arXiv 2602.024752026-02cs.AI

AgentRx: 从执行轨迹诊断AI代理故障

AgentRx: Diagnosing AI Agent Failures from Execution Trajectories

Microsoft Research（微软研究院）；Microsoft（微软）

Shraddha Barke, Arnav Goyal, Alind Khare, Avaljot Singh, Suman Nath, Chetan Bansal

备注：: Accepted to EMNLP Findings 2026

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

AgentRx通过自动化诊断框架，从执行轨迹中定位AI代理的关键故障步骤，并提升跨领域的故障归因能力。

摘要：

AI代理常常以难以定位的方式失败，因为执行过程是概率性的、长周期的、多代理的，并且受到噪声工具输出的中介。我们通过手动注释失败的代理运行，填补这一空白，并发布了一个包含115条失败轨迹的新基准，涵盖结构化API流程、事件管理以及开放式的网络/文件任务。每条轨迹都被注释为一个关键故障步骤和一个从扎根理论推导出的跨领域故障分类。为了减轻失败归因的人力成本，我们提出了AGENTRX，一个自动化的、领域无关的诊断框架，能够定位失败代理轨迹中的关键故障步骤。它合成约束，逐步评估这些约束，并生成可审计的约束违反验证日志，附带相关证据；一个基于LLM的裁判使用此日志来定位关键步骤和类别。我们的框架在三个领域中比现有基线在步骤定位和故障归因上有所提升。

英文摘要：

AI agents often fail in ways that are difficult to localize because executions are probabilistic, long-horizon, multi-agent, and mediated by noisy tool outputs. We address this gap by manually annotating failed agent runs and release a novel benchmark of 170 trajectories across 11 diverse task settings, including structured API workflows, incident management, and open-ended web/file tasks. Each trajectory is annotated with a critical failure step and a category from a grounded-theory derived, cross-domain failure taxonomy. To mitigate the human cost of failure attribution, we present AgentRx, an $\textit{automated diagnostic framework}$ that pinpoints the critical failure step in a failed agent trajectory. It synthesizes constraints, evaluates them step-by-step, and produces an auditable validation log of constraint violations with associated evidence; an LLM-based judge uses this log to localize the critical step and category. AgentRx improves step localization by 75% on average over prior work, while providing failure category attribution.

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。

AI总结：

摘要：

英文摘要：

相关论文