Hot storiesContinuously updated
多智能体漏洞利用框架DoGNAVY-Exploit发布
1 reports1 reporting sourcesUpdated 9 hours ago
Start here
AI overview
2026年11月4日,arXiv发布关于DoGNAVY-Exploit框架的研究成果。该成果提出了一个面向漏洞利用的多智能体框架,主要通过阶段分解与分层并行探索机制,来提升开源权重模型的任务表现。测试数据显示,当搭配开源权重模型GLM-5.2时,该框架在网络安全基准Cybench上达到了97.5%的Pass@1表现;在CVE-Bench测试中,其零日漏洞Pass@1达到97.5%,一日漏洞Pass@1达到100%。
AI-generated from reports · Updated 9 hours ago
Latest developmentsNov 4 13:00
DoGNAVY-Exploit框架发布,搭配GLM-5.2在CVE-Bench取得极高Pass@1。Report timeline
Follow the reports to see different sides of the story.
Nov 4
- arXiv 人工智能DoGNAVY-Exploit:利用并行开源权重智能体实现阶段感知漏洞利用
DoGNAVY-Exploit 提出面向漏洞利用的多智能体框架,通过阶段分解与分层并行探索提升开源权重模型表现。搭配 GLM-5.2 时,该框架在 Cybench 达到 97.5% Pass@1,在 CVE-Bench 取得 97.5% 零日与 100% 一日 Pass@1。
Story heat trend
Current heat 8·Comparable peak 10(Nov 4 14:00)·Comparable change, last 24 hours–
Trends compare only the same parties observed continuously, so the scope can be narrower than current heat counts. Move the pointer or tap the chart to see hourly heat; use the left and right arrow keys on a keyboard.