Jargonautica
Agents and Tools

Browser Automation

Driving a web page programmatically to fill forms, click through flows and read results. It is how an agent reaches systems with no proper interface for machines. It is fragile because pages change, and risky because everything on the page is content the model will read and may follow.

Technical view

Implemented over a browser control protocol with actions expressed against accessibility trees, document structure or screen coordinates, the first being more robust than the last. Failure modes include layout changes, dynamic content and timing, anti-automation measures, and authentication. The security position is that page content is untrusted input capable of carrying instructions, so session scope, credential handling and confirmation for consequential actions require explicit design.


Part of AI Terminology Index, section: Agents and Tools.

How it connects

Browser Automation links to 5 other terms. The faint lines are links between those terms. See it in the full graph.

AgentAgentLeast PrivilegeLeast PrivilegeTool Use and Function CallingTool Use and Functi…Computer UseComputer UseIndirect Prompt InjectionIndirect Prompt Inj…Browser Automation