A penetration test is intended to provide reasonable assurance within the scope that it's defined. The length of the engagement, testers' knowledge/experience, and other limiting factors (i.e. removing critical systems from the scope, using a test/dev network instead of production, disallowing social engineering/client-side attacks, etc.) all factor into the equation. Your lawyer/legal team should explicitly state that this is a "best-effort" service within the contract, and you (and/or sales reps/PMs) should also clearly communicate the level of assurance such an engagement will provide to the client.

Also, a PT is simply a test at a point in time. Things can change the second you walk out the door so the language in the contract should stress that its simply a point in time.

I see that seawolve1 had the answer, and while reading other guys posts an old question came back to me "When are you good enough to go out and perform penetration testing ?" . Sil's link confirmed my problem.

Doing "penetration testing" for your own company is a thing. You are their colleague, you can screw things up, but.. they are not perfect, too, so you can play around and do and learn. If you have a senior guy to supervise and teach you... you are the luckiest man in the industry.

But calling yourself a penetration tester and going out and asking for the big buck... is a totally different animal. Years ago a read a book about the differences between the visual and auditory people. What stroked me is that the auditory ones will not start a project unless they have at least 80 - 90 % of the information and skills. Visual people only need 10 - 20% to consider themselves good enough to do the job. While there are advantages to both types, I wold definitely don't want a visual guy come and test my security.

So, question number two, when a security pro is ready to become a penetration testing consultant??


PS The more I study the more I consider myself... no so skilled   

I always got the idea, that you're going to be comfortable with one system and you will begin to test according what you learned and experience, from there you will build up slowly until you be a better pentester.

I will share with you guys a gig I did about 3 weeks ago. Went to another state to perform an assessment slash test against a videoconferencing system. Client is a financial trading information powerhouse who's revenue is in the billions. Premise for the test: "We get on conf calls with the SEC, we want to make sure our conference is secure, untappable, etc. we are using X system" Nothing else was given to me.

I was NOT able to arbitrarily plug in anything without their IT staff getting a whiff of things and literally running to the location were a device was plugged in. I had zero knowledge of the infrastructure outside of: "this is the vendor we use, this is how we make these teleconference calls...."

Under 5 minutes... Trusted laptop on the network, bootable operating system, no DHCP, sniffing the network. Seriously? ... Nice MAC addresses flying by in tcpdump, think I will take one. No MIS guys running to find a rogue device. Teleconferencing? Game over. Credentials were horrible. Gone in under 3-5 minutes. Could I have escalated - sure, but I was only there focusing on the video/voip side of the equation but I mentioned it to them.

Moral of the story: Know your systems and protocols. Had I not understood how voice and video worked, I would likely be intimidated and not known where to begin. Had I not understood how switching, routing, VLANs work - I would not have been able to sniff, hijack a MAC and get on the network. Had I not understood matters of timing, any password cracking would have been detected from excess packets flooding the network. Had I not the ingenuity to created a quick targeted wordlist, I would not have gotten the password and credentials. I sat down and in less than a maximum of about 15 minutes, I had access to do whatever an admin did to their teleconferencing system. As an attacker I could have re-routed the registrar to a rogue server, recorded the calls, took pictures of anyone in a call and so on. What's the big deal you ask? Imagine a conference call before earnings are reported where I was recording. One could make millions, take a company out of business, and so on and so forth.

End of the day, I made my report based on 2 days at the client. I was not allowed to perform a full blown penetration test as many departments had to be involved and the original individual tasked with the test was out of office so the coordination to do the test never came to fruition. They however were spooked enough to understand I needed to really go no further from there. On a conf call with an entire security team, many of whom are visible in the industry (I know of them, the books they've written, what colleges they TEACH at, etc.), not one challenged me on anything I said. I was able to explain the technical risk and swap into the management scope of risk management.

Experience is everything. Not a cert, not a college. When you're comfortable standing your ground with any security engineer, then you're ready to do consulting on your own. When you don't necessarily need to do any research in a quick scenario like this, then you're at your at the top of the game. I am fortunate enough to be such a pain in the ... that I have been able to collaborate with, talk with, learn from some of the top in the industry (and I mean top). This comes from years upon years upon years of studying and dabbling in the industry.

As for money, comes with the territory. Its not everything, there are times I am more curious and in a tinkering kind of mode for the sake of STILL learning something. Those times I can lower a price if I see a benefit (learning something new, testing a unique environment, etc.) Last thing I do though, is ever bite off more than I can chew. If I have trouble understanding a concept, technology, I take a step back rather than make an idiot out of myself pretending to be able to do something I can't. I had to avoid a test that was out of my league that involved satellites, yachts (really big mega millionaire type yachts) and a whole bunch of marine communications. I had to avoid an ATM (airline traffic management) test because its a whole different ballgame. Know your limits, be truthful with yourself. If you have to ask one too many questions and are shaky going into an environment, you might not be ready for this type of work yet.

Zero switchport security (their entire innards were Cisco down)

Man the "CISSP" and "Backtrack CD" in the same sentence made me chuckle.  It is not just security that they screw up.  I have dealt with some as "Infrastructure Architects" spouting off about how iSCSI is an up-and-coming tech for the SAN storage on a VMware infrastructure.  Sadly the world is full of charlatans and snake oil salesmen.  The best thing one can do is learn to smell the BS.  Also some have just been doing the same job for so long that they feel they don't need to educate themselves.  I was on a con call with an IR company and when asked about the network, the senior manager for the networking team spoke up and said, "yes we are segmented, we have different domains between the business units"  /facepalm