Phish Detection Agency

Tools/Skills: Identifying Phishing Emails, DMARC, DKIM-Signature

SYNOPSIS

A phishing identifier isn't working perfectly, so I needed to go in and verify that phishing emails were correctly marked and that non-phishing emails were not mis-labeled.

By reviewing the return-paths, received domains, problems with the DKIM-Signature and when the DMARC failed, I correctly identified all emails as either phishing or not.

Recommendation: A program should be able to be made or updated to correctly capture the phishing emails I identified. If the program automatically labeled the following cases as phishing emails, it would have had 100% success (based on this example - further testing would be required): 1. A different return-path 2. A different received domain 3. Problems with the DKIM-Signature 4. Failing the DMARC.

SOLUTION & PROCESS

  1. Reviewed all emails

  2. Other techniques

    • I set up my search bar (command + f) to always looking for the domain “geeseislands”

      • It didn’t end up revealing any phishing attempts, but I did because this would be an easy typosquat to miss

    • I was prepared to also look for unusual dates and times, as well as content that didn’t seem to match the domain, but I didn’t end up needing to.

*I used a couple of ChatGPT prompts to help ensure I was on the right track.


Example Prompt:

  • What clues in this email could I look to decide if it's a phishing email or safe?

Email Details

From: laura.green@geeseislands.com

Reply-to: admin.security@geeseislands.com

Date: 2023-07-20 09:15:00

Subject: Security Protocol Briefing

Email Headers

Return-Path: <laura.green@unauthorized.com>

Received: from unauthorized.com

DKIM-Signature: v=1; a=rsa-sha256; d=unauthorized.com; s=default; b=HJgZP0lGJb8xK3t18YsOUpZ+YvgcCj2h3ZdCQF/TN0XQlWgZt4Ll3cEjy1O4Ed9BwFkN8XfOaKJbnN+lCzA8DyQ9PDPkT9PeZw2+JhQK1RmZdJlfg8aIlXvB2Jy2b2RQlKcY0a5+j/48edL9XkF2R8jTtKgZd9JbOOyD4EHD6uLX5;

DMARC: Pass

Last updated