There is no value in this "attestation" for me as a user. I want to be able to do whatever I want with the browser (for example, remove ads or block access to canvas and webgl) and I want sites to be unable to know this. And probably this attestation will provide additional fingerprinting signals which is what I don't want.
That said, the concept is seemingly aimed at blocking ad blockers and preventing browsers like Brave from impersonating Chrome so it can block ads without the need for extensions and such.
The only user-positive use case I can think of for this is for self-hosted software. Maybe it can be used to detect MitM attacks or malware messing with the browser? In practice this will just mean "no Firefox, no Linux, no adblockers".
In theory one could imagine a scenario like a bank website refusing to be accessed unless the entire OS & browser stack pass attestation - as that would rule out things like keyloggers, malicious browser extensions, and session hijacking.
In practice it'll just be used to lock down content and force unskippable ads on users, of course.
I'm not interested in being hobbled for either of those problems. I remember when banks used to reject my browser because it wasn't IE in Windows. I remember when I had to look at webpages that were 50% advertising.
Screw that.
The important part is that "malicious" isn't up to you to decide anymore; if you have any "unapproved" software that acts in your interests and not others', this could theoretically be used to lock you out too.
And no curl, no yt-dlp or youtube-dl, no alternative YouTube frontends, no scraping the web to build an alternative search engine.
Even that use case leads to bad outcomes. I already have to jump through hoops to get banking apps to run on my rooted phone. Banking websites refusing to run on anything but Chrome on Windows is a likely scenario here, and that's awful.
I don't want them to have a say in how I run my devices.
I'm also sure it'll end up with things like "your browser is too up-to-date" or crap like that.
But software already exists to do this kind of thing for private networks. I really, strongly believe that this kind of functionality has no place on the open web.
This proposal is user-hostile, and could be very dangerous to the future of the web.