Penetration tests are among the blunter tools CISOs use to find out weaknesses in their cyber security program. Whether conducted by an internal or external team, a well-conducted pen test will expose ugly realities.
But what’s a well-conducted test? By now most infosec pros know that a test with too many limitations (‘We don’t need to test that: No one would think of attacking us that way’ or ‘We’ve sure this is covered’…) won’t find out very much. However, realistically budget and other restraints may mean some parameters have to be set.
Which is why a column this week by Edward Amoroso, former CSO of AT&T and now chief executive of TAG Cyber LLC, a cyber security consulting firm, will be of interest. He offers five tips for getting the most from a pen test which make a lot of sense.
One of the obvious ones is not allowing the testing team full rein and setting proper limits — like don’t do a denial of service test on a production system.
I won’t run through all Amoroso’s pointers, which are listed here. But I will quote his number one warning. Don’t think that mitigating discovered flows has removed all of them. “Penetration testing – like all testing – is great is showing the presence of errors, but is a terrible means for proving their absence,” he writes. “Never confuse fixes to some penetration test-obtained flaws with security. This is a rookie mistake to avoid at all costs.”
And while pen tests can give a wealth of information about the strength of defences, they are not for every organization. While preparing this I recalled this advice given at a session during last year’s SecTor conference: “If you have a very immature security program and you know it – which most clients do – then that’s a very clear indication you should probably put your money into the building blocks that make you secure rather than a shot in the dark,” said one speaker. “You already know your network is insecure because you haven’t put any effort into it.”
That’s a tip every infosec leader should consider.