On Thu, Jun 04, 2026 at 06:36:21AM -0600, Nathanael Ritz wrote:
> 
> Of course, symbolic models cannot reason over implementation challenges and
> so the next question may be fall more to pure speculation. I am curious if
> you might anticipate potential implementation errors moving from classical
> DHE or pure PQ-KEM towards a hybrid that risks compromising both methods
> due to complexity?

Any kind of Undefined Behavior in implementation would compromise both.

The TLS hybrid combiner itself is extremely simple, and can be
basically implemented via reading/writing correct places in array.

And as for non-UB flaws in component implementations, not reusing keys
makes exploiting such flaws much harder. 




-Ilari

_______________________________________________
TLS mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to