Both Google and Apple have MCP-like interface frameworks for agents that developers can enable so their apps are not dependent on touch or vioce or keyboard or

Both Google and Apple have MCP-like interface frameworks for agents that developers can enable so their apps are not dependent on touch or vioce or keyboard or pointing device or D-pad interactions. This is the correct way to do it because it puts the interaction responsibility into the agent, not the app. That enables the agent to work across multiple apps to achieve tasks by interacting with the user. That's where both sides of the interaction should live. That's where the context of the interaction exclusively lives because no individual app holds the full context. That also makes it the place where solving adaptive interface challenges is best implemented.