A new framework, MASTIF, evaluates Agentic Web technologies, including LLMs, protocols, and agent frameworks. The baseline assessment uses 300 benchmark tasks and shows that task-specific tech stacks can improve completion rates and efficiency.