Minos: A Multi-Agent Collaborative Framework for Provenance-Based Backward Tracking
2026-07-01 • Cryptography and Security
Cryptography and Security
AI summaryⓘ
The authors developed Minos, a system that uses large language models (LLMs) to help trace the steps of cyber attacks after they happen. Unlike previous methods that rely on simple features and exhaustive searching, Minos uses different specialized agents and smart guessing to understand and follow attack traces more efficiently. Tests showed Minos is better at finding attack details while keeping the results simpler and easier to understand, which helps in investigating attacks. The system also explains its reasoning, making it useful for further analysis and improvements.
Advanced Persistent Threats (APTs)forensic analysisprovenance-based trackinglarge language models (LLMs)backward trackingfinite state machine (FSM)graph explorationhierarchical contextretrieval-augmented reasoningcybersecurity
Authors
Jiahui Wang, Zhenyuan Li, Zhengkai Wang, Xiangmin Shen, Fan Zhang
Abstract
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.