bugtraq - Re: [FD] Strength and Weakness of Methods to Confirm SSH Host Key

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20140924173613.GB55782@gwolf.org>
Date: Wed, 24 Sep 2014 12:36:13 -0500
From: Gunnar Wolf <gwolf@...lf.org>
To: John Leo <johnleo@...ckssh.com>
Cc: bugtraq@...urityfocus.com, fulldisclosure@...lists.org
Subject: Re: [FD] Strength and Weakness of Methods to Confirm SSH Host Key

John Leo dijo [Mon, Sep 22, 2014 at 03:51:57PM +0800]:
> Monkeysphere
> (advice from maxigas)
> "verify your SSH key through the OpenPGP web of trust"
> Strength: OpenPGP is cool if you REALLY know how to use it.
> Weakness: "vote counting scheme" does not sound too cool.

The "vote counting" goes against knowing whether the signing key is
valid or not. When you are asserting the identity of a site you
control, or a site you trust, this would only become a *second* chain
of trust, if I understand you right. And, of course, the signer
*should* be the same as the site operator!

> "use of an organization's own HTTPS site"
> (advice from Stephanie Daugherty)
> In my personal opinion, this is the best solution.
> Weakness: basically nothing - it's very secure.

A PKI is just the same as the vote counting you mention for OpenPGP,
but with money involved and a single point of failure. That is, having
the key in a HTTPS site will just mean the organization paid the PKI
cartel for a certificate strong enough for a given purpose, not that
it is the legitimate organization.

> "use DNSSEC to validate SSH fingerprints"
> (advice from Micha Borrmann / Jeroen van der Ham / john)
> This is a good solution.
> Weakness: HTTPS is more mature than DNSSEC(in my personal opinion).

The three above are +- the same — different out-of-band channels to
establish a given message (the key fingerprint) is genuine.

> "ssh-keyscan -p 22 domain.com ..."
> (advice from Busindre)
> It's the same as running "ssh" directly.

Right. We wil also do it implicitly every time we connect to said
host, unless our ssh client is *very* badly configured.

> Check SSH(https://checkssh.com/)
> (we made it)
> Strength: this definitely stops ALL local bad boys.
> Weakness:
> While it's open source(and source code is less than 100 lines)...
> We simply won't give you root password of the server(you don't own the server).
> If adversary is EXTREMELY powerful:
> It's better to set up your own Check SSH.

Humh, still... the heart of your site is:

      shell_exec("ssh-keyscan -p ".$p." ".$h." > ".$f);
      $r=shell_exec("ssh-keygen -l -f ".$f." 2>&1");

So, what difference would that make WRT running ssh-keyscan from a
host we currently trust already?

Download attachment "signature.asc" of type "application/pgp-signature" (837 bytes)