[论文] SoK: Assessing the Evolution of LLM Capability in Software Vulnerabili...
研究领域: ML 作者: Amine Lbath, Manan Suri, Aurelien Delaitre, Vadim Okun, Massih-Reza Amini, Dinesh Manocha 发布时间: 2026-11-19 arXiv: 2611.26601
论文概要
研究领域: ML 作者: Amine Lbath, Manan Suri, Aurelien Delaitre, Vadim Okun, Massih-Reza Amini, Dinesh Manocha 发布时间: 2026-11-19 arXiv: 2611.26601
中文摘要
基于LLM的智能体在发现、验证和修补真实软件漏洞方面能力日益增强,厂商和研究人员报告称这已经改变了漏洞发现和修复的方式。然而,这些声明散落在论文、厂商报告和披露中,很少可比较。我们引入一个框架,参照驾驶自动化等级对能力进行自主性评级,并按正确性、可用性和证据强度对每个声明进行限定。我们用此框架系统化LLM漏洞研究能力的已发表证据,发现智能体在2026年受限前沿模型之前就已无需人工干预地发现、验证和修补了真实软件漏洞,而开源权重模型随后在发现和验证方面达到了相当的自主性。接着我们在十年公共漏洞记录、公告和利用数据中衡量LLM安全能力演化的后果。2026年披露的漏洞数量翻了一倍以上,被评为严重或高危的比例增加,在广泛使用的代码中存活十年以上的严重bug被报告的频率比以前高出数倍。然而,漏洞生命周期各阶段之间的时间——从报告到修复、从披露到攻击——迄今尚未明显变化。我们按明确量规对文献中提出的缓解措施进行评级,并从证据最充分的研究中为各利益相关方群体提炼指导。
原文摘要
LLM-based agents are becoming increasingly capable at discovering, validating and patching vulnerabilities in real software, and vendors and researchers report that this is already changing how vulnerabilities are found and fixed. These claims, however, are scattered across papers, vendor reports and disclosures and are rarely comparable. We introduce a framework that rates these capabilities on an autonomy scale modeled on the levels of driving automation, and qualifies each claim by its correctness, its availability and the strength of its evidence. We use this framework to systematize the published evidence on LLM capability in vulnerability research. We find that agents discovered, validated, and patched vulnerabilities in real-world software without human intervention before the restr...
*自动采集于 2026-11-21*
#论文 #arXiv #ML #小凯