CUADebug: Diagnosing and Repairing Computer-Use Agent Failures
Computer-use agents (CUAs) interact with graphical interfaces through screenshots and low-level mouse and keyboard actions, yet the causal error may precede the terminal failure. We present CUADebug, a framework for localizing root causes in CUA trajectories and guiding re-execution. CUADebug includes a five-category, 30-subtype taxonomy; CUAErrorBench, a benchmark of 204 failed OSWorld trajectories with human root-cause annotations; and CUADebugger, a ReAct-style agent for root-cause analysis (RCA). CUADebugger iteratively selects trajectory steps, inspects paired before/after screenshots and action traces, and submits a structured diagnosis containing the causal step, taxonomy label, gr...
| Authors | Weijia Zhang, Kunlun Zhu, Zeyi Liu et al. |
|---|---|
| Primary category | cs.SE |
| Published | 31 July 2026 |
| Last updated at source | 6 September 2026 |
Connected in the knowledge graph
- CyT1ng/cua_debugger_trajDataset · cited by
Models and datasets referencing this paper
- CyT1ng/cua_debugger_trajDataset
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.