[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Pine.LNX.4.44.0509151946580.7897-100000@bugsbunny.castlecops.com>
Date: Thu, 15 Sep 2005 19:50:05 -0400 (EDT)
From: Paul Laudanski <zx@...tlecops.com>
To: jim999@....net
Cc: r.verton@...il.com, <bugtraq@...urityfocus.com>,
<bugs@...uritytracker.com>, <moderators@...db.org>,
<news@...uriteam.com>, <vuln@...unia.com>
Subject: Re: PHP Nuke <= 7.8 Multiple SQL Injections
On Fri, 16 Sep 2005, Matthias Jim Knopf wrote:
> What do you gain from that? In what way would you think your advice did
> ANYTHING GOOD?
> You did neither issue a "addslashes()" as appropriate for SQL-commands,
> nor did you explain, why a variable set by a POST or a COOKIE could be
> worse than anything you could give any URL by appending '?name=...' or
> '&name=...' (->GET vars)
>
If you read the code and original advisory, there is filtering already in
place within the PHP-Nuke framework that monitors HTTP GETs. As such,
this 'exploit' makes of variables which should only be passed via HTTP
GETs and not POSTs nor cookies.
Proper data sanitization for this is simply to retrieve what is expected:
$name = $_GET['name'];
The function you specify like http://php.net/addslashes is neither here
nor there. Why use that function for a variable in POST or cookies when
it is clearly expected to be returned via GET alone?
Ergo:
register_globals off !
--
Paul Laudanski, Microsoft MVP Windows-Security
CastleCops(SM), http://castlecops.com
________ Information from Computer Cops, L.L.C. ________
This message was checked by NOD32 Antivirus System for Linux Mail Server.
part000.txt - is OK
http://castlecops.com
Powered by blists - more mailing lists