Re: [Full-disclosure] HTTP AUTH BASIC monowall.

I think that we've lost focus of my original question. My question
refined is, does anyone else agree with me that using HTTP BASIC AUTH
for important applications is a security risk/vulnerability (regardless
of SSL)? Or, is everyone here telling me that they "feel safe" if the
connections are SSL'ed and are not worried that the HTTP BASIC AUTH is
only creating a base64 hash of their usernames and passwords that can
easily be reversed? My personal opinion, I feel like we're painting over
the rust on an old car... I don't feel like we're fixing the risks.

Keith wrote:
Does this console have to face the Internet?

Why not put the management console in a protected environment with a
VPN doing the authentication to the subnet that would allow you to
manage it? You should be able to protect the web interface and still
allow the managed devices to report to it.

Of course if it is as weak as you say that may not help - you could
probably attack the interface that receives reports from the client

Good luck with that,


Simon Smith wrote:
SSL is not a fix for the problem, SSL is just a way of evading the
issue or hiding the hole. I can bypass SSL with a man in the middle
attack (which I've already done several times). Once I bypass SSL I am
able to capture the http headers and extract the auth string. The auth
string is vulnerable because it is only a base64 hash. I just reverse
the hash, then presto, I have firewall access... or better still....

Lets take this a step further. There is a tool that I have been
researching for some time. This tool doesn't even use SSL (which really
scares me) and is used for centralized web based computer system
management. This tool enables the administrators to perform tasks such
as mass software installation, mass software removal, record emails, and
even record keystrokes. This tool is a standard tool used by IT
companies around the world to manage their clients networks.

The console for this tool exists on the Internet and is PHP driven.
Login to the console is also plain text and basic auth. If an attacker
can successfully compromise the console (not difficult at all), then the
attacker is in a prime position to extort companies being managed by
this tool. This is possible because the exposure and damage caused to
the company by going after the attacker would be far greater than just
paying the attacker off. (Don't bother asking me what tool this is, I am
not going to tell anyone because that would cross my ethical

So, I guess I've really answered my own question, perhaps I should
release some sort of an advisory on all of these products that are using
basic auth. Basic auth is not really providing anyone with any security.
Maybe they feel good because they need to type in a username and a
password? Would they feel so good if they knew what was really

What is the solution to this problem? Is there a solution that does
not require a different auth type?

Jeremy Bishop wrote:
On Monday 13 March 2006 11:56, Matthijs van Otterdijk wrote:

except for that SSH uses RSA, which uses a public and private key. If
the password is encrypted during the transfer to the site, and can
only get decrypted there, then it can't possibly be sniffed with some
computer inbetween, can it?

As Tim mentioned, the question isn't about the information getting
to a site securely, it's about whether that site is the correct one
and not an impostor.

(I think the original poster was referring to SSL, not SSH, but that
is really immaterial to the question.)


Full-Disclosure - We believe in it.
Hosted and sponsored by Secunia -


Adriel T. Desautels
Harvard Security Group

Full-Disclosure - We believe in it.
Hosted and sponsored by Secunia -