Current learning-based Automated Vulnerability Repair (AVR) approaches, while promising, often fail to generalize effectively in real-world scenarios. Our diagnostic analysis reveals three fundamental weaknesses in state-of-the-art AVR approaches: (1) limited cross-repository generalization, with performance drops on unseen codebases; (2) inability to capture long-range dependencies, causing a performance degradation on complex, multi-hunk repairs; and (3) over-reliance on superficial lexical patterns, leading to significant performance drops on vulnerabilities with minor syntactic variations like variable renaming.
To address these limitations, we propose SeCuRepair, a semantics-aligned, curriculum-driven, and reasoning-enhanced framework for vulnerability repair. At its core, SeCuRepair adopts a reason-then-edit paradigm, requiring the model to articulate why and how a vulnerability should be fixed before generating the patch. This explicit reasoning enforces a genuine understanding of repair logic rather than superficial memorization of lexical patterns. SeCuRepair also moves beyond traditional supervised fine-tuning and employs semantics-aware reinforcement learning, rewarding patches for their syntactic and semantic alignment with the oracle patch rather than mere token overlap. Complementing this, a difficulty-aware curriculum progressively trains the model, starting with simple fixes and advancing to complex, multi-hunk coordinated edits.
We evaluate SeCuRepair on strict, repository-level splits of BigVul and newly crafted PrimeVul_AVR datasets. SeCuRepair significantly outperforms all baselines, surpassing the best-performing baselines by 34.52% on BigVul and 31.52% on PrimeVul\textsubscript{AVR} in terms of CodeBLEU, respectively. Comprehensive ablation studies further confirm that each component of our framework contributes to its final performance.
The Java libraries JCA and JSSE offer cryptographic APIs to facilitate secure coding. When developers misuse some of the APIs, their code becomes vulnerable to cyber-attacks. To eliminate such vulnerabilities, people built tools to detect security-API misuses via pattern matching. However, most tools do not (1) fix misuses or (2) allow users to extend tools' pattern sets. To overcome both limitations, we created Seader-an example-based approach to detect and repair security-API misuses. Given an exemplar <insecure, secure>code pair, Seader compares the snippets to infer any API-misuse template…
Java platform provides various APIs to facilitate secure coding. However, correctly using security APIs is usually challenging for developers who lack cybersecurity training. Prior work shows that many developers misuse security APIs; such misuses can introduce vulnerabilities into software, void security protections, and present security exploits to hackers. To eliminate such API-related vulnerabilities, this paper presents SEADER -- our new approach that detects and repairs security API misuses. Given an exemplar, insecure code snippet, and its secure counterpart, SEADER compares the snippet…
Vul-R2: A Reasoning LLM for Automated Vulnerability Repair
Tencent Inc.（腾讯公司）；Department of Computer Science, City University of Hong Kong（香港城市大学计算机科学系）；Department of Computer Science and Engineering, The Chinese University of Hong Kong（香港中文大学计算机科学与工程系）
K. N. Toosi University of Technology（肯恩·托基大学）；Ferdowsi University of Mashhad（马什哈德费尔多西大学）；Simon Fraser University（西蒙菲莎大学）；Amirkabir University of Technology（阿米尔卡比尔理工大学）
Nanjing University of Posts and Telecommunications（南京邮电大学）；National Interdisciplinary Research Center of Engineering Physics（工程物理国家 interdisciplinary 研究中心）；Alibaba Qwen（阿里巴巴通义千问）；Beijing Jiaotong University（北京交通大学）；Zhejiang University（浙江大学）；Yangzhou University（扬州大学）；Nanjing University（南京大学）；Singapore Management University（新加坡管理大学）
KTH Royal Institute of Technology（瑞典皇家理工学院）；Colorado State University（科罗拉多州立大学）
In this paper, we address the problem of automatic repair of software vulnerabilities with deep learning. The major problem with data-driven vulnerability repair is that the few existing datasets of known confirmed vulnerabilities consist of only a few thousand examples. However, training a deep learning model often requires hundreds of thousands of examples. In this work, we leverage the intuition that the bug fixing task and the vulnerability fixing task are related and that the knowledge learned from bug fixes can be transferred to fixing vulnerabilities. In the machine learning community, …
Users around the world rely on software-intensive systems in their day-to-day activities. These systems regularly contain bugs and security vulnerabilities. To facilitate bug fixing, data-driven models of automatic program repair use pairs of buggy and fixed code to learn transformations that fix errors in code. However, automatic repair of security vulnerabilities remains under-explored. In this work, we propose ways to improve code representations for vulnerability repair from three perspectives: input data type, data-driven models, and downstream tasks. The expected results of this work are…
