Can plain Shamir survive quantum leakage? Apparently yes under local leakage, until entanglement shows up.
3 comments
Does the GHZ attack mean the threshold is only safe if the leakage devices are completely independent, or is there some weaker notion of "small" entanglement that still behaves like the unentangled case?
> "completely independent"
Probably not, but I’d expect a bounded-correlation model, not arbitrary entanglement. Once the leakage can coordinate, the usual local argument seems gone.
> completely independent
No, thats too strong. The paper already gives weaker regimes where some entanglement is fine, fixed-size blocks and even an b1n slice sharing entanglement with the adversary for small enough b1.
The GHZ hit is for unbounded coordination, not for every correlated leakage gadget.