Thanks for the detailed response Susan. Some comments inline below
If you can't nail an RTM Windows 2000 in say... oh... what..
5 minutes or less? I'd be surprised. I'm not sure that's
testing those pool shots (and what is it with security and
people who play pool?) and exercising anything when that's
sooooo vulnerable it's not funny. You don't even have to do
anything.. just build it and stick it on the internet. What
kind of pool shot is that?
While I agree that the degree of difficulty to compromise an RTM w2k image
is practically nil, I don't see practicing on it as completely useless,
especially for beginners. I think it all depends on how you practice. The
reason even great pool players practice simple shots is to hone their craft
to amazing levels of understanding and intuition. A straight-in shot to a
corner pocket is easy. A straight-in shot to a corner pocket where your cue
ball consistently ends up at the same exact spot every time isn't as easy. A
straight-in shot where you vary draw, follow, or english (uh, places where
you hit the ball that will affect how it rolls for you non-pool players) and
*still* getting the cue to stop at that same exact spot... That's real
mastery of skill. Solid repeatable results 99.9% of the time regardless of
I view pen-testing practice much the same way. Repeat over and over until
it's second nature to you... And then change something and try to get the
exact same results.
John could easily compromise an RTM w2k image with an IIS 5.0 exploit. But I
don't think owning the box should be the only point of practice if he wants
to expand his knowledge and get better. Take a simple known exploit. Use it.
Use it many times until you are thoroughly comfortable with it. Now take a
deeper look. How exactly does the exploit work? Buffer overflow? What is the
diff between the patched version and the unpatched? What does the traffic
look like on the wire? How exactly does the target change or react when the
exploit hits? How would I hide any telltale signs?
Now try mixing things up a little. If there was an IPS in the way how would
you fragment the packets to still get that exploit through? What is the
minimum level of fragmentation that would still work and what is the
difference in the amount of time it takes? Heck, what about different types
of IPS? If you modify the exploit payload how does the target box react? Can
you modify it enough so that the standard signatures on the IPS don't
trigger? How many different variables can you work through and *still* reach
your objective? Solid repeatable results 99.9% of the time regardless of the
variables should be the goal.
This kind of practice isn't going to make RTM w2k IIS any less exploitable,
solve some great unknown, or cause women to swoon. But I can almost
guarantee that it will give him a more thorough understanding and background
of how and why it works. The research, testing and understanding needed to
tackle just the suggestions above for a beginner will greatly enhance their
skills. Building on a foundation of knowledge and then adding layer upon
layer isn't a waste of time IMHO.
Go to the metasploit site and see if some of the oldies but
goodies are there. Any of the IIS5 stuff will work....
I have the feeling that HD Moore uses a lot of the same tactics I describe
above for practice when looking for new holes and exploits. Poking and
prodding and seeing what happens when things are changed around. Of course,
he may be too hung over after Defcon and Blackhat to reply right now :)
SecurityFocus pen-test list
"Do Not Taunt Happy-Fun Ball"