"User agent" but do users run LLMs behind their agents on their hardware? No, it's run in an AI datacenter. They can easily make a deal to route that without involving user's machine, or they can route this via user's machine but attach a signature. Of course both parties must agree on that, but technically there's nothing stopping them

Soon they might require a vendor specific api key to access it and that requires support from the big players (anthropic,openai). They’re laying down the “framework” now.