full-disclosure - Re: NTLM HTTP Authentication is insecure by design

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-next>] [day] [month] [year] [list]

Date: Tue Jul 19 20:38:44 2005
From: aksecurity at hotpop.com (Amit Klein (AKsecurity))
Subject: Re: NTLM HTTP Authentication is insecure by
 design - a new writeup by Amit Klein

Dear 3APA3A,

On 19 Jul 2005 at 13:35, 3APA3A wrote:

> Dear Amit Klein (AKsecurity),
> 
> 
> 
> --Monday, July 18, 2005, 9:40:32 PM, you wrote to bugtraq@...urityfocus.com:
> 
> AKA>                      NTLM HTTP Authentication
> AKA>               (and possibly other connection-oriented
> AKA>           HTTP authentication and authorization protocols)
> AKA>                       is insecure by design
> 
>  NTLM  authentication is insecure by design for external authentication,
>  because  of  "single  signon" ideology. NTLM authentication can be used
>  for  NTLM  proxy  attack. For example, attacker who can hijack or spoof
>  server  connection  with NTLM authentication can use this connection to
>  access  different  Web server, mail server or file server with client's
>  credentials.  This  is  known for many years and was discussed for many
>  times.  

While not downplaying this remark, I'd say that:

1. By this argument, any non-cryptographically strong protocol should be considered 
insecure by design. I don't think that's what people have in mind. I think that per each 
protocol, there's a set of expectations. For example, no-one expects NTLM auth to protect 
data in transit. But people do expect NTLM not to by bypassed by silly methods.

2. I see a big difference between spoofing and hijacking attacks, which require non trivial 
skill set, and the attack I described, which, given the right preconditions, is so easy to 
mount that it can be done so accidentally.

Few  years  ago  Internet  Explorer  was  patched  to  use NTLM
>  authentication  only  for  local  network zone. Local network are hosts
>  with  NetBIOS name (for example WEBSRV, excluded by default from proxy)
>  and list of proxy exclusions.
>

Uh, I don't think so. From my experiments with IE 6.0, it happily engages in NTLM 
authentication on non local network sites. In fact, there are many sites on the Internet 
which require NTLM authentication. For example, OWA 2000/2003...

>  So,  under default configuration, NTLM will not be used through proxy
>  server, at least in Internet Explorer.
> 

As a result of my former comment, I have to disagree. There ARE websites that require NTLM 
authentication, IE DOES perform it, so there's no theoretic reason why there shouldn't be 
proxy servers in between.

Thanks,
-Amit