AI agents that automate web workflows operate within the browser’s web layer, the DOM that Playwright and the Chrome DevTools Protocol (CDP) expose. AgentCore Browser provides a secure, isolated browser environment for this, and it works well for the vast majority of automation: navigating pages, filling forms, clicking elements, extracting content. But the web layer has a hard boundary. Anything that the operating system renders (native dialogs, security prompts, certificate choosers, context menus, even Chrome settings) sits outside the DOM entirely. CDP can’t see it, and Playwright can’t interact with it.
When a web application calls window.print() and a system print dialog appears, Playwright has no DOM to interact with. When a workflow requires a keyboard shortcut or a right-click context menu, CDP has no mechanism to issue those commands at the OS level. When a browser session encounters a macOS privacy dialog, a Windows Security prompt, or a certificate…
