The day after my battle cry to Rally the Anti-Spammers here comes ShoeMoney with some great suggestions for stopping blog spam. Everything ShoeMoney posted is very solid advice but some spammers have already been evolving past some of those patches which is why I use my draconian anti-spam methods. Basically, ShoeMoney's advice will stop the majority of your garden variety spammers, but not all as they are constantly adapting, so as you improve your defenses they improve their ability to bypass those defenses.
Remember, security is built in layers and the more layers you pile on, the more the spammers will chip away at your security so building the better spamtrap just results in smarter spammers and they're already here which I'll address with examples below.
Let's examine ShoeMoney's anti-spam advice, see what some state of the art spammers are already doing, and add a few more tricks here and there for even better security.
Starting with the first item he listed:
5) Deny Access to No Referrer Requests
The approach does work on most spammers but I had about 10 requests today where it would've failed. Not that you shouldn't implement this, it's a good trick to stop a lot of spam, just be aware it won't stop everything.
Example:
My bounced spam log shows the following:
IP: 84.110.248.226Take a look at what's in my server log:
User Agent: "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1)"
Subject: "Viagra"
URL: http://anol.webhosting.gs/viagrageneric.html#viagra
84.110.248.226 - "POST /formsubmit.html HTTP/1.0" 200 11918 "http://www.mysite.com/formsubmit.html" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1)"Yup, that's right, a referrer, and I had about 10 of those and they were all from spambots.
Stopping the poorly coded spambots is easy, but they won't be vulnerable for long as the patch to add the domain name being spammed into the referrer is trivial so I expect this anti-spam advantage to be short-lived but I use it too, you should still do this.
Now, let's tackle the next item, which is VERY good advice:
4) Kill tor anonymous proxies
I block many proxies on my servers, which does stop a lot of spam, but don't think that all spammers use known proxies. This is the reason I also block dedicated server hosting facilities because a series of $2 webhosting accounts can be used to effectively spam and bypass the proxy lists.
Example of 4 sample spams (out of many) today that all had referrers mentioned above and came from some ISP/Host called bezeqint.net:
09/29/2006 84.110.248.226Use this with caution:
"Viagra" http://anol.webhosting.gs/viagrageneric.html#viagra
09/29/2006 84.110.244.240
"Viagra" http://gerda.forospace.com/#viagra
09/29/2006 84.110.243.107
"Cialis" http://borea.forospace.com/#cialis
09/29/2006 84.110.241.163
"Cialis" http://kaizer.webhosting.gs/cialisbuy.html#cialis
2) Blacklist Repeat Offenders:
First off, blacklist on the FIRST offense so there is no second time. However, you really need to know what you're doing and lookup who the IP address belongs to so you aren't blocking IP addresses from places like the AOL IP pool (reused every 15 minutes or so) or any other shared proxy dial-up IP pools as those IP assignments are very temporary and the next access is probably a different visitor, not a spammer, so be very careful with this.
This is a gem and we can make it better:
1) Rename your comment file
Excellent advice as I've done that on some websites but don't be shocked when it's short-lived as spammers also have crawlers looking for these comment pages and the fact that you're still linking it under the keyword "comments" is a dead giveaway.
If you're going to change the file name, also change the word that links to the file name to "discussion", "verbal intercourse", or "rants", anything but "comments" to throw them off.
Additionally, move the actual FORM into obfuscated javascript document writes. How this works is the spambot scanning your website can't even find the webform to submit comments as most bots don't use javascript, so only an actual visitor would see an actual webform written into the web page via javascript.
Don't forget the CAPTCHA!
Now, the one thing ShoeMoney didn't mention which works wonders is a simple CAPTCHA and it's keeping a few of my sites spam free without ANY other work involved. Yes, there are ways to bypass a captcha but it's not easy for the spammer. So far most captcha protected sites are safe with such simple protection, but I expect that situation to escalate soon.
Kudos to ShoeMoney for spreading the word, we need more anti-spam information spreading and more people jumping on the anti-spam bandwagon so we can rid the 'net of this scourge as soon as possible and move on to more productive activity.