I currently have a script for blocking non-business email addresses on my website form (Marketo) yet bots are still finding a way to bypass it. Today I got another "gmail" and "hotmail.fr" submission. Whenever I test my form it works but the bots are finding a way to bypass the script. I have also tried the honeypot method but it has not worked. They are also submitting with two-letter names, the past month has been really bad and I am desperate for help, I am not very good in JS so any help would be very much appreciated.
This is my script:
(function (){
// Please include the email domains you would like to block in this list
var invalidDomains = ["#gmail.","#yahoo.","#hotmail.","#live.","#aol.","#outlook.","#icloud.","#zoho.","#hubspot.","#gmx.","#yandex.","#mail.","#email.","#tutanota.","#trashmail.","#lycos.","#tutanota.","#protonmail."];
MktoForms2.whenReady(function (form){
form.onValidate(function(){
var email = form.vals().Email;
if(email){
if(!isEmailGood(email)) {
form.submitable(false);
var emailElem = form.getFormElem().find("#Email");
form.showErrorMessage("Must be Business email.", emailElem);
}else{
form.submitable(true);
}
}
});
});
function isEmailGood(email) {
for(var i=0; i < invalidDomains.length; i++) {
var domain = invalidDomains[i];
if (email.indexOf(domain) != -1) {
return false;
}
}
return true;
}
})();
</script>```
Some bots when they run, they don't even execute the JavaScript on the page. They just take the fields, find the post request the form is submitting, and submit the fields to them with pre-defined values. Thus, ignoring your validation completely.
So now the solution would be checking the post request values on the backend. Some people use the fact that some bots are dumb, and they include a honeypot field in their fields. They mark it hidden with CSS on the frontend, but again, some bots are dumb, and they will fill it out regardless and send it in the post request. Now you can have your sever throw that out right away.
so I received an obvious phising email today with this js code in it:
<script type="text/javascript" language="Javascript1.1">
<!-- Begin
var bCancel = false;
function validateRegistrationDetails(form) {
hmrc.portal.clearFieldValidationErrors(form);
if (bCancel) {
return true;
} else {
var registrationDetailsPageMessage = new String("<p>ERROR: This page contains one or more errors. See details below.</p>")
var formValidationResult;
formValidationResult = validateRequired(form) & validateMask(form) & validateIdenticalEmailAddresses(form);
if (!formValidationResult){
var formName=form.name;
var ele=document.getElementById('pageError.registrationDetails');
if(ele){
ele.innerHTML = registrationDetailsPageMessage;
ele.style.display = ''; }
}
return (formValidationResult == 1);
}
}
function registrationDetails_required () {
this.a0 = new Array("selectedServices", "<p>ERROR: Please select at least one online service.</p>", new Function ("varName", " return this[varName];"));
}
function registrationDetails_mask () {
}
function registrationDetails_identicalEmailAddresses () {
}
//End -->
</script>
Is it malicious in anyway, what exactly does it do with the form data. I am not that versed in vanilla javascript. Any explanation would be helpful.
Thanks
In all likelihood, whoever sent you this simply lifted a section of HTML and inline JavaScript from the site they were trying to pretend to be. A few lines in the code such as:
hmrc.portal.clearFieldValidationErrors(form);
suggest that they were trying to be HMRC, with the rest of the code being simple validation of the information being entered; I'm going to guess that the content was taken from the 'Registration' section of that site
So you've already established that it's a phishing email.
Typically phishing emails try to make themselves look legitimate by copying large chunks of code from the original website that they're trying to pretend to be (ie your bank's site or whatever). They'll then alter that code so that it sends the relevant data to the phisher rather than to the bank. They may also add fields that weren't in the original, such as asking for your PIN, etc.
However, the main point here is that the bulk of the original code is generally retained, in order to maintain the look and feel of the original site.
Therefore the chances are that the code you're seeing has actually been copied by the phishers from the original site.
There's nothing explicitly malicious about this code in itself -- it has a lot of badly written code, but it isn't trying to do anything wrong in this code.
Where the problem lies for the phishers here is that Javascript code is blocked by most email clients; ie regardless of its intent, the chances are that that this code won't actually work in your mail client.
But I would guess that the phishers have just taken the original form wholesale from the website and dumped it into an email without bothering to take out any javascript that might have been embedded in it.
So the short answer is: Don't worry about this code in particular, but please do delete the email.
As far as I can see, there's nothing malicious with it, unless some script has been included outside of this script itself.
I've spent the past few days working on updating my personal website. The URL of my personal website is (my first name).(my last name).com, as my last name is rather unusual, and I was lucky enough to pick up the domain name. My e-mail address is (my first name)#(my last name).com. So really, when it comes down to guessing it, it's not very hard.
Anyways, I want to integrate a mailto: link into my website, so people can contact me. And, despite my e-mail address not being very hard to guess, I'd rather not have it harvested by spam bots that just crawl websites for e-mail address patterns and add them to their database.
What is the best way for me to obfuscate my e-mail address, preferably in link form? The methods I know of are:
e-mail me
It works, but it also means that as soon as my website hits Google, I'll be wading through spam as spam bots easily pick out my e-mail address.
<img src="images/e-mail.png" />
This is less desirable, because not only will visitors be unable to click on it to send me an e-mail, but smarter spam bots will probably be able to detect the characters that the image contains.
I know that there is probably no perfect solution, but I was just wondering what everyone thought was best. I'm definitely willing to use JavaScript if necessary, as my website already makes use of tons of it.
I encode the characters as HTML entities (something like this). It doesn't require JS to be enabled and seems to have stopped most of the spam. I suppose a smart bot might still harvest it, but I haven't had any problems.
Personally, I've given up on hiding my email address. I find it easier to look into better spam-filtering solutions than worry about obfuscating. You could spend days trying to find the best way to obfuscate your address, and then all it takes is one person to sell your address to a spammer and all that work was useless.
The current accepted solution is to create a contact form that allows users to email you. If you receive a lot of spam from that (I don't on my site), then you can add a captcha for good measure, and you'll be far from the "low hanging fruit" at that point.
The fact of the matter is that if you are providing a link that a user can click on to pop open their email client with your address in the To: field, then the computer is able to decipher the email address from the page and so can a spam bot.
You mentioned this is for your personal website. On my personal site (for example, bobsomers.com) I just have a paragraph that says this:
The best way to get in contact with me
before the new site is up is to send
me an email. My email address is my
first name at this website. If you
can't figure it out from that hint,
well, you might find email more of a
challenge than figuring out my
address.
People seem to be able to figure that out just fine, as I get legitimate email all the time. Sometimes the best solutions don't require writing any code. :)
A lightweight way to obfuscate the href of an anchor is to base64-encode it:
> btoa('mailto:email#example.com')
< "bWFpbHRvOmVtYWlsQGV4YW1wbGUuY29t"
And then include it hardcoded:
E-Mail
Or dynamically server-side e.g. in PHP:
E-Mail
In combination with string reversion it could be pretty spam-save:
<?= strrev("email#example.com") ?>
Apparently using CSS to change the direction of your text works pretty well. That link has a test of a bunch of other obfuscation methods as well.
Whatever you use is inevitably going to be defeated. Your primary aim should be to avoid annoying the heck out of your users.
Don't use any obfuscation techniques here because it's probably the first place the email harvesters will look to find out how people are obfuscating emails. If you have to have your email address visible on the site don't just copy verbatim someone else's method; obfuscate it in some unique way that no other site has used so that your method won't be known to harvesters before they visit your site.
mine is actually simple:
<h3 id="email">hello#gmail.com</h3><!-- add a fake email -->
$(document).ready(function(){
//my email in reverse :)
var s = 'moc.elibomajninbew#htiek';
var e = s.split("").reverse().join("");
$('#email').html(''+e+'');
});
You could do as Google do on Google Code (and Groups). Display a par tof the email, and a clickable portion ("..."). Clicking that indicates you want to know the email, and you are asked to fill in a captcha. Afterwards the email (and others?) are visible to you.
One website I maintain uses a somewhat simplistic JavaScript means of (hopefully) keeping spambots out.
Email links call a JS function:
function sendEmail(name, domain) {
location.href = 'mailto:' + name + '#' + domain;
}
To make sure only users who have JS enabled can see the link, write them out with this:
function writeEmailLink(realName, name, domain) {
document.write('<a href="javascript:sendEmail(\''
+ name + '\', \'' + domain + '\')">');
document.write(realName);
document.write('</a>');
}
The use of one JS function to write out a link that calls another means that there are two layers of protection.
I use JavaScript obfuscation, take a look at this one for example:
http://www.jottings.com/obfuscator/
As a poster above said, I also use JavaScript obfuscation from the jottings website.
The web page generates some JavaScript which can be improved on. The mailto: text string is in the clear and identifiable by robots (which could spot this and unobfuscate this string), but if one enters into the jottings.com webpage an email address of the form mailto:addr#site.tld instead of addr#site.tld and then removes the text mailto: from the JavaScript that is generated, one suddenly has some JavaScript that does not look as though it has anything to do with email at all - just random JavaScript of which the web is full. One can improve this still further by getting rid of the link text - I replaced mine by an image of my email address that is in a fairly obscure font. Then just in case this method on jottings.com becomes popular, I randomized the variable names in the output JavaScript to make it hard for a robot to spot an instance of jottings generated JavaScript code.
Obviously some of these improvements could be built into the mechanism on jottings itself, and since the code is openly available this would be relatively easy.
An example may make this a bit more clear. I used the Jottings Obfuscator at the link above to obscure mailto:foo#bar.com (note I am cheating on the original intent of the jottings website by entering the string mailto:foo#bar.com instead of foo#bar.com) with text "Send Me Email", which jottings turned into this Javascript:
<script type="text/javascript" language="javascript">
<!--
// Email obfuscator script 2.1 by Tim Williams, University of Arizona
// Random encryption key feature by Andrew Moulden, Site Engineering Ltd
// This code is freeware provided these four comment lines remain intact
// A wizard to generate this code is at http://www.jottings.com/obfuscator/
{ coded = "3A1OTJ:rJJ#VAK.GJ3"
key = "J0K94NR2SXLupIGqVwt8EZlhznemfaPjs7QvTB6iOyWYo3rAk5FHMdxCg1cDbU"
shift=coded.length
link=""
for (i=0; i<coded.length; i++) {
if (key.indexOf(coded.charAt(i))==-1) {
ltr = coded.charAt(i)
link += (ltr)
}
else {
ltr = (key.indexOf(coded.charAt(i))-shift+key.length) % key.length
link += (key.charAt(ltr))
}
}
document.write("<a href='mailto:"+link+"'>Send Me Email</a>")
}
//-->
</script><noscript>Sorry, you need Javascript on to email me.</noscript>
After I get that back, I paste it into an editor and:
remove the mailto:
replace link text with pointer to an image of my email address
rename all the variables
replace the "noscript" section with another link to the email address image
I end up with this:
<script type="text/javascript" language="javascript">
<!--
// Email obfuscator script 2.1 by Tim Williams, University of Arizona
// Random encryption kkeoy feature by Andrew Moulden, Site Engineering Ltd
// This kudzu is freeware provided these four comment lines remain intact
// A wizard to generate this kudzu is at http://www.jottings.com/obfuscator/
{ kudzu = "3A1OTJ:rJJ#VAK.GJ3"
kkeoy = "J0K94NR2SXLupIGqVwt8EZlhznemfaPjs7QvTB6iOyWYo3rAk5FHMdxCg1cDbU"
shift=kudzu.length
klonk=""
for (variter=0; variter<kudzu.length; variter++) {
if (kkeoy.indexOf(kudzu.charAt(variter))==-1) {
lutu = kudzu.charAt(variter)
klonk += (lutu)
}
else {
lutu = (kkeoy.indexOf(kudzu.charAt(variter))-shift+kkeoy.length) % kkeoy.length
klonk += (kkeoy.charAt(lutu))
}
}
document.write("<a href='"+klonk+"'><img src='contactaddressimage.png' alt='Send Me Email' border='0' height='62' width='240'></a>")
}
//-->
</script>
<noscript>
<img src="contactaddressimage.png" border="0" height="62" width="240">
<font face="Arial" size="3"><br> </font></p>
</noscript>
I don't how well this would work. Could you not leave your email address out and make it load using an AJAX call once the page has finished loading. Not sure if spam bots can pick up the altered HTML or if they are clever enough to listen on other HTTP traffic to try and pick email addresses or if they just scan the page as it is received the first time.
One guy tested nine different ways of presenting an email address on a page and then published results on his blog.
His three best ways were:
Changing the code direction with CSS
Using CSS display:none
ROT13 Encryption
Caveat -- this was posted two years ago. Spam bots might've gotten smarter.
If you work with PHP, you can grab a free script that does that automatically. It's called "Private Daddy" and we use it for our own online audio streaming service. Just one line of code and it works out of the box... you can grab it here
Another approach could be by using a JavaScript framework and binding the data/model to the HTML elements. In the case of AngularJS, the HTML elements would be written as:
<span>{{contactEmail}}</span>
The interpolation {{data}} binding uses a scope variable that contains the actual email value. In addition, a filter could also be used that handles the decoding of the email as follows:
<span>{{contactEmail | decode}}</span>
The benefits are in the way the HTML is written. The downside is that it requires scripting support which some for may be a no no.
just another approach.
Using JQuery, but can easily be ported to plain JS if needed. Will take the following HTML block. This example I provided is also for tel: links for phone calls.
<a class="obfuscate"
href="mailto:archie...trajano...net">
archie...trajano...net
</a>
<a class="obfuscate"
href="tel:+One FourOneSix-EightFiveSix-SixSixFiveFive">
FourOneSix-EightFiveSix-SixSixFiveFive
</a>
and convert it to the proper links using Javascript.
$(".obfuscate").each(function () {
$(this).html($(this).html()
.replace("...", "#").replace(/\.\.\./g, ".")
.replace(/One/g, "1")
.replace(/Two/g, "2")
.replace(/Three/g, "3")
.replace(/Four/g, "4")
.replace(/Five/g, "5")
.replace(/Six/g, "6")
.replace(/Seven/g, "7")
.replace(/Eight/g, "8")
.replace(/Nine/g, "9")
.replace(/Zero/g, "0"))
$(this).attr("href", $(this).attr("href")
.replace("...", "#").replace(/\.\.\./g, ".")
.replace(/One/g, "1")
.replace(/Two/g, "2")
.replace(/Three/g, "3")
.replace(/Four/g, "4")
.replace(/Five/g, "5")
.replace(/Six/g, "6")
.replace(/Seven/g, "7")
.replace(/Eight/g, "8")
.replace(/Nine/g, "9")
.replace(/Zero/g, "0"))
})
I documented it in more detail here https://trajano.net/2017/01/obfuscating-mailto-links/
The de/obfuscation algorithm is pretty simple so its not too taxing to write either (no need for base64 parsing)
The Ajax call solution
The best is to have a form on the website and not to show email address, because all robots are more intelligent day after day, but if you need to show email address on the website, so, you can make it with ajax call on your server, and show it on click.
HTML
<a class="obfmail" href="#" rel="info">click here to show email address</a>
or
<a class="obfmail" href="#" rel="info">
<img src="img/click-to-show-email.jpg">
</a>
jQuery
$(document).one'click', '.obfmail', function(e) {
e.preventDefault();
a = $(this);
addr = a.attr('rel');
$.ajax({
data: {
email: addr
},
url : "/a/getemail",
type: "POST",
dataType: 'json',
success: function(data) {
a.html(data.addr);
a.attr('href', 'mailto:' + data.addr);
}
});
});
PHP
if($_POST['email']) {
...
return json_encode(array(
code => '200',
response => 'success',
addr => 'info#domain.ltd'
));
}
For more security, you can change .on by .one like this $(document).one('click', '.obfmail', function(e) { or even work with a PHP generated token that you pass into data on ajax call, to accept only one call of the ajax function like this :
html: <a class="obfmail" href="#" rel="info" token="w3487ghdr6rc">
jquery:
...
addr = a.attr('rel');
tkn = a.attr('token');
$.ajax({
data: {
email: addr,
token: tkn
}, ...
.
It is possible to encode the returned email address too or invert it.
.
Working fine for phone numbers too !
Cloudflare is now offering free email obfuscation service. This might be an option if you use Cloudlfare.
I agree with #srobinson in that using an online form for encoding to html entities seems a little shady. A few lines of python will do it for you:
def htmlEntities( string ):
return ''.join(['&#{0};'.format(ord(char)) for char in string])
htmlEntities("barnstable#example.com")
the above returns:
'barnstable#example.com'
Which is barnstable#example.com encoded to HTML Entities (surrounded by single quotes).
Honestly, your problem may be moot if you asked the question of whether or not a mailto is really what you want to use. A lot of people who use web mail, for example, or do not have the proper mail client setup in their browser are not going to benefit from a mailto. You are exposing your email address for a function that isn't going to work for a large portion of your users.
What you could do instead is use a form to send the e-mail behind the scenes so that the e-mail address is hidden and you don't have to worry about the poor saps who won't benefit from a mailto.
If you say on your site that "My e-mail address is (my first name)#(my last name).com.", and your first name and last name are pretty darn obvious, that seems to be the best spam protection you're going to get.
If anyone's using Rails, they can use the actionview-encoded_mail_to gem. (https://github.com/reed/actionview-encoded_mail_to)
There are a few options:
:encode - This key will accept the strings "javascript" or "hex".
Passing "javascript" will dynamically create and encode the mailto
link then eval it into the DOM of the page. This method will not show
the link on the page if the user has JavaScript disabled. Passing
"hex" will hex encode the email_address before outputting the mailto
link.
:replace_at - When the link name isn't provided, the
email_address is used for the link label. You can use this option to
obfuscate the email_address by substituting the # sign with the string
given as the value.
:replace_dot - When the link name isn't provided,
the email_address is used for the link label. You can use this option
to obfuscate the email_address by substituting the . in the email with
the string given as the value.
<!-- Multi-Email Obfuscator -->
<!-- step 1: # = # -->
<!-- step 2: a scrap element -->
<!-- step 3: ROT13 encode for .com -->
info<!-- step 1 -->#<!-- step 2 --><b style="display:none">my</b>domain<!-- step 3 --><script>document.write(".pbz".replace(/[a-zA-Z]/g,function(c){return String.fromCharCode((c<="Z"?90:122)>=(c=c.charCodeAt(0)+13)?c:c-26);}));</script>
Since this solution is not mentioned anywhere, but works well for me:
I do this:
create a mailto link with fake email. I like admin#localhost.localdomain for obvious reasons: Spammer might spam his own botnet when using this address unchecked.
cypher real email address and put it in an unrelated but findable hidden span or whatever element you like. Obviously to obfuscate the email and hide it from the harvester. Depending on your project structure, you might even want to put it in a JS or Session variable.
create a click handler for these links after a second that decyphers and write the correct email address into the fake mailto link not preventing defaults.
I do not think that crawlers click on mailto links, but if they would, they probaby won't wait a second, while a human being would have to be extremly fast to click a link in the first second after pageload.
Now you have a fully functional but obfuscated, honeypoted and timesecured mailto link.
Working example php file:
<html>
<head>
<title>E-Mail Obfuscating</title>
</head>
<body>
<?php
$email = "example#email.org";
echo "<a class='emailLink' href='mailto:admin#localhost.localdomain' >Send me an e-mail!</a>"
."<span style='display:none' data-hash='" . base64_encode($email) . "' />";
?>
<script>
<!--
var emailLinks = document.getElementsByClassName("emailLink");
setTimeout(function() {
for(var i=0; i <emailLinks.length; ++i){
emailLinks[i].addEventListener("click", function(){
let encodedEmail = this.nextSibling.getAttribute('data-hash');
let decodedEmail = atob(encodedEmail);
this.href = "mailto:" + decodedEmail;
this.text = decodedEmail;
});
}
}, 1000);
-->
</script>
</body>
</html>
May the code be with you.
What if creating a link "Contact me" pointing to a directory protected by a password? Of course, you have to give the pass to access.
"Contact me" > ••••••••••• > contact/index.html
Once accessed, the contact/index.html page reveals the email, a mailto for instance.
My solution is to rearrange the characters using css and replacing the element on hover. No change is visible to the user.
const obscureHoverReverseMailTo = input => `<span style="display: inline-flex; color: rgb(0, 0, 238); cursor: pointer; text-decoration: underline;" onmouseover="const newContent = [...this.children].sort((a, b) => a.style.order - b.style.order).map(el => el.innerText).join('');this.outerHTML = \`<a href='mailto: \${newContent}'>\${newContent}</a>\`">${input.split("").map((char, index) => `<span style="order: ${index}">${char}</span>`).sort(() => 0.5 - Math.random()).join("")}</span>`;
const obscureHoverReverseMailTo = input => `<span style="display: inline-flex; color: rgb(0, 0, 238); cursor: pointer; text-decoration: underline;" onmouseover="const newContent = [...this.children].sort((a, b) => a.style.order - b.style.order).map(el => el.innerText).join('');this.outerHTML = \`<a href='mailto: \${newContent}'>\${newContent}</a>\`">${input.split("").map((char, index) => `<span style="order: ${index}">${char}</span>`).sort(() => 0.5 - Math.random()).join("")}</span>`;
document.getElementById("testRoot").innerHTML = obscureHoverReverseMailTo("hello#example.com")
<div id="testRoot"></div>
<input type="text" onkeyup="document.getElementById('testOut').innerHTML = obscureHoverReverseMailTo(this.value)">
<div id="testOut"></div>
here is the function if you have something else to hide:
const obscureHoverReverse = input => `<span style="display: inline-flex" onmouseover="this.outerHTML = [...this.children].sort((a, b) => a.style.order - b.style.order).map(el => el.innerText).join('')">${input.split("").map((char, index) => `<span style="order: ${index}">${char}</span>`).sort(() => 0.5 - Math.random()).join("")}</span>`;
Simple JavaScript-unaware bots typically look for mailto: and/or # in HTML page contents. Obfuscating these keywords will dramatically decrease chances of email address scraping.
A Base-64 encoded URL template mailto:%user%#%domain% can be employed:
function contact(user, domain = location.hostname) {
const template = atob('bWFpbHRvOiV1c2VyJUAlZG9tYWluJQ==');
location.href = template
.replace('%user%', user)
.replace('%domain%', domain);
return false;
}
Where 'bWFpbHRvOiV1c2VyJUAlZG9tYWluJQ==' is btoa('mailto:%user%#%domain%').
HTML links would need to be updated as follows:
e-mail me
Furthermore, javascript: addresses can be hidden from the users:
e-mail me
The return statements prevent the page navigation to the # anchor.
I use a PHP function to generate some javascript to output the email on page load. Note that you don't need PHP to generate the JS at runtime, you can generate the JS once locally and then include the static JS in your page.
You can also use the linked function with this snippet below to automatically obfuscate email addresses in some given HTML (where $processedContent is the HTML):
$emailMatches = array();
$matchCount = preg_match_all('/(?:[a-zA-Z0-9_\.\-])+\#(?:(?:[a-zA-Z0-9\-])+\.)+(?:[a-zA-Z0-9]{2,4})+/', $processedContent, $emailMatches);
if($matchCount > 0) {
$emailMatches = $emailMatches[0];
foreach($emailMatches as $email) {
$replacement = createJSMailLink($email);
$processedContent = str_replace($email, createJSMailLink($email), $processedContent);
}