Mailing List Archive

Web of Trust spam prevention (was: Re: recommendation for key servers)
Andrew Gallagher via Gnupg-devel wrote:
> I think the third-party sig issues raised in this post are best
> tackled with attestations, as discussed already. The trick is to get
> the end-user workflow cleaned up and into as many clients as possible.

As I see the problem, links in the Web of Trust should be symmetric: if
Alice has verified Bob's key, Bob should have also verified Alice's
key. Enforcing this would eliminate spam signatures, but would also
require some way for the system to recognize the intermediate state
where Bob has uploaded his signature for Alice's key but Alice has not
yet uploaded her signature for Bob's key. Perhaps you have 30 days to
upload your signature after certifying a key? Unidirectional signatures
would not be publicly shown until the "other half" of the link is
uploaded and would be dropped after the keyservers have held them for 30
days if the link is not completed?

There would still be possibility to build an entire fake "troll" Web of
Trust with fake keys cross-certifying each other, but I do not have any
ideas to solve that issue yet.


-- Jacob


_______________________________________________
Gnupg-devel mailing list
Gnupg-devel@gnupg.org
http://lists.gnupg.org/mailman/listinfo/gnupg-devel
Re: Web of Trust spam prevention (was: Re: recommendation for key servers) [ In reply to ]
> On 1 Jul 2021, at 01:01, Jacob Bachmeyer <jcb62281@gmail.com> wrote:
>
> As I see the problem, links in the Web of Trust should be symmetric: if Alice has verified Bob's key, Bob should have also verified Alice's key. Enforcing this would eliminate spam signatures, but would also require some way for the system to recognize the intermediate state where Bob has uploaded his signature for Alice's key but Alice has not yet uploaded her signature for Bob's key.

The intermediate state doesn’t have to use keyservers at all. Bob signs Alice’s key and simply emails it back to her. She then has the option of attesting Bob’s signature and publishing it, or not, as she ses fit.

A
_______________________________________________
Gnupg-devel mailing list
Gnupg-devel@gnupg.org
http://lists.gnupg.org/mailman/listinfo/gnupg-devel
Re: Web of Trust spam prevention (was: Re: recommendation for key servers) [ In reply to ]
> As I see the problem, links in the Web of Trust should be symmetric:  if
> Alice has verified Bob's key, Bob should have also verified Alice's
> key.

"Should" virtually always means "I can't personally think of a use case
to the contrary", which is a lot different from there actually being no
use cases to the contrary.

Hint: if Alice issues Bob his certificate, Alice obviously has verified
Bob's certificate. But has Bob verified anything about Alice?

_______________________________________________
Gnupg-devel mailing list
Gnupg-devel@gnupg.org
http://lists.gnupg.org/mailman/listinfo/gnupg-devel