Email address challenge script for KMail

[Thomas Zander]

On Wed, Sep 10, 2003 at 09:05:18PM +0200, Ingo Kl?cker wrote:
> On Wednesday 10 September 2003 18:26, Andras Mantia wrote:
> > On Wednesday 10 September 2003 16:04, Antonio Larrosa Jim?nez wrote:
> > > El Wednesday 10 September 2003 11:12, Matthias Kalle Dalheimer 
> escribi?:
> > > > On Wednesday 10 September 2003 10.26, Andras Mantia wrote:
> > > > The key itself will, but not that additional user ID, AFAIK. So
> > > > people will consider your freemail.hu address still valid, but
> > > > not your kde.org address.
> > >
> > > I don't have any experience at all with this (just started using
> > > pgp for the signing party), but ... if he sends a signed mail from
> > > his valid uid (the freemail.hu one) saying that his kde.org address
> > > is also from him, wouldn't that be enough to make us recognize that
> > > address as valid for him? so we could also sign his other uid,
> > > isn't it?
> >
> > This was also my way of thinking. As I would like to use the @kde.org
> > address for KDE related issues, hereby I say that the amantia at kde.org
> > address belongs to me.
> 
> If your key would have been compromised then anyone could have added a 
> new user id and then written and signed the above message.

I dismiss your problems to this approuch based on the fact that this is
a very silly idea. Allow me to explain:
When the key is compromised (including the password since the email was
signed) absolutely no assumptions based on any identity can be made anymore.
i.e. total (electronic) identity theft.
I feel that if you take the standpoint that a key _can_ be compromised; you
should not even be using and trusting PGP to begin with.

Andras is right, if a signed email coming from a source you trust
(verifyable via the signature) everything in that signed part of the email
can be trusted to come from that source.
The email said the amantia at kde is his; and after getting an updated public
key (either from a server or via the attachment, which is irrelevant) all
you have to do is verify that:
1) the (new) email address in the public key is the same as the one
mentioned in the email.
2) the userID (in the public key) is attached to a trusted userID, which
is also directly attached to the _signature_ of the originating email.


> > This message is coming from the e-mail address 
> > which was also on the printed paper at the keysigning party and it's
> > signed using that key. My public key is also attached to this mail.
> 
> Anyone can enter any e-mail address in the From: header and anyone who 
> has compromised your key could attach "your" public key (which isn't 
> even signed because you didn't use the PGP/MIME plugin).

What this proves is that the content (the signed part) caries the sender
information and not the 'From' header.  This is indeed something that you
should be aware of, they don't have to be equal.  But is certainly does
not make it impossible to define trust to the new email.

-- 
Thomas Zander
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://mail.kde.org/pipermail/novehrady/attachments/20030911/cb474b85/attachment.bin