<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Amarachi Obiajulu on Medium]]></title>
        <description><![CDATA[Stories by Amarachi Obiajulu on Medium]]></description>
        <link>https://medium.com/@amarajulia75?source=rss-b0603482b85d------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*qH8oy1rtb9zeZD9NPHCbQw.jpeg</url>
            <title>Stories by Amarachi Obiajulu on Medium</title>
            <link>https://medium.com/@amarajulia75?source=rss-b0603482b85d------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 01:36:25 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@amarajulia75/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[A Beginner’s Guide To Investigating Suspicious URL]]></title>
            <link>https://medium.com/@amarajulia75/a-beginners-guide-to-investigating-suspicious-url-2a1614ff387c?source=rss-b0603482b85d------2</link>
            <guid isPermaLink="false">https://medium.com/p/2a1614ff387c</guid>
            <category><![CDATA[smishing]]></category>
            <category><![CDATA[phishing-awareness]]></category>
            <category><![CDATA[virustotal]]></category>
            <category><![CDATA[url-analysis]]></category>
            <category><![CDATA[cybersecurity-awareness]]></category>
            <dc:creator><![CDATA[Amarachi Obiajulu]]></dc:creator>
            <pubDate>Mon, 26 Oct 2026 07:22:08 GMT</pubDate>
            <atom:updated>2026-10-26T07:22:08.988Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*X3edK80XNRk1tC2gFm06Dw.png" /></figure><p>Cybercriminals still use a lot of deceptive links (phishing) via SMS, WhatsApp, and email to steal credentials and money from people. There has been a widespread of these links across social media lately. An average internet user often lack the technical knowledge to verify these links before clicking.</p><p>These article is a step by step guide and simple ways to verify the authenticity of any URL before clicking on them no matter how urgent the message being passed across seem to be.</p><p>Regardless of your background in IT, If you follow through with this, you will be able to learn how to spot red flags, safely inspect links, use free verification tools, and make safe decisions.</p><p>Here are some Social Engineering tricks to look out for to ensure you’re not about to get scammed.</p><ul><li><strong>A Manufactured Urgency:</strong> Be highly suspicious of messages that manufacture a crisis or sense of urgency (e.g., “final notice,” “your account will be suspended”)</li><li><strong>Requests for Verification Codes:</strong> No legitimate bank, IT department, or service will ever contact you and ask you to read out a one-time passcode or Multi-Factor Authentication (MFA) code. If someone asks for this code, they are actively trying to break into your account.</li><li><strong>Domain Spoofing:</strong> Always expand the sender’s details to view the actual domain, not just the display name.</li><li><strong>Out-of-Character Requests from Known Contacts:</strong> If a familiar contact or vendor suddenly asks you to do something they have never requested before, treat it as a red flag. Their account may have been compromised.</li></ul><p><strong>Case Study 1</strong></p><p>This first case study was sent to a multiple users impersonating UBA. Imagine receiving a text message saying your UBA account has an issue with account transactions and you have to click this link to verify your identity. The urgency is your first clue to pause.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/505/1*Eao-Vj3HH25aTwYa60bsNA.png" /></figure><p>URL- <a href="https://ubagroupid.com/">hxxps://ubagroupid[.]com/</a></p><p>The first tool I would be recommending for this analysis will be VirusTotal (<a href="https://www.virustotal.com/gui/home/upload">https://www.virustotal.com/gui/home/upload</a>) VirusTotal is a URL reputation checker. It asks dozens of different antivirus engines and security databases if a link is known to be malicious.</p><p>Defang the URL (hxxps://ubagroupid[.]com/[.]nigeria/) or Long-press (press and hold) the link until a menu pops up. <strong>Do not tap it quickly.</strong> Copy the link and paste on VirusTotal.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_S4wPUuE8EOxlr5ydjONqg.png" /></figure><p>As at the time I checked this URL on VirusTotal, security vendors haven’t flagged this URL as malicious. From here you can also see the IP address (8.222.190.93) upon analyzing the IP address on VirusTotal, you can see also that a security vendor (alphaMountain.ai ) has flagged the IP address as Suspicious.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*THag15ZTzF1Q6yA-v5Bcwg.png" /></figure><p>To get more details of the domain, I will be using whois (<a href="https://www.whois.com/">https://www.whois.com/</a>) to check for their registration details.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/815/1*lhZerbICpKeliVOLQIQoVg.png" /></figure><p>From the domain registration Information, It’s very obvious this domain was created for the sole aim of exploiting people. The Domain was registered on 19th October, 2026 while the phishing SMS was sent the next day being 20th October, 2026.</p><p>Another tool we will be using for this URL Analysis is<strong> URLScan.io </strong>(urlscan.io) This tool works as a behavioral analyzer and sandbox. It visits the website for you and takes a screenshot, allowing you to see what the site looks like without actually putting your device at risk.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rvRfxBHvIlYNXGQLA7hMvA.png" /></figure><p>URLScan.io also provided more details about the domain. By the left is a screenshot of what the site looks like when you click on the link. This helps with verifying the site activities without having to click on the link.</p><p><strong>Url2Png</strong> (<a href="https://www.url2png.com/">https://www.url2png.com/</a>) is a powerful tool for capturing snapshots of any website, right in your app, quickly and reliably.</p><p>From the screenshot below the site just loads endlessly while displaying a white background but it will most likely have a different effect when someone with an actual account with UBA clicks on the link.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Y0KMuVDZ-2shTyo4JERRpw.png" /></figure><p><strong>Case Study 2: The Fake FRSC Service Link</strong></p><p><strong>URL:</strong> hxxps://www[.]fctevreg[.]cam/z</p><p>The threat actors tried to impersonate The Federal Road Safety Corps (FRSC) which is Nigeria’s lead agency for road traffic administration, safety management, and highway law enforcement. One of the major red flag on this URL is the unusual .cam extension. legitimate government sites in Nigeria use .gov.ng.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/841/1*aVYCOkCkjq4mPHOTR3wQNA.jpeg" /></figure><p>Upon submitting the URL to VirusTotal for analysis the link, the result showed that about 5 different vendors have flagged the link as both phishing and a malicious link.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WuGPz7zCzO-1xvYlamJY5w.png" /></figure><p>The IP Address 43.165.173.182</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LV4saxwAxwBUi-wkHwHiBw.png" /></figure><p>Also, used Whois to get more details of the domain. The domain registration details is also a red flag, the domain was registered few weeks before the message was sent across to people.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/831/1*LpGRGciVmqyqgpfQ2YkiqQ.png" /></figure><p>In Conclusion, based on the URL’s we investigated, always look for typos or extra words (e.g., ubagroupid instead of ubagroup). Look at the ending. Is a Nigerian government service using .cam instead of .gov.ng Pausing to spot the red flags, inspecting the URL for subtle misspellings, and verifying links with some free tool I shared in the walkthrough are some of the recommended best practices for staying protected online. Stay alert, trust your instincts, and think before you link.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2a1614ff387c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Configuring GoPhish for Phishing Simulation and Awareness Training]]></title>
            <link>https://medium.com/@amarajulia75/configuring-gophish-for-phishing-simulation-and-awareness-training-700e44b2909e?source=rss-b0603482b85d------2</link>
            <guid isPermaLink="false">https://medium.com/p/700e44b2909e</guid>
            <category><![CDATA[phishing-awareness]]></category>
            <category><![CDATA[gophish]]></category>
            <category><![CDATA[soc-analyst-training]]></category>
            <category><![CDATA[cybersecurity-awareness]]></category>
            <dc:creator><![CDATA[Amarachi Obiajulu]]></dc:creator>
            <pubDate>Mon, 12 Oct 2026 17:19:48 GMT</pubDate>
            <atom:updated>2026-10-14T10:28:31.664Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/711/1*GUH94QTqotniHieNnESKvw.png" /></figure><p>Majority of data breaches still bypass firewalls by targeting the human element, most commonly through phishing. October is Cybersecurity awareness month and this project is to provide a walkthrough for security analysts who wants to perform phishing simulation campaing to strenghthen their security posture.</p><p>In the course of this project, I will be deploying GoPhish which is an open-source phishing simulation framework that helps security team conduct phishing campaigns thereby transforming abstract security policies into measurable, real-world readiness.</p><p>While traditional awareness training relies on passive compliance videos, simulated campaigns actively test your organization’s resilience against the exact tactics adversaries use for intrution.</p><h3>Project Requirements</h3><p>To Start and complete this project we need the following:</p><ul><li>A Windows computer.</li><li>A Gmail account with 2FA enabled.</li><li>Administrator access on your windows machine.</li></ul><h3>Phase 1: Installing and downloading GoPhish</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Z2n4xhYqAvPjClY4oYvnYw.png" /></figure><p>From the github asset select the windows zip file ( gophish-v0.12.1-windows-64bit.zip)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KWmQvlPjyW_wpYgwlyfr2w.png" /></figure><p>Once the download is completed, locate the zip file on the download section and extract all afterwards.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/615/1*a6AKrndgAL1YRdhhY97AGA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/908/1*DoAfhKZObtpRu7EiVudAGg.png" /></figure><h3>Phase 2: GoPhish Configuration</h3><p>Find to the extracted folder and Right-click the gophish.exe file and choose <strong>Run as administrator</strong></p><p>If Windows asks whether to allow the application through the firewall, click Allow.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/994/1*TC9g82FrScOe_KV_nmmH3g.png" /></figure><p>After Gophish installation has completed successfully, the terminal window will display the default administrator credentials:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/965/1*kkzP5Thr4C82dy2kvbOexg.png" /></figure><ul><li>Username: admin.</li><li>Password: A 16-character generated password displayed in the terminal. There’s an option to change it when you get to the user interface.</li><li>Gophish Server URL: <a href="https://127.0.0.1:3333/">https://127.0.0.1:3333</a>.</li></ul><p>To interface with GoPhish, Open your web browser and navigate to the gophish Server using the server URL <a href="https://127.0.0.1:3333/">https://127.0.0.1:3333</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/691/1*SAR3f21_0q9yMKbb5THj7Q.png" /></figure><p>Click on proceed and change the password after the first login.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_jnD6DkiiJs_xp3v4KMidA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7Zp5E2wgAXD0MYnWZydoBA.png" /></figure><h3>Phase 3: Setting up the Sending Profile</h3><p>This <strong>profile</strong> stores the specific SMTP details required for GoPhish to deliver the campaign emails. To Set this up: Go to [Sending Profiles] → [New Profile].</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/937/1*UYJQDBDhVgQW36jjvI87Aw.png" /></figure><p>To setup the sending profile, we will need to provide the following information.</p><p><strong>Name:</strong> An internal label used to identify this specific configuration within your GoPhish dashboard (e.g. “Phishing Simulation”).</p><p><strong>Interface Type:</strong> The protocol used to send the emails, which is set to “SMTP” by default.</p><p><strong>SMTP From:</strong> The email address that will appear to the recipients in the “From” field of their email client.</p><p><strong>Host:</strong> The address and port of your outgoing mail server. smtp.gmail.com:465.</p><p><strong>Username:</strong> The login email or credential required by the SMTP server to authenticate your session.</p><p><strong>Password:</strong> The corresponding password for the SMTP server. For Gmail and many modern providers, this must be a specific “App Password” generated from the account’s security settings, not the standard login password.</p><p><strong>Ignore Certificate Errors:</strong> A checkbox that, when selected, forces GoPhish to bypass strict SSL/TLS certificate validation during the SMTP handshake. This is often necessary when testing or using self-signed certificates.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/631/1*9HeGg6eU8vHJAXSLApWEsg.png" /></figure><p>The next step here is to test the sending configuration we just setup.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/694/1*QwKrdwcs4MvIE0L_ZnLb_Q.png" /></figure><p>To do that: Click on Send Test Email.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/709/1*q5SEew6A5jDMLPMjDbYWQQ.png" /></figure><p>Add the required details and click Send</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/701/1*ogB4uc9o-XUYR1QlSwcZ7A.png" /></figure><p>After that, check your inbox to be sure you received the email and the SMTP configuration is working.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/470/1*Ooa560i4EaXkHGUdZ3HzPA.png" /></figure><p>Click on the Save profile option to save the email.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/708/1*QX_gTSWWZhrwqwaAMrlxrw.png" /></figure><h3>Phase 4: Setting up the Landing Page</h3><p>A <strong>landing page</strong> is a standalone web page created specifically for a targeted the phishing campaign that a user arrives at after clicking a link in the email.</p><p>To Set this up for the purpose of this project. Click on the landing page option by the left and then click on add new page.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/807/1*fKBpbzWVU1iHiqQfV2QJkg.png" /></figure><p>Fill in the following details,</p><ul><li><strong>Name:</strong> An internal label used to identify and select this specific landing page when configuring a campaign (e.g., “Fake Facebook Simulation”).</li><li><strong>Import Site:</strong> A feature that allows you to automatically clone the HTML markup of an existing website by providing its web address.</li></ul><p><em>Note:</em> Imported sites often require manual HTML/CSS adjustments to ensure relative asset paths (images, stylesheets) and form actions display and function correctly.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/696/1*rn5NSibKZO6wFqRh5RSrQQ.png" /></figure><ul><li><strong>HTML / Code Editor:</strong> The editor window where you build, paste, or refine the raw HTML, CSS, and JavaScript for the landing page. By Clicking the <strong>Source</strong> button toggles between raw HTML code view and rendered preview.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/701/1*SrBtVwiI6zcDJnIUeJvd8A.png" /></figure><p>After that Enable Data capture by selecting <strong>Capture Submitted Data</strong> and <strong>Capture Passwords</strong> so all form submissions are recorded in the campaign results.</p><p><strong>Set Post-Submission Destination:</strong> In the <strong>Redirect To</strong> box, provide the landing URL where participants will be forwarded after submitting (for example: [https://www.facebook.com/login]</p><p><strong>Then Save Configuration:</strong> Click <strong>Save Page</strong> to complete the setup.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/1*BYhVaQuKFM7QKI9_PwFWNA.png" /></figure><p>After clicking on the Save option below, then you’ve successfully set up a landing page for the simulation exercise.</p><h3>Phase 5: Configuring the Email Template</h3><p>An <strong>email template</strong> in GoPhish is the pre-designed blueprint for the phishing content sent to target users. It contains the email’s subject line, visual formatting (HTML/CSS), text copy, and dynamic placeholders that personalize the message and embed tracking functionality.</p><p>Click on the Email Template and select New Template</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/912/1*bNFfNoeBhLQf5m3OydtmMQ.png" /></figure><p>Then fill in the required details for the setup.</p><ul><li><strong>Name:</strong> A unique internal label used to identify and select the template when launching a campaign (e.g., Fake <em>Facebook Simulation</em>).</li><li><strong>Subject:</strong> The exact subject line that appears in the recipient’s inbox (e.g., <em>Urgent: Unauthorized access detected on your Facebook</em>).</li><li><strong>Envelope Sender:</strong> Specifies the Return-Path header address for bounce processing. Type in the Sender’s name and email address the recipient will on their end. (e.g. Facebook&lt;security@mail.facebook.com&gt;)</li><li><strong>HTML Editor:</strong> The main code or visual editor where you build the email structure using HTML, CSS, inline images, and styled buttons to match authentic organizational or third-party communications.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/702/1*zuDSbZw2541GX-FmScZfbQ.png" /></figure><p><strong>Click on Add Tracking Image</strong> to track and see when the email is opened. Click on the Save Template and the Template will be added successfully.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sOlvBzcPAQHA8BODIinhbw.png" /></figure><h3>Phase 6: Users &amp; Group Setup</h3><p>In GoPhish, <strong>Users &amp; Groups</strong> represent the target audience for your simulation. A “Group” is essentially the mailing list or distribution list containing the details of the individuals who will receive the simulated email.</p><p>To create one, click on the Users &amp; Groups option and then click on New Group.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/878/1*tejXBfUDKulxogheDK4IMA.png" /></figure><p>Provide the needed information</p><ul><li><strong>Group Name:</strong> A descriptive internal label to identify the target audience (e.g., <em>IT Department</em>, <em>New Hires</em>, <em>Q4 Security Test Group or Fake facebook Simulation Group</em>).</li><li><strong>First Name:</strong> The recipient’s first name (e.g., Amarachi) Capturing this allows you to use personalization tags (like {{.FirstName}}) in your email template to make the simulation more realistic.</li><li><strong>Last Name:</strong> The recipient’s last name (e.g., Julia)</li><li><strong>Email:</strong> The recipient’s email (e.g. <a href="/service/mailto:amarajulia22@gmail.com">amarajulia@gmail.com</a>)</li><li><strong>Position:</strong> The recipient’s position (e.g., CTO) Including this is highly recommended, as it allows you to filter your final campaign dashboard to see which departments or roles are most susceptible to phishing</li></ul><p>Click on Add and Save Changes afterwards</p><p>If you’re not importing a CSV, you can add targets one by one using the following fields</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/701/1*ZBUCUJ8P4q9ssOzQ_i5O-A.png" /></figure><p><strong>For Bulk Import (CSV Upload):</strong> The feature allows administrators to upload a large list of targets all at once using a CSV (Comma-Separated Values) file. The CSV file must include specific column headers for GoPhish to parse it correctly: First Name, Last Name, Email, and Position.</p><h3>Phase 7: Launching a Campaign</h3><p>The <strong>Campaigns</strong> tab is where the simulation is officially assembled and launched. By selecting your prepared Email Template, Landing Page, Sending Profile, and Target Group, you create the final operation.</p><p>Click on Campaigns and choose add New Campaign</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/882/1*LDf-hogOqHBDMBrNFe2k8Q.png" /></figure><ul><li><strong>Name:</strong> A clear descriptive name for the campaign to help identify it in the dashboard later (e.g.<em>October Cybersecurity Awareness — Facebook Scenario</em>).</li><li><strong>Email Template:</strong> A dropdown menu where you select the pre-configured simulation email (e.g., your Facebook Security Alert scenario) that will be sent to the targets.</li><li><strong>Landing Page:</strong> A dropdown menu to select the destination page targets will see if they click the link in the email.</li><li><strong>URL:</strong> The web address that will replace the {{.URL}} variable in your email template. This must be the IP address or domain name where your GoPhish server is reachable by the targets (e.g., [http://192.168.2.19]</li><li><strong>Launch Date:</strong> Allows you to schedule the campaign to start immediately or at a specific date and time in the future.</li><li><strong>Sending Profile:</strong> A dropdown to select the SMTP configuration you created earlier (which handles the actual email delivery).</li><li><strong>Groups:</strong> A dropdown to select the specific target audience (the Users &amp; Groups list) that will receive the simulation.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/696/1*SDluBS6JpJkzrqlbb_Dhmg.png" /></figure><p>After the setup there’s an option to test the campaign by first sending it a another Gmail account you have access to, before launching the campaign. This allows you to be sure everything is working just fine.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/688/1*TK3lADRnEDVqrMPIn03U0w.png" /></figure><p>After that, confirm you got the email and everything is set to be sent out.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/974/1*crNR6W7C0KO25GFLMnH7MQ.png" /></figure><p>Here’s what the landing page looks like when you click on the link.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/712/1*SnaPd4Tx4044Ryrp2d-F2w.png" /></figure><p>After that go back to Campaign interface and click on Launch Campaign:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/531/1*eVaM8bKO2GanYrthDtELDQ.png" /></figure><p>Click on Launch and Select OK to Confirm the campaign has been scheduled for launch</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/543/1*3nygWy2L-WQOLr4mReMK9g.png" /></figure><h3>Phase 8: Track the Phishing Campaign</h3><p>From the dashboard you can get an overview of the result of the phishing campaign you launched. Get the details of the number of people that received the email, those that opened the email, those that clicked on the link as well as those that keyed in there credential on the site.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Lu8ObLEizL8BtPoPdQX3Dw.png" /></figure><p>To get more details of the person that clicked on the phishing link and submitted there details, click on the drop down close to the person name.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/959/1*3Il7Wdwi2KtU9QPad0R6yQ.png" /></figure><p>This simply shows how a phishing simulation can reveal the exact information an attacker aims to collect and how vulnerable certain information can be during a real phishing attack.</p><h3>In Conclusion</h3><p>Proactive security measures requires us to test our human defenses just as rigorously as our technical ones. Phishing simulations provide a safe, controlled environment for employees to experience real-world attacker tactics. By running these campaigns, we are building the muscle memory they need to recognize and report threats, ultimately transforming everyone on the team into our strongest line of defense.</p><p><strong>Disclaimer</strong></p><blockquote>The information presented in this guide is intended solely for educational purposes and authorized internal security testing. Executing phishing campaigns against any individual, network, or organization without explicit written permission is strictly prohibited and illegal. Unauthorized testing violates individual rights and corporate policies. Always operate with proper authorization.</blockquote><p>I hope you found this helpful as well.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=700e44b2909e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Phishing Triage & IoC Extraction.]]></title>
            <link>https://medium.com/@amarajulia75/phishing-triage-ioc-extraction-11c0d546f0dd?source=rss-b0603482b85d------2</link>
            <guid isPermaLink="false">https://medium.com/p/11c0d546f0dd</guid>
            <category><![CDATA[phishing-analysis]]></category>
            <category><![CDATA[security-operation-center]]></category>
            <category><![CDATA[ioc]]></category>
            <category><![CDATA[phishing-email]]></category>
            <category><![CDATA[malware-analysis]]></category>
            <dc:creator><![CDATA[Amarachi Obiajulu]]></dc:creator>
            <pubDate>Tue, 29 Sep 2026 10:12:56 GMT</pubDate>
            <atom:updated>2026-09-29T10:12:56.465Z</atom:updated>
            <content:encoded><![CDATA[<p>From Inbox to IoC: A Safe Methodology for Triage and Threat Extraction</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/697/1*elQXTSw8IVwGomcZEyOIAw.png" /></figure><p>This project simulates a real-world scenario for a Security Operations Center (SOC) Analyst. It presents a hands-on simulation of live phishing triage, how to safely analyze real-world email artifacts, document key Indicators of Compromise (IoCs), and defend an enterprise network.</p><h4>Requirements:</h4><p>A VirtualBox ( To create a disposable virtual machine so you don’t accidentally infect your actual computer).</p><p>Sample Dataset: GitHub repository rf-peixoto/phishing_pot (a frequently updated collection of real, defanged phishing .eml files caught by honeypots).</p><p>Email Client: Mozilla Thunderbird (for safely viewing .eml files offline without triggering tracking pixels).</p><p>Analysis Web Tools: MXToolbox (header parsing), <a href="http://urlscan.io/">URLScan.io</a> (safe link clicking), VirusTotal (malware checking), and CyberChef (decoding hidden text).</p><h3>Step 1: Setup a Sandbox environment.</h3><ol><li>Download and install VirtualBox.</li><li>Setup a Linux Virtual Machine (VM).</li><li>Take a “Snapshot” of the VM in its clean state. This allows you to instantly revert the machine to a safe state if you accidentally detonate a malicious payload during the project.</li></ol><p>To take a snapshot</p><ol><li>Focus the VM Window: Click onto your active Kali VM window.</li><li>Access the Menu: Click Machine on the top menu bar and Take Snapshot.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/527/1*6hYOhNFKJJGz4KfR84ypDQ.png" /></figure><ol><li>Click <strong>OK</strong>. Once completed, you will see your baseline snapshot listed under the machine’s history tree.</li></ol><h3>Step 2: Installing Mozilla Thunderbird</h3><p>sudo apt update &amp;&amp; sudo apt install thunderbird -y</p><ul><li>Launch Thunderbird by typing thunderbird &amp; in the terminal or finding it in the Kali Application Menu.</li><li>When Thunderbird opens, an <strong>“Account Setup”</strong> wizard will pop up asking for your name, email address, and password.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/576/1*r0Dvi1sr_h4_k8RHHUAFSA.png" /></figure><ul><li><strong>Do NOT sign in.</strong> Click <strong>Cancel</strong> at the bottom left.</li><li>Thunderbird will open to a blank home interface. We are using it strictly as a local .eml reader, so no email account is required.</li></ul><p><strong>Configure Security settings on thunderbird.</strong></p><p>This is to ensure that Thunderbird never automatically executes code or downloads tracking assets:</p><ol><li><strong>Disable Remote Content (Tracking Pixels):</strong></li></ol><ul><li>Click the Gear Icon (Settings) at the bottom left (or Go to Menu &gt; Settings).</li><li>Navigate to Privacy &amp; Security.</li><li>Scroll to the Web Content section and ensure “Allow remote content in messages” is UNCHECKED. <em>(This blocks external images/trackers from pinging the attacker’s server).</em></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/702/1*TtkmRQsAPCJlcdcEcl8ZuQ.png" /></figure><p><strong>2. Set Message Display to Plain Text / Simple HTML:</strong></p><ul><li>Go to the top menu bar (or press Alt to reveal it) and select <strong>View &gt; Message Body As</strong>.</li><li>Change the setting from <em>Original HTML</em> to <strong>Simple HTML</strong> or <strong>Plain Text</strong>. This prevents hidden HTML scripts from running when you preview an email.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/996/1*phV5w6Qp96AgFNu8umjH_A.png" /></figure><h3>Step 3: Getting samples of .eml file from an open repository on Github</h3><p>The phishing_pot repository consists of raw .eml files, which are fundamentally plain-text documents containing email headers, body text, and encoded strings. In their raw state, they are completely inert data. Again we are using sandbox environment so everything is trapped entirely inside Kali Linux Virtual Machine. Your actual host operating system remains completely isolated and protected.</p><p>Bash</p><pre>git clone https://github.com/rf-peixoto/phishing_pot.git</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/652/1*JzgDLzzjBUIF1V6hq8JdaA.png" /></figure><p>Change the network adapter to host- only adapter</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/826/1*BvT1MxF-zNdwgSdzwsOxfw.png" /></figure><h3>Step 4: Email Header Analysis</h3><p>cd to the directory that contains the phishing email and list out the content of the directory to see where the different email’s for investigations are housed.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/633/1*O3qzY7m1e4yrrou6Gza1mw.png" /></figure><p>First, move into the directory containing the actual .eml files and pick one to analyze.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/598/1*5d_dKh9o_Dyn85BXH2kpUg.png" /></figure><p>You can choose to do the analysis in two ways either using the command line or through GUI using Thunar</p><p>To use Command line to extract email header from one of the email. I will be using sample-1001.eml for the static analysis.</p><p>Bash</p><pre>grep -i -E &quot;^(From|To|Subject|Date|Received|Return-Path):&quot; sample-1001.eml</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/654/1*W2vhQSvBcPTezLR1lAXR5g.png" /></figure><p>To extract and Hash Attachments we will be using Thunar</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/410/1*pu-zDkHLKXaQl0uQMu6x-w.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/647/1*UZLKm8MO5T8EkQMLjGwdWw.png" /></figure><p>Open the .eml File in Thunderbird: Inside the graphical folder window (Thunar) that just opened:</p><p>Right-click on sample-1.eml</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8j1XIHIRLTSrLcOLVhSL6Q.png" /></figure><p>Hover over Open With, then select Thunderbird Mail (if it isn’t listed, click <em>Open With Other Application</em> and select <em>Thunderbird</em>). Thunderbird will launch, opening the email message safely in offline mode.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/718/1*mexgaoFtKEUzm6IsYsqhPg.png" /></figure><p>Open the view source option and collect the email header artifact in details.</p><p><strong>IOCs from the email header.</strong></p><p>Originating Sender IP: 137.184.34.4 (The true Public IP address that connected to the receiving gateway. It belongs to a DigitalOcean cloud server (SFO3 Region) This will be added to the blocklist)</p><p>Originating Hostname: ubuntu-s-1vcpu-1gb-35gb-intel-sfo3–06PNG ( Thus indicates the mail was sent from a compromised Linux virtual private server (VPS)</p><p>Originating sender ID: Postfix, from userid 0 PNG (This shows that this message was generated directly on the Linux server by the root user (userid 0) through the local Postfix mail transfer agent</p><p>Header Sender Domain: atendimento.com.br</p><p><strong>Spoofing Assessment:</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/642/1*P7V_hKxMaZT1BFGdUnBTZA.png" /></figure><p>The sender attempted to impersonate atendimento.com.br. The email failed authentication checks (compauth=fail) because it originated from an unauthenticated cloud server (137.184.34.4) with no valid DKIM signature.</p><h3>Step 5: URL And Attachment Analysis</h3><p>The next step is to analyze the url attached to the email.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9fDKGSznxPCchVaKT5Kugw.png" /></figure><ol><li>Copy Destination URL from Thunderbird: In Thunderbird, right-click the hyperlink or action button in the email body and select Copy Link Location.</li></ol><p>2. Defang the URL Text: Modify the URL string inside your text editor so it cannot accidentally be rendered or clicked as an active hyperlink.<br> Change http to hxxp and https to hxxps.<br> Replace dots with [.] or [dot].</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/647/1*JGKemai_9tQn_vnsDYf4-g.png" /></figure><p>3.Transfer the URL to Your Host System: Copy the defanged string from your Kali text editor and paste it onto your host operating system using VirtualBox Shared Clipboard (or type the clean domain name manually on your host machine).</p><p>4.Perform Passive Threat Intel Scans: From your host machine’s browser, analyze the URL using online analysis platforms: URLScan.io, VirusTotal, WHOIS Search</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/894/1*6A2LwbV_9VHzUL3osz-zuw.png" /></figure><p>Key Findings</p><p>The domain blog1seguimentmydomaine2bra.me has been active and involved in automated security submissions for at least one year. The automated dynamic scanners classified the landing page as benign/unknown (3/100), which is typical for static social engineering/credential harvesting sites.</p><h3>Step 6: Analyzing the second email with attachment.</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/706/1*sDXXpgc7DL0bIGpr3-Ln5A.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/643/1*R1O6UTkFztcaScr3QKxmCw.png" /></figure><p>Isolate and extract the payload.</p><pre> mkdir sample-5599-analysis<br> cd sample-5599-analysis<br></pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/630/1*sSzUpngPM98E5W9teyy-gQ.png" /></figure><p>Run a script to extract the file without executing.</p><pre>python3 -c &quot;<br> import email<br> with open(&#39;../sample-5599.eml&#39;, &#39;rb&#39;) as f:<br> msg = email.message_from_binary_file(f)<br> for part in msg.walk():<br> fn = part.get_filename()<br> if fn:<br> with open(fn, &#39;wb&#39;) as out:<br> out.write(part.get_payload(decode=True))<br> print(f&#39;Extracted: {fn}&#39;)<br> &quot;<br></pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/636/1*mt_v4U5LCnNU9FKlzLkclw.png" /></figure><p>To get the file type and file hash</p><pre>file lnvoice-Number002346234.docx<br>sha256sum lnvoice-Number002346234.docx<br></pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/627/1*BtaI2RZD_XbsFnunokLLSQ.png" /></figure><p>Run the file hash through VirusTotal</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kQTSlqZhoaQvcSYKNGMbSQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JT6VxIn_gEDutSWO7wiDiQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FpOI-MFOcHG1MEIoTWju8w.png" /></figure><p>Key Findings:</p><p>The attachment was named lnvoice-Number002346234.docx. The use of a lowercase &quot;L&quot; instead of an &quot;I&quot; in &quot;Invoice&quot; is a common evasion technique used to bypass basic string-matching filters and deceive end users.</p><ul><li><strong>Static Analysis:</strong> The base64-encoded payload was safely extracted using a Python MIME parsing script. The file utility confirmed it as a valid Microsoft Word OpenXML document (ZIP archive).</li><li><strong>Dynamic Analysis Review:</strong> Sandbox reports highlighted suspicious behavior, including WMI queries and outbound network connections initiated by the Office application.</li><li>Zero external malicious URLs or macros were found within the XML structure. The network traffic observed in the sandbox (metadata.templates.cdn.office.net, c.pki.goog) consisted entirely of benign Microsoft Office background processes, rendering the sandbox behavioral alerts false positives.</li></ul><h3>Step 7: Remediation and Recommendations</h3><ol><li>Perimeter Blocking: Immediately add the domain, blog1seguimentmydomaine2bra[.]me and the IP address 137.184.34.4 to the organization’s firewall, web proxy, and email gateway blocklists.</li><li>Conduct a retroactive search across the email environment for messages originating Sender IP’s or containing the malicious URL, and purge them from user inboxes.</li><li>Email Gateway Hardening: Ensure strict enforcement of SPF, DKIM, and DMARC policies to drop incoming emails that fail authentication, mitigating the sender spoofing technique observed in the first email.</li><li>Security Awareness: Distribute a brief notification to staff highlighting the visual spoofing technique (e.g., “lnvoice” vs “Invoice”) observed in the second email.</li></ol><p>Conclusion:</p><p>I hope this project serves as a guide to security operation analyst and enthusiast who are intrested in email security and investigating phishing emails.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=11c0d546f0dd" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Phishing Analysis Tools- A Tryhackme Walkthrough]]></title>
            <link>https://medium.com/@amarajulia75/phishing-analysis-tools-a-tryhackme-walkthrough-25493c4ce0f5?source=rss-b0603482b85d------2</link>
            <guid isPermaLink="false">https://medium.com/p/25493c4ce0f5</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[phishing-analysis]]></category>
            <category><![CDATA[malware-analysis]]></category>
            <category><![CDATA[phishtool]]></category>
            <category><![CDATA[securityoperationsanalyst]]></category>
            <dc:creator><![CDATA[Amarachi Obiajulu]]></dc:creator>
            <pubDate>Thu, 17 Sep 2026 15:32:26 GMT</pubDate>
            <atom:updated>2026-09-17T15:32:26.247Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/613/1*Uv0a9hKlTxhaZ8hh1TqJug.png" /></figure><p>In this phishing analysis walkthrough, you will be learning more about the tools that enable deeper analysis, then apply your skills by investigating three real-world phishing cases.</p><h3>Objectives</h3><ul><li>Understand what information should be collected during an email analysis</li><li>Learn about the tools that enable effective email investigation</li><li>Explore IP and URL analysis, including reputation lookups</li><li>Investigate phishing cases to apply your understanding and strengthen your analysis skills</li></ul><p><strong>Task 2 — Identifying Artifacts</strong></p><p>When analyzing a suspicious or malicious email, your first goal is to collect key artifacts that can help determine its legitimacy and intent. These artifacts provide the foundation for deeper investigation, such as reputation checks, threat intelligence lookups, and behavioral analysis.</p><h3>Header Artifacts</h3><p>When you begin your analysis with the email header, you should take note of the following points:</p><ul><li><strong>Sender email address:</strong> Where did the email originate from?</li><li><strong>Sender IP address:</strong> What is the source IP, and what does a reverse lookup reveal?</li><li><strong>Email subject line:</strong> Does it contain urgency or a call to action?</li><li><strong>Recipient email address:</strong> Who is the intended recipient (To/CC/BCC)?</li><li><strong>Reply-To email address:</strong> Where are responses being directed?</li><li><strong>Date and time:</strong> When was the email sent?</li></ul><h3>Body Analysis</h3><p>Next comes the email body, which has its own set of artifacts to be aware of:</p><ul><li><strong>URLs and hyperlinks:</strong> Identify all links and expand shortened URLs to reveal their true destination</li><li><strong>Attachment name(s):</strong> What files are included, and do their names or extensions appear suspicious</li><li><strong>Attachment hash:</strong> Generate a hash value for threat intelligence lookups</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/791/1*sGFawtJHqaErRJtiq3qjbg.png" /></figure><p><strong>Task 3: Email Header Analysis</strong></p><p>Some of the key information we need to collect can be viewed directly within an email or web client, such as Gmail or Yahoo!. However, other details, such as the sender’s IP address and Reply-To information, can only be obtained from the email header. In <a href="https://tryhackme.com/room/phishingemails1tryoe">Phishing Analysis Fundamentals</a>, we explored how to manually extract this data by analyzing the email’s source code. Now, we will look at tools that can help streamline and automate this process.</p><h3>Mail Header Analysis</h3><p><a href="https://toolbox.googleapps.com/apps/messageheader/analyzeheader">Messageheader</a>, part of the Google Admin Toolbox, helps analyze email headers. By simply pasting the full header into the tool, you can quickly extract key details such as the sender’s IP address, routing path, and potential misconfigurations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gMH_oJz-lOVMZUHaBpsQuQ.png" /></figure><p>Another great tool, <a href="https://mha.azurewebsites.net/">Message Header Analyzer</a>, can perform the same type of analysis. Check it out in the screenshot below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VLj7x-oHx8YPxlRx-UD5Vg.png" /></figure><h3>IP and URL Reputation Analysis</h3><p>When investigating suspected phishing emails, it is important to determine the origin of any identified IP addresses and assess the reputations of any associated IPs or URLs. These indicators can reveal whether the infrastructure is legitimate or tied to known malicious activity. Fortunately, several free tools are available that allow analysts to quickly perform these checks and gain valuable insight into potential threats.</p><p><a href="https://ipinfo.io/">IPinfo</a> is a simple and effective tool for gathering information about an IP address. By entering an IP, you can quickly view details such as its geographic location and associated organization. This helps analysts determine whether an IP is potentially malicious.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/913/1*u4yXRve4JzRzDAn7HTXJ2A.png" /></figure><p><a href="https://urlscan.io/">URLScan.io</a> is a tool that enables analysts to safely investigate websites without visiting them directly. By submitting a URL, the service simulates a real user browsing session and records all activity generated by the page. It also captures a screenshot of the site and provides insight into its behavior, helping analysts quickly identify suspicious or potentially malicious content.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/863/1*9IcsDlDNGLCEwmvxXrYwMQ.png" /></figure><p>Talos <a href="https://talosintelligence.com/reputation_center/">IP &amp; Domain Reputation Center </a>is a threat intelligence tool from Cisco that enables analysts to assess the reputation of IP addresses, domains, and networks. By submitting an indicator, you can quickly assess whether it has been associated with malicious activity and view its classification.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/831/1*mRxsvHJijWvAqWUiWEj4Bg.png" /></figure><p>Question 1: According to the Messageheader analysis shown in this task, what is the SPF result for the email?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/745/1*bRT9xqlNuth-6PoiDx9xaw.png" /></figure><p><strong>Answer: Pass</strong></p><p><strong>Task 4: Email Body Analysis</strong></p><p>Now it’s time to shift your focus to the email body, where the true intent of a phishing message is often revealed. This is typically where the malicious payload is delivered, either as a hyperlink designed to lure users to a phishing site or as an attachment intended to compromise the system. During analysis, links can be extracted manually from the visible email content or by examining the underlying HTML and raw source to uncover hidden or obfuscated URLs.</p><h3>Mail Body Analysis</h3><p>Determining the destination of a link within an email can be as easy as right-clicking the link and selecting Copy link address (the option&#39;s name may vary depending on your browser or email provider). From here, you can safely assess the URL without actually following it to the destination page.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*11qjZnw3JSONf1GAnsvRxw.png" /></figure><p>Another effective way to identify URLs in an email is to use a <a href="https://www.convertcsv.com/url-extractor.htm">URL extraction tool</a>. These tools let you paste raw email content and automatically parse all embedded links, saving time and reducing the risk of missing hidden or obfuscated URLs. Tools like <a href="https://gchq.github.io/CyberChef/#recipe=Extract_URLs(false,false,false)">CyberChef</a> can also perform this function, making them a versatile option for email analysis.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FHbDE0XiMGuOlraT4xSU5w.png" /></figure><p><strong>Email Attachments</strong></p><p>If the email you are investigating contains an attachment, it may be tempting to download it directly using your email client. However, attachments should only be downloaded in a controlled environment, such as a lab machine or sandbox, to reduce the risk of accidental execution. Once safely obtained, you can use tools like the sha256sum command in a Linux environment to generate a hash value for further analysis and reputation checks.</p><p>Obtain SHA256 Hash</p><pre>user@tryhackme$ sha256sum shady_attachment.pdf<br>025ba9ce4a2118a9ca7b115c8869ff73bc16bad3732ba359cef1e60ad8f961f9 shady_attachment.pdf</pre><p>Once you’ve obtained the hash value, several sites can safely assist with analyzing the attached file. Let’s return to Talos <a href="https://talosintelligence.com/reputation_center/">IP &amp; Domain Reputation Center</a> from the previous task and analyze the hash value from above. In this case, the file is labeled as <em>Phishing</em>, <em>Malicious</em>, and <em>Spam</em>, indicating that this file is likely harmful.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Yj-sLSueLFBlWFYA2xuinA.png" /></figure><p><a href="https://www.virustotal.com/gui/">VirusTotal</a> is a widely used tool that allows analysts to check the reputation of files, URLs, IP addresses, and domains using data from dozens of security vendors. By submitting a file or indicator, you can quickly determine whether it has been flagged as malicious and review detailed detection results. VirusTotal also allows users to upload files and submit URLs for analysis, making it a powerful resource for investigating suspicious artifacts.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Juv23drhzxpUy2jYeLumUg.png" /></figure><p>Question 1: What command can you use in a Linux environment to obtain the SHA256 hash value of an attachment?</p><p><strong>Answer: sha256sum</strong></p><p><strong>Task 5 : Malware Sandboxes</strong></p><p>Fortunately, as defenders, we don’t need advanced malware analysis skills to fully understand a malicious attachment. Instead, we can use online tools and services known as malware sandboxes, which allow files to be uploaded and analyzed in a controlled environment. This enables us to safely observe the behavior of a potentially malicious file without risking our own systems. For example, by uploading an attachment from a suspicious email, we can identify the URLs it attempts to contact, any additional payloads it downloads, and other indicators of compromise (IOCs).</p><p><strong>ANYRUN</strong></p><p><a href="https://app.any.run/">ANY.RUN</a> is an interactive malware sandbox that allows analysts to safely execute and observe suspicious files and URLs in real time. Unlike traditional sandboxes, it provides a hands-on experience in which you can interact with the environment, monitor processes, view network activity, and analyze system changes as they happen. This makes it a powerful tool for understanding how malware behaves and identifying key indicators of compromise in a controlled setting.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UHcRfdYba03SbMPSwJ4HXg.png" /></figure><p><strong>Hybrid Analysis</strong></p><p>Similar to the tool above, <a href="https://hybrid-analysis.com/">Hybrid Analysis</a> is a free malware analysis sandbox that allows analysts to upload and examine suspicious files in a controlled environment. It provides detailed insights into file behavior, including system changes, network activity, and indicators of compromise, helping analysts better understand the potential impact of a malicious file.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/680/1*VDCfcV8Pd3oE17ElCCzr7A.png" /></figure><p><strong>JOESandbox</strong></p><p>JOESecurity created the <a href="https://www.joesandbox.com/">JOESandbox</a> for advanced malware analysis. It performs both static and dynamic analysis on suspicious files and URLs. It generates comprehensive reports that highlight behavior, indicators of compromise, and threat classifications.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eIh-jkaBsiydSaaZTm0Osg.png" /></figure><p><strong>Task 6: Using PhishTool</strong></p><p>A powerful platform that streamlines phishing investigations is <a href="https://www.phishtool.com/">PhishTool</a>. It is designed for security professionals across multiple disciplines. PhishTool automates much of the manual work involved in analyzing suspicious emails. Whether you are a SOC analyst triaging user-reported messages, a threat intelligence analyst collecting indicators, or a researcher investigating phishing kits, PhishTool brings everything together in one place. By combining threat intelligence, OSINT, email metadata, and automated analysis workflows, PhishTool provides a centralized view of phishing activity, helping you quickly identify malicious intent and extract valuable indicators of compromise.</p><h3>Identifying Artifacts</h3><p>Upon uploading an email to PhishTool, you are presented with many of the valuable artifacts that we discussed in the second task. On the right-hand side, the tool provides the rendered HTML view, similar to what you’d see in your inbox, the raw HTML, and also the message source. Pretty useful!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lP34_jxl8F0i_sHmc5qtLQ.png" /></figure><h3>Further Analysis</h3><p>By navigating the tabs across the top of PhishTool, you can analyze key components of the email, including authentication results, transmission paths, and any embedded URLs. If the email contains attachments, they can also be reviewed and analyzed directly within the platform, giving you a comprehensive view of the message in one place.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bhhovOZS1bzy606NnRf3AA.png" /></figure><p>One of PhishTool’s standout features is its integration with VirusTotal, allowing you to view reputation and detection results without leaving the tool. This seamless integration helps speed up investigations by bringing external intelligence directly into your workflow</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Eb_jVILZQ-nYbQ7ChZj9qg.png" /></figure><h3>Resolving the Case</h3><p>Once your analysis is complete, PhishTool allows you to formally document your findings. As shown below, the email can be marked as malicious, with key artifacts, such as sender addresses, originating IP, and embedded URLs, to be flagged. You can then add investigation notes and click <strong>Resolve</strong>, mirroring the documentation and case closure process used in a real SOC environment.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/615/1*1xSPwA5MLK4rrYEUk7T9oQ.png" /></figure><p>Question 1 — According to the VirusTotal analysis from above, which vendor categorized the URLs as phishing?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cIHwV9gjL2d6PQQN_XuHMg.png" /></figure><p><strong>Answer: SafeToOpen</strong></p><p><strong>Task 7: Your Account is on Hold</strong></p><p>In this scenario, you are a level 1 SOC analyst. As the first line of defense, you have to identify phishing and spam emails reported by end users. In the following tasks, you will step into that role and analyze multiple suspicious emails that have been escalated by coworkers. Your objective is to carefully review each email, extract key indicators, and identify any signs of malicious activity. The findings you uncover will help your team develop detection rules to prevent similar threats from reaching other users in the future.</p><p>To begin, click the <strong>Start Lab Machine</strong> button to deploy the lab machine. Once the environment has loaded in the split-screen view, navigate to the desktop and open the file Phish3Case1.eml. From there, begin your investigation by reviewing the email header and body, and apply the analysis techniques covered throughout this room.</p><p><strong>Note</strong>: The domains, URLs, and IP addresses referenced in this task have hosted real malicious content and may still pose a risk. Do not access or interact with them outside of a controlled lab environment.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/925/1*aeo6f2kaT9NgFw21yd0DQQ.png" /></figure><p>Question 1: What reputable brand is this email tailored to impersonate?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/744/1*KDZKqx0zynvzs-KrAiIVkg.png" /></figure><p>Open the Phish3Case1-.eml using Moxila Thunderbird</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/745/1*rYoMOKrDsK8AQ1qsLywAuQ.png" /></figure><p><strong>Answer: Netflix</strong></p><p><strong>Question 2: Based on the email headers, who is the intended recipient of this message.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/745/1*IEf_yxKV8t1mc-oc5m-rLA.png" /></figure><p><strong>Answer: redacted@yahoo.com</strong></p><p>Question 3: In Thunderbird mail use View → Message Source.<br>What is the Received: from IP address?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/744/1*1-2zL63pxAlUi61mTyZOyA.png" /></figure><p><strong>Answer: 10.197.37.234</strong></p><p>Question 4: Check out the Return-Path field in the message source.<br>What would you consider a domain of interest based on this field?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/744/1*jtMQpt_tMWUiiZfF7LcwBA.png" /></figure><p><strong>Answer: etekno.xyz</strong></p><p>Question 5: Investigate the UPDATE ACCOUNT NOW button.<br>What is the shortened URL?</p><p>Hover around the attachment and see the shortened URL</p><p><strong>Answer: </strong><a href="https://t.co/yuxfZm8KPg?amp=1"><strong>https://t.co/yuxfZm8KPg?amp=1</strong></a></p><p><strong>Task 8: Update Payment Details</strong></p><p>In this task, you will investigate a malicious attachment using <strong>ANYRUN</strong>. This is a suspected phishing email masquerading as a Netflix payment update notification. Using the provided sandbox, review the file’s behavior, network activity, and any indicators of compromise. Pay close attention to how the attachment executes and interacts with the system.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*f7n8hmG6kOLYxqa7MjQWxg.png" /></figure><p>Question 1: How does ANYRUN classify this suspected phishing email?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5hakdA8xlv5hDprm5VfdMg.png" /></figure><p>Answer: Suspicious activity</p><p>Question 2: What is the name of the PDF attachment?</p><p><strong>Answer: Payment-updateid.pdf</strong></p><p>Question 3:Investigate the email attachment.<br>What is the SHA256 hash of the PDF file?</p><p><strong>Answer: cc6f1a04b10bcb168aeec8d870b97bd7c20fc161e8310b5bce1af8ed420e2c24</strong></p><p>Question 4: Check out the ANYRUN text report on the phishing email.<br>Which IP address associated with the process AcroRd32.exe is flagged as malicious?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/991/1*qD8fRi1N_nbsYvHyG75RNw.png" /></figure><p><strong>Answer: 2.16.107.24</strong></p><p>Question 5: Continue investigating the text report.<br>Which Windows process is classed as Potentially Bad Traffic?</p><p><strong>Answer: svchost.exe</strong></p><p><strong>Task 9: Excel Executable</strong></p><p>Question 1: How does ANYRUN classify the .xlsx attachment?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/617/1*C7F2Af-Dx60JyEVxxav34g.png" /></figure><p><strong>Answer: Malicious activity</strong></p><p>Question 2: What is the file name of the Excel attachment?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cMmi7EFGNtySFEQG33cc3Q.png" /></figure><p><strong>Answer: CBJ200620039539.xlsx</strong></p><p>Question 3: Investigate the Excel attachment.<br>What is the SHA256 hash value?</p><p><strong>Answer: 5f94a66e0ce78d17afc2dd27fc17b44b3ffc13ac5f42d3ad6a5dcfb36715f3eb</strong></p><p>Question 4: Check out the ANYRUN text report.<br>What IP address is associated with the malicious domain biz9holdings.com</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/770/1*tlg1poigYTXe_QJcl86YZA.png" /></figure><p><strong>Answer: 204.11.56.48</strong></p><p>Question 5: Which other domain is classified as malicious?</p><p><strong>Answer: findresults.site</strong></p><p>Question 6: What vulnerability does this malicious attachment attempt to exploit?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/615/1*_BI43T-27kz6x7GF6P6riQ.png" /></figure><p><strong>Answer:CVE-2017–11882</strong></p><p><strong>Conclusion:</strong></p><p>In this walkthrough, you’ve explored a structured approach to phishing analysis by learning what information should be collected during an investigation and how to effectively examine suspicious emails. And a range of tools that support email analysis and gained experience performing IP and URL reputation checks to validate potential threats. I hope this helps cybersecurity defenders and enthusiasts and strengthen your ability to analyze real-world phishing cases.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=25493c4ce0f5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Phishing Analysis Fundamentals Tryhackme Walkthrough.]]></title>
            <link>https://medium.com/@amarajulia75/phishing-analysis-fundamentals-tryhackme-walkthrough-e5ebc48280ca?source=rss-b0603482b85d------2</link>
            <guid isPermaLink="false">https://medium.com/p/e5ebc48280ca</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[soc-analyst]]></category>
            <category><![CDATA[phishing]]></category>
            <category><![CDATA[tryhackme-walkthrough]]></category>
            <category><![CDATA[phishing-analysis]]></category>
            <dc:creator><![CDATA[Amarachi Obiajulu]]></dc:creator>
            <pubDate>Mon, 14 Sep 2026 17:39:12 GMT</pubDate>
            <atom:updated>2026-09-14T17:39:12.772Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6HPDARLx0SbZGl-Wkb6K6Q.png" /></figure><p>Spam and Phishing remain the most common social engineering threats facing modern organizations. Phishing refers to an attempt to steal sensitive information, typically in the form of usernames, passwords, credit card numbers, bank account information, or other important data in order to utilize or sell the stolen information.</p><p>In this walkthrough, the objective involves</p><ul><li>Learning the basics of email delivery</li><li>Exploring email header analysis</li><li>Investigating and analyzing email bodies</li><li>Learning about the different types of phishing</li><li>Analyzing emails to identify potential security threats</li></ul><p>Every phishing email you’ll investigate starts with something simple: an email address. If you don’t understand how it’s structured, you may overlook the very clues attackers rely on.</p><p><strong>Task 2: The Email Address</strong></p><p><strong>ANATOMY OF AN EMAIL ADDRESS:</strong></p><ol><li>Username: User mailbox that identifies the specific recipient’s mailbox on the email server</li><li>@ symbol: Separates the username from the domain and tells the system where to route the email</li><li>Domain name: Specifies the mail server responsible for receiving the message.</li></ol><p>Question 1</p><p>Identify the domain used in the following email address.hatsalesman@tryhatme.com</p><p><strong>Answer: tryhatme.com</strong></p><p><strong>Task 3: EMAIL DELIVERY</strong></p><p>When you send an email, several protocols work together behind the scenes to deliver your message from sender to recipient, and each protocol has a specific role:</p><ul><li><strong>Simple Mail Transfer Protocol</strong> (SMTP): Sends emails</li><li><strong>Post Office Protocol</strong> (POP3): Downloads emails to a device</li><li><strong>Internet Message Access Protocol</strong> (IMAP): Syncs emails across devices.</li></ul><p>A Brief walkthrough of how email goes from the sender to the reciever.</p><ol><li>User sends an email: The sender’s email client sends the message to their mail server using SMTP</li><li>Mail server queries DNS: The sending server asks DNS for the recipient domain’s mail server</li><li>DNS responds: DNS returns the address of the recipient’s mail server</li><li>Email is delivered: The message is sent across the Internet to the recipient’s server</li><li>The recipient checks their mailbox: The recipient’s email client connects to their mail server</li><li>Email is retrieved: The message is downloaded (POP3) or synced (IMAP) to the recipient’s device</li></ol><p>Question 1: Which protocol is responsible for sending an email from a client to a mail server?</p><p><strong>Answer: SMTP (Simple Mail Transfer Protocol)</strong></p><p>Question 2: Which service is used to look up the recipient domain’s mail server ?</p><p><strong>Answer: DNS (Domain Name Server)</strong></p><p>Question 3: Bob wants to access his email from multiple devices, including his phone and laptop. Which protocol should he use?</p><p><strong>Answer: IMAP (Internet Message Access Protocol)</strong></p><p><strong>Task 4 : Email Headers</strong></p><p>An email consists of two main parts:</p><ul><li><strong>Email header</strong>: Contains metadata about the message, such as sender and the servers involved in delivery</li><li><strong>Email body</strong>: Contains the actual message content, which may be plain text or HTML</li></ul><p>The email header consist of the following.</p><ol><li><strong>From</strong>: The sender’s email address</li><li><strong>To</strong>: The receiver’s email address</li><li><strong>Reply to</strong>: Address where replies are sent (not required)</li><li><strong>Subject</strong>: The email’s subject line</li><li><strong>Date</strong>: The time and date that the email was sent</li></ol><p>Question 1: What is the full subject line of email1.eml?</p><p><strong>Answer: Help protect your budget by protecting your home</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/679/1*n6eompTMwOygiePgVfAHHA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/680/1*R_41a1VyEL7GskVZUrjV1Q.png" /></figure><p>Right Click and open with thunderbird mail.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/853/1*T3ihPMB6jjwvjfljA0aukQ.png" /></figure><p>Question 2: View the message source of email1.eml using Thunderbird in your VM.<br>What the IP address listed as the X-Originating-Ip?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/852/1*98F44wRfdOw4e87aTKdO9w.png" /></figure><p><strong>Answer: 43.255.56.161</strong></p><p><strong>Task 5: Email Body</strong></p><p>Emails are sent as text only or formatted in HTML. HTML supports elements such as images, links, and styling. Most email clients show you the rendered content. You can also inspect the underlying source to see how the message is structured, spot embedded elements, and look for signs of phishing or malicious content.</p><p><strong>Viewing HTML Source Code</strong></p><p>Just as we can inspect an email’s source to analyze its headers, we can also examine the source of the email body. This allows us to see the raw HTML that renders the message.</p><p><strong>Reconstructing Attachment</strong></p><p>Emails can also include attachments, such as documents, images, or other file types. Just as the email body can be analyzed by viewing the message source, attachments can be analyzed as well.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-KZG1C4tzalWvNW92fVWIw.png" /></figure><p>In the example below, you’ll see an HTML-formatted email containing a PDF attachment. The rendered view shows the attachment as it appears in the email client, while the source view shows how it is actually stored within the message. When we inspect the source, we can identify several important headers associated with the attachment:</p><ul><li><strong>Content-Type</strong> indicates the file type application/pdf</li><li><strong>Content-Disposition</strong> specifies that the file is an attachment and includes its filename</li><li><strong>Content-Transfer-Encoding</strong> shows that the file is base64 encoded</li></ul><p>The encoded base64 that follows represents the file itself. This data can be decoded to reconstruct the original attachment using either <a href="https://gchq.github.io/CyberChef/#recipe=From_Base64(&#39;A-Za-z0-9%2B/%3D&#39;,true,false)">CyberChef(opens in new tab)</a> or a base64 to PDF <a href="https://www.apivoid.com/tools/base64-to-pdf/">converter(opens in new tab)</a>.</p><p><strong>Question 1: Open up the </strong><strong>email2.txt file to view the source of an attachment.<br>What is the </strong><strong>Content-Type of the attachment?</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/745/1*XPhBXmr7kXiJI6k3n121-w.png" /></figure><p><strong>Answer: application/pdf</strong></p><p><strong>Question 2: What is the name of the attachment from the previous question?</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/745/1*esdFPOzjlRvfJUV8t10CLg.png" /></figure><p><strong>Answer: zmqpalgh.pdf</strong></p><p>Question 3: Decode the base64 string using either a <a href="https://www.apivoid.com/tools/base64-to-pdf/">PDF converter (opens in new tab)</a> or <a href="https://gchq.github.io/CyberChef/#recipe=From_Base64(&#39;A-Za-z0-9%2B/%3D&#39;,true,false)">CyberChef (opens in new tab)</a>.<br>What is the hidden flag value?</p><p>Open email sample on a note pad.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/747/1*FBDNayRUxthEZ1_Q5PjzCQ.png" /></figure><p>Copy the base64 strings to cyberchef to decode from Base64</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*al4-h5iOel0KhjST6dNJ_Q.png" /></figure><p>Check the saved output file for the Flag.</p><p><strong>Answer: THM{BENIGN_PDF_ATTACHMENT}</strong></p><p><strong>Task 6: Types of Phishing</strong></p><p>Different types of malicious emails can be classified as one of the following:</p><ul><li>Spam: Unsolicited bulk emails sent to a large number of recipients. A more malicious form of spam is often called malspam.</li><li>Phishing: Emails that impersonate a trusted entity to trick recipients into revealing sensitive information.</li><li>Spear Phishing: A targeted form of phishing aimed at a specific individual or organization, often using personalized information.</li><li>Whaling: A type of spear phishing that specifically targets high-level executives (CEO, CFO) to obtain sensitive data or financial access.</li><li>Smishing: Phishing attacks conducted via SMS or text messages, targeting users on mobile devices.</li><li>Vishing: Phishing attacks carried out through voice calls, where attackers use social engineering over the phone.</li></ul><h3>Anatomy of a Phishing Email</h3><p>When it comes to phishing, attackers often rely on similar techniques regardless of their goal. Their objective may be to harvest credentials, deliver malware, or gain unauthorized access to a system, but the methods used to trick the recipient are often the same.</p><p>Below are some common characteristics of phishing emails:</p><ul><li><strong>Spoofed From Address</strong>: The sender’s email is spoofed to appear as a trusted entity (noreply@microsof.com)</li><li><strong>Urgent Subject or Message</strong>: The email creates a sense of urgency (“Your account will be locked in 24 hours”)</li><li><strong>Brand Impersonation</strong>: The email is designed to mimic a legitimate organization (logos or colors matching a real company)</li><li><strong>Grammar &amp; Spelling Issues</strong>: The message may contain errors, though with AI these are now less common (awkward phrasing or unnatural wording)</li><li><strong>Generic Content</strong>: The message lacks personalization (“Dear Customer” instead of your name)</li><li><strong>Hidden or Shortened Links</strong>: Hyperlinks may disguise their true destination (bit.ly/secure-login)</li><li><strong>Malicious Attachments</strong>: Attachments are included and disguised as legitimate files (invoice.pdf.exe)</li></ul><p>Question 1: Which reputable organization is being spoofed in this phishing attempt?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/747/1*bKpTqzyh2-JVflnZvwaM0A.png" /></figure><p><strong>Answer: Home Depot</strong></p><p>Question 2: What is the sender’s email address?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/747/1*h33fKUPUOlklsnfYYCKKEw.png" /></figure><p><strong>Answer: </strong><a href="/service/mailto:support@teckbe.com"><strong>support@teckbe.com</strong></a></p><p>Question 3: Inspect the email message source.<br>What is the <a href="https://gchq.github.io/CyberChef/#recipe=Defang_IP_Addresses()">defanged (opens in new tab)</a> X-Originating-IP?</p><p>First get the X-Originating-IP from the source.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/641/1*pRfOr38sHneYrhAtUAZlsg.png" /></figure><p>Use Cyberchef to find the Defang IP Address</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/896/1*QNycQN55mGlugCN9U6iYVg.png" /></figure><p><strong>Answer: 103[.]234[.]236[.]83</strong></p><p>Question 4: Continue analyzing the email message source.<br>Which mail server generated the Authentication-Results header?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/639/1*xdA5cKQDdtnP2UU9LlEeYw.png" /></figure><p><strong>Answer: atlas102.free.mail.gq1.yahoo.com</strong></p><p>Task 7:</p><p><strong>Question 1: What attack, signified by the acronym BEC, uses a compromised email to trick employees into fraud?</strong></p><p><strong>Answer: Business Email Compromise (BEC)</strong></p><p><strong>Conclusion</strong></p><p>This walkthrough is meant to help cybersecurity defenders and enthusiast better understand the fundamentals of Phishing analysis and to develop the technical skills needed to investigate suspicious emails by extracting and analyzing both the header and body source code. I hope you found it as helpful as I did.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e5ebc48280ca" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Setting Up Wazuh and n8n for SOC Automation.]]></title>
            <link>https://medium.com/@amarajulia75/setting-up-wazuh-and-n8n-for-soc-automation-11fe04a18fa4?source=rss-b0603482b85d------2</link>
            <guid isPermaLink="false">https://medium.com/p/11fe04a18fa4</guid>
            <category><![CDATA[automation]]></category>
            <category><![CDATA[security-operation-center]]></category>
            <category><![CDATA[wazuh]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Amarachi Obiajulu]]></dc:creator>
            <pubDate>Wed, 02 Sep 2026 12:47:46 GMT</pubDate>
            <atom:updated>2026-09-18T11:25:11.524Z</atom:updated>
            <content:encoded><![CDATA[<p>Security Operations Centers (SOC) are no longer just defending against threat actors, it is increasingly facing automated, adaptive threats capable of operating at scale and speed. The recent incident with hugging face showed how an attacker leveraged malicious data inputs to exploit code execution vulnerabilities, steal credentials, and move laterally across systems with minimal resistance.</p><p>What makes this incident particularly significant for SOC teams is not just the entire attack process, but the failure in response. Despite detecting early reconnaissance activity, the security workflow failed to assign the appropriate severity to trigger timely human intervention. This gap allowed the threat to persist and establish command-and-control mechanisms.</p><p>Detection alone is not enough, rapid correlation, prioritization, and response must happen alongside.</p><p>In this project, I outlined a step by step process of setting up Wazuh and n8n and integrateing it to build an automated SOC workflow. The goal is to bridge the gap between detection and response ensuring that high-risk alerts are not only identified but acted upon immediately. This first part focuses on integrating Wazuh and n8n.</p><p><strong>Part 1 — WAZUH INSTALLATION</strong></p><p>To setup Wazuh on Linux using Docker your machine should have atleast 6GB RAM.</p><p><strong>Step 1: Updating Kali Linux Machine</strong></p><p>Started by updating my kali using the command; sudo apt-get update kali</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*U2vBHYlOgWvyMf7nEmCTWQ.png" /></figure><p><strong>Step 2: Installation of Docker and Docker compose</strong></p><p>I installed Docker to create an isolated, containerized environment for the Wazuh deployment. By packaging the SIEM components alongside their required dependencies, this helped to ensure a consistent setup and prevents any software conflicts with the Kali system it’s been hosted on .</p><p>Wazuh relies on a multi-tier architecture, specifically the Wazuh manager, indexer, and dashboard. I installed Docker Compose which helps to streamline the process by allowing all these interconnected services to be configured and launched simultaneously from a single file.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/614/1*gd391Jd4oPe0k_yZsUn4AA.png" /></figure><p><strong>Step 3: Starting and Enabling Docker</strong></p><p>After the installation process was successful, I used the command systemctl start docker to activate the service to activate the service. This helped to initialize the docker daemon in the background, which is required to pull images, create volumes and actively run the Wazuh containers right away.</p><p>Also, I enabled the docker using the command systemctl enable docker to configure the service to launch automatically whenever the Kali VM boots up.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/618/1*i744AKGjuL817dYYum7nmQ.png" /></figure><p><strong>Step 4: Installing and configuring wazuh Docker stack</strong></p><p>I deployed Wazuh Manager, Indexer, and Dashboard as containers using Docker Compose from the official <a href="https://documentation.wazuh.com/current/deployment-options/docker/wazuh-container.html#single-node-stack">Wazuh Docker Repository.</a> This will create a single-node Wazuh SIEM environment.</p><p>From the offical Wazuh Docker Repository, using the Docker compose, I deployed Wazuh Manager, Indexer, and Dashboard as containers</p><ul><li><strong>Wazuh Manager:</strong> This will serve as the central engine responsible for analyzing data received from endpoints, applying security rules, and triggering alerts.</li><li><strong>Wazuh Indexer:</strong> This component act as the highly scalable, full-text search and analytics engine that securely stores and indexes all generated security data.</li><li><strong>Wazuh Dashboard:</strong> This will serve as the web-based user interface for data visualization, threat hunting, and overall system management.</li></ul><p>Run: git clone <a href="https://github.com/wazuh/wazuh-docker.git">https://github.com/wazuh/wazuh-docker.git</a> -b v4.14.4</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/623/1*qmLBNi26wg4-UucDUl9LGg.png" /></figure><p>By orchestrating these containers together, I created a complete single-node Wazuh SIEM environment. In this setup, all three critical components operate collectively on my Kali VM. This approach is highly efficient for resource-constrained environments like a Virtual Machine, it delivers a fully functional, comprehensive security monitoring platform optimized for testing, learning, and local development without the heavy infrastructure overhead required by a multi-node, distributed cluster.</p><p><strong>Step 5:</strong></p><p>I navigated to the single-node directory,cd wazuh-docker/single-node which contains a Docker Compose file that automatically starts the Wazuh Manager, Indexer, and Dashboard as containers. This approach eliminates manual dependency installation and isolates all services in Docker containers for cleaner management.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/594/1*wjYZMvkAgLDwziaHzIo4dA.png" /></figure><p><strong>Step 6: Generate certificate using the command from official Wazuh website</strong></p><p>As part of the security configuration process, there’s a need to establish a secure foundation for the environment. Following the official Wazuh documentation, I generated the required TLS certificates by running a temporary utility container using the following command:<br> sudo docker compose -f generate-indexer-certs.yml run --rm generator</p><p>run: sudo docker compose -f generate-indexer-certs.yml run — rm generator</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/622/1*jrQz-pebeZXW9Ew4WqYpeQ.png" /></figure><p>I provisioned these certificates specifically because the Wazuh indexer mandates them to establish encrypted communication channels across the entire architecture (manager, indexer, and dashboard).</p><p>In any security-focused infrastructure, protecting data in transit is paramount. A SIEM handles highly sensitive network telemetry, vulnerability data, and security alerts. By generating and implementing these TLS certificates, I ensured that all internal traffic flowing between the Wazuh containers is fully encrypted. This prevents potential eavesdropping, man-in-the-middle attacks, or data interception within the Docker environment, maintaining the integrity and confidentiality of the security logs.</p><p>This generator also automatically creates several certificate files used to secure internal communication within the Wazuh stack.</p><p><strong>Step 7: Deploying Wazuh Containers</strong></p><p>I deployed Wazuh by running the command, sudo docker compose up -d. This command automates the entire provisioning process by instructing Docker Compose to parse the YAML configuration file, pull the specified official images from the repository, and build the isolated environments for the Wazuh Manager, Indexer, and Dashboard.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/612/1*jTwFtbtC9pktpTwyhThbKw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/601/1*HQhZ-tNBanaOGIv8LHSPzg.png" /></figure><p><strong>Step 8: Confirm the running processes in the docker</strong></p><p>sudo docker ps</p><p>After starting the deployment, I ran the sudo docker ps command to make sure everything went smoothly. This command gives a quick look at all the active containers currently running on the system. The output showed that the Wazuh Manager, Indexer, and Dashboard containers were all active and listed with a status of &quot;Up.&quot;</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/618/1*uqgtac9y_yTytGKk53Pq_w.png" /></figure><p><strong>Step 9: Accessing the Wazuh Dashboard.</strong></p><p>First of using the ifconfig, I got the docker IP which is where the Wazuh is being hosted on.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/563/1*iab3TWEl-y1FaCpDFSBTaQ.png" /></figure><p>To access wazuh dashboard on your web browser use <a href="https://docker/">https://docker</a> IP address</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/567/1*C5oahb0V2VfpcuCuqw5Rdw.png" /></figure><p>The default administrator credentials for authenticating the dashboard is,</p><p>Username: admin</p><p>Password: SecretPassword.</p><p>With the Wazuh dashboard activated now, I now have a centralized platform that gives me visibility into alerts, endpoint/agent activity, and indexed security events collected by the Wazuh platform.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wThkqOn6FhG_t3lNFuOAIw.png" /></figure><p><strong>Part 2:</strong></p><p><strong>Deploying Wazuh agent on Windows endpoint for Monitoring:</strong></p><p>Now that the Wazuh SIEM platform is fully operational on my Kali VM, it is ready to receive security data. The next phase involves installing the Wazuh agent on a Windows machine and connecting it to the server so it can begin collecting logs.</p><p>The Wazuh agent acts as a sensor on the endpoint device. It allows the central Wazuh server to monitor important activities on the Windows machine, which includes; User login attempts, changes to critical files or system settings, Phishing and malware detection events and General suspicious behavior.</p><p><strong>Step 1: Confirming the two Virtual Machines are on the same network</strong></p><p>It’s important to note that before installing the agent, ensure that the Windows VM could communicate with the Kali VM. To do this, I changed the network settings for both machines in VirtualBox to use a “Host-only Network.” This creates a private network that allows the two virtual machines to talk to each other without being connected to the outside internet.</p><p><strong>Step 2: Download and Install the Wazuh Agent</strong></p><p>I downloaded the official Windows Wazuh agent installer directly from the Wazuh package repository (<a href="https://documentation.wazuh.com/current/installation-guide/wazuh-agent/wazuh-agent-package-windows.html">https://documentation.wazuh.com/current/installation-guide/wazuh-agent/wazuh-agent-package-windows.html</a>)</p><p>I saved the file on my local Downloads folder in preparation for the deployment phase.</p><p>The installer is essential because it registers the Windows endpoint with the central Wazuh manager on my Kali VM, creating the secure communication link required to transmit security logs. To ensure the agent has the deep system permissions needed to monitor operating system activities, I executed the installation using an administrative Command Prompt.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/755/1*qpejkaxLRqlXf59MwylzEQ.png" /></figure><p>From the Downloads folder, I executed the deployment command and configured the agent to communicate directly with the Wazuh manager using the Kali VM’s Host-only IP address.</p><p>Step 3: Start the agent from CMD:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/652/1*Z0NYF1h4ATDL6yPJHFcf0Q.png" /></figure><p>Step 4: Confirm Agent Registration in the Wazuh Dashboard</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Wa3T5rng48gWMHb9npeJVA.png" /></figure><p>After starting the agent on my kali, I opened the Wazuh dashboard to confirm that the Windows endoint has been activated on the Wazuh Manger. On the Wazuh dashboard, I simply navigated to the Agents section and saw the endpoint has successfully been registered.</p><p><strong>Part 3:</strong></p><p><strong>Setting up n8n on my kali docker</strong></p><p>I set up n8n locally using Docker on Kali Linux to serve as the automation engine for my SOC environment. This setup lays the foundation for integrating alert data, streamlining ticketing processes, and simulating real SOC analysts’ workflows.</p><p>Step 1</p><p>Run:</p><pre>mkdir n8n_docker<br><br>cd n8n_docker</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/634/1*QYRAMa5oC6tMrXHeKw7DFA.png" /></figure><p>Step 2:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/632/1*Sf8dcKELyG-3SX6TNVL_nA.png" /></figure><p>Step 3:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/603/1*Azz3uXHkqe1usNNSHBk9nw.png" /></figure><p>on the text editor add this,</p><p>POSTGRES_USER=n8n_db_user<br> POSTGRES_PASSWORD=Enter_A_Strong_Password_Here<br> POSTGRES_DB=n8n<br> N8N_ENCRYPTION_KEY=Enter_Your_Generated_Hex_Key_Here</p><p>create a strong passowrd and enter the generated for Hex_Key you generated then save afterwards.</p><p>step 4:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/548/1*pdMZON3o0Y9E9StbhX8UUQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/608/1*hRi1cM_AljgDCviF2St5cQ.png" /></figure><p>the Docker Compose configuration defines how n8n is deployed and managed within the lab environment. It specifies the Docker image (n8nio/n8n) to be automatically downloaded and used, assigns a fixed container name (n8n) for easier management, and maps port 5678from the container to the host system to enable web access.</p><p><strong>Step 5: Deploying and Verifying n8n Container</strong></p><ol><li><strong>Deploy the stack by running</strong> sudo docker compose up -d.</li><li><strong>Verify the deployment using the command</strong> sudo docker ps.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/630/1*LfPysJYdHDVKh0zluadRNA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/606/1*J7t4IRdU9s4fqr3gxwMD-A.png" /></figure><p><strong>Step 6: Initial UI Access and Security Best Practices</strong></p><p>With the containers successfully running, I can now access the platform locally.</p><p>To Access the Dashboard I openned my web browser on the Kali VM and navigated to http://local address:5678</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*m0pqKbZvmhIj9LiTw6mF-A.png" /></figure><p>Upon first access, n8n will prompt you to configure the initial owner account. Provide a secure email and strong password.</p><p>My n8n instance is now securely deployed, backed by a persistent database, and ready to automate your security workflows.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*p11lv9DSLWTT98qGz_aReQ.png" /></figure><p><strong>Step 7: Connecting Wazuh and n8n</strong></p><p>To connect Wazuh and n8n, I used a universal web standard known as <strong>Webhook</strong>. Think of a Webhook as a digital mailbox. n8n opens a mailbox and waits; whenever Wazuh detects a threat, it instantly mails a copy of the alert (in JSON format) directly into that mailbox.</p><p><strong>Step 8</strong></p><p>Open n8n dashboard and click on add workflow to create a blank canvas.<br> Click the <strong>+</strong> button and search for <strong>Webhook</strong>. Add it to the canvas.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/634/1*mj17OVfJyBN99fCOXp9F0w.png" /></figure><ul><li>Double-click the Webhook node. Change the <strong>HTTP Method</strong> to POST (because Wazuh is <em>posting</em> data to it).</li><li>Click on <strong>Test URL</strong> and copy the web address it gives you (it will look something like <a href="http://localhost:5678/webhook-test/xxxx-xxxx-xxxx).">http://localhost:5678/webhook-test/xxxx-xxxx-xxxx).</a></li></ul><p>In Conclusion</p><p>As modern threats continue to evolve, SOC teams must shift from reactive monitoring to proactive, automated defense strategies.</p><p>In the part two of the project I will walkthrough the process of automating alert enrichment, prioritization, and response actions by connecting threat intelligent platform and a messaging platform to the workflow.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=11fe04a18fa4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[LLM SECURITY TRYHACKME WALKTHROUGH]]></title>
            <link>https://medium.com/@amarajulia75/llm-security-tryhackme-walkthrough-43a5e7a75ca9?source=rss-b0603482b85d------2</link>
            <guid isPermaLink="false">https://medium.com/p/43a5e7a75ca9</guid>
            <category><![CDATA[llm-security]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[ai-security]]></category>
            <dc:creator><![CDATA[Amarachi Obiajulu]]></dc:creator>
            <pubDate>Wed, 19 Aug 2026 13:04:49 GMT</pubDate>
            <atom:updated>2026-08-19T13:04:49.125Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/380/1*h9axzrVG0zCdYO8T8FbdrA.png" /></figure><p>Large language models (LLMs) are AI models trained on enormous collections of text, including books, articles, websites, and research papers to process, predict, and generate human language. By learning statistical patterns across this data, LLMs produce responses that closely resemble human communication.</p><p>This room aims to break down LLMs as an attack surface. Turning these question marks in your brain into security considerations you are actively aware of when engaging with an LLM in any capacity.</p><p>Learning Objectives</p><ol><li>Understand LLM-specific vulnerabilities that do not apply to traditional ML systems.</li><li>Recognise how LLMs create new attack surfaces through natural language interfaces and memory/context mechanisms.</li><li>Identify and categories LLM threat types: data-based, model-based, system-based, user-based.</li><li>Understand the realistic risks LLMs introduce in production systems.</li></ol><p>The LLM threat landscape is broken down into four categories: data-based threats, model-based threats, system-based threats and user-based threats. In this walkthrough, we will see how each of these key areas can be exploited.</p><p><strong>Section 1 — Data-Based Threats</strong></p><p>LLMs are fundamentally data-driven; they learn from a corpus of training data and generate outputs based on it. This is one of LLMs’ greatest strengths. But that strength can be turned into a weakness.</p><p><strong>Training Data Extraction</strong></p><p>Training data extraction attacks aim to recover actual sequences from the model’s original training data by interacting with the model.</p><p>Training data extraction attacks generate large amounts of text from an LLM and analyse those outputs to identify sequences that show behavioural signs of memorisation, such as unusually high likelihood/confidence, deterministic regeneration (a model’s ability to reproduce the same output), or realistic structured content</p><p>The training dataset which is supposed to be confidential can be used by the adversary by crafting prompts designed to trigger memorised content and the LLM on the other hand provides an output of verbatim or near-verbatim of it’s training dataset (text, PII, secrets)</p><p><strong>Membership Inference</strong></p><p>Membership inference attacks ask whether the model ever recorded a specific data sample. The attacker tests the model’s reaction to that exact sample, looking for unusually confident or familiar responses that indicate it was part of training. The adversary doesn’t necessarily obtain the full content of a training example, but can detect the presence of certain data in the training set by observing the model’s behaviour.</p><p>In Membership Inference, the training dataset membership (privacy metadata) is the attack surface. The attacker uses a known candidate data sample they already possessed and the output from the model is usually a yes/No(or probability) decision indicating whether the sample was used in training.</p><p><strong>Prompt leakage (LLM07:2025 — System prompt leakage)</strong></p><p>LLMs like ChatGPT, Claude, Gemini, etc., don’t just operate using the learnings from their training data; they also use hidden instructions known as system or developer prompts. This system prompt is kept hidden in many instances, especially when what sits in front of the model is some form of “intellectual property,” such as an application feature or product. Therefore, the leaking of such a prompt could be compared to the leaking of sensitive company data such as application source code.</p><p>For prompt Injection, the system prompt/developer instructions is the attack surface. When a user prompts and ask the model to reveal instructions the output is partial or full disclosure of hidden system or developer prompts.</p><p>Ask your chatbot assistant to give you the Task 2 demonstration. You’ll then need to use a membership inference attack to determine which of the three placeholder samples is a member.</p><p><strong>Question 1</strong></p><p><strong>Which sample is a member?</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/608/1*Xpwc5nGqGt2Gurumjm1sjw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/610/1*LqjG8N7Uwrlo2yrm9n7GLQ.png" /></figure><p><strong>Answer: MI_SAMPLE_ALPHA</strong></p><p><strong>Question 2.</strong></p><p><strong>Which attack determines whether a known data sample was part of an LLM’s training set?</strong></p><p><strong>Answer: Membership inference</strong></p><p><strong>Question 3</strong></p><p><strong>Which data-based threat involves the model reproducing memorised snippets of its training data?</strong></p><p>Training data extraction attacks aim to recover actual sequences from the model’s original training data by interacting with the model.</p><p><strong>Answer: Training data extraction</strong></p><p><strong>Task 3 — Model-Based Threats</strong></p><p>Model-based threats exploit the model itself as the attack surface, abusing how information is encoded within its parameters and representations. As a consequence, these attacks may expose intellectual property (model weights) or sensitive training data that the model has memorised.</p><p>Model Extraction: Model extraction is the process of illicitly copying a machine learning model’s functionality or parameters without authorisation.An attacker can do this if they can interact with an LLM through its public API and send a large number of prompts; the responses to these prompts are then stored in a sort of input-output pair.</p><p>Model Inversion: Model inversion attacks exploit a model’s output to reveal information about its training data.</p><p>Model inversion attacks treat the model as a source of stored information rather than a classifier to be probed.</p><p><strong>Question 1</strong></p><p><strong>What is the employee ID?</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/629/1*iWJ6YWtwd1BaFxDhtMHr3g.png" /></figure><p><strong>Answer: Employee ID: 7814</strong></p><p>Question 2</p><p><strong>Which model-based threat attempts to reconstruct sensitive information encoded within a model’s internal representations?</strong></p><p><strong>Answer: Model inversion</strong></p><p><strong>Task 4: System-Based Threats</strong></p><p>Unlike traditional software, LLMs process all input (system instructions, user prompts, etc.) as a single concatenated context without a built-in security boundary separating trusted content (i.e., the system instructions) from untrusted content (i.e., user prompts).</p><p><strong>Prompt injection</strong></p><p>Prompt injection is one of the most well-known and widely studied threats to LLMs. At a system level, it is enabled by what can be described as context-window poisoning: the manipulation of the model’s input context to override or subvert its intended behaviour.</p><p><strong>Context Overflow (LLM10:2025 — Unbounded Consumption)</strong></p><p>This context window contains both the initial input and the model’s output. This token limit can be abused to either force important information out of the context (to circumvent safeguards) or to overwhelm the model’s processing capacity (causing delays or denial-of-service).</p><p>This attack happens when an attacker supplies an extremely long input or continuously appends content until the context is overfull.</p><p>Memory Poisoning:</p><p><strong>Question 1: Did you convince the model? Whats the flag?</strong></p><p><strong>Answer: THM{MEMORY_POISONED}</strong></p><p><strong>Question 2:</strong></p><p><strong>Which system component combines system instructions, retrieved data, and user input into a single sequence?</strong></p><p><strong>Answer: The Context Window</strong></p><p><strong>Task 5 — User-Based Threats</strong></p><p>This section looks at how AI can be used to manipulate human trust and judgment.</p><p><strong>LLM Powered Social Engineering</strong></p><p>An LLM can now generate spear-phishing emails that read exactly like a colleague or executive.</p><p>Imagine an attacker has compromised a locally deployed LLM within an organisation, revealing customer or project data it had been trained on. This information is only available to people within the company. Using that on top of any OSINT available on any high-ranking member of this organisation could result in an email indistinguishable from a real one, underscoring the need to cover LLM Security across all areas of the attack surface.</p><p><strong>Trust Exploitation (LLM09:2025 — Misinformation)</strong></p><p>LLMs introduce new trust risks. Because they often answer with confident, authoritative-sounding text, users may place too much trust in their outputs.</p><p>In this phase hallucination stops being a reliability problem and becomes an active attack vector, one that attackers can deliberately engineer by registering packages they know a model is likely to hallucinate.</p><p>Practical: Alright, time for the last demonstration. Ask your assistant to begin with Task 5. Your job this time is to validate the LLM’s advice.</p><p><strong>Question 1</strong></p><p><strong>Which package should you NOT download?</strong></p><p><strong>Answer: robbco-llm-audit</strong></p><p><strong>Question 2</strong></p><p><strong>LLM-powered social engineering primarily amplifies which existing attack category?</strong></p><p><strong>Answer: Phishing</strong></p><p>In conclusion:</p><ul><li>LLMs introduce a unique attack surface distinct from traditional ML systems, driven by natural language interaction, context handling, and emergent behaviour.</li><li>Data-based threats exploit how LLMs learn from and memorise training data, enabling attacks such as training data extraction, membership inference, and system prompt leakage.</li><li>Model-based threats target the model itself, including model extraction (theft of model behaviour or weights) and model inversion (reconstructing sensitive training data).</li><li>System-based threats arise from how LLMs process all inputs as a single context, enabling prompt injection, context window overflow, and memory poisoning.</li><li>User-based threats leverage LLMs as force multipliers for social engineering, increasing the effectiveness of phishing, scams, and trust exploitation.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=43a5e7a75ca9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Model and Data: A Tryhackme Walkthrough]]></title>
            <link>https://medium.com/@amarajulia75/ai-model-and-data-a-tryhackme-walkthrough-8dba98f00335?source=rss-b0603482b85d------2</link>
            <guid isPermaLink="false">https://medium.com/p/8dba98f00335</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[ai-model]]></category>
            <category><![CDATA[ai-security]]></category>
            <category><![CDATA[data-provenance]]></category>
            <category><![CDATA[federated-learning]]></category>
            <dc:creator><![CDATA[Amarachi Obiajulu]]></dc:creator>
            <pubDate>Thu, 28 May 2026 12:16:15 GMT</pubDate>
            <atom:updated>2026-05-28T12:58:34.881Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/998/1*tzKWL-spMjKdjjYJh7WzwQ.png" /></figure><p>Understanding AI Models and Data</p><p>Room: <a href="https://tryhackme.com/room/aimodelsdata">https://tryhackme.com/room/aimodelsdata</a></p><p>By the end of this room you will understand</p><ul><li>Where AI training data comes from and security risks introduced by poor data provenance.</li><li>How PII and sensitive credentials can become permanently embedded in model weights through large-scale web scraping.</li><li>How key model-building decisions can introduce security risks.</li><li>And the inheritance problem and what organizations unknowingly take on when fine-tuning pre-trained models.</li></ul><p><strong>Task 1</strong></p><p>Simply understand the learning objectives no answer needed</p><p><strong>Task 2</strong> — <strong>Training Data</strong></p><p>Ever wondered where the data for training large language models come from? GPT-3 was trained on roughly 570GB of filtered text. To get this vast amount of data developers draw data from Web Scraping, Licensed datasets, synthetic data and internal corpora</p><p>The most widely used training dataset is Common crawl, a free, publicly available archive of web crawl data. DeepSeek -V2 was pretrained on it.</p><p>A model’s behavior is a direct product of what it was trained on. If that data was scraped without audit, contaminated with PII, or manipulated upstream, those characteristics become part of the model and there’s no reliable way for the organization deploying it to know.</p><p>What term describes the ability to answer where data came from, when it was collected, and whether it has been modified?</p><ul><li><strong>Answer: Data Provenance</strong></li></ul><p>What is the name of the most widely used public corpus that underpins essentially every major model family?</p><ul><li><strong>Answer: Common Crawl</strong></li></ul><p>What is the equivalent of a software Bill of Materials(SBOM), used to document dataset sources, licenses and filtering decisions?</p><ul><li><strong>Answer: ML-BOM</strong></li></ul><p><strong>Task 3 — Building the Model</strong></p><p>Once a model is trained, it often goes through compression steps before deployment. The two of the most common are pruning and quantisation.</p><p>What term describes one complete pass of the training algorithm through the entire dataset?</p><ul><li><strong>Answer: Epoch</strong></li></ul><p>What problem occurs when a model memorises training data rather than learning general patterns.</p><ul><li><strong>Answer: Overfitting</strong></li></ul><p>What post-training optimisation technique reduces the numerical precision of model weights to cut memory and compute requirements?</p><ul><li><strong>Answer: Quantisation</strong></li></ul><p>What training approach trains a model across decentralised devices, sending only weight updates rather than raw data to a central server?</p><ul><li><strong>Answer: Federated learning</strong></li></ul><p><strong>Task 4 — The Inheritance Problem</strong></p><p>A pre-trained model is one that has already been trained on a large, general-purpose dataset. These base models learn broad language understanding: grammar, facts, reasoning patterns, and world knowledge. Fine-tuning is the process of continuing to train one of these pre-trained models on a smaller, task-specific dataset.</p><p>What is the process of taking a pre-trained model and continuing to train it on a smaller, task-specific dataset?</p><ul><li><strong>Answer: Fine-tuning</strong></li></ul><p>What term describes a model that has already been trained on a large general-purpose dataset?</p><ul><li><strong>Answer: Pre-trained model</strong></li></ul><p><strong>Task 5- The Black Box Problem</strong></p><p>The documentation artefact designed to address this is the model card: a structured document that accompanies a model and describes what it is, how it was built, and where it falls short.</p><p>What documentation artifact accompanies a model to describe what it is, how it was built, and where it falls short?</p><ul><li><strong>Answer: model card</strong></li></ul><p>What are the billions of floating-point numbers that make up a trained model collectively referred to as?</p><ul><li><strong>Answer: Weight</strong></li></ul><p><strong>Task 6 — Practical</strong></p><p>Complete the exercise to get the flag!</p><p>This task involved performing a model audit, looking through the model card, the metadata, the file listings and the training details for warning signs and severity of each risk found.</p><p>Publicly available web sources and forums, Q&amp;A Sites are both<strong> high risk </strong>because the training data source is vague, no provenance whatsoever and there was also no mention of PII filtering from scraped forums.</p><p>Fine-tuned from enterprise-base-v1.1, the model achieved a micro-averaged F1 score of 0.91 across all classification categories, the licenced listed as custom and it had no links or terms attached to it. The model size is smaller than the base model with no documentation of what was applied after post training. This are all <strong>medium risk</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/992/1*jybRqBinmBd8j0Eg1r0zoQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/966/1*q3qvNHnuej7oJeOTLX4mrw.png" /></figure><ul><li><strong>Answer: THM{A_m0del_Stud3nt}</strong></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8dba98f00335" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Model Context Protocol (MCP) from the lens of a Security Operation Analyst.]]></title>
            <link>https://medium.com/@amarajulia75/model-context-protocol-mcp-from-the-lens-of-a-security-operation-analyst-5a4c5450ef64?source=rss-b0603482b85d------2</link>
            <guid isPermaLink="false">https://medium.com/p/5a4c5450ef64</guid>
            <dc:creator><![CDATA[Amarachi Obiajulu]]></dc:creator>
            <pubDate>Wed, 13 Aug 2025 08:01:48 GMT</pubDate>
            <atom:updated>2025-08-13T08:01:48.838Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/644/1*6dLVzcXBOURKq-yPQoPP4A.png" /></figure><p>Now imagine a system or a program that is capable of Independently performing tasks on your behalf by designing a workflow and utilizing available tools- that’s what Artificial intelligence (AI) agents are built to perform. It becomes a multiagent systems when multiple AI agents are working collectively to perform a task on behalf of the user.</p><p>Lately, a lot of organizations are rapidly adopting the use AI agents like Model Context Protocol (MCP) and Copilots — while this things are actively transforming the way this organizations operate there are also concerns as to how the security posture of this organizations are being altered as a result of this technologies. Some of this Models when enabled can execute and complete complex task with little human assistance and integrate seamlessly with existing enterprise tools.</p><p>Incase you’re wondering, Model Context Protocol (MCP) is a kind of API, but instead of allowing one computer program to talk to another computer program in a standardized way, it allows an AI agent or copilot to talk to databases, tools, and other resources. <a href="https://www.anthropic.com/news/model-context-protocol">Anthropic</a> came up with MCP to help developers build against a standard protocol instead of maintaining separate connectors for each data source.</p><p>While listening to the Microsoft <a href="https://www.youtube.com/watch?v=ceV3RsG946s">build 2025</a>, I found out that with Copilot tuning you can build an MCP fine-tuned on your company’s data, workflows and style. This copilot can learn your company’s unique tone, fine-tuned for your customers use. what this mean is that the tuned MCP server that is tailored to your organization can interact with your tools in a way that reflects your company’s brand and operational style.</p><p>Looking at it from the security perspective a Model Context Protocol connected to sensitive systems like databases, finance platforms and ticketing tools creates more API endpoint that could be potentially exploited. Understanding MCP security is going to be key for enterprises going forward, especially if they are deploying AI agents in any significant way.</p><p>MCP servers are still catching up in this security maturity cycle, making them particularly vulnerable during this adoption <a href="https://equixly.com/blog/2025/03/29/mcp-server-new-security-nightmare/">phase</a>. Supply chain risk is a key concern when it comes to MCP, the MCP severs are often built by Independent developers with no established standards for secure development — this creates a plugin-like ecosystem where unvetted code, generally pulled from a random GitHub repo, may gain access to sensitive data or local system capabilities.</p><p>Before implementing MCP server, organizations should carefully consider the security implications and implement robust security controls. The standardization efforts around MCP must prioritize security alongside functionality to prevent creating a new generation of vulnerable AI systems. Constantly monitor Copilot interactions for abnormal model outputs and use an MCP client whose default is to show you every tool call and its input before approving it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5a4c5450ef64" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ANALYSING A BITMAP IMAGE TO FIND COMMAND AND CONTROL SERVER.]]></title>
            <link>https://medium.com/@amarajulia75/analysing-a-bitmap-image-to-find-command-and-control-server-34a716e8a6ed?source=rss-b0603482b85d------2</link>
            <guid isPermaLink="false">https://medium.com/p/34a716e8a6ed</guid>
            <dc:creator><![CDATA[Amarachi Obiajulu]]></dc:creator>
            <pubDate>Mon, 12 May 2025 12:18:02 GMT</pubDate>
            <atom:updated>2025-05-12T12:18:02.397Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/646/1*2tElnto3tfFnp4ay6oiT6g.png" /></figure><p>To evade detection, adversaries use lots of techniques to compromise their target(s). In this post, I will focus on how to uncover obfuscation technique that involves hiding payload in an image with a bitmap extension.</p><p>BMP (Bitmap) files are image formats used to store graphics. Threat actors have begun embedding malicious payloads inside BMP files. These aren’t just images ,they’re carriers of malware, cleverly disguised to bypass detection.</p><p>The Lazarus APT group are popularly known to employ this technique in its operation. Lazarus group use .bmp files embedded with malicious HTML Application (HTA) object to drop it’s loader.</p><p>This is quite different from steganography (which hides data in pixel values), this technique often abuses BMP file structure to insert executable code or script objects that can later be extracted. It goes through the process of converting a PNG file format into a BMP file format that automatically decompresses the embedded malicious zlib object.</p><p>I started by using static analysis to get file type, integrity and available metadata. To achieve this, I used command-line tools like file, exiftool, binwalk and sha256.</p><p>The aim of this is to look for anomalies like compression, embedded GPS, or comments that might contain URLs or IPs. Most of the information can be found on the screenshots below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hZLv_Kl9QTTs0fUWePA8Yw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/773/1*frSqZODFVYedrXhA4Opx6w.png" /></figure><p>After getting the hash value using sha256sum, I uploaded it on VirusTotal to see if its been flagged malicious by security vendors.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*l2nLftvpfUmfo1QPOMCh_w.png" /></figure><p>Most of these results identify it as a “Generic Metasploit Exploit”, we can assume the malware embedded in the image is in-fact a Metaploit payload.</p><p>I will be going further to run a dynamic analysis on this image, to see how the malware works. For this phase, I tried using a lot of malware tool and sandbox but most of them didn’t give me a detailed information I needed on the C2 server.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/1*wS-QS6GqRZ3IJqIjIUk_xQ.png" /></figure><p>I found “NativePayload_Image.cs” which is a script that executed malware embedded in images and gave it a go. It is a C# script that extracted the payload but then executes the script by changing its own instruction pointer to the beginning of the payload, forcing the kernel to execute it as part of the script.</p><p>I used my windows server on a Flare-VM to first convert the the Nativepayload_image.cs to Nativepayloadimage.exe before executing the script.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jUXEIMVab_lS-YA5u5kw1w.png" /></figure><p>On running the command, I setup my Wireshark to capture and monitor all communication and after some time I terminated it to check through the packets.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AUwXcrRRNqlX3lk5t221qA.png" /></figure><p>From the Wireshark packets, I could see the C2 server used for this image by filtering, tcp.port == 4444 to get all Metasploit payload communication, as port 4444 is the default port for all communication on Metasploit TCP payloads.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RLk8wTBeDQnmi9h8mXOUqw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*otNYK3T6J0vYrrhz2lsyqw.png" /></figure><p>The payload communicates to the IP 217.182.54.224 on port 4444 but seeing as all SYN packets were not acknowledged shows that the server is down and the threat is no longer active.</p><p>While further investigating the IP address, I found few files that are associated with the IP address. The IP address has been cleaned as well.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/1*wXL7BmEiYvftfpnqQ0CWmA.png" /></figure><p>In conclusion, while threat actors are consistently looking for ways to evade detection and obfuscate, it very important that organization use good endpoint detection and response tool (EDR) to monitor suspicious activities on the network. Apply advanced email and content filtering to ensure deep scanning of non-executable files like bitmap images and known steganography patterns. Unexpected file types sent as attachment should also be blocked and quarantined from having access to the network.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=34a716e8a6ed" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>