Detect up to 1327 disposable email providers with MailChecker | Hacker News Reader

Detect up to 1327 disposable email providers with MailChecker

github.com

44 pointsfgribreau45 comments

If your site will be sending something of value over email, then people will want to use a real address, and you don't need this. On the other hand, if you just want a user ID, then it doesn't matter what they use, and you don't need this, either.

So, if you use this, you're a site who wants a real deliverable address for some reason, but which doesn't offer enough benefit to the user to naturally compel them to share one. Put differently, if you need this, you're precisely the kind of site that shouldn't get my real address.

All that this script means is that I'm going to leave your site unregistered and frustrated, and will mentally bookmark your domain as "those twits who force you to sign in and creepily want an actual address".

Except you probably do need this if your site allows posting of user generated content, because spammers love these throwaway email addresses.

Spammers who bother with email based registration sites at all, often take the time to routinely acquire one address from a major free provider and use it everywhere they can't use a throwaway as well as get naive people to send email to it.

For other people who kind of don't trust your site, its future, or its security, but are too lazy to create extra accounts, you are forcing them to do a full evaluation in a way that adds additional weight to the possibility the site is an actual data broker today instead of an existential spam threat.

Personally, if I decided a site was bellow the threshold for my real email on an initial encounter and then I saw it perform this kind of detection, I wouldn't touch the site again.

I built this kind of check into my ecommerce sites for high-value items. They're disproportionally susceptible to fraud and carding for some reason, and implementing a "no free email provider" check has cut fraud on those items to zero.

I work at a financial institution and disposable e-mails are used by fraudsters all of the time. This is certainly better than my current blacklist.

Do you blacklist gmail and hotmail addresses as well? It takes 5 minutes to put a throwaway gmail account together.

Yes, this. Processing fraudulent cards can result in a costly sum of chargebacks for the seller, and a fake email address is a very strong indicator that the person is a fraudster. Other services like Maxmind minFraud consider the email address and many other attributes when ranking the likelihood that the order is coming from a fraudster. It doesn't mean that you have to disallow the order from being saved (or whatever), just that you might want to flag it for manual review before processing the CC to capture payment.

This point is orthogonal to the OP's.

OP's point as I understand it is "there is no circumstance where you need this, because you're either providing value or you're shit."

My point is that there are legitimate circumstances where you are providing value but in a way that people take advantage of for unrelated fraudulent purposes.

I expect blocking GMail, Hotmail, & Yahoo Mail also cut your conversion rate close to a factor of zero of what it would otherwise be.

Here's how that might happen: http://mailinator.blogspot.com/2011/05/how-to-get-gmailcom-b...

Nope. The blocking is only active on certain products over $500 that don't make sense for non-companies to buy.

I've encountered sites that do this kind of check before, and I usually just don't bother signing up.

It's usually the case that I didn't /really/ want to sign up anyway, but I needed something (that with a little effort I could find elsewhere)

Seriously, if any site owner has even a half of a brain, he should stay away from this shit.

Disposable email exists for a reason.

That is not true at all.

If your site is just a chat forum and you don't mind spammers, sure let people use disposable emails.

If your site involves money or auctions or credit cards.. letting disposable emails in can be very costly.

If a quick-and-dirty domain check on an email address is more secure for fraud detection than credit card data, then that says a lot about how broken the whole credit card system is. No surprise, of course. Sad that Bitcoin blew its chance to replace this crap.

You don't replace, you use it in addition. Layers of security.

Now if you want to spam 500 million credit cards on my system, you are going to need to at least make them from some common domains, which I can block down the road. If you are using an email hiding service, way easier for you.

Not sure bitcoin matters, if I had users paying with bitcoin in an app I would STILL blacklist certain email domains.

You probably have encountered even more sites that use this, and didn't notice because you didn't try giving them a throwaway address. Which might be exactly the point of using the filter.

Searching in a hard-coded list of strings is not the best idea, and why try to detect them anyway? You can block all the free email services as well because they can also be used as disposable mail providers with a few more steps to get going. Not to mention that it takes around 10 minutes to set up a disposable email service if you have a spare domain and when it's in this list so you let it expire, even if this repo gets updated, there'll be many sites using an older version, causing too much problems for the new owners.

I agree with the hard-coding being not-too-great-of-an-idea. However, I'm going to argue that a @gmail has a way higher chance of being a legitimate email out of a random sample space than a domain from @tempx.com.

The larger issue here is if someone is entering false information, it's likely because you don't have a good UX and are soliciting information before value is delivered.

With passport.js, it's so easy to implement authentication providers against existing FB/Twitter/LinkedIn accounts, which should suffice for your SaaS 14-day trial. When it expires, they'll enter in legitimate information if they see value added to it.

I understand everyone wants to capture e-mails for drip marketing purposes, but if you have a bunch of people entering fake information because you walled off something critical (Oracle, I'm looking at you and your JRE...), it's not an engineering problem - its a social one.

Site owners that need to use this shit probably also ignores users need for privacy, like ignoring unsubscribe requests, the right to be removed from their databases, or hiding the fact that they distribute or sell user info.

> This will be very helpful when you have to contact your users

Someone using a disposable address probably does not even care about your "need". Why block them?

Agree, they will go to the competitor instead.

Why allow them if you think they probably are more trouble than they are worth? (no way to contact them if there are payment issues, easier to take-over accounts, ...)

IMHO it really depends on the kind of service if this makes sense or not.

Totally depends but in the majority of cases I see no issues. Why do you assume they are up to cause trouble? I used throw-away mails hundreds of times and the only thing I usually did was inflate their user table with another single time visitor.

hushmail.com is on that list, this is an increased privacy email with built-in PGP and encryption support, and it's a paid service this is hardly a throwaway email, funny enough their alias domains (e.g. nym.hush.com) aren't on that list, if any one will use it as a throwaway email (i do) they'll use the alias function and delete or suspended the alias after the signup (that's what i do, if i need to reset a password i recreate the alias). That list seem also to contain ISP's from eastern Europe and Asia so I would go over that list very carefully before implementing it because it might break your site.

China's largest freemail provider (which requires a chinese national identity number to register for) is also on that list.

Lol.

Please don't do this, unless you are being attacked and have exhausted all other options (such as fingerprinting the attackers in a more targeted way).

I appreciate all the efforts that went into making this but I wish you didn't release it.

I wish that you didn't write this comment

So much vitriol in here for something I immediately thought: "damn, that'd be handy." So often we have a potential student email us, and our reply disappears into a spam folder because it came from Zendesk or Postmark, or because their own words (echoed back) trigger it. And when we finally reach those people, people who actually and legitimately want to talk to us, they're thankful that we tried to find them, and asked us to change their email entry in our database to a more reliable one. Then there are those who get angrier and angrier at our failure to respond, while our responses are piling up in a spam folder somewhere. Well THAT's a great start to what is typically a years-long relationship. Or those who just take their business elsewhere.

Incidentally, ours is a B2B market, remember that when judging how we - and others - roll. Also: we check RBLs and test deliverability frequently, use SPF/DKIM allowing the providers we use to deliver mail on behalf of our domain, and we have DMARC configured so we can better monitor deliverability issues.

I'd use this to show a brief warning on our "contact us" form: "we've found that using an email provider such as example.com may result in our reply getting filed away as spam, and we recommend using your real work address to be sure you see our reply without having to dig through all the spam on your personal email account. You don't have to change it, but we wanted to warn you in advance."

Neat, a list of 1327 disposable e-mail providers.

In fact, there are now 1979 disposable e-mail providers

I think this would be more usable if the lists weren't hard-coded into the functions. Passing JSON as a parameter would be possible for any language worth using, no "templates" necessary. Requiring Node.js to rebuild an array inside a PHP function seems kind of weird.

Also, list.json doesn't appear to be a valid json file since it uses comments.

I'm not going to comment on whether or not I think it's useful.

An argument that I haven't seen yet is that some of these providers allow to choose the mail address used, allowing anybody who knows or can guess the address used to take over the account.

I signup with bugmenot:bugmenot whenever I can to explicitly allow that.

I wonder if all the critical commenters here feel the same about sites sending a verification e-mail before completing your registration. After all, it serves the same purpose of preventing usage of bogus e-mail addresses.

If you're against this kind of library, I suppose you shouldn't bother with verifying e-mail addresses either.

Verification emails serve the purpose of ensuring that people don't sign up other people's email addresses.

And if they do, why is that bad?

Because it makes signing people up to all kinds of things they don't want to be signed up to trivial.

If you are actually going to use the address given, it makes sense to verify it to prevent abuse.

Not least because you risk having your ability to deliver e-mail severely jeopardised by abuse if you don't.

Because of the script that destroys email addresses by signing them up for 5000 mailing lists.

What script is that? Never heard of such a thing.

There were several going around in the "warez" scene in the mid-90s.

It didn't require a script, either. When mailing lists and other automatic email sources let you add destination addresses without closing the confirmation loop by sending a test email, you can denial-of-service email addresses with just an SMTP client.

The really nasty part about this attack is that it's not just bandwidth amplification. Normal amplification attacks go away when the attacker decides to stop sending packets. With mailing subscriptions, the badly-configured mailing lists keep sending the attack on their own.

People can create fake accounts with other person email. For example, a last year case of a fake account for Linus Torvalds in change.org

https://plus.google.com/+LinusTorvalds/posts/DPY7H4a9Ma5

> Somebody signed a Change.Org petition in my name, and using a really old email address of mine.

> So since I apparently had an "account", I reset the password, and made a petition of my own.

> Change.Org - please change your dickish ways. Ok?

There is a difference in detecting if someone actually owns an email address and detecting if an email address is the persons primary email address.

Reply on news.ycombinator.com