All posts tagged tradecraft

This primer covers sending spoofed emails from an online service with a link to a clone credential harvesting site.  SET provides a clean, menu-driven interface for website cloning and automates the process. Using sendmail directly is also an option in SET; it requires a single change to the configuration and a mail relay to function correctly.

We will again use the Hackerone directory to identify a company but WILL NOT be sending phishing emails to them. This would be really bad form and potentially illegal. For this, we are going to pick on a known antivirus and security company, Kaspersky. Kaspersky was basically chosen because it is a large enough organization that we should be able to find a decent page to clone, and there should be enough email addresses in the wild to generate a list from a few different places.

Email Target List

The heart of any successful phishing campaign is the list of targets. Normally we would use recon-ng to build this list, but in this tutorial, we will do a few manual processes to show other methods. These can absolutely be automated, but for now lets do it the hard way.

From, I searched for “ -license”. This is because the top pastes were all license key dumps, and I was specifically looking for emails.

Pastebin Link 1 -
Pastebin Link 2 -

I threw the entire raw contents of the data in a text file, lets look in a few other places. We have already used theHarvester to find email addresses in the recon-ng tutorial, so let’s also use and put the results in the same text file.

root@kali:~/Desktop# theharvester -d -b all

We need to quickly pull out and de-duplicate the email addresses. This isn’t an issue with theHarvester but the pastebin data isn’t structured.

root@kali:~/Desktop# grep kasperskyRawEmails.txt > kasperskyEmail.txt
root@kali:~/Desktop# leafpad kasperskyEmail.txt
root@kali:~/Desktop# sort -d kasperskyEmail.txt > kasperskyEmailSorted.txt
root@kali:~/Desktop# sort -u kasperskyEmailSorted.txt > kasperskyEmail.txt

I used leafpad to clean up a few lines that had extra data. If there were more than a handful, I would have written a sed or awk script to clean it up, but with just a few, it was just as quick to do it manually. I sorted it with the -d option to put them in alphabetical order then -u to get rid of any duplicates. More importantly, by looking at the email addresses, we can easily guess anyone’s email in the company because they follow the A quick LinkedIn search of people who work at Kaspersky would provide us with another list. Personally, I would remove Eugene Kaspersky from the list since he is the founder, but hey, do what you want. In a normal penetration test, most companies would ask to approve the list and remove or add people as they see fit. For reporting reasons, running a wc -l against this list once approved will give you the number of emails. From these three searches, we have 142 unique targets.

While in the list, it is always good to look through and remove any email that is generic or does not direct to an actual person. Examples would be or Also and more importantly, remove anything that will get you caught. I hate to say this, but do not attempt to phish spam@ or abuse@ addresses; if it does work, you probably get extra h4x0r cred, though.

Baiting the Hook

To figure out who to spoof, I went on LinkedIn and searched for IT Support in the United States for Kaspersky. I’ve redacted the person’s name, but it is plausible that this person would send out an email about a website,  and since the email will be in English, that shouldn’t be a huge red flag. Sending out an email from an employee in the Russian Federation written in English should raise questions. Writing it in English and attempting to emulate a translation from Russian may work, as well.

Spoofed From Account

Spoofed From Account

Next, we need to identify a website to clone. We need a login on this page to spoof, so first we will use fierce to identify potential sub-domains.

root@kali:~/Desktop# fierce -dns

I looked through the fierce results and settled on for the page to be cloned. I like the simple page layout, so there are less chances for the clone to go wrong and log into a support page based on the title. VPN and Outlook Web Access (OWA) are normally my favorite pages, but occasionally, the clone needs to be massaged to make it look normal.

Next, let’s work on the Subject and Body of the email message.

Subject: Support Bot Login Page Test

All, We are testing the new login page for support bot at The old page had some certificate errors that prevented some users from reaching it so please let us know if this is still happening and also test that your password still works before the new systems goes live ( Let me know if you have any problems.


Spoofed IT Guy

Reading the email, we need to address a few things. First, is the certificate error that we get going to the real site? This is one of the many reasons why it is important to NOT train your users to click through certificate errors. Also, mixing a little bit of truth in with the lie helps to make it more effective.

Certificate Error

Certificate Error

Second, ask them to log into the page and contact us if there are any errors. Because we will be using a spoofed email and not a fake email address, this adds a little bit of risk that someone will respond and either confuse the IT staff or alert them to the phishing campaign that is underway. Spend some time thinking and crafting the email message to work for the company you’re testing.

We will use SET to clone the website and harvest credentials. The ***** indicates that I’ve redacted the screen text and have only shown the option to choose.

root@kali:/opt/social-engineer-toolkit# setoolkit
 Select from the menu:

1) Social-Engineering Attacks
 2) Penetration Testing (Fast-Track)
 3) Third Party Modules
 4) Update the Social-Engineer Toolkit
 5) Update SET configuration
 6) Help, Credits, and About

99) Exit the Social-Engineer Toolkit

set> 1
 Select from the menu:

1) Spear-Phishing Attack Vectors
 2) Website Attack Vectors
 3) Infectious Media Generator
 4) Create a Payload and Listener
 5) Mass Mailer Attack
 6) Arduino-Based Attack Vector
 7) Wireless Access Point Attack Vector
 8) QRCode Generator Attack Vector
 9) Powershell Attack Vectors
 10) SMS Spoofing Attack Vector
 11) Third Party Modules

99) Return back to the main menu.
 set> 2
 1) Java Applet Attack Method
 2) Metasploit Browser Exploit Method
 3) Credential Harvester Attack Method
 4) Tabnabbing Attack Method
 5) Web Jacking Attack Method
 6) Multi-Attack Web Method
 7) Full Screen Attack Method
 8) HTA Attack Method

99) Return to Main Menu

 1) Web Templates
 2) Site Cloner
 3) Custom Import

99) Return to Webattack Menu


[-] Credential harvester will allow you to utilize the clone capabilities within SET
 [-] to harvest credentials or parameters from a website as well as place them into a report
 [-] This option is used for what IP the server will POST to.
 [-] If you're using an external IP, use your external IP for this
 set:webattack> IP address for the POST back in Harvester/Tabnabbing: 

The address is either for the local host (in this case) or for the server that will be hosting the web servers external interface. (NOTE: We are currently watching this thread, if a fix gets posted for the OpenSSL/PEM file issue we will update this)

[-] SET supports both HTTP and HTTPS
 [-] Example:
 set:webattack> Enter the url to clone:

[*] Cloning the website:
 [*] This could take a little bit...
 Python OpenSSL wasn't detected or PEM file not found, note that SSL compatibility will be affected.
Cloned Site

Cloned Site

Real Site

Real Site

If you wanted to add SSL support to improve the quality of the attack, SET absolutely supports it with a few changes. And will allow you to get free SSL certificates. I’ve highlighted the differences in the preceding images.

[*] Printing error: zipimporter() argument 1 must be string, not function

The best way to use this attack is if username and password form
 fields are available. Regardless, this captures all POSTs on a website.
 [*] The Social-Engineer Toolkit Credential Harvester Attack
 [*] Credential Harvester is running on port 80
 [*] Information will be displayed to you as it arrives below: - - [27/Apr/2017 16:11:47] "GET / HTTP/1.1" 200 -

Now, we send out bait and see who we can catch.

 Set the Hook

We have used a couple different email spoofing services and have recently settled on They work well, and they don’t get blacklisted in mail servers. It is a paid server; though, if you want premium features such as removing the tag line at the bottom or sending SMS messages. Below is a screen showing how to send messages in Sharpmail; there is an argument to be made that putting a few of the addresses together in the To: line would make it more believable, but for now, they are all in the BCC: line.

Sharpmail Example

Sharpmail Example

Finally, we would send this email out and simply wait to see who got hooked.

SET has a nice live update when credentials are captured and also packages up a report at the end.

Live Output

Live Output

HTML Report

HTML Report

From here it would depend on what the client requests. Do you use these credentials to attempt further exploitation or just produce the report? Your testing might be done, or you could have potentially generated a bunch of additional work for yourself.


I’ve conducted phishing campaigns at many different companies. Overall, I probably have a 10% success rate. Some were a little higher and some a little lower. That doesn’t sound too impressive, right? How many successes does it take to compromise a network? One. One user clicking on a link in an email exposes the entire network. So, for most companies, I got significantly more than that one success. How did I do it? More importantly, what tricks do I have up my sleeve that other penetration testers could steal? At almost every conference you will see a talk on some super sweet post-exploitation tool or privilege escalation technique if you can talk to the speaker 9 times out of 10 they gained initial network access through phishing. Phishing is the dirty pen testing secret that we all do but nobody wants to talk about because it isn’t nearly as cool as remote code execution.

Finding Targets

Generally, there are two methods for generating lists for phishing campaigns: either the client will provide you a list (which is boring) or you can find valid targets and get the list approved. Where can you find valid targets? I consider a valid target any email address already exposed on the internet.

  • Recon-ng
  • Linkedin
  • Pastebin
  • Google
  • FOCA
  • Web content

In the future, we will look at each of these methods in depth, but for now, let’s just assume you have a list.

Sending Emails

The FROM line in the address is just as important as the TO. Are you sending a fairly generic phishing email hoping to get a few clicks? If so, your success rate is going to be fairly low. For a company without security awareness training in place, this might be appropriate, but most tests are meant to be more sophisticated. I am going to show you how I make the sausage; fair warning, it isn’t pretty.

If you are going for a generic attempt without a spoofed email address, you can try to get an email address from any of the normal providers like Gmail or Outlook. Registering an email that looks at least semi-plausible will help. Outlook has a built-in limitation for new accounts to restrict the number of emails sent until the account ages or milestones are met (such as phone number verification). Also, filling in the display name and information to seem legitimate will increase the chance of success. CompanyHelpDesk@gmail is better than PhishingAttempt6@yahoo.

If you are allowed to spoof email addresses, a few better options are available. Setting up sendmail and sending everything through Social-Engineering Toolkit (SET) is a great option. Using a webmail service that allows spoofed emails is also a great option and protects your fixed IP from being banned for email abuse. It is also smart to pay the small fee that allows the footer to be removed. If you are performing a penetration test, it is the cost of doing business. I personally like using Sharpmail out of the UK but have used a couple other servers, as well. Sharpmail has SMS functionality, which I have used on assessments in the past.

Everyone has seen poorly crafted phishing emails signed Help Desk, so you need to step up your game and do some research. Find the company on Linkedin, and figure out who the IT person is. Getting an email from a help desk address signed Gary when employees know there is an IT person named Gary is way more convincing. The correct tone is important too, as a busy help desk person sending a curt email stating, ‘We are testing a new web server for email, can you log in and test it? -Gary‘ is more believable than a two-paragraph, formal-sounding email. I rarely even hide my URL behind a link for the same reason; I wouldn’t do that as a systems administrator in a company and want them to be believable.

Some clients will also want you to get the email text and targets approved. I’ve had to add typos and dumb down my emails for clients who wanted to make it easier to be spotted. Those assessments are the best because you can almost guarantee success if even their employers think they are going to click on anything. Most of these assessments will come shortly after the company has been breached by a phishing attack.

Microsoft Outlook Web Access or a VPN login page are my two favorite sites to clone in SET. We will conduct a primer on SET soon, but for now, just know that I use the clone website function with the capture credentials module. I’ve used Browser Exploitation Framework (BeEF) in the past, but keeping it simple usually works better.

Now What?

The first time I was assigned a phishing campaign, I had no idea what to do. I fired up SET but didn’t have sendmail installed and configured. The client for that assessment wanted multiple tests done. Not only was it testing employee awareness, it was testing the email security appliance in place. Sendmail took me most of the day to get set up and start sending emails. Let’s just say that it did not go well; the appliance blocked all my spoofing attempts and having an included URL hidden behind link text tripped the heuristics, with the end result being the end users didn’t even get the attempts. Not only that, but because I worked from home, the IP I paid for from my ISP got blacklisted for sending spam.

What lessons did I learn? One, I rarely use my own sendmail account anymore. Two, I’ve gotten simpler in my messages. Three, I respond to replies. What? That’s right, when your login fails on the credential harvesting site I’ve created and you reply to the email complaining, I’ll tell you I’m working on it and that I will let you know when it’s fixed. Why? So that you don’t tell other people you’re having a problem and potentially prevent them from giving me their credentials. Sneaky right?

Reporting on phishing is simple; we normally produce a statistics-based report that shows how many credentials were gathered versus the number of emails sent. We avoid giving specific names, which clients always want, because it is normally a systemic issue, not a user issue. We have performed custom redirects, after credential harvesting, to a site that forces users to complete a short training on phishing awareness.

Pulling it All Together

Now that you have read this tradecraft on phishing, you may be asking, “what are the next steps?” Next, we are going to create some primers on setting up phishing campaigns using sendmail and Sharpmail and using SET to clone a website and harvest credentials. This simply gave you a glimpse into the mindset of how we think about attacks and some of the pitfalls encountered.