asked on

Preventing Credit Card Form Abuse Captcha Solution

We have a donation form on our charity site. We would like to accept donations without requiring users to set up an account. We have PayPal set but we also have a form for entering credit card info. Previously we had a very difficult captcha. Later we changed this to human readable question, e.g. "is the sun hot or cold" (answer, cold) We have about 15 questions that a called by random. But it is being hacked. Someone keeps using the form to test credit cards, charging $1 to $50.00 onthe same and different cards. Obviously he is testing and when one goes there he use it to purchase things or whatever. So we need to set up again, a strong defense for this abuse, but we are not sure the best method. Some new captcha methods are good, but so difficult that you have to be an artist or super visual brain to do them. We want a system that works, but will not be so hard as to cause resistance to users to donate.

ASKER CERTIFIED SOLUTION

Amick

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

Sivakatirswami

ASKER

@amick Thanks for the very clear advice. I asked the accounting team (who see the charges being made) the time between charges. sometimes it was a few minutes. So, I have heard of the IP wait, but we don't think it will help. Either he was running dictionary attacks on the form (to get words that would answer our human readable questions) which would take some time (until he had the answers and logged those and then he could reduce his dictionary to the answers that worked before) OR he was manually sitting there madly filling in the form and entering CC numbers by hand. Hard to know... We took down the form, but forgot to disable the CGI, so he already had a copy of our form and continuing POSTing directly to the CGI without going thru our form, we finally had to turn disable the CGI itself.

I found another system we are going to try: it generates a random set of colored digits that are relatively easy to read along with a MD5 hash string in a hidden field that must be submitted back along with the form.

So this will a) be relatively easy for users to see and enter but b) impossible to attack programatically. (we think)

So, this will leave him no choice but to manually fill in the form. Hopefully this will reach critical mass for "difficult and time-consuming..."

we will see

Amick

It sounds as if you've got a good plan for the challenge.

If this were my site, I would keep an eye on the "top IP" reports from my web server, and archive the server logs. Also, don't hesitate to get the authorities involved - In the US wire fraud is a Federal offense, but the penalties range from a fine and probation to several years in prison depending upon the severity of the crime and the history of the offender. There is a website that accepts reports of Internet crime, http://www.ic3.gov, and there may already be a wealth of information available from the FBI to help you prevent further attacks.

It is good that you are taking the problem seriously, and with a combination of luck and skill you'll have this problem controlled very soon.