CloudCruise
BATS (Browser Agent Test Suite) — an open-source harness for stress-testing browser agents.
Core trick
GPT-5 generates a single-file test page and agent prompt, then an adversarial agent injects confusing elements to stress-test browser agents.
Description
BATS (Browser Agent Test Suite) — an open-source harness for stress-testing browser agents.
You describe a scenario (“checkout process with credit card info in iframes + random popup”), and GPT-5 generates a realistic single-file test site and an agent prompt. Agents try to fulfill the task on the generated page with the instructions and the platform tracks the time they need, the type and amount of interactions, and if it fulfilled the task successfully. Additionally, the platform has an adversarial agent that tries to make the task more difficult for browser agents by e.g. injecting buttons with confusing labels.
README excerpt
# BATS (Browser Automation Testing Suite) BATS is a specialized platform designed to generate dynamic UIs and websites that serve as testing environments for browser automation agents. The system creates realistic, interactive web pages with various UI elements, popups, and dynamic behaviors to challenge and validate browser automation tools. ## Purpose This project addresses the need for comprehensive testing environments for browser automation agents by: - **Generating Dynamic Test Sites**: Creates websites with complex UI patterns, modals, forms, and interactive elements - **Agent Mode Testing**: Provides an integrated agent that can modify websites in real-time to test automation resilience - **Disruption Simulation**: Intentionally moves buttons, adds decoy elements, and creates challenging scenarios for automation tools - **Real-time Feedback**: Offers visual feedback and activity monitoring for agent interactions ## Tech Stack ### Frontend - **Next.js 15** - React framework with App Router - **TypeScript** - Type-safe development - **Tailwind CSS** - Utility-first CSS framework - **Shadcn/ui** - Modern UI component library - **Lucide React** - Icon library ### AI & Automation - **AI SDK** - Vercel's AI SDK for streaming and tool calling - **OpenAI** - Language model for intelligent agent behavior - **Langfuse** - LLM observability and prompt management - **Zod** - Schema validation for tool inputs ### Development Tools - **ESLint** - Code linting - **PostCSS