CUAWright: A Minimal Unified Interface for Digital Agents
Yadong Lu and Ahmed Hassan Awadallah (Microsoft Research) with Yu Su, Huan Sun and colleagues at Ohio State present CUAWright, a harness in which computer-use agents act on web, desktop and CAD tasks only through terminal commands.
Ask this paper
Design. Every action is a terminal command. Browser tasks are driven by agent-written code against the debugging protocol, desktop tasks by CLI tools or file edits, and keyboard and mouse actions are used only when a surface cannot be controlled programmatically. Screenshots are taken on demand.
File system as context. The agent stores notes and writes, tests and reuses its own tools as files; ablations show the file system is required for the gains.
OSWorld 2.0. CUAWright improves partial reward by 33.2% relative to the published GPT-5.5 baseline while cutting estimated cost by 37.5%, about $6.0 less per task.
Web and CAD. It beats a GUI-native harness by 4.7 points on Online-Mind2Web and 44.0 points on the long-horizon Odysseys benchmark, where GPT-5.4 reaches 77.5% against 33.5%, and gives 8.1% to 41.6% relative gains on CAD benchmarks.
Abstract
The prevailing approach to computer-use agents couples a model with a domain-specific harness: a browser or desktop environment equipped with human engineered tools that are fixed before task execution. As models' coding capabilities improve, the GUI native and static harness prevents them from direct programmatic operation on system state, as well as flexible construction of tools. To this end, we introduce CUAWright, a minimal terminal harness of roughly 3K lines of code that uses bash commands as its sole action interface, and a file system as its evolvable space for dynamically creating tools and managing the context. We conduct comprehensive experiments across a wide range of digital tasks, and demonstrate that by giving the agent a minimal, programmable interface, it achieves substantially stronger results compared to their GUI or hybrid CLI interface across a wide range of tasks. On OSWorld 2.0, CUAWright delivers a 33.2% relative improvement in partial reward while reducing estimated cost by 37.5% compared with the published GPT-5.5 baseline. On Online-Mind2Web and the long horizon Odysseys benchmark, CUAWright substantially outperforms GUI native harness by 4.7% and 44.0% in success rate, respectively. Furthermore, we found the gains extend to CAD applications that require accurate visual understanding and CLI interaction: on CADGenBench and BenchCAD, our unified harness yields 8.1%-41.6% relative improvements over other CLI based harnesses with GPT-5.5. Together, these results suggest digital environments are far more programmable than their GUI interfaces imply, and a minimal terminal-focused harness is the key for better performance and efficiency.