Ramparts - Open Source Information Detection

Careguide Team

Fri Jan 12 2018

Take the following message from a spammer to another user on our site:

Hello Anna.kindly have your cell number and email sent to my cell phone (five- five- five-nine -five-two-nine-six-eight-eight )to enable us discuss the job duties and pay scale ...

For a while we've had a basic method in place to detect phone numbers, emails, and spammer-esque urls. The former example was however, beyond its scope. We needed a solution that could provide us better results while at the same time be easily updatable to spammers tactics.

On that note, CareGuide is happy to introduce Ramparts, an open source spam detection Ruby gem. There's no shortage of gems that parse blocks of text to find emails, phone numbers, and urls. However, what was lacking in the space was a solution to find these attributes (eg. emails, phone numbers) when the user purposely tries to obfuscate them. For Ramparts, the example noted at the beginning of the post is entirely within its scope, which can go as far to detect l33t speak numbers with spaces eg. 'F 0 r F l v 3'.

CareGuide keeps confidential a user's phone number and emails so that they don't get harassed by spammers or sold to third parties. To contact other users on all our sites, users have the ability to message through the platform and make calls using TeleSafe. Telesafe shows a toll-free number in place of your own number when you make a call to another user on CareGuide.

While this works great for most users, spammers and scammers still want to be able to post phone numbers, emails and scammer-esque urls. Why? A few reasons. Number one is they want to get potential victims off our platform and onto a platform where they have much more freedom and control. Also, if they post an email or phone number, there's a chance the recipient will contact them. Once contacted they'll have the recipient's phone number or email which they can then resell to third parties if the scam attempt fails.

Spammers, or anyone for that matter, trying to post phone numbers and email in areas where it is restricted isn't new. Airbnb and countless other sites have users posting in phonetics and even l33t speak to bypass their filters. While we realize it's going to be an arms race, we thought this was a useful starting point others could use to begin to protect against spammers.

As you read this, Ramparts is being used by both NannyLane and all our constituent CareGuide sites.