Securer Contexts?

Hey folks,

In the context of the set of side-channel-mitigating isolation primitives
that we've discussed on the past few calls, I've been thinking about the
notion of Secure Contexts. TL;DR: I think limiting the scope of that
mechanism to the transport layer was a great idea in 2015; I think 2020 is
a great time to revisit and expand it to include the threats we care deeply
about today.

I sketched out a proposal for an updated threat model in
https://github.com/mikewest/securer-contexts/, which includes
COOP/COEP/CORP on the one hand, and hand-waves at injection mitigation on
the other.

I'd appreciate feedback, either here on the list, in the GitHub repository,
or on the design review request at
https://github.com/w3ctag/design-reviews/issues/471. :)

Thanks!

-mike