Large language models (LLMs) demonstrate considerable proficiency in numerous coding-related tasks; however, their capabilities in detecting software vulnerabilities remain limited. This limitation primarily stems from two factors: (1) the absence of reasoning data related to vulnerabilities, which hinders the models' ability to capture underlying vulnerability patterns; and (2) their focus on learning semantic representations rather than the reason behind them, thus failing to recognize semantically similar vulnerability samples. Furthermore, the development of LLMs specialized in vulnerability detection is challenging, particularly in environments characterized by the scarcity of high-quality datasets. In this paper, we propose a novel framework ReVD that excels at mining vulnerability patterns through reasoning data synthesizing and vulnerability-specific preference optimization. Specifically, we construct forward and backward reasoning processes for vulnerability and corresponding fixed code, ensuring the synthesis of high-quality reasoning data. Moreover, we design the triplet supervised fine-tuning followed by curriculum online preference optimization for enabling ReVD to better understand vulnerability patterns. The extensive experiments conducted on PrimeVul and SVEN datasets demonstrate that ReVD sets new state-of-the-art for LLM-based software vulnerability detection, e.g., 12.24\%-22.77\% improvement in the accuracy. The source code and data are available at https://github.com/Xin-Cheng-Wen/PO4Vul.
As Large Language Models (LLMs) evolve in understanding and generating code, accurately evaluating their reliability in analyzing source code vulnerabilities becomes increasingly vital. While studies have examined LLM capabilities in tasks like vulnerability detection and repair, they often overlook the importance of both structure and semantic reasoning crucial for trustworthy vulnerability analysis. To address this gap, we introduce SV-TrustEval-C, a benchmark designed to evaluate LLMs' abilities for vulnerability analysis of code written in the C programming language through two key dimensions: structure reasoning - assessing how models identify relationships between code elements under varying data and control flow complexities; and semantic reasoning - examining their logical consistency in scenarios where code is structurally and semantically perturbed. Our results show that current LLMs are far from satisfactory in understanding complex code relationships and that their vulnerability analyses rely more on pattern matching than on robust logical reasoning. These findings underscore the effectiveness of the SV-TrustEval-C benchmark and highlight critical areas for enhancing the reasoning capabilities and trustworthiness of LLMs in real-world vulnerability analysis tasks. Our initial benchmark dataset is publicly available.
Large Language Model (LLM) agents can automate cybersecurity tasks and can adapt to the evolving cybersecurity landscape without re-engineering. While LLM agents have demonstrated cybersecurity capabilities on Capture-The-Flag (CTF) competitions, they have two key limitations: accessing latest cybersecurity expertise beyond training data, and integrating new knowledge into complex task planning. Knowledge-based approaches that incorporate technical understanding into the task-solving automation can tackle these limitations. We present CRAKEN, a knowledge-based LLM agent framework that improves cybersecurity capability through three core mechanisms: contextual decomposition of task-critical information, iterative self-reflected knowledge retrieval, and knowledge-hint injection that transforms insights into adaptive attack strategies. Comprehensive evaluations with different configurations show CRAKEN's effectiveness in multi-stage vulnerability detection and exploitation compared to previous approaches. Our extensible architecture establishes new methodologies for embedding new security knowledge into LLM-driven cybersecurity agentic systems. With a knowledge database of CTF writeups, CRAKEN obtained an accuracy of 22% on NYU CTF Bench, outperforming prior works by 3% and achieving state-of-the-art results. On evaluation of MITRE ATT&CK techniques, CRAKEN solves 25-30% more techniques than prior work, demonstrating improved cybersecurity capabilities via knowledge-based execution. We make our framework open source to public https://github.com/NYU-LLM-CTF/nyuctf_agents_craken.
Code security and usability are both essential for various coding assistant applications driven by large language models (LLMs). Current code security benchmarks focus solely on single evaluation task and paradigm, such as code completion and generation, lacking comprehensive assessment across dimensions like secure code generation, vulnerability repair and discrimination. In this paper, we first propose CoV-Eval, a multi-task benchmark covering various tasks such as code completion, vulnerability repair, vulnerability detection and classification, for comprehensive evaluation of LLM code security. Besides, we developed VC-Judge, an improved judgment model that aligns closely with human experts and can review LLM-generated programs for vulnerabilities in a more efficient and reliable way. We conduct a comprehensive evaluation of 20 proprietary and open-source LLMs. Overall, while most LLMs identify vulnerable codes well, they still tend to generate insecure codes and struggle with recognizing specific vulnerability types and performing repairs. Extensive experiments and qualitative analyses reveal key challenges and optimization directions, offering insights for future research in LLM code security.
Software vulnerabilities such as buffer overflows and SQL injections are a major source of security breaches. Traditional methods for vulnerability detection remain essential but are limited by high false positive rates, scalability issues, and reliance on manual effort. These constraints have driven interest in AI-based approaches to automated vulnerability detection and secure code generation. While Large Language Models (LLMs) have opened new avenues for classification tasks, their complexity and opacity pose challenges for interpretability and deployment. Sparse Autoencoder offer a promising solution to this problem. We explore whether SAEs can serve as a lightweight, interpretable alternative for bug detection in Java functions. We evaluate the effectiveness of SAEs when applied to representations from GPT-2 Small and Gemma 2B, examining their capacity to highlight buggy behaviour without fine-tuning the underlying LLMs. We found that SAE-derived features enable bug detection with an F1 score of up to 89%, consistently outperforming fine-tuned transformer encoder baselines. Our work provides the first empirical evidence that SAEs can be used to detect software bugs directly from the internal representations of pretrained LLMs, without any fine-tuning or task-specific supervision.
Large Language Models (LLMs) can achieve strong performance on everyday coding tasks, but they can fail on complex tasks that require non-trivial reasoning about program semantics. Finding training examples to teach LLMs to solve these tasks can be challenging. In this work, we explore a method to synthetically generate code reasoning training data based on a semantic inequivalence game SInQ: a generator agent creates program variants that are semantically distinct, derived from a dataset of real-world programming tasks, while an evaluator agent has to identify input examples that cause the original programs and the generated variants to diverge in their behaviour, with the agents training each other semi-adversarially. We prove that this setup enables theoretically unlimited improvement through self-play in the limit of infinite computational resources. We evaluated our approach on multiple code generation and understanding benchmarks, including cross-language vulnerability detection (Lu et al., 2021), where our method improves vulnerability detection in C/C++ code despite being trained exclusively on Python code, and the challenging Python builtin identifier swap benchmark (Miceli-Barone et al., 2023), showing that whereas modern LLMs still struggle with this benchmark, our approach yields substantial improvements. We release the code needed to replicate the experiments, as well as the generated synthetic data, which can be used to fine-tune LLMs.
The application of Artificial Intelligence has become a powerful approach to detecting software vulnerabilities. However, effective vulnerability detection relies on accurately capturing the semantic structure of code and its contextual relationships. Given that the same functionality can be implemented in various forms, a preprocessing tool that standardizes code representation is important. This tool must be efficient, adaptable across programming languages, and capable of supporting new transformations. To address this challenge, we build on the existing SCoPE framework and introduce SCoPE2, an enhanced version with improved performance. We compare both versions in terms of processing time and memory usage and evaluate their impact on a Large Language Model (LLM) for vulnerability detection. Our results show a 97.3\% reduction in processing time with SCoPE2, along with an improved F1-score for the LLM, solely due to the refined preprocessing approach.
Static analysis plays a crucial role in software vulnerability detection, yet faces a persistent precision-scalability tradeoff. In large codebases like the Linux kernel, traditional static analysis tools often generate excessive false positives due to simplified vulnerability modeling and overapproximation of path and data constraints. While large language models (LLMs) demonstrate promising code understanding capabilities, their direct application to program analysis remains unreliable due to inherent reasoning limitations. We introduce BugLens, a post-refinement framework that significantly enhances static analysis precision for bug detection. BugLens guides LLMs through structured reasoning steps to assess security impact and validate constraints from the source code. When evaluated on Linux kernel taint-style bugs detected by static analysis tools, BugLens improves precision approximately 7-fold (from 0.10 to 0.72), substantially reducing false positives while uncovering four previously unreported vulnerabilities. Our results demonstrate that a well-structured, fully automated LLM-based workflow can effectively complement and enhance traditional static analysis techniques.
Large language models (LLMs) have shown promising performance in software vulnerability detection (SVD), yet their reasoning capabilities remain unreliable. Existing approaches relying on chain-of-thought (CoT) struggle to provide relevant and actionable security assessments. Additionally, effective SVD requires not only generating coherent reasoning but also differentiating between well-founded and misleading yet plausible security assessments, an aspect overlooked in prior work. To this end, we introduce R2Vul, a novel approach that distills structured reasoning into small LLMs using reinforcement learning from AI feedback (RLAIF). Through RLAIF, R2Vul enables LLMs to produce structured, security-aware reasoning that is actionable and reliable while explicitly learning to distinguish valid assessments from misleading ones. We evaluate R2Vul across five languages against SAST tools, CoT, instruction tuning, and classification-based baselines. Our results show that R2Vul with structured reasoning distillation enables a 1.5B student LLM to rival larger models while improving generalization to out-of-distribution vulnerabilities. Beyond model improvements, we contribute a large-scale, multilingual preference dataset featuring structured reasoning to support future research in SVD.
As AI advances, broader consensus is needed to determine research priorities. This endeavor discusses offensive AI and provides guidance by leveraging Sustainable Development Goals (SDGs) and interpretability techniques. The objective is to more effectively establish priorities that balance societal benefits against risks. The two forms of offensive AI evaluated in this study are vulnerability detection agents, which solve Capture- The-Flag challenges, and AI-powered malware.
Automating software vulnerability detection (SVD) remains a critical challenge in an era of increasingly complex and interdependent software systems. Despite significant advances in Large Language Models (LLMs) for code analysis, prevailing evaluation methodologies often lack the \textbf{context-aware robustness} necessary to capture real-world intricacies and cross-component interactions. To address these limitations, we present \textbf{VulnSage}, a comprehensive evaluation framework and a dataset curated from diverse, large-scale open-source system software projects developed in C/C++. Unlike prior datasets, it leverages a heuristic noise pre-filtering approach combined with LLM-based reasoning to ensure a representative and minimally noisy spectrum of vulnerabilities. The framework supports multi-granular analysis across function, file, and inter-function levels and employs four diverse zero-shot prompt strategies: Baseline, Chain-of-Thought, Think, and Think & Verify. Through this evaluation, we uncover that structured reasoning prompts substantially improve LLM performance, with Think & Verify reducing ambiguous responses from 20.3% to 9.1% while increasing accuracy. We further demonstrate that code-specialized models consistently outperform general-purpose alternatives, with performance varying significantly across vulnerability types, revealing that no single approach universally excels across all security contexts. Link to dataset and codes: https://github.com/Erroristotle/VulnSage.git
Identifying vulnerabilities in source code is crucial, especially in critical software components. Existing methods such as static analysis, dynamic analysis, formal verification, and recently Large Language Models are widely used to detect security flaws. This paper introduces CASTLE (CWE Automated Security Testing and Low-Level Evaluation), a benchmarking framework for evaluating the vulnerability detection capabilities of different methods. We assess 13 static analysis tools, 10 LLMs, and 2 formal verification tools using a hand-crafted dataset of 250 micro-benchmark programs covering 25 common CWEs. We propose the CASTLE Score, a novel evaluation metric to ensure fair comparison. Our results reveal key differences: ESBMC (a formal verification tool) minimizes false positives but struggles with vulnerabilities beyond model checking, such as weak cryptography or SQL injection. Static analyzers suffer from high false positives, increasing manual validation efforts for developers. LLMs perform exceptionally well in the CASTLE dataset when identifying vulnerabilities in small code snippets. However, their accuracy declines, and hallucinations increase as the code size grows. These results suggest that LLMs could play a pivotal role in future security solutions, particularly within code completion frameworks, where they can provide real-time guidance to prevent vulnerabilities. The dataset is accessible at https://github.com/CASTLE-Benchmark.
Smart contracts, closely intertwined with cryptocurrency transactions, have sparked widespread concerns about considerable financial losses of security issues. To counteract this, a variety of tools have been developed to identify vulnerability in smart contract. However, they fail to overcome two challenges at the same time when faced with smart contract bytecode: (i) strong interference caused by enormous non-relevant instructions; (ii) missing semantics of bytecode due to incomplete data and control flow dependencies. In this paper, we propose a multi-teacher based bytecode vulnerability detection method, namely \textbf{M}ulti-\textbf{T}eacher \textbf{V}ulnerability \textbf{Hunter} (\textbf{MTVHunter}), which delivers effective denoising and missing semantic to bytecode under multi-teacher guidance. Specifically, we first propose an instruction denoising teacher to eliminate noise interference by abstract vulnerability pattern and further reflect in contract embeddings. Secondly, we design a novel semantic complementary teacher with neuron distillation, which effectively extracts necessary semantic from source code to replenish the bytecode. Particularly, the proposed neuron distillation accelerate this semantic filling by turning the knowledge transition into a regression task. We conduct experiments on 229,178 real-world smart contracts that concerns four types of common vulnerabilities. Extensive experiments show MTVHunter achieves significantly performance gains over state-of-the-art approaches.
Pre-training a language model and then fine-tuning it has shown to be an efficient and effective technique for a wide range of code intelligence tasks, such as code generation, code summarization, and vulnerability detection. However, pretraining language models on a large-scale code corpus is computationally expensive. Fortunately, many off-the-shelf Pre-trained Code Models (PCMs), such as CodeBERT, CodeT5, CodeGen, and Code Llama, have been released publicly. These models acquire general code understanding and generation capability during pretraining, which enhances their performance on downstream code intelligence tasks. With an increasing number of these public pre-trained models, selecting the most suitable one to reuse for a specific task is essential. In this paper, we systematically investigate the reusability of PCMs. We first explore three intuitive model selection methods that select by size, training data, or brute-force fine-tuning. Experimental results show that these straightforward techniques either perform poorly or suffer high costs. Motivated by these findings, we explore learning-based model selection strategies that utilize pre-trained models without altering their parameters. Specifically, we train proxy models to gauge the performance of pre-trained models, and measure the distribution deviation between a model's latent features and the task's labels, using their closeness as an indicator of model transferability. We conduct experiments on 100 widely-used opensource PCMs for code intelligence tasks, with sizes ranging from 42.5 million to 3 billion parameters. The results demonstrate that learning-based selection methods reduce selection time to 100 seconds, compared to 2,700 hours with brute-force fine-tuning, with less than 6% performance degradation across related tasks.
Code vulnerability detection (CVD) is essential for addressing and preventing system security issues, playing a crucial role in ensuring software security. Previous learning-based vulnerability detection methods rely on either fine-tuning medium-size sequence models or training smaller neural networks from scratch. Recent advancements in large pre-trained language models (LLMs) have showcased remarkable capabilities in various code intelligence tasks including code understanding and generation. However, the effectiveness of LLMs in detecting code vulnerabilities is largely under-explored. This work aims to investigate the gap by fine-tuning LLMs for the CVD task, involving four widely-used open-source LLMs. We also implement other five previous graph-based or medium-size sequence models for comparison. Experiments are conducted on five commonly-used CVD datasets, including both the part of short samples and long samples. In addition, we conduct quantitative experiments to investigate the class imbalance issue and the model's performance on samples of different lengths, which are rarely studied in previous works. To better facilitate communities, we open-source all codes and resources of this study in https://github.com/SakiRinn/LLM4CVD and https://huggingface.co/datasets/xuefen/VulResource.
Cryptographic protocols play a fundamental role in securing modern digital infrastructure, but they are often deployed without prior formal verification. This could lead to the adoption of distributed systems vulnerable to attack vectors. Formal verification methods, on the other hand, require complex and time-consuming techniques that lack automatization. In this paper, we introduce a benchmark to assess the ability of Large Language Models (LLMs) to autonomously identify vulnerabilities in new cryptographic protocols through interaction with Tamarin: a theorem prover for protocol verification. We created a manually validated dataset of novel, flawed, communication protocols and designed a method to automatically verify the vulnerabilities found by the AI agents. Our results about the performances of the current frontier models on the benchmark provides insights about the possibility of cybersecurity applications by integrating LLMs with symbolic reasoning systems.
Binary analysis is a core component of many critical security tasks, including reverse engineering, malware analysis, and vulnerability detection. Manual analysis is often time-consuming, but identifying commonly-used or previously-seen functions can reduce the time it takes to understand a new file. However, given the complexity of assembly, and the NP-hard nature of determining function equivalence, this task is extremely difficult. Common approaches often use sophisticated disassembly and decompilation tools, graph analysis, and other expensive pre-processing steps to perform function similarity searches over some corpus. In this work, we identify a number of discrepancies between the current research environment and the underlying application need. To remedy this, we build a new benchmark, REFuSE-Bench, for binary function similarity detection consisting of high-quality datasets and tests that better reflect real-world use cases. In doing so, we address issues like data duplication and accurate labeling, experiment with real malware, and perform the first serious evaluation of ML binary function similarity models on Windows data. Our benchmark reveals that a new, simple basline, one which looks at only the raw bytes of a function, and requires no disassembly or other pre-processing, is able to achieve state-of-the-art performance in multiple settings. Our findings challenge conventional assumptions that complex models with highly-engineered features are being used to their full potential, and demonstrate that simpler approaches can provide significant value.
Recently, deep learning (DL) approaches to vulnerability detection have gained significant traction. These methods demonstrate promising results, often surpassing traditional static code analysis tools in effectiveness. In this study, we explore a novel approach to vulnerability detection utilizing the tools from topological data analysis (TDA) on the attention matrices of the BERT model. Our findings reveal that traditional machine learning (ML) techniques, when trained on the topological features extracted from these attention matrices, can perform competitively with pre-trained language models (LLMs) such as CodeBERTa. This suggests that TDA tools, including persistent homology, are capable of effectively capturing semantic information critical for identifying vulnerabilities.
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.
Recently, deep learning has demonstrated promising results in enhancing the accuracy of vulnerability detection and identifying vulnerabilities in software. However, these techniques are still vulnerable to attacks. Adversarial examples can exploit vulnerabilities within deep neural networks, posing a significant threat to system security. This study showcases the susceptibility of deep learning models to adversarial attacks, which can achieve 100% attack success rate (refer to Table 5). The proposed method, EaTVul, encompasses six stages: identification of important samples using support vector machines, identification of important features using the attention mechanism, generation of adversarial data based on these features using ChatGPT, preparation of an adversarial attack pool, selection of seed data using a fuzzy genetic algorithm, and the execution of an evasion attack. Extensive experiments demonstrate the effectiveness of EaTVul, achieving an attack success rate of more than 83% when the snippet size is greater than 2. Furthermore, in most cases with a snippet size of 4, EaTVul achieves a 100% attack success rate. The findings of this research emphasize the necessity of robust defenses against adversarial attacks in software vulnerability detection.