Wonyoung Kim, Seunggi Min, Minjae Gwon, Haein Lee, Dowoo Baik, Hyeon Heo, Minjae Lee, Min Woo Baek, Yonghwi Jin, Younggi Park, Yunjae Choi, Taesoo Kim, Sangdon Park, Insu Yun

December 2026

Abstract

Continuous fuzzing platforms such as OSS-Fuzz uncover large numbers of vulnerabilities, yet the subsequent repair process remains largely manual. Unfortunately, existing Automated Vulnerability Repair (AVR) techniques—including recent LLM-based systems—are not directly applicable to continuous fuzzing. This is because these systems are designed and evaluated on a static, single-run benchmark setting, making them ill-suited for the diverse, noisy, and failure-prone environments in continuous fuzzing.

To address these issues, we introduce PatchIsland, a system for Continuous Vulnerability Repair (CVR) that tightly integrates with continuous fuzzing pipelines. Our key idea for PatchIsland is to employ an ensemble of diverse LLM agents. By leveraging multiple LLM agents, PatchIsland can cover a wider range of settings (e.g., different projects, bug types, and programming languages) and also improve operational robustness. In addition, PatchIsland utilizes a two-phase deduplication to mitigate duplicate crashes and patches, which can be problematic in continuous fuzzing.

We evaluate PatchIsland in both internal and external settings. In our internal evaluation, PatchIsland successfully repaired 84 of 92 vulnerabilities, outperforming state-of-the-art CVR systems. We also applied PatchIsland in the official AIxCC final. During this competition, PatchIsland patched 31 of 43 vulnerabilities while operating fully autonomously for more than a week. This result represents the largest number of patches and the second-highest repair rate (72.1%) among participating systems. More surprisingly, PatchIsland could produce a zero-day patch that exactly matched with the official one. This shows that PatchIsland can be a practical solution for real-world CVR.