Anthropic has been doing this sort of stuff for a while already. I mean, who remembers when Claude would just consume all your remaining usage if it read anything indicating that Openclaw had been used on your codebase? Because I remember. Two months ago btw https://news.ycombinator.com/item?id=47963204 Then there was the whole debacle of Fable silently downgrading to other models if it detected wrong think, or worse, outright sabotaging your codebase if you were working on language models lol