We propose VerifyMAS, a hypothesis verification framework for agent failure attribution. Instead of directly predicting faulty agents and error types, VerifyMAS formulates and verifies failure hypotheses against full trajectories. This verification-based approach decomposes attribution into trajectory-level error validation and fine-grained agent localization, providing an error-first attribution approach that captures global failure patterns while substantially reducing the search space. We further introduce a hypothesis-based data construction strategy grounded in a structured error taxonomy and fine-tune a specialized LLM verifier model for trajectory-level failure verification and agent attribution. Experiments on Aegis-Bench and Who&When show that VerifyMAS consistently improves diverse backbone models, including open-source Qwen and API-based GPT models, outperforming prior methods without sacrificing inference efficiency for long multi-agent trajectories.
The code will be uploaded soon.
If you find this work useful, please cite our paper:
@article{qiao2026verifymas,
title={VerifyMAS: Hypothesis Verification for Failure Attribution in LLM Multi-Agent Systems},
author={Qiao, Hezhe and Tong, Hanghang and Lim, Ee-Peng and Liu, Bing and Pang, Guansong},
journal={arXiv preprint arXiv:2605.17467},
year={2026}
}