arXiv subjects

Haipeng Cai

Publications and source records attributed to Haipeng Cai.

At least 19 recordsLinked to original sources

arXiv2026

The Data Problem in Software Vulnerability Analysis: Artifacts, Quality, and Consumption

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

cs.SE↗

arXiv2026

POLYFLOW: A Neuro-Symbolic Framework for Static Cross-Language Information Flow Analysis

Modern software systems are commonly constructed in multiple, interacting programming languages. This construction leads to additional, often stealthy vulnerabilities buried in complex information flow due to language interactions. Existing static analyzers are impeded by the heterogeneous semantics of different languages, whereas dynamic approaches suffer from the limited coverage of (available and/or generated) test inputs. In this paper, we develop PolyFlow, a neural-symbolic framework for statically reasoning about information flow across language boundaries, combining large language models (LLMs) and static analysis synergistically. Governed by the control-flow representation of a given multi-language system, PolyFlow leverages LLMs to identify implicit flow facts due to challenging language features, hence augmenting the base representation and then propagating data flow through the system. It tackles inherent barriers (e.g., token limit and hallucination) of LLMs by putting them under careful guidance (e.g., static-analysis-guided scoping, context management, and fact checking), along with a multi-LLM expert panel for negotiated validation. Our experiments on real-world Python-C and Java-C systems show that PolyFlow is cost-effective and superior to various kinds of state-of-the-art baselines, revealing previously unknown cross-language vulnerabilities that are missed by all the baselines.

cs.CR↗

arXiv2026

A Comprehensive Study of Native Code Bugs in Python Applications

The impact of Python applications has been evidenced by their widespread presence in some of the most impactful software domains, such as machine learning frameworks and scientific computing platforms. These applications often integrate native code components written in a lower-level programming language like C. This multilingual construction brings various benefits such as greater performance efficiency and easier interoperability with diverse runtime environments. However, bugs in the native code (i.e., native code bugs), which are usually stealthy, also constitute a major additional challenge to the quality of the Python applications as a whole. Yet despite existing relevant studies, there remains a lack of comprehensive understanding of native code bugs in Python applications. In this paper, we aim to mitigate this knowledge gap through the first in-depth study of such bugs, dissecting their common symptoms, introducing locations, manifestation characteristics, root causes, and fixes. Based on our extensive automated and manual analyses of 216 native code bugs in real-world Python projects on GitHub, we obtained novel findings about and new insights into the occurrence mechanisms and resolution strategies of those bugs.

cs.SE↗

arXiv2026

LLM-Based Agents for Software and Systems Security: Approaches, Applications, and Assessment

Software and systems security workflows are typically procedural: analysts inspect heterogeneous artifacts, form hypotheses, invoke tools, interpret outputs, and revise plans. Large language model (LLM)-based agents, which can plan, use tools, retain state, and revise actions across multi-step workflows, are being rapidly adopted to automate this work. Given the consequences of delegating security decisions to autonomous systems, understanding how such agents are built, used, and assessed is crucial. Yet to this date, there remains a lack of systematic understanding of what has been done and how far we are in this field: the term "agent" is applied inconsistently, applications differ sharply in risk, and assessment protocols are often incomparable. To gain a comprehensive and coherent view of this area hence inform relevant future research, this paper provides a systematic literature review of the (1) technical approaches, including agent architecture, perception, memory, reasoning and planning, action space, orchestration, and self-improvement, (2) applications, with respect to the security tasks served, and (3) assessment, including the datasets, outcome and trajectory metrics, safety measures, and baselines considered, over the peer-reviewed literature spanning the emergence of this area (2023--2026). Our synthesis reveals a field that has built agents able to act but not yet agents whose authority is bounded or whose behavior is auditable. In addition to knowledge systematization, we also extend our insights into the limitations of and challenges faced by current approach, application, and assessment designs, which shed light on potentially promising future research directions.

cs.CR↗

arXiv2026

How Reliable Are NVD CWE Labels? A Large-Scale Semantic Audit with Seclometry

CWE labels in the National Vulnerability Database (NVD) are widely treated as ground truth for vulnerability search, scanner evaluation, benchmark construction, learning-based security tools, and vulnerability prioritization. Yet their reliability has not been systematically measured at scale, despite growing concerns about NVD's enrichment backlog and anecdotal reports of inaccurate, ambiguous, or missing labels. This paper presents a large-scale, code-semantics-grounded measurement of CWE labeling quality in NVD. We build CWEAgent, a validated auditing instrument based on seclometry, a structured representation of vulnerability semantics that captures the root cause, trigger condition, violated security property, exploit mechanism, and impact of vulnerable code. On a manually curated benchmark of 100 open-source CVEs, CWEAgent achieves 85% top-1 accuracy and 92% ambiguity-aware accuracy. Applying CWEAgent to 15,556 open-source CVEs disclosed from 2017-2026, we find that only 49.70% of NVD CWE labels exactly match the code-grounded label. Another 31.37% are defensible alternatives under taxonomy ambiguity, while 3.63% are evidence-inconsistent likely errors. Label reliability varies sharply by assigning organization and weakness type, and apparent project- or language-level differences are largely composition effects of those underlying weakness types. Evidence-inconsistent labels have also increased over time. Through manual review of 434 confirmed mislabels, we identify six recurring error patterns, showing that CWE noise is a structural problem in vulnerability metadata rather than isolated annotation mistakes.

cs.CR↗

arXiv2026

Dissecting Software Graphs: Structural Insights for Driver-Guided Fuzzing

Many software systems expose multiple execution modes through command-line options, subcommands, and configuration flags. For such programs, fuzzing depends on both mutated inputs and the invoked mode. Yet evaluations still focus on coverage and bug counts, leaving unclear how execution modes partition, overlap, and miss software structure, and how these differences affect effectiveness. We present an empirical study of software structure under multi-driver fuzzing. We propose a structural abstraction that uses a static call graph as a shared backbone and projects driver-specific dynamic coverage onto it to derive driver-induced subgraphs. Based on this abstraction, we develop a four-phase methodology for backbone construction, fuzzing and profiling, graph-based analysis, and research-question-driven evaluation. We apply it to 27 OSS-Fuzz-derived C/C++ projects, spanning 43 executables and 854 driver configurations. Under the same total budget, multi-driver fuzzing outperforms the best single-driver baseline, increasing covered call-graph nodes by 27.9% and CFG-edge coverage by 73.5%, and revealing 11 unique bugs and abnormal behaviors largely missed by single-driver fuzzing. However, driver contributions are uneven, subgraphs differ substantially in cohesion, fragmentation, modularity, overlap, and residual under-exploration follows recurring regimes rather than a homogeneous tail. These results show that multi-driver fuzzing is fundamentally a structural exploration problem.

cs.SE↗

arXiv2026

Neuro-Symbolic Proof-of-Vulnerability Generation with Open-Weight Models

Software vulnerabilities are persistent, but validating them remains difficult: a Proof-of-Vulnerability (PoV) requires a concrete input that triggers the vulnerable behavior, yet public triggering inputs are often unavailable for disclosed vulnerabilities. Existing techniques make different tradeoffs in effectiveness, scalability, cost, and controllability, leaving room for complementary designs. To complement them, we present POVGEN, a low-cost neuro-symbolic framework that makes PoV generation cost-effective via semantic focusing and LLM-guided constraint reasoning using open-weight models. POVGEN first localizes vulnerability-relevant regions (utilizing patch information if available), then performs path-sensitive reachability analysis, and finally generates PoVs by extracting and solving constraints with LLM-guided reasoning backed by an SMT solver. POVGEN successfully generates PoVs for 78.98% of vulnerabilities in a recent benchmark, outperforming fuzzing (up to 50.20%) and symbolic execution (2.45%). On 250 real-world CVEs without public PoVs, it generates valid PoVs for 74.80% of cases and reproduces 65.1% when without patch information. The fine-tuned open-weight models match frontier commercial LLMs on key sub-tasks (i.e., the core constraint-reasoning steps) while running locally at no per-sample API cost. Applying the generated PoVs revealed six flawed patches in disclosed CVEs (all subsequently fixed) and five previously unreported vulnerabilities (of which four have been confirmed and fixed by the developers).

cs.SE↗

arXiv2026

Representation Matters: An Empirical Study of Program Representations for LLM Vulnerability Reasoning

Large Language Models (LLMs) are increasingly used for automated vulnerability detection, but it remains unclear how program structure and semantics should be represented for LLM-based reasoning. Most prompting-based approaches provide raw source code, implicitly assuming that more source-level context gives the model better evidence. This paper challenges that assumption through RepBench, an empirical benchmark comparing raw source code with static-analysis-based program representations. RepBench converts real-world C/C++ vulnerability testcases into multiple representations: raw source, Abstract Syntax Trees (ASTs), Control-Flow Graphs (CFGs), Program Dependence Graphs (PDGs), their combinations, and an auxiliary track of enriched PDGs (ePDGs). Using a curated PrimeVul-derived corpus of 107 Joern-based testcases across five CWE categories, we evaluate ten representation variants under a fixed Chain-of-Thought and structured-output protocol, plus 19 additional ePDG cases generated through VulChecker/Hector. Representation choice substantially affects LLM vulnerability reasoning. The strongest variant, AST+PDG, achieves 83.2% accuracy, compared with 53.5% for raw source. At the family level, graph-only prompts outperform both source-only and source-plus-graph prompts while requiring far less prompt overhead. These results reveal a context dilution effect: adding raw source code to compact structural graph evidence can degrade reasoning by making vulnerability-relevant evidence less salient. Overall, our findings show that carefully selected structural representations offer a better accuracy-overhead tradeoff than simply giving LLMs more raw input, and suggest that static analysis can serve as an effective prompt-construction layer for security-focused LLM reasoning.

cs.CR↗

arXiv2026

SoK: AI Secure Code Generation: Progress, Pitfalls, and Paths Forward

The increasing use of AI systems for code generation raises a central security question: what can today's models and coding agents actually do to produce secure code, where do they still fail, and what would move the field forward? Existing work has explored prompting, fine-tuning, reinforcement learning, and agentic workflows for secure code generation, but the field still lacks a systematic understanding of how these techniques improve security and why substantial failures persist. In this SoK, we systematize the progress, pitfalls, and paths forward for AI secure code generation. We introduce a three-level framework that measures models' natural-language understanding of secure coding principles, their code-level actuation of those principles during generation, and the knowledge--actuation gaps between the two. We instantiate this framework across models and coding agents on benchmarks covering both isolated function-level security and full web-application security. Our results show that secure-coding-principle understanding is a statistically strong predictor of code-level outcomes, including functional correctness, security, and joint functional-security correctness. Yet substantial knowledge--actuation gaps remain: models can recognize relevant security principles but still fail to translate them into secure and functional code. These findings offer a principle-centered account of where AI secure code generation stands today and identify concrete paths forward through principle-guided generation, evaluation, benchmarking, and agentic workflows.

cs.CR↗

arXiv2026

Local LLM Agents as Vulnerable Runtimes:A Source-Code Audit of the Agent Runtime Layer

Local LLM agents such as OpenClaw and Nanobot run on end-user machines and act on host resources - the shell, filesystem, browser, stored credentials, and messaging applications - through natural-language goals. These agents have become privileged software runtimes that mediate between user intent, model outputs, and host-level actions. Existing research characterizes the landscape through prompt injection, malicious skills, marketplace risks, or black-box evaluation of agents. But the implementation layer that performs this mediation, the prompt builder, parser, tool dispatcher, skill loader, memory writer, network client, and permission gate, has remained an unexamined safety boundary. To our knowledge, no prior work has examined the agent's source tree to audit these components for implementation-level security weaknesses. We present CLAWAUDIT, a static auditing framework for measuring vulnerability exposure in local LLM agent runtimes. CLAWAUDIT derives a five-category vulnerability taxonomy from STRIDE and develops custom static-analysis rules that target agent-specific patterns absent from established rule sets for vulnerability analysis. We instantiate the taxonomy in two backends, 47 Semgrep YAML rules and 30 CodeQL queries, and evaluate on OPENCLAWBENCH, a benchmark of 446 source-code-level advisories from the OpenClaw repository and split temporally into 229 rule-derivation (train) and 217 held-out (test) advisories. On the held-out test, CLAWAUDIT raises Semgrep recall from 21.7% (Pro baseline) to 66.8%, and CodeQL recall from 13.8% (security-extended) to 75.1%. Train/test gaps remain within 4 percentage points for all four configurations, indicating that the rules generalize to vulnerabilities unseen during rule writing. A preliminary live-code audit shows that these recall-oriented rules require manual triage, motivating semantic filtering before production deployment.

cs.CR↗

arXiv2025

An Exploratory Study of Bayesian Prompt Optimization for Test-Driven Code Generation with Large Language Models

We consider the task of generating functionally correct code using large language models (LLMs). The correctness of generated code is influenced by the prompt used to query the given base LLM. We formulate the problem of finding the appropriate prompt as combinatorial search process and propose a Bayesian optimization (BO) approach referred to as {\em BO for Code GENeration (BODE-GEN)}. BODE-GEN performs an adaptive data-driven search over prompts guided by training data in the form of prompts tried and the functional accuracy of the generated code over a set of given test cases. The key insight is to perform BO in continuous embedding space by using an auxiliary LLM to bridge the gap between discrete prompt space and continuous embedding space. We leverage two synergistic ideas, namely, random projections and dimensionality scaled priors, to build effective Gaussian process based surrogate models over the high-dimensional embedding space. Our experiments on the HumanEval+ benchmark using multiple base LLMs show that BODE-GEN can improve performance in terms of code generation accuracy compared to fixed prompts and manual prompt engineering. Additionally, we demonstrate that BODE-GEN is sample-efficient, requiring relatively few iterations of BO to demonstrate improvements in code accuracy.

cs.SE↗

arXiv2025

DualGauge: Automated Joint Security-Functionality Benchmarking of Specification-Only Code Generation by LLMs and Coding Agents

Large language models (LLMs) and LLM-based coding agents are now used to generate code from natural-language specifications, yet ensuring such code is both functionally correct and secure remains a challenge. We present DualGauge, the first fully automated framework for jointly evaluating correctness and security of specification-only code generation, supported by DualGauge-Bench, a language-agnostic benchmark of 307 coding tasks each paired with functional and security tests derived from the same specification. Evaluating 10 representative LLMs across Python, C++, and JavaScript, we find that functional correctness substantially overestimates reliable code generation: even the strongest model remains below 15% joint security-functionality success in every language. Common model-side factors--scale, extended thinking, quantization, instruction tuning, and code specialization--do not reliably improve joint performance, suggesting secure-and-correct code generation does not simply emerge from stronger coding capability. Evaluation of 3 leading agentic coding systems (Codex, OpenHands, and Claude Code) shows that iterative scaffolding provides no advantage over direct (LLM-based) generation on specification-only tasks. A qualitative audit reveals failures concentrate at the output contract boundary and in guards that exist but are insufficient--patterns that only joint benchmarking reliably exposes.

cs.SE↗

arXiv2025

Fortifying LLM-Based Code Generation with Graph-Based Reasoning on Secure Coding Practices

The code generation capabilities of Large Language Models (LLMs) have transformed the field of software development. However, this advancement also presents significant security challenges, as LLM-generated code often contains vulnerabilities. One direction of research strengthens LLMs by injecting or refining security knowledge through curated datasets, model tuning, or static analyzers. While effective in certain settings, these methods can be resource-intensive, less adaptable to zero-day vulnerabilities, and often inapplicable to proprietary models. To address these challenges, we introduce GRASP, which explores a new direction that focuses on structured reasoning over Secure Coding Practices(SCPs) rather than additional training or external feedback. GRASP comprises two key ideas: (1) an SCP graph that organizes SCPs into a Directed Acyclic Graph (DAG) capturing dependencies and relationships, and (2) a graph-based reasoning process that systematically guides LLMs through relevant SCPs for code generation. This design enables interpretable, model-agnostic, and scalable security improvements, particularly for previously unseen vulnerabilities. Our evaluation shows that GRASP consistently achieves Security Rates (SR) exceeding 80% across multiple LLMs, and delivers up to 88% improvements over baselines on zero-day vulnerabilities.

cs.CR↗

arXiv2025

System Prompt Poisoning: Persistent Attacks on Large Language Models Beyond User Injection

Large language models (LLMs) have gained widespread adoption across diverse applications due to their impressive generative capabilities. Their plug-and-play nature enables both developers and end users to interact with these models through simple prompts. However, as LLMs become more integrated into various systems in diverse domains, concerns around their security are growing. Existing studies mainly focus on threats arising from user prompts (e.g. prompt injection attack) and model output (e.g. model inversion attack), while the security of system prompts remains largely overlooked. This work bridges the critical gap. We introduce system prompt poisoning, a new attack vector against LLMs that, unlike traditional user prompt injection, poisons system prompts hence persistently impacts all subsequent user interactions and model responses. We systematically investigate four practical attack strategies in various poisoning scenarios. Through demonstration on both generative and reasoning LLMs, we show that system prompt poisoning is highly feasible without requiring jailbreak techniques, and effective across a wide range of tasks, including those in mathematics, coding, logical reasoning, and natural language processing. Importantly, our findings reveal that the attack remains effective even when user prompts employ advanced prompting techniques like chain-of-thought (CoT). We also show that such techniques, including CoT and retrieval-augmentation-generation (RAG), which are proven to be effective for improving LLM performance in a wide range of tasks, are significantly weakened in their effectiveness by system prompt poisoning.

cs.CR↗

arXiv2024

Enhancing Android Malware Detection: The Influence of ChatGPT on Decision-centric Task

With the rise of large language models, such as ChatGPT, non-decisional models have been applied to various tasks. Moreover, ChatGPT has drawn attention to the traditional decision-centric task of Android malware detection. Despite effective detection methods proposed by scholars, they face low interpretability issues. Specifically, while these methods excel in classifying applications as benign or malicious and can detect malicious behavior, they often fail to provide detailed explanations for the decisions they make. This challenge raises concerns about the reliability of existing detection schemes and questions their true ability to understand complex data. In this study, we investigate the influence of the non-decisional model, ChatGPT, on the traditional decision-centric task of Android malware detection. We choose three state-of-the-art solutions, Drebin, XMAL, and MaMaDroid, conduct a series of experiments on publicly available datasets, and carry out a comprehensive comparison and analysis. Our findings indicate that these decision-driven solutions primarily rely on statistical patterns within datasets to make decisions, rather than genuinely understanding the underlying data. In contrast, ChatGPT, as a non-decisional model, excels in providing comprehensive analysis reports, substantially enhancing interpretability. Furthermore, we conduct surveys among experienced developers. The result highlights developers' preference for ChatGPT, as it offers in-depth insights and enhances efficiency and understanding of challenges. Meanwhile, these studies and analyses offer profound insights, presenting developers with a novel perspective on Android malware detection--enhancing the reliability of detection results from a non-decisional perspective.

cs.CR↗

arXiv2024

APPATCH: Automated Adaptive Prompting Large Language Models for Real-World Software Vulnerability Patching

Timely and effective vulnerability patching is essential for cybersecurity defense, for which various approaches have been proposed yet still struggle to generate valid and correct patches for real-world vulnerabilities. In this paper, we leverage the power and merits of pre-trained language language models (LLMs) to enable automated vulnerability patching using no test input/exploit evidence and without model training/fine-tuning. To elicit LLMs to effectively reason about vulnerable code behaviors, which is essential for quality patch generation, we introduce vulnerability semantics reasoning and adaptive prompting on LLMs and instantiate the methodology as APPATCH, an automated LLM-based patching system. Our evaluation of APPATCH on 97 zero-day vulnerabilities and 20 existing vulnerabilities demonstrates its superior performance to both existing prompting methods and state-of-the-art non-LLM-based techniques (by up to 28.33% in F1 and 182.26% in recall over the best baseline). Through APPATCH, we demonstrate what helps for LLM-based patching and how, as well as discussing what still lacks and why.

cs.CR↗

arXiv2024

VulScribeR: Exploring RAG-based Vulnerability Augmentation with LLMs

Detecting vulnerabilities is vital for software security, yet deep learning-based vulnerability detectors (DLVD) face a data shortage, which limits their effectiveness. Data augmentation can potentially alleviate the data shortage, but augmenting vulnerable code is challenging and requires a generative solution that maintains vulnerability. Previous works have only focused on generating samples that contain single statements or specific types of vulnerabilities. Recently, large language models (LLMs) have been used to solve various code generation and comprehension tasks with inspiring results, especially when fused with retrieval augmented generation (RAG). Therefore, we propose VulScribeR, a novel LLM-based solution that leverages carefully curated prompt templates to augment vulnerable datasets. More specifically, we explore three strategies to augment both single and multi-statement vulnerabilities, with LLMs, namely Mutation, Injection, and Extension. Our extensive evaluation across four vulnerability datasets and DLVD models, using three LLMs, show that our approach beats two SOTA methods Vulgen and VGX, and Random Oversampling (ROS) by 27.48%, 27.93%, and 15.41% in f1-score with 5K generated vulnerable samples on average, and 53.84%, 54.10%, 69.90%, and 40.93% with 15K generated vulnerable samples. Our approach demonstrates its feasibility for large-scale data augmentation by generating 1K samples at as cheap as US$ 1.88.

cs.SE↗

arXiv2024

From LLMs to LLM-based Agents for Software Engineering: A Survey of Current, Challenges and Future

With the rise of large language models (LLMs), researchers are increasingly exploring their applications in var ious vertical domains, such as software engineering. LLMs have achieved remarkable success in areas including code generation and vulnerability detection. However, they also exhibit numerous limitations and shortcomings. LLM-based agents, a novel tech nology with the potential for Artificial General Intelligence (AGI), combine LLMs as the core for decision-making and action-taking, addressing some of the inherent limitations of LLMs such as lack of autonomy and self-improvement. Despite numerous studies and surveys exploring the possibility of using LLMs in software engineering, it lacks a clear distinction between LLMs and LLM based agents. It is still in its early stage for a unified standard and benchmarking to qualify an LLM solution as an LLM-based agent in its domain. In this survey, we broadly investigate the current practice and solutions for LLMs and LLM-based agents for software engineering. In particular we summarise six key topics: requirement engineering, code generation, autonomous decision-making, software design, test generation, and software maintenance. We review and differentiate the work of LLMs and LLM-based agents from these six topics, examining their differences and similarities in tasks, benchmarks, and evaluation metrics. Finally, we discuss the models and benchmarks used, providing a comprehensive analysis of their applications and effectiveness in software engineering. We anticipate this work will shed some lights on pushing the boundaries of LLM-based agents in software engineering for future research.

cs.SE↗