arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2607.00440cs.CR
收藏

Minos:一种基于溯源的反向追踪多智能体协作框架

Minos: A Multi-Agent Collaborative Framework for Provenance-Based Backward Tracking

Zhejiang University(浙江大学);Ningbo Key Laboratory of Quantum Software and Security(宁波量子软件与安全重点实验室);Hofstra University(霍夫斯特拉大学)

Jiahui Wang, Zhenyuan Li, Zhengkai Wang, Xiangmin Shen, Fan Zhang

AI总结:

提出多智能体框架Minos,利用LLM驱动推理进行基于溯源的反向追踪,通过分层上下文管理、检索增强推理和有限状态机协调四个智能体,有效缓解依赖爆炸,在5个数据集上平均召回率0.92、精确率0.64,攻击子图压缩49%。

摘要:

复杂的网络攻击,特别是高级持续性威胁(APT),需要有效的入侵后取证分析。基于溯源的反向追踪通过从安全告警中追踪因果关系来重建攻击场景,但现有方法依赖低层统计特征和刚性遍历策略,限制了捕获高层攻击意图的能力,并遭受依赖爆炸问题。我们提出Minos,一个多智能体框架,将反向追踪形式化为LLM驱动的推理过程。Minos采用两层架构:在事件级分析中,它结合了分层上下文管理、带引用验证的检索增强推理以及对抗性审议,以提高推理质量;在图探索中,它在有限状态机(FSM)下协调四个专用智能体,用假设引导推理和计数优先查询协议替代穷举遍历,以高效剪枝搜索空间。在五个公开数据集的14个攻击场景上的实验表明,Minos实现了平均召回率0.92和精确率0.64,显著优于最先进的基线,同时生成的攻击子图紧凑度提高了49%。此外,Minos在整个追踪过程中生成可解释的推理,便于取证审计和系统改进。这些结果证明了LLM驱动的推理在自动化基于溯源的反向追踪中的有效性。

英文摘要:

Sophisticated cyber attacks, particularly Advanced Persistent Threats (APTs), require effective post-intrusion forensic analysis. Provenance-based backward tracking reconstructs attack scenarios by tracing causality from security alerts, but existing methods rely on low-level statistical features and rigid traversal strategies, limiting their ability to capture high-level adversarial intent and suffering from dependency explosion. We present Minos, a multi-agent framework that formulates backward tracking as an LLM-driven reasoning process. Minos adopts a two-tiered architecture: for event-level analysis, it combines hierarchical context management, retrieval-augmented reasoning with citation verification, and adversarial deliberation to improve reasoning quality; for graph exploration, it coordinates four specialized agents under a finite state machine (FSM), replacing exhaustive traversal with hypothesis-guided reasoning and count-first query protocols to efficiently prune the search space. Experiments on 14 attack scenarios across five public datasets show that Minos achieves an average recall of 0.92 and precision of 0.64, significantly outperforming state-of-the-art baselines while producing attack subgraphs that are 49% more compact. Moreover, Minos generates interpretable reasoning throughout the tracking process, facilitating forensic auditing and system refinement. These results demonstrate the effectiveness of LLM-driven reasoning for automated provenance-based backward tracking.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑