ST-WebAgentBench: A Benchmark for Evaluating Safety and Trustworthiness in Web Agents
For researchers and developers deploying web agents in enterprise settings, this benchmark provides the first systematic way to measure safety and trustworthiness, addressing a prerequisite for adoption.
ST-WebAgentBench is a benchmark for evaluating safety and trustworthiness of web agents across 222 tasks with six orthogonal dimensions. Testing three state-of-the-art agents showed their average Completion Under Policy (CuP) score was less than two-thirds of their nominal completion rate, revealing critical safety gaps.
Autonomous web agents solve complex browsing tasks, yet existing benchmarks measure only whether an agent finishes a task, ignoring whether it does so safely or in a way enterprises can trust. To integrate these agents into critical workflows, safety and trustworthiness (ST) are prerequisite conditions for adoption. We introduce \textbf{\textsc{ST-WebAgentBench}}, a configurable and easily extensible suite for evaluating web agent ST across realistic enterprise scenarios. Each of its 222 tasks is paired with ST policies, concise rules that encode constraints, and is scored along six orthogonal dimensions (e.g., user consent, robustness). Beyond raw task success, we propose the \textit{Completion Under Policy} (\textit{CuP}) metric, which credits only completions that respect all applicable policies, and the \textit{Risk Ratio}, which quantifies ST breaches across dimensions. Evaluating three open state-of-the-art agents reveals that their average CuP is less than two-thirds of their nominal completion rate, exposing critical safety gaps. By releasing code, evaluation templates, and a policy-authoring interface, \href{https://sites.google.com/view/st-webagentbench/home}{\textsc{ST-WebAgentBench}} provides an actionable first step toward deploying trustworthy web agents at scale.