MAS-FIRE: Fault Injection and Reliability Evaluation for LLM-Based Multi-Agent Systems

2026-02-23

AI Summary

This work addresses the vulnerability of existing large language model (LLM)-based multi-agent systems to semantic failures stemming from their reliance on natural language communication, as well as the absence of effective reliability evaluation methodologies. To this end, the authors propose MAS-FIRE, a novel framework that introduces the first comprehensive taxonomy of 15 failure types specific to LLM multi-agent systems and a four-layer fault-tolerant behavior model. The framework enables fine-grained failure diagnosis through three non-intrusive fault injection mechanisms: prompt modification, response rewriting, and message routing manipulation. Experimental results demonstrate that iterative closed-loop architectures can mitigate over 40% of failures that would otherwise cause linear pipelines to collapse, further revealing that system topology exerts an influence on robustness comparable to that of the underlying LLM capabilities.

Technology Category

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

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

As LLM-based Multi-Agent Systems (MAS) are increasingly deployed for complex tasks, ensuring their reliability has become a pressing challenge. Since MAS coordinate through unstructured natural language rather than rigid protocols, they are prone to semantic failures (e.g., hallucinations, misinterpreted instructions, and reasoning drift) that propagate silently without raising runtime exceptions. Prevailing evaluation approaches, which measure only end-to-end task success, offer limited insight into how these failures arise or how effectively agents recover from them. To bridge this gap, we propose MAS-FIRE, a systematic framework for fault injection and reliability evaluation of MAS. We define a taxonomy of 15 fault types covering intra-agent cognitive errors and inter-agent coordination failures, and inject them via three non-invasive mechanisms: prompt modification, response rewriting, and message routing manipulation. Applying MAS-FIRE to three representative MAS architectures, we uncover a rich set of fault-tolerant behaviors that we organize into four tiers: mechanism, rule, prompt, and reasoning. This tiered view enables fine-grained diagnosis of where and why systems succeed or fail. Our findings reveal that stronger foundation models do not uniformly improve robustness. We further show that architectural topology plays an equally decisive role, with iterative, closed-loop designs neutralizing over 40% of faults that cause catastrophic collapse in linear workflows. MAS-FIRE provides the process-level observability and actionable guidance needed to systematically improve multi-agent systems.

Problem

• LLM-based Multi-Agent Systems
• reliability evaluation
• semantic failures
• fault injection
• coordination failures

Innovation

• fault injection
• multi-agent systems
• LLM reliability
• semantic failures
• systematic evaluation