Monitoring LLM-based Multi-Agent Systems Against Corruptions via Node Evaluation

2025-10-22

AI Summary

Large language model (LLM)-driven multi-agent systems (MAS) face escalating corruption attacks under dynamic communication, while existing static graph-based defenses lack adaptability to evolving threats. Method: This paper proposes a dynamic defense paradigm leveraging graph representation learning and node behavioral assessment to detect anomalous agent interactions in real time and dynamically reconfigure the communication topology to sever malicious connections. Contribution/Results: The core innovation is the first “monitor–assess–reconstruct” closed-loop mechanism, enabling adaptive identification and response to complex, time-varying attacks. Extensive experiments across diverse dynamic scenarios demonstrate that the approach significantly outperforms state-of-the-art defenses, effectively mitigating heterogeneous corruption attacks and substantially enhancing MAS robustness and trustworthiness—establishing a novel, deployable security paradigm for LLM-MAS.

Technology Category

• Multiagent Systems: Adversarial Agents
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Safety and Robustness

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Large Language Model (LLM)-based Multi-Agent Systems (MAS) have become a popular paradigm of AI applications. However, trustworthiness issues in MAS remain a critical concern. Unlike challenges in single-agent systems, MAS involve more complex communication processes, making them susceptible to corruption attacks. To mitigate this issue, several defense mechanisms have been developed based on the graph representation of MAS, where agents represent nodes and communications form edges. Nevertheless, these methods predominantly focus on static graph defense, attempting to either detect attacks in a fixed graph structure or optimize a static topology with certain defensive capabilities. To address this limitation, we propose a dynamic defense paradigm for MAS graph structures, which continuously monitors communication within the MAS graph, then dynamically adjusts the graph topology, accurately disrupts malicious communications, and effectively defends against evolving and diverse dynamic attacks. Experimental results in increasingly complex and dynamic MAS environments demonstrate that our method significantly outperforms existing MAS defense mechanisms, contributing an effective guardrail for their trustworthy applications. Our code is available at https://github.com/ChengcanWu/Monitoring-LLM-Based-Multi-Agent-Systems.

Problem

• Defend multi-agent systems against dynamic corruption attacks
• Monitor communication and adjust graph topology dynamically
• Disrupt malicious communications in evolving multi-agent environments

Innovation

• Dynamic defense paradigm for MAS graph structures
• Continuously monitors communication within MAS graph
• Dynamically adjusts graph topology against attacks