ajay Hotchandani, Akanksha Shukla, Akhiad Bercovich, Aleksander Ficek, Aleksandr Shaposhnikov, Alex Gronskiy, Alex Kondratenko, Alex Neefus, Alex Steiner, Alex Yang , et al. (522 additional authors not shown)

Abstract: We describe the pre-training, post-training, and quantization of Nemotron 3 Super, a 120 billion (active 12 billion) parameter hybrid Mamba-Attention Mixture-of-Experts model. Nemotron 3 Super is the first model in the Nemotron 3 family to 1) be pre-trained in NVFP4, 2) leverage LatentMoE, a new Mixture-of-Experts architecture that optimizes for both accuracy per FLOP and accuracy per parameter, a… ▽ More We describe the pre-training, post-training, and quantization of Nemotron 3 Super, a 120 billion (active 12 billion) parameter hybrid Mamba-Attention Mixture-of-Experts model. Nemotron 3 Super is the first model in the Nemotron 3 family to 1) be pre-trained in NVFP4, 2) leverage LatentMoE, a new Mixture-of-Experts architecture that optimizes for both accuracy per FLOP and accuracy per parameter, and 3) include MTP layers for inference acceleration through native speculative decoding. We pre-trained Nemotron 3 Super on 25 trillion tokens followed by post-training using supervised fine tuning (SFT) and reinforcement learning (RL). The final model supports up to 1M context length and achieves comparable accuracy on common benchmarks, while also achieving up to 2.2x and 7.5x higher inference throughput compared to GPT-OSS-120B and Qwen3.5-122B, respectively. Nemotron 3 Super datasets, along with the base, post-trained, and quantized checkpoints, are open-sourced on HuggingFace. △ Less

Submitted 14 April, 2026; originally announced April 2026.

Can LLMs Deobfuscate Binary Code? A Systematic Analysis of Large Language Models into Pseudocode Deobfuscation

Authors: Li Hu, Xiuwei Shang, Jieke Shi, Shaoyin Cheng, Junqi Zhang, Gangyang Li, Zhou Yang, Weiming Zhang, David Lo

Abstract: Deobfuscating binary code remains a fundamental challenge in reverse engineering, as obfuscation is widely used to hinder analysis and conceal program logic. Although large language models (LLMs) have shown promise in recovering semantics from obfuscated binaries, a systematic evaluation of their effectiveness is still lacking. In this work, we present BinDeObfBench, the first comprehensive benchm… ▽ More Deobfuscating binary code remains a fundamental challenge in reverse engineering, as obfuscation is widely used to hinder analysis and conceal program logic. Although large language models (LLMs) have shown promise in recovering semantics from obfuscated binaries, a systematic evaluation of their effectiveness is still lacking. In this work, we present BinDeObfBench, the first comprehensive benchmark for assessing LLM-based binary deobfuscation across diverse transformations spanning pre-compilation, compile-time, and post-compilation stages. Our evaluation shows that deobfuscation performance depends more on reasoning capability and domain expertise than on model scale, and that task-specific supervised fine-tuning consistently outperforms broad domain pre-training. Reasoning models can maintain robustness under severe obfuscation, generalize across different instruction set architectures (ISAs) and optimization levels. In-context learning benefits standard models but yields limited gains for reasoning models. Overall, our study highlights the importance of task-specific fine-tuning and reasoning-driven strategies, and positions BinDeObfBench as a basis for future work in binary deobfuscation. △ Less

Submitted 9 April, 2026; originally announced April 2026.

Evaluating LLM-Based 0-to-1 Software Generation in End-to-End CLI Tool Scenarios

Authors: Ruida Hu, Xinchen Wang, Chao Peng, Cuiyun Gao, David Lo

Abstract: The evolution of Large Language Models (LLMs) has catalyzed a paradigm shift towards intent-driven software development, where autonomous agents are expected to design and deliver complete, runnable software systems from scratch. However, existing benchmarks fail to adequately assess this 0-to-1 generation capability due to two fundamental limitations. First, they rely on predefined structural sca… ▽ More The evolution of Large Language Models (LLMs) has catalyzed a paradigm shift towards intent-driven software development, where autonomous agents are expected to design and deliver complete, runnable software systems from scratch. However, existing benchmarks fail to adequately assess this 0-to-1 generation capability due to two fundamental limitations. First, they rely on predefined structural scaffolds, which reduces the task to mere file-filling. Second, they depend on rigid white-box unit testing, which forces generated code to conform to specific internal implementations rather than validating end-to-end user-centric behavior. To bridge this gap, we introduce CLI-Tool-Bench, a novel, structure-agnostic benchmark designed to evaluate the ground-up generation of Command-Line Interface (CLI) tools. Powered by an automated black-box differential testing framework, the benchmark comprises 94 high-quality, real-world repositories spanning diverse programming languages and complexity levels. For each task, agents are provided with an empty workspace, forcing them to autonomously handle repository planning and dependencies. We evaluate the generated software by executing it in isolated sandboxes. The system-level side effects and terminal outputs are then compared against human-written oracles using a rigorous multi-tiered equivalence metric. Extensive evaluation of seven state-of-the-art LLMs reveals that the top-tier models achieve a maximum overall success rate of only 43.8%, highlighting that 0-to-1 software generation remains a highly challenging frontier. Furthermore, we discover that agents exhibit a strong tendency to generate monolithic code structures, and that higher token consumption does not necessarily yield better task performance. △ Less

Submitted 17 July, 2026; v1 submitted 8 April, 2026; originally announced April 2026.

Comments: Data link: https://github.com/kinesiatricssxilm14/CLI-Tool-Bench

AgentSZZ: Teaching the LLM Agent to Play Detective with Bug-Inducing Commits

Authors: Yunbo Lyu, Jieke Shi, Hong Jin Kang, Ratnadira Widyasari, Junda He, Yuqing Niu, Chengran Yang, Junkai Chen, Zhou Yang, Julia Lawall, David Lo

Abstract: The SZZ algorithm is the dominant technique for identifying bug-inducing commits and underpins many software engineering tasks, such as defect prediction and vulnerability analysis. Despite numerous variants, including recent LLM-based approaches, performance remains limited on developer-annotated datasets (e.g., recall of 0.552 on the Linux kernel). A key limitation is the reliance on git blame,… ▽ More The SZZ algorithm is the dominant technique for identifying bug-inducing commits and underpins many software engineering tasks, such as defect prediction and vulnerability analysis. Despite numerous variants, including recent LLM-based approaches, performance remains limited on developer-annotated datasets (e.g., recall of 0.552 on the Linux kernel). A key limitation is the reliance on git blame, which traces line-level changes within the same file, failing in common scenarios such as ghost and cross-file cases-making nearly one-quarter of bug-inducing commits inherently untraceable. Moreover, current approaches follow fixed pipelines that restrict iterative reasoning and exploration, unlike developers who investigate bugs through an interactive, multi-tool process. To address these challenges, we propose AgentSZZ, an agent-based framework that leverages LLM-driven agents to explore repositories and identify bug-inducing commits. Unlike prior methods, AgentSZZ integrates task-specific tools, domain knowledge, and a ReAct-style loop to enable adaptive and causal tracing of bugs. A structured compression module further improves efficiency by reducing redundant context while preserving key evidence. Extensive experiments on three widely used datasets show that AgentSZZ consistently outperforms state-of-the-art SZZ algorithms across all settings, achieving F1-score gains of up to 27.2% over prior LLM-based approaches. The improvements are especially pronounced in challenging scenarios such as cross-file and ghost commits, with recall gains of up to 300% and 60%, respectively. Ablation studies show that task-specific tools and domain knowledge are critical, while compression tool outputs reduce token consumption by over 30% with negligible impact. The replication package is available. △ Less

Submitted 2 April, 2026; originally announced April 2026.

TestDecision: Sequential Test Suite Generation via Greedy Optimization and Reinforcement Learning

Authors: Guoqing Wang, Chengran Yang, Xiaoxuan Zhou, Zeyu Sun, Bo Wang, David Lo, Dan Hao

Abstract: With the rapid evolution of LLMs, automated software testing is witnessing a paradigm shift. While proprietary models like GPT-4o demonstrate impressive capabilities, their high deployment costs and data privacy concerns make open-source LLMs the practical imperative for many academic and industrial scenarios. In the field of automated test generation, it has evolved to iterative workflows to cons… ▽ More With the rapid evolution of LLMs, automated software testing is witnessing a paradigm shift. While proprietary models like GPT-4o demonstrate impressive capabilities, their high deployment costs and data privacy concerns make open-source LLMs the practical imperative for many academic and industrial scenarios. In the field of automated test generation, it has evolved to iterative workflows to construct test suites based on LLMs. When utilizing open-source LLMs, we empirically observe they lack a suite-level perspective, suffering from structural myopia-failing to generate new tests with large marginal gain based on the current covered status. In this paper, from the perspective of sequences, we formalize test suite generation as a MDP and demonstrate that its objective exhibits monotone submodularity, which enables an effective relaxation of this NP-hard global optimization into a tractable step-wise greedy procedure. Guided by this insight, we propose TestDecision, which transforms LLMs into neural greedy experts. TestDecision consists of two synergistic components: (1) an inference framework which implements test suite construction following a step-wise greedy strategy; and (2) a training pipeline of reinforcement learning which equips the base LLM with sequential test generation ability to maximize marginal gain. Comprehensive evaluations on the ULT benchmark demonstrate that TestDecision significantly outperforms existing advanced methods. It brings an improvement between 38.15-52.37% in branch coverage and 298.22-558.88% in execution pass rate over all base models, achieving a comparable performance on 7B backbone with a much larger proprietary LLM GPT-5.2. Furthermore, TestDecision can find 58.43-95.45% more bugs than vanilla base LLMs and exhibit superior generalization on LiveCodeBench, proving its capability to construct high-quality test suites. △ Less

Submitted 17 April, 2026; v1 submitted 2 April, 2026; originally announced April 2026.

Comments: 22 pages, 4 figures; corrected metadata; marked corresponding author

Executing as You Generate: Hiding Execution Latency in LLM Code Interpreters

Authors: Zhensu Sun, Zhihao Lin, Zhi Chen, Chengran Yang, Mingyi Zhou, Li Li, David Lo

Abstract: Current LLM systems are increasingly equipped with a code interpreter that executes generated code to obtain results. This works serially: the model first generates the complete code, then an interpreter executes it. This sequential workflow leaves the executor idle during generation and the generator idle during execution, resulting in unnecessary end-to-end latency. Our key observation is that a… ▽ More Current LLM systems are increasingly equipped with a code interpreter that executes generated code to obtain results. This works serially: the model first generates the complete code, then an interpreter executes it. This sequential workflow leaves the executor idle during generation and the generator idle during execution, resulting in unnecessary end-to-end latency. Our key observation is that an LLM, unlike a human developer, emits code tokens left to right and does not backtrack over what it has already written. This makes it possible to start executing a piece of code while later tokens are still being generated. We formalize this parallel execution paradigm, modeling it as a three-stage pipeline of generation, detection, and execution, and derive closed-form latency bounds that characterize its speedup potential and operating regimes. We then present EAGER, a concrete implementation featuring AST-based chunking, dynamic batching with gated execution, and early error interruption. We evaluate EAGER across four benchmarks, seven LLMs, and three execution environments. The overlap mechanism hides almost all execution behind generation, reducing the non-overlapped portion of execution time by up to 99.8% and cutting end-to-end latency by up to 37.3% on error-free runs. △ Less

Submitted 22 June, 2026; v1 submitted 1 April, 2026; originally announced April 2026.

Comments: 10 pages

Compiling Code LLMs into Lightweight Executables

Authors: Jieke Shi, Junda He, Zhou Yang, Chengran Yang, Mykhailo Klymenko, Thong Hoang, Xiwei Xu, Zhenchang Xing, David Lo

Abstract: The demand for better prediction accuracy and higher execution performance in neural networks continues to grow. The emergence and success of Large Language Models (LLMs) have produced many cloud-based tools for software engineering tasks such as code suggestion. Although effective, cloud deployment raises concerns over privacy, latency, and reliance on network connectivity. Running LLMs locally o… ▽ More The demand for better prediction accuracy and higher execution performance in neural networks continues to grow. The emergence and success of Large Language Models (LLMs) have produced many cloud-based tools for software engineering tasks such as code suggestion. Although effective, cloud deployment raises concerns over privacy, latency, and reliance on network connectivity. Running LLMs locally on personal devices such as laptops would address these issues, because it enables offline use and reduces response time. However, local deployment is challenging, since commodity devices lack high-performance accelerators such as GPUs and are constrained by limited memory and compute capacity, which makes it hard to execute large models efficiently. We present Ditto, a framework that optimizes both the model size of Code LLMs and the inference programs that execute them. Our approach integrates two components. The first is a quantization technique inspired by product quantization, which groups model parameters into per-block codebooks via K-Means clustering and stores each weight as a bit-packed low-bitwidth index. The second component is a compilation pass integrated into LLVM that automatically detects and replaces unoptimized General Matrix-Vector Multiplication (GEMV) operations, with calls into Basic Linear Algebra Subprograms (BLAS) libraries that are highly optimized for the target hardware. The output of Ditto is a compiled executable that runs the selected Code LLM on commodity hardware. We evaluate Ditto on three popular Code LLMs, namely Code Llama, MagicCoder, and OpenCodeInterpreter, achieving up to 10.5$\times$ faster inference, 6.4$\times$ lower memory usage, and 10.5$\times$ lower energy consumption compared with their original inference pipelines, while preserving accuracy close to the full-precision models, with an average loss of only 0.27% in pass@1. △ Less

Submitted 24 June, 2026; v1 submitted 31 March, 2026; originally announced March 2026.

Comments: Accepted at the 34th ACM International Conference on the Foundations of Software Engineering (FSE 2026), 25 pages

Debt Behind the AI Boom: A Large-Scale Empirical Study of AI-Generated Code in the Wild

Authors: Yue Liu, Ratnadira Widyasari, Yanjie Zhao, Ivana Clairine Irsan, Junkai Chen, David Lo

Abstract: AI coding assistants are now widely used in software development. Software developers increasingly integrate AI-generated code into their codebases to improve productivity. Prior studies have shown that AI-generated code may contain code quality issues under controlled settings. However, we still know little about the real-world impact of AI-generated code on software quality and maintenance after… ▽ More AI coding assistants are now widely used in software development. Software developers increasingly integrate AI-generated code into their codebases to improve productivity. Prior studies have shown that AI-generated code may contain code quality issues under controlled settings. However, we still know little about the real-world impact of AI-generated code on software quality and maintenance after it is introduced into production repositories. In other words, it remains unclear whether such issues are quickly fixed or persist and accumulate over time as technical debt. In this paper, we conduct a large-scale empirical study on the technical debt introduced by AI coding assistants in the wild. To achieve that, we built a dataset of 302.6k verified AI-authored commits from 6,299 GitHub repositories, covering five widely used AI coding assistants. For each commit, we run static analysis before and after the change to precisely attribute which code smells, correctness issues, and security issues the AI introduced. We then track each introduced issue from the introducing commit to the latest repository revision to study its lifecycle. Our results show that we identified 484,366 distinct issues, and that code smells are by far the most common type, accounting for 89.3% of all issues. We also find that more than 15% of commits from every AI coding assistant introduce at least one issue, although the rates vary across tools. More importantly, 22.7% of tracked AI-introduced issues still survive at the latest version of the repository. These findings show that AI-generated code can introduce long-term maintenance costs into real software projects and highlight the need for stronger quality assurance in AI-assisted development. △ Less

Submitted 26 April, 2026; v1 submitted 30 March, 2026; originally announced March 2026.

Finding Memory Leaks in C/C++ Programs via Neuro-Symbolic Augmented Static Analysis

Authors: Huihui Huang, Jieke Shi, Bo Wang, Zhou Yang, David Lo

Abstract: Memory leaks remain prevalent in real-world C/C++ software. Static analyzers such as CodeQL provide scalable program analysis but frequently miss such bugs because they cannot recognize project-specific custom memory-management functions and lack path-sensitive control-flow modeling. We present MemHint, a neuro-symbolic pipeline that addresses both limitations by combining LLMs' semantic understan… ▽ More Memory leaks remain prevalent in real-world C/C++ software. Static analyzers such as CodeQL provide scalable program analysis but frequently miss such bugs because they cannot recognize project-specific custom memory-management functions and lack path-sensitive control-flow modeling. We present MemHint, a neuro-symbolic pipeline that addresses both limitations by combining LLMs' semantic understanding of code with Z3-based symbolic reasoning. MemHint parses the target codebase and applies an LLM to classify each function as a memory allocator, deallocator, or neither, producing function summaries that record which argument or return value carries memory ownership, extending the analyzer's built-in knowledge beyond standard primitives such as malloc and free. A Z3-based validation step checks each summary against the function's control-flow graph, discarding those whose claimed memory operation is unreachable on any feasible path. The validated summaries are injected into CodeQL and Infer via their respective extension mechanisms. Z3 path feasibility filtering then eliminates warnings on infeasible paths, and a final LLM-based validation step confirms whether each remaining warning is a genuine bug. On eight real-world C/C++ projects totaling over 3.6M lines of code, MemHint detects 54 unique memory leaks (53 confirmed/fixed) at approximately $1.7 per detected bug, compared to 19 by vanilla CodeQL and 3 by vanilla Infer. △ Less

Submitted 31 May, 2026; v1 submitted 28 March, 2026; originally announced March 2026.

Comments: 16 pages

Revisiting Vulnerability Patch Identification on Data in the Wild

Authors: Ivana Clairine Irsan, Ratnadira Widyasari, Ting Zhang, Huihui Huang, Ferdian Thung, Yikun Li, Lwin Khin Shar, Eng Lieh Ouh, Hong Jin Kang, David Lo

Abstract: Attacks can exploit zero-day or one-day vulnerabilities that are not publicly disclosed. To detect these vulnerabilities, security researchers monitor development activities in open-source repositories to identify unreported security patches. The sheer volume of commits makes this task infeasible to accomplish manually. Consequently, security patch detectors commonly trained and evaluated on secur… ▽ More Attacks can exploit zero-day or one-day vulnerabilities that are not publicly disclosed. To detect these vulnerabilities, security researchers monitor development activities in open-source repositories to identify unreported security patches. The sheer volume of commits makes this task infeasible to accomplish manually. Consequently, security patch detectors commonly trained and evaluated on security patches linked from vulnerability reports in the National Vulnerability Database (NVD). In this study, we assess the effectiveness of these detectors when applied in-the-wild. Our results show that models trained on NVD-derived data show substantially decreased performance, with decreases in F1-score of up to 90\% when tested on in-the-wild security patches, rendering them impractical for real-world use. An analysis comparing security patches identified in-the-wild and commits linked from NVD reveals that they can be easily distinguished from each other. Security patches associated with NVD have different distribution of commit messages, vulnerability types, and composition of changes. These differences suggest that NVD may be unsuitable as the \textit{sole} source of data for training models to detect security patches. We find that constructing a dataset that combines security patches from NVD data with a small subset of manually identified security patches can improve model robustness. △ Less

Submitted 17 March, 2026; originally announced March 2026.