ExploreRepair：多智能体协作框架提升漏洞修复成功率
ExploreRepair: Evidence-Specialized Agents for Vulnerability Repair under Information Constraints
一篇写代码安全的论文：三个探索智能体分工分析漏洞再汇总修复，缺 PoV 测试也能修，成绩比 EvoRepair 高出最多 25.4 个百分点。
ExploreRepair 是一个自动化漏洞修复框架，专门应对漏洞分类不明确、缺少 PoV 测试时的修复难题。框架先由三个探索智能体分别从不同维度分析漏洞，再由 Lesson Aggregator 和 Patch Aggregator 汇总观察与实现，最后由一个 Aggregator 在无基准反馈的情况下敲定补丁。在 PatchEval 基准上，配合 GPT-5.6-Luna 达到 39.6% 的修复成功率，配合 DeepSeek-V4.1-Flash 达到 71.6%，分别领先现有最优方法 EvoRepair 16.0 和 25.4 个百分点。在 VulnLoc+ 上成功率为 51.4% 和 62.9%，比不使用 PoV 测试的 LoopRepair 变体分别高出 20.0 和 17.1 个百分点。
ExploreRepair: Evidence-Specialized Agents for Vulnerability Repair under Information Constraints
Automated vulnerability repair must often operate with incomplete vulnerability information: developers may lack an accurate vulnerability classification or a provided proof-of-vulnerability (PoV) test, yet still need to produce a repair. Without detailed information, a key challenge is to explore multiple plausible interpretations of the vulnerability before committing to a repair. To address this challenge, we propose ExploreRepair, a vulnerability repair framework that reconciles the vulnerability observations and repair implementations from a team of exploration agents. We first collect these observations and repair implementations from three agents, each tasked with exploring a specific dimension of the vulnerability. Then, a Lesson Aggregator and a Patch Aggregator independently synthesize repairs from these observations and concrete implementations before a final Aggregator reconciles their proposals without benchmark feedback. On PatchEval, ExploreRepair achieves repair success rates of 39.6% and 71.6% with GPT-5.6-Luna and DeepSeek-V4.1-Flash, respectively, exceeding EvoRepair, a state-of-the-art vulnerability repair approach, by 16.0 and 25.4 percentage points. On VulnLoc+, it achieves success rates of 51.4% and 62.9%, respectively, exceeding LoopRepair's variant without PoV tests by 20.0 and 17.1 percentage points.