Re: broken RSA keys

[Simon McVittie <smcv () debian org>]

On Wed, 04 May 2016 at 21:18:26 -0400, Stanislav Datskovskiy wrote:
> 3) The 'mirrored' keys found thus far in no case have valid
> self-signatures. (A number of the remaining phuctored keys - do.) Thus
> it does not follow from the facts at hand that these particular keys
> were generated /by the people and organizations whose names appear in
> the user string/ !

Even if these keys had valid self-signatures, that wouldn't imply anything
about whether they were generated by the people or organizations named
in the uids; anyone could generate a PGP key right now that claimed
to be yours or mine or anyone else's. That's why we have the "web
of trust", along with competing identity-claiming mechanisms like
keybase.io - the generated key wouldn't have (reputable) third-party
signatures, unless its generator was able to do some social engineering
to obtain them.

I would have expected that an attacker trying for things like evil32 would
want to have a valid self-signature, and the self-signature isn't magic
(it's just an ordinary signature made with the private certification
key as far as I know), so I'm a bit confused by why these "mirrored"
keys would lack them?

    S