<?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 Scott Plastine on Medium]]></title>
        <description><![CDATA[Stories by Scott Plastine on Medium]]></description>
        <link>https://medium.com/@plantainstan?source=rss-7b5ee9155ea7------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*4igdDdpti0jEF47okf6azg.jpeg</url>
            <title>Stories by Scott Plastine on Medium</title>
            <link>https://medium.com/@plantainstan?source=rss-7b5ee9155ea7------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 07:17:23 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@plantainstan/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[Detection Ratio Metrics]]></title>
            <link>https://detect.fyi/detection-ratio-metrics-d7e5fdf87d47?source=rss-7b5ee9155ea7------2</link>
            <guid isPermaLink="false">https://medium.com/p/d7e5fdf87d47</guid>
            <category><![CDATA[blue-team]]></category>
            <category><![CDATA[metrics]]></category>
            <category><![CDATA[detection-engineering]]></category>
            <dc:creator><![CDATA[Scott Plastine]]></dc:creator>
            <pubDate>Wed, 20 May 2026 02:03:46 GMT</pubDate>
            <atom:updated>2026-05-20T06:35:30.545Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MGSXa6bFP4fB57nGE6zY9w.png" /></figure><p>A detection engineering team monitors an environment that evolves over time. Generally, that environment grows as the enterprise hires more employees and expands its cyber footprint. To adjust for this we must measure a detection engineering team’s output in a way that incorporates enterprise growth. We are going to build upon the two previous posts (<a href="https://medium.com/detect-fyi/detection-visibility-metrics-577fa1d63696">Detection Visibility Metrics</a> &amp; <a href="https://medium.com/@plantainstan/detection-pipeline-metrics-ae6757d7e6b2">Detection Pipeline Metrics</a>). These three posts intentionally align since it’s critical we keep the same terminology when referencing the entire detection ecosystem.</p><p><strong>Basic Detection Metrics</strong></p><p>We need a way to measure our success beyond basic alert counts which are just numbers without any context. <strong>The only way to show improvement for numbers without any context would be to reduce alert counts.</strong> We all know it is not that simple when most environments are growing. As the enterprise expands alert counts will also increase. Let’s start with an example:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/548/1*gUshT1UAl56VRRMKkK_LRw.png" /></figure><p><strong>At first glance it appears we are getting worse!</strong> It will look that way if we strictly look at the alert totals. There is much more to the story when we incorporate the growth of the environment we are monitoring.</p><p><strong>The Detection Ratio</strong></p><p><strong>We can incorporate environment change into our metrics by creating a ratio for triggered alerts to monitored objects.</strong> This will allow us to both measure our improvement as the enterprise grows and even forecast alert output to ensure we have the appropriate staffing in place to monitor ahead of time. I will share a preview of the final metrics before I walk through the steps. The chart below is specific to endpoints.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*618htC2sph2uv9-WmD9Jcg.png" /></figure><p>From the chart we can make the following conclusions:</p><ul><li>From 2023 to 2025 we reduced our triggered alert counts per endpoint from 0.026 to 0.015. The monthly average triggered alerts increased due to the amount of endpoints increasing within the enterprise.</li><li>We can forecast triggered alert counts when the enterprise grows in 2026 since we have the alert ratio per endpoint for 2025. Forecasting will allow the monitoring team to prepare ahead of time.</li><li>We can easily speak to the metrics with other teams and leadership by saying we have X alerts per month per 100 endpoints. For example in 2025 we generated approximately 1.5 alerts per 100 endpoints per month.</li></ul><p><strong>Detection Ratio Categories</strong></p><p>We will need to create high level buckets to define a monitored object. I will propose some but you should adapt these to better fit your environment.</p><ul><li>Endpoint — count of endpoints being monitored within your organization.</li><li>Network — sum of monitored network throughput within your organization.</li><li>Identity — count of monitored users within your organization.</li></ul><p><strong>Detection Ratio Data Sources</strong></p><p>The next step will be to split out all triggered alerts by the categories above. I have listed some example data sources below.</p><p>Endpoint — sum of triggered alerts from endpoint related data sources.</p><ul><li>Antivirus, Endpoint Detection &amp; Response (EDR), Endpoint Protection Platforms (EPP), Application Control, Operating System Events, PowerShell events, etc.</li></ul><p>Network — sum of triggered alerts from network related data sources.</p><ul><li>Web Proxies, Network Intrusion Detection Systems (IDS/IPS), Web Application Firewalls, VPN, Netflow, etc.</li></ul><p>Identity — sum of triggered alerts from identity related data sources.</p><ul><li>Email, Directory Services (Active Directory), Privileged Access Monitoring, Entra ID Protection, etc.</li></ul><p>NOTE: This will get trickier with more complex detection correlation across data sources but realistically for metrics we need to summarize the data. If we do not summarize our monitoring when reporting up we will be deep in weeds with leadership that simply does not have the time.</p><p><strong>Detection Ratio Calculations</strong></p><p>Below are the generic formulas to calculate the ratios. The objects will be what makes the most sense for your organization.</p><blockquote>Alerts per Object = Object Alert Totals / Object Count</blockquote><blockquote>Alerts per 100 Objects = Object Alert Totals / (Object Count / 100)</blockquote><blockquote>Next Year Forecast = Object count * Alerts per Object [Current Year Ratio]</blockquote><p>Here are some examples with fabricated alert counts for the categories I listed above.</p><p><strong>Endpoint — Alerts Per Endpoint</strong></p><blockquote>Alerts per Endpoint = Endpoint Alert Totals / Endpoint Count</blockquote><blockquote>Alerts per 100 Endpoints = Endpoint Alert Totals / (Endpoint Count / 100)</blockquote><blockquote>Year 2026 Forecast = Endpoint count * Alerts per Endpoint [Year 2025 Ratio]</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*618htC2sph2uv9-WmD9Jcg.png" /></figure><p><strong>Network — Alerts per GB transferred</strong></p><blockquote>Alerts per GB Transferred = Network Alert Totals / (Network Ingress + Network Egress)</blockquote><blockquote>Alerts per 100 GB transferred = Network Alert Totals / ((Network Ingress + Network Egress) / 100)</blockquote><blockquote>Year 2026 Forecast = (Network Ingress + Network Egress) * Alerts per GB Transferred [Year 2025 Ratio]</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IUxr6Qkx614AV_6xTee0WQ.png" /></figure><p><strong>Identity — Alerts per User</strong></p><blockquote>Alerts per User = Identity Alert Totals / User Count</blockquote><blockquote>Alerts per 100 User = Identity Alert Totals / (User Count / 100)</blockquote><blockquote>Year 2026 Forecast = User Count * Alerts per User [Year 2025 Ratio]</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6OZo3vcZvRiQZOHQAWRr-g.png" /></figure><p><strong>Detection Ratio Visualization</strong></p><p>Next lets graph an example showing how these numbers fit together. Below you can see clearly that the Detection Engineering team is improving by reducing the detection ratio (green line) even while the enterprise is growing (blue line). The alert counts (red line) fluctuates which will certainly be noticed by leadership. If we only had the alerts totals (red line) there would be some frantic explaining to do.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/773/1*7GIyHd5w_zJS1y7H_LE1mQ.png" /></figure><p>The detection landscape is always changing so we need to ensure when we measure it we incorporate that factor. If we do not then we may be improving but the metrics will not reflect it. As long as the final numbers are simple to communicate we are free to calculate them in a way that accounts for the environment change. Give these ratios a shot to see how your detections have improved over time.</p><p><strong>Sources and Additional Reading</strong></p><p><a href="https://detect.fyi/detection-visibility-metrics-577fa1d63696">https://detect.fyi/detection-visibility-metrics-577fa1d63696</a></p><p><a href="https://medium.com/@plantainstan/detection-pipeline-metrics-ae6757d7e6b2">https://medium.com/@plantainstan/detection-pipeline-metrics-ae6757d7e6b2</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d7e5fdf87d47" width="1" height="1" alt=""><hr><p><a href="https://detect.fyi/detection-ratio-metrics-d7e5fdf87d47">Detection Ratio Metrics</a> was originally published in <a href="https://detect.fyi/">Detect FYI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Detection Pipeline Metrics]]></title>
            <link>https://medium.com/@plantainstan/detection-pipeline-metrics-ae6757d7e6b2?source=rss-7b5ee9155ea7------2</link>
            <guid isPermaLink="false">https://medium.com/p/ae6757d7e6b2</guid>
            <category><![CDATA[metrics]]></category>
            <category><![CDATA[threat-hunting]]></category>
            <category><![CDATA[blue-team]]></category>
            <category><![CDATA[detection-engineering]]></category>
            <dc:creator><![CDATA[Scott Plastine]]></dc:creator>
            <pubDate>Thu, 23 Apr 2026 15:50:19 GMT</pubDate>
            <atom:updated>2026-04-23T17:22:06.822Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2oaq3-IJ6FDu0-eNz6xaEQ.png" /></figure><p>This post is inspired by the awesome <a href="https://specterops.io/blog/2019/11/20/introducing-the-funnel-of-fidelity/">SpecterOps Funnel of Fidelity</a>! <strong>A detection engineering team’s product is finding the signal (evil) within the noise (events).</strong> The noise includes commercial security tool alerts too!</p><p>Too often the conversation is how do we reduce triggered alerts since leadership and response teams only see the end result. They do not realize how much effort goes into getting to that point.<strong> We are always driving towards improvement but we must also showcase how much we do filter since it’s astonishing.</strong></p><p>These metrics build from two of my previous blog posts which are the <a href="https://medium.com/detect-fyi/detection-pipeline-maturity-model-076984779651">Detection Pipeline Maturity Model</a> and the <a href="https://medium.com/detect-fyi/detection-visibility-metrics-577fa1d63696">Detection Visibility Metrics</a>. The Detection visibility numbers will be the starting point of the funnel as that’s the data we have to sift through to identify evil. Then we will report the amount of events at different stages of the pipeline.</p><p>Let’s start with our base detection pipeline from my <a href="https://medium.com/detect-fyi/detection-pipeline-maturity-model-076984779651">previous blog post</a>. Your pipeline may be slightly different so apply the same thought process at the core filtering points. I have simplified it a bit since less is more with metrics and fabricated some numbers to show how this could look. Pick a standard time frame that aligns with the reporting of metrics.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*NTEsEYdP2Pj9ikG-" /></figure><ul><li>Data Sources — The amount of events residing within the analytics platform for a specific time frame.</li><li>Analytics — The amount of events our analytics search for that same time frame. <em>Detections do not generally search every log source in the platform.</em></li><li>Risk Engine — All of the suspicious events discovered by the detection analytics that feed the Risk engine. <em>This is what the Response team would have to deal with if the detection engineering team did not exist.</em></li><li>Alerts — The actual triggered alerts from the Risk Engine that would be reviewed by an analyst.</li></ul><p><strong>This changes the conversation from why is the Detection Engineering team burdening us with so many alerts to showing how the Detection Engineering team is continuing to reduce the amount of alerts that would require an analyst response.</strong></p><p>This doesn&#39;t mean there isn’t ever room for improvement but if the pipeline metrics are not shared then we are not advertising our value! We then put a spotlight on the true noise generators which is the myriad of closed source signature tools.</p><p>It’s very easy for the Detection Engineering teams’ hard work to be overlooked so I encourage everyone to start gathering these numbers. If you’re feeling discouraged it’s fun to look at to remind yourself about the awesome work you’re doing!</p><p><strong>Sources and Additional Reading</strong></p><p><a href="https://specterops.io/blog/2019/11/20/introducing-the-funnel-of-fidelity/">https://specterops.io/blog/2019/11/20/introducing-the-funnel-of-fidelity/</a></p><p><a href="https://detect.fyi/detection-pipeline-maturity-model-076984779651">https://detect.fyi/detection-pipeline-maturity-model-076984779651</a></p><p><a href="https://detect.fyi/detection-visibility-metrics-577fa1d63696">https://detect.fyi/detection-visibility-metrics-577fa1d63696</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ae6757d7e6b2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Detection Visibility Metrics]]></title>
            <link>https://detect.fyi/detection-visibility-metrics-577fa1d63696?source=rss-7b5ee9155ea7------2</link>
            <guid isPermaLink="false">https://medium.com/p/577fa1d63696</guid>
            <category><![CDATA[blue-team]]></category>
            <category><![CDATA[threat-detection]]></category>
            <category><![CDATA[metrics]]></category>
            <category><![CDATA[detection-engineering]]></category>
            <dc:creator><![CDATA[Scott Plastine]]></dc:creator>
            <pubDate>Mon, 20 Apr 2026 13:49:51 GMT</pubDate>
            <atom:updated>2026-04-20T17:30:17.442Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_dIdKHFnrKntYf7BYEU5rQ.png" /><figcaption>I only write blog posts so I can make a sweet AI photo</figcaption></figure><p>Metrics are always on the top of everyone&#39;s mind and for good reason.</p><ul><li>How do we show improvement year over year?</li><li>How do we show improvement month over month?</li><li>What are we even driving towards?</li><li>Is a Detection Engineering team needed?</li><li>When are we done?</li><li>What are we monitoring?</li><li>Why does this cost so much?</li></ul><p>These are all excellent questions that we need to prove to leadership and ourselves really. In this post I will go over ways to measure how much we monitor every day. After compiling these numbers for your own organization it will be impressive to see and in my experience very eye-opening to anyone outside of Detection Engineering.</p><p>Metrics need some sort of base context. Lets take this common example:</p><ul><li>We had 1000 alerts in March and 1200 in April.</li></ul><p>The next logical questions are:</p><ul><li>Why did it go up?</li><li>Is that even good?</li><li>What should it be?</li></ul><p>These can be very difficult questions to answer in a growing or shrinking enterprise with many new tools and users.</p><p>So first things first, what are we monitoring? Answering this question is fun and you can admire your hard work!</p><p><strong>These visibility metrics will be measuring our organizational shift over time.</strong> Organizations grow by acquisition or just general expansion and those are critical numbers to track as your detection engineering program matures.</p><p>Below is a template with some base metrics to track your enterprise cyber footprint. <strong>Most tools track these numbers since they are critical for licensing!</strong> These base metrics would include user counts, workstation counts, mobile device counts, and network throughput to name a few. Your business may be different so track whatever is relevant for your detection program.</p><p><strong>Users</strong></p><p>A great place to gather user counts can be in your corporations financial reports. These numbers are often reported to regulators quarterly so they are as accurate as it gets.</p><ul><li>Customer Totals</li><li>Employee Totals</li><li>Contractor Totals — might be trickier but could be valuable to delineate</li></ul><p><strong>Technology</strong></p><p>There are many places to gather asset counts and it’ll be important to keep it consistent month over month. Endpoint Security tools generally have very good coverage in many organizations and realistically if a device doesn&#39;t have a security agent you are not really monitoring it anyway.</p><ul><li>Workstations by Major OS (Prod/QA/Dev)</li><li>Servers by Major OS (Prod/QA/Dev)</li><li>Mobile Devices by Major OS</li></ul><p><strong>Identity</strong></p><p>We should track our user base and the amount of security groups it takes to manage all of those users. Every group or account (even non-prod) is an opportunity for an attacker to escalate privileges.</p><ul><li>Accounts (Prod/QA/Dev)</li><li>Groups (Prod/QA/Dev)</li><li>Service Accounts (Prod/QA/Dev)</li><li>Privileged Groups (Prod/QA/Dev)</li><li>Privileged Users (Prod/QA/Dev)</li></ul><p><strong>Logging Footprint</strong></p><p>The amount of data an analytics platform ingests is truly impressive. Then when you multiply that by how long data is retained it goes up even more. This shows leadership what they are paying for. It could be a double edged sword since every organisation wants to save money on data retention but that’s still a good conversation to have! Tracking this will also lead to conversations around data retention consistency for similar types of data.</p><ul><li>Ingestion — TB/GB/Day</li><li>Searchable (X days) — PB/TB — this would be searchable events that detections query</li></ul><p><strong>Data Source Samples</strong></p><p>This is where we can show how much data we collect from the largest data generators. It will be very interesting to anyone who does not live in the analytics platform like the Detection Engineering team. I will list a few common log sources but this will be different for every organization. Remember that not all telemetry is in the main analytics platform. If your enterprise uses different platforms call those out as well. If you use it for detections then track it.</p><ul><li>Active Directory Events —TB/GB/Day</li><li>Web Application Firewall Events — TB/GB/Day</li><li>Windows Server OS Events — TB/GB/Day</li><li>Linux Server OS Events — TB/GB/Day</li><li>Web Proxy Events — TB/GB/Day</li><li>AWS CloudTrail Events — TB/GB/Day</li><li>Azure EventHub Events — TB/GB/Day</li><li>DNS Events — TB/GB/Day</li><li>EDR Events — TB/GB/Day</li><li>Email Events — TB/GB/Day</li></ul><p><strong>Email Throughput</strong></p><ul><li>Inbound emails with attachments per day/week/month</li><li>Inbound emails without attachments per day/week/month</li><li>Outbound emails with attachments per day/week/month</li><li>Outbound emails without attachments per day/week/month</li><li>Emails from new senders per day/week/month</li></ul><p><strong>Network Throughput</strong></p><ul><li>VPN Throughput — GB/Day</li><li>Network Proxy Throughput — GB/Day</li><li>Externally Facing Web Application Firewall Throughput — GB/Day</li><li>Internet Network Egress Throughput — GB/Day</li><li>Internet Network Ingress Throughput — GB/Day</li></ul><p><strong>Storage</strong></p><ul><li>Storage Transfer Throughput — TB/Day</li><li>Total Storage Retained — TB/PB</li></ul><p><strong>Tracking Organizational Shift</strong></p><p>Gathering these numbers is the first step. The next is to continue to collect them regularly to show overall growth within the enterprise. <strong>As the enterprise grows the Detection Engineering team must grow as well since there will be more to monitor.</strong> These numbers are generally point in time so if we are not recording them we cannot go backwards to gather them.</p><p>Below I created an example graph showing how tracking the data will tell a useful story.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/891/1*vJT9s-htYZlka0uG4A1sBg.png" /><figcaption>Asset Growth within an Enterprise</figcaption></figure><p>At the start of the graph there is approximately 18,000 assets and through 2 acquisitions and general growth of the enterprise there are approximately 38,000 assets. More assets require more effort to monitor and these numbers can be used to justify more resources and explain higher alert counts.</p><p>These numbers are truly impressive and eye-opening in any organization so I encourage everyone to start gathering them!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=577fa1d63696" width="1" height="1" alt=""><hr><p><a href="https://detect.fyi/detection-visibility-metrics-577fa1d63696">Detection Visibility Metrics</a> was originally published in <a href="https://detect.fyi/">Detect FYI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Detection Researcher Mindset — DCSYNC T1003.006]]></title>
            <link>https://detect.fyi/a-detection-researcher-mindset-dcsync-t1003-006-2532bddefbf5?source=rss-7b5ee9155ea7------2</link>
            <guid isPermaLink="false">https://medium.com/p/2532bddefbf5</guid>
            <category><![CDATA[active-directory-security]]></category>
            <category><![CDATA[detection-engineering]]></category>
            <category><![CDATA[threat-detection]]></category>
            <category><![CDATA[threat-hunting]]></category>
            <category><![CDATA[blue-team]]></category>
            <dc:creator><![CDATA[Scott Plastine]]></dc:creator>
            <pubDate>Fri, 27 Mar 2026 15:07:59 GMT</pubDate>
            <atom:updated>2026-03-27T17:51:52.231Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*F_J4f4tcdY4h4rBhYzJMXg.png" /><figcaption>Gemini drew this in one take and it’s pretty good so I went with it</figcaption></figure><p>This is part 2 of the Detection Researcher Mindset series. If you have not seen the <a href="https://medium.com/@plantainstan/a-detection-researcher-mindset-f2ed045480c5">first post</a> where I discuss the process then start there. In this post I will walk through an example where we will develop a detection for a Remote <a href="https://attack.mitre.org/techniques/T1003/006/">DCSync attack</a> following those steps.</p><p>As a refresher here is the process from the previous post.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*D0wqQTa8Zu3gymlF" /></figure><h3>Understand the attack</h3><p><em>1. What is the goal of the attack technique?</em></p><p>Obtain password hashes from a domain controller. From there an adversary can impersonate that user and even create a <a href="https://www.huntress.com/cybersecurity-101/topic/what-is-golden-ticket-attack">golden ticket</a> if they replicate the krbtgt password hash.</p><p><em>2. Why does it work?</em></p><p>If an attacker has control of an account with replication permissions they can instruct a domain controller to sync password hashes to their system using special tools.</p><p><em>3. At what phase of the attack would it be leveraged?</em></p><p>This is the credential access phase of <a href="https://attack.mitre.org/techniques/T1003/006/">Mitre ATT&amp;CK</a> so generally an attacker would have a foothold within the environment and has elevated their privileges significantly.</p><p>If they have the account privileges to perform a DCSync attack they have quite a bit of access within an environment already. A successful DCSync would allow them to have even more resilient access into the environment depending on the sensitive data they replicate from the Domain Controller. Let’s assume they would go for the gold since they may not get that shot again. The DCSync as we will see later, regardless of what they request will generate the same logging evidence to detect them.</p><p>After successfully grabbing the krbtgt account password hash they would be able to forge tickets and move laterally pretty quickly within the environment. They would still want to evade detection since their mission is not yet complete.</p><p><em>4. What artifacts would an adversary require to perform this attack?</em></p><ul><li>They would need access to a domain controller from a host they could run specific attack tools from.</li><li>They would need an account with replication rights which is generally but not always domain admin level credentials.</li></ul><p><em>5. Does this attack work in our organization?</em></p><p>Yes, we have 2 domain controllers in the lab that perform replication. As you can see below the attack worked just fine and we were able to grab the krbtgt hashes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*3LV-d4hR325bWwdan8-MiA.png" /><figcaption>Mimikatz performing a DCSync for the krbtgt account hash from the Win10 system</figcaption></figure><p>This was run from a Windows 10 VM in the lab with defender services disabled. <strong>Remember, we are not trying to detect mimikatz.</strong> We do not want to focus on the tool that is used by the adversary since we are detecting the malicious DCSync technique itself. If the detection is limited to the specific tool then an adversary could run a custom attack tool that we would likely not detect. We want to be at the top of the <a href="https://center-for-threat-informed-defense.github.io/summiting-the-pyramid/">pyramid</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/801/0*bMlXwZfN0TWIvRhG" /><figcaption>Simple lab design</figcaption></figure><h3>Understand normal system and user behavior</h3><p><em>1. How is this service used normally? What is it used for?</em></p><p>Domain controllers will replicate directory changes to each other constantly (new groups, password updates, etc). Typically an enterprise will have a few to hundreds of domain controllers in different regions to ensure redundancy and good response times. Larger enterprises will even have multiple domains. A distributed infrastructure is important so logins to your workstation in the morning don’t take forever while you authenticate to the domain controller on the other side of the world. Below is a very basic architecture showing domain controllers in different regions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/0*JxrLQYJTUMpq6L5R" /></figure><p><em>2. Is this “exploited” feature needed in the enterprise? And why?</em></p><p>DCSync is a critical function if an organization has more than 1 domain controller. If you’re in the rare situation where there is only 1 Domain Controller then any DCSync activity would be suspicious!</p><p><em>3. Is this behavior outside of a standard process?</em></p><p>For our DCSync example knowing what’s normal will be critical to creating a detection that is not noisy as DCSync’s occur regularly in healthy environments. It may not be a standard process since it’s a windows function but there is a way to identify typical DCSync’s. That is, it will always originate from a Domain Controller!</p><p><em>4. Do administrators or power users perform this operation?</em></p><p>Very high privileges (domain admin+) are required to perform this operation.</p><p>The permissions specifically required to perform a DCSync are these [<a href="https://threathunterplaybook.com/library/windows/active_directory_replication.html">reference1</a>, <a href="https://www.ired.team/offensive-security-experiments/active-directory-kerberos-abuse/dump-password-hashes-from-domain-controller-with-dcsync">reference 2</a>]:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/489/1*QRm3ARQJMVbF1b6PHNtGkg.png" /><figcaption>Extended Rights</figcaption></figure><p><a href="https://happycamper84.medium.com/who-can-execute-dcsync-in-your-ad-environment-495f704eded0">Here’s</a> a great post with a script on how to query which users/computers in your environment have these rights. In my lab here are the users/computers that can perform a DCSync.</p><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*USHJSAQAWCsp_ci2_aukAQ.png" /></figure><p>It’s important to limit the amount of users/computers that can perform a DCSync but that will not affect the final detection. Controlling these privileges is beneficial and will make it more difficult for an adversary to compromise an account that can perform a DCSync.</p><h3>Visibility into the attack</h3><p><em>1. Which steps from the attack can we see within telemetry?</em></p><p>We’d start by looking at which events occur when there’s a DCSync. It’s not clearly apparent from the logs however the internet has the answer to at least get started.</p><p><a href="https://attack.mitre.org/detectionstrategies/DET0594/">Mitre: Detection of Unauthorized DCSync Operations via Replication API Abuse</a></p><p><a href="https://threathunterplaybook.com/library/windows/active_directory_replication.html">Threat Hunter Playbook: Active Directory Replication</a></p><p>Per these sources it’s pretty clear that the event we need to start with is <a href="https://learn.microsoft.com/en-us/previous-versions/windows/it-pro/windows-10/security/threat-protection/auditing/event-4662">Event Code 4662</a>.</p><p>Lets look in our lab environment to see what happens when a DCSync occurs then break down all details of the event. The Microsoft link and the few others posted above help to describe the event. Red team blogs also are very helpful when digging into the details of events and specific objects within AD.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/783/1*Q6ZPiwpGfFji-U3BQb6Zdg.png" /><figcaption>Event Code 4662</figcaption></figure><p>Section 1 Event:</p><p>Event Code 4662 — An operation was performed on an object.</p><p>ComputerName — where the event occurred, it is important to know which DC fulfilled the DCSync request.</p><p>Section 2 Subject:</p><p>The operations were performed by accounts with the privileges to do so which makes sense. That does not tell us however if the action was authorized.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xoEkLz3ekumNqPK58TpFdw.png" /><figcaption>Summary of accounts performing 4662 activity in the lab</figcaption></figure><p>Section 3 Object:</p><p>The object server is always <em>DS</em> per the Microsoft <a href="https://learn.microsoft.com/en-us/previous-versions/windows/it-pro/windows-10/security/threat-protection/auditing/event-4662">documentation</a>. I suspect it stands for Domain Services.</p><p>Object type is the SchemaGUID for the <a href="https://learn.microsoft.com/en-us/windows/win32/adschema/c-domaindns">DomainDNS class</a>.</p><p>Object name is the objectGUID for the domain.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/425/1*-tqWgMqjt5dy86e8tmKkNg.png" /><figcaption>Object Name</figcaption></figure><p>Section 4 Operation:</p><p>The type of access will be important. The Microsoft 4662 documentation lists the different access masks. The statement below also makes sense since we know DC replication is an Extended right.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/864/1*61Usm8bGxnIL8Ej41hn8nQ.png" /><figcaption>Control Access Mask</figcaption></figure><p>Then finally the access right we saw earlier is in the event! Below is a LDAP query you could run to grab them. These are critical to incorporate into the detection.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*GvFGOxUsOLwwdtSm8zcfvA.png" /><figcaption>Extended Rights</figcaption></figure><p>As you have noticed nothing in this event shows the system where the DCSync request came from. This is why it’s important to break down every part of a log. There is however the Login ID which is used to map different windows events to the same session. What if we look within the analytics platform to find other events associated with that Login ID?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/784/1*Cmcee6Rnf8bQ45HKDKkIaQ.png" /><figcaption>Associated 4624 Logon Event</figcaption></figure><p>So this log shows the network address the logon came from which is very interesting. Also remember that Logon IDs are unique to each host system so there could be collisions across many domain controllers.</p><p><em>2. Does this telemetry exist in your organization?</em></p><p>Lets recap what important artifacts we have collected so far.</p><ul><li>Event Code 4662 from Domain Controllers — shows a DCSync has occurred</li><li>Event Code 4624 from Domain Controllers — correlated by the Logon ID it shows the source IP address of the DCSync request.</li></ul><p>All of these events are within our lab and very likely available within large enterprises. Yay!</p><h3>Is prevention possible?</h3><p>In this case we cannot prevent this attack. There will always be accounts with replication privileges within an organization. These can be limited and guarded (and should be) however they do need to exist.</p><p>Endpoints need access to domain controllers so that opens the possibility for an attacker to compromise a system to access a domain controller. If we could lock a DC in a closet without network access it may be safe but also it wouldn&#39;t be useful to the organization.</p><h3>Isolate malicious behavior</h3><p>At this point after digging into these events it may be very apparent how we plan on isolating malicious behavior from normal.</p><ul><li>Domain controllers sync between each other.</li><li>A DCSync from an IP address that isn’t a domain controller would be a concern!</li></ul><p>Let’s update our list of requirements for a detection.</p><ul><li>Event Code 4662 from Domain Controllers — shows a DCSync has occurred</li><li>Event Code 4624 from Domain Controllers — correlated by the Logon ID/host it shows the source IP address of the DCSync request.</li><li>List of the IP Addresses of all Domain Controllers per Domain</li></ul><h3>Document</h3><p>Documentation is very important but that will double the size of this blog post so I have the luxury of skipping that here (but a good Detection Researcher does not!). Refer to the <a href="https://medium.com/@plantainstan/a-detection-researcher-mindset-f2ed045480c5">first post</a> on a template that can be used. Our notes up until this point would be the majority of what is needed anyway.</p><h3>Development</h3><p>All of the details and requirements from the research phase will feed into the development phase which allows for a modular detection engineering process which enables flexibility in fast paced environments. Also, the developer does not need to be the same resource as the researcher since the skill sets for each are slightly different.</p><p>We have the 3 requirements from above so lets break the query into 3 parts. In my lab I use Splunk so if you use something else then you can likely translate this to your platform.</p><p><em>Event Code 4662</em></p><p>This is straight forward since we are looking for the 4662 event codes that apply to the fields we identified above.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wFEh7yEuVMcNxCTchH6O8w.png" /><figcaption>Splunk query for 4662 Event Code</figcaption></figure><p>NOTE: The LIKE in the query looking for the rightsGuid in the entire Message field is not efficient. I’d recommend extracting the rightsGuid field from the event properly in Production.</p><p><em>Event Code 4624</em></p><p>This next step needs to be a bit fancier. We need to take all of the Logon_IDs from these 4662 events and look back through the events and correlate them to 4624 events. To do this in Splunk we need to use a <a href="https://help.splunk.com/en/splunk-enterprise/get-started/search-tutorial/10.2/part-4-searching-the-tutorial-data/use-a-subsearch">sub-search</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fq58MT9GUuqZa0X4G9rIRQ.png" /><figcaption>Splunk query for 4662 correlated with 4624 Logon ID/Host</figcaption></figure><p><em>Domain Controller IP Address List Lookup</em></p><p>There are many ways to grab a Domain controller list. I chose to use this neat <a href="https://splunkbase.splunk.com/app/1151">ldapsearch</a> app from Splunk. I can query LDAP for the configuration parameters for a Domain Controller. Then we will plug the output into a lookup to be used in the search logic for the detection.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6TkHelLcY-zLBT1wmOuZ4Q.png" /><figcaption>DC Lookup List</figcaption></figure><p>Let’s first break down the LDAP query. I used this presentation for the <a href="https://fintech.global/cybertechforum/wp-content/uploads/2020/11/ALSID-ACADEMY-EPI-8-DCShadow.pdf">DCShadow attack</a> heavily. That presentation outlined research by Benjamin Delpy and Vincent Le Toux identifying core SPNs for a DC to have.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/838/1*YwHYUqnXAuC6j_-pVfe8LA.png" /></figure><p>In addition to that the Domain Controllers reside within the Domain Controllers group which has the <a href="https://learn.microsoft.com/en-us/windows-server/identity/ad-ds/manage/understand-security-identifiers">primaryGroupID of 516</a>. As you can see above the LDAP query checks those parameters. This detection is not meant to discover a DCShadow attack so if a rogue DC is added to the domain this detection would not fire. I will add some notes about that at the end of this post.</p><p>LDAP returns the dNSHostName so we then instruct Splunk to query DNS for the client IP address as well.</p><p>That info is added to a lookup table dc.csv within Splunk. More details could be added to the lookup to make this more robust such as time added and time removed (DC is decommissioned).</p><p>Finally, this should be scheduled as a report that can run prior to the alert running. If the alert runs every hour on the hour then this report can run every hour at 55mins so the lookup has the most up to date data.</p><p><em>Detection Query including the DC IP Lookup</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mmvR5cDYPGGh0o6EhfekTQ.png" /></figure><p>Now we’ve added the lookup and you can see on the right hand column the IP address is matching the DC hostname from the lookup.</p><p><em>Detection Query Filtering out known Domain Controllers — Remote DCSync Detection</em></p><p>All that’s left is to remove known DCs and the suspicious DCSyncs from my Win10 (192.168.1.139) system where I ran mimikatz from will remain.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZKf3yAS5UPLeI17Yq2TPDA.png" /><figcaption>PANIC!</figcaption></figure><h3>Final Thoughts</h3><p>We’ve built a query that will identify Suspicious Remote DCSync events leveraging generally available logs within a large enterprise. There is still more that could be done to make this even better but I think this is a good walk through for applying a Research Mindset to discover innovative detections.</p><p>I had mentioned that to identify a Domain Controller I used documentation from a DCShadow Attack. This detection is not meant to detect a DCShadow Attack so does that mean this is all useless? NOPE! <strong>With this detection we’ve successfully raised the bar for an attacker to go unnoticed.</strong> Adversaries and red teams have tight timelines too and forcing them to fit more in makes it less likely they will succeed in their mission. And we can add the <a href="https://attack.mitre.org/techniques/T1207/">DCShadow attack</a> to the Detection Research backlog to dig into next!</p><p>There are 2 other more rare but feasible scenarios (if an adversary gains access to a DC) that should be added to the backlog as well since at some point I have to end this post!</p><ul><li>Running a DCSync from the DC itself against itself</li><li>Running a DCSync from the DC itself against a different DC.</li></ul><h3>Sources and Additional Reading</h3><p><a href="https://threathunterplaybook.com/library/windows/active_directory_replication.html">https://threathunterplaybook.com/library/windows/active_directory_replication.html</a></p><p><a href="https://www.ired.team/offensive-security-experiments/active-directory-kerberos-abuse/dump-password-hashes-from-domain-controller-with-dcsync">https://www.ired.team/offensive-security-experiments/active-directory-kerberos-abuse/dump-password-hashes-from-domain-controller-with-dcsync</a></p><p><a href="https://fintech.global/cybertechforum/wp-content/uploads/2020/11/ALSID-ACADEMY-EPI-8-DCShadow.pdf">https://fintech.global/cybertechforum/wp-content/uploads/2020/11/ALSID-ACADEMY-EPI-8-DCShadow.pdf</a></p><p><a href="https://learn.microsoft.com/en-us/previous-versions/windows/it-pro/windows-10/security/threat-protection/auditing/event-4662">https://learn.microsoft.com/en-us/previous-versions/windows/it-pro/windows-10/security/threat-protection/auditing/event-4662</a></p><p><a href="https://learn.microsoft.com/en-us/windows/win32/adschema/c-domaindns">https://learn.microsoft.com/en-us/windows/win32/adschema/c-domaindns</a></p><p><a href="https://www.ultimatewindowssecurity.com/securitylog/encyclopedia/event.aspx?eventid=4662">https://www.ultimatewindowssecurity.com/securitylog/encyclopedia/event.aspx?eventid=4662</a></p><p><a href="https://adsecurity.org/?p=3377">https://adsecurity.org/?p=3377</a></p><p><a href="https://learn.microsoft.com/en-us/windows-server/identity/ad-ds/manage/understand-security-identifiers">https://learn.microsoft.com/en-us/windows-server/identity/ad-ds/manage/understand-security-identifiers</a></p><p><a href="https://center-for-threat-informed-defense.github.io/summiting-the-pyramid/">https://center-for-threat-informed-defense.github.io/summiting-the-pyramid/</a></p><p><a href="https://attack.mitre.org/techniques/T1207/">https://attack.mitre.org/techniques/T1207/</a></p><p><a href="https://attack.mitre.org/techniques/T1003/006/">https://attack.mitre.org/techniques/T1003/006/</a></p><p><a href="https://happycamper84.medium.com/who-can-execute-dcsync-in-your-ad-environment-495f704eded0">https://happycamper84.medium.com/who-can-execute-dcsync-in-your-ad-environment-495f704eded0</a></p><p><a href="https://evilsaint.com/article/dsquery.html">https://evilsaint.com/article/dsquery.html</a></p><p><a href="https://blog.joeware.net/2018/05/25/5824/">https://blog.joeware.net/2018/05/25/5824/</a></p><p><a href="https://dirkjanm.io/abusing-forgotten-permissions-on-precreated-computer-objects-in-active-directory/">https://dirkjanm.io/abusing-forgotten-permissions-on-precreated-computer-objects-in-active-directory/</a></p><p><a href="https://github.com/gentilkiwi/mimikatz">https://github.com/gentilkiwi/mimikatz</a></p><p><a href="https://www.huntress.com/cybersecurity-101/topic/what-is-golden-ticket-attack">https://www.huntress.com/cybersecurity-101/topic/what-is-golden-ticket-attack</a></p><p><a href="https://www.zeropointsecurity.co.uk/course/red-team-ops">https://www.zeropointsecurity.co.uk/course/red-team-ops</a></p><p><a href="https://medium.com/30-days-of-red-team/30-days-of-red-team-day-25-dcsync-domain-admin-compromise-d34a557920db">https://medium.com/30-days-of-red-team/30-days-of-red-team-day-25-dcsync-domain-admin-compromise-d34a557920db</a></p><p><a href="https://tools.thehacker.recipes/mimikatz/modules/lsadump/dcsync">https://tools.thehacker.recipes/mimikatz/modules/lsadump/dcsync</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2532bddefbf5" width="1" height="1" alt=""><hr><p><a href="https://detect.fyi/a-detection-researcher-mindset-dcsync-t1003-006-2532bddefbf5">A Detection Researcher Mindset — DCSYNC T1003.006</a> was originally published in <a href="https://detect.fyi/">Detect FYI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Detection Researcher Mindset]]></title>
            <link>https://detect.fyi/a-detection-researcher-mindset-f2ed045480c5?source=rss-7b5ee9155ea7------2</link>
            <guid isPermaLink="false">https://medium.com/p/f2ed045480c5</guid>
            <category><![CDATA[threat-detection]]></category>
            <category><![CDATA[detection-engineering]]></category>
            <category><![CDATA[threat-hunting]]></category>
            <category><![CDATA[blue-team]]></category>
            <dc:creator><![CDATA[Scott Plastine]]></dc:creator>
            <pubDate>Fri, 27 Mar 2026 15:06:46 GMT</pubDate>
            <atom:updated>2026-04-11T19:57:42.135Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*I9V7A1S3HNZNg8XzyuP9xA.png" /></figure><p>As detection researchers we are frequently asked where do our detection ideas come from (and to build a backlog for them, and when will it all be done, etc.). At some point I needed to stop referencing Demetri Martins’ <a href="https://youtu.be/5CYPee5bi6U?t=903">stand up</a> where he describes how his jokes are delivered by a delicate fairy from a magical shire (the AI drawing may make more sense now…or not).</p><p>It’s important to think through where these ideas come from and how they are formed so we have the time, resources and the correct mindset to continue to extract more meaning from the data.</p><p>I have outlined a high level thought process and throughout this post there are questions to ask at each phase. I’ll share some insights on why these questions are useful when it’s not extremely obvious. Part 2 of this post walks through this process using the DCSync attack example <a href="https://detect.fyi/a-detection-researcher-mindset-dcsync-t1003-006-2532bddefbf5">here</a>.</p><h3><strong>The Starting Point</strong></h3><p>This is often the most difficult piece to identify since there are so many choices! We can keep it simple for now and say the starting point is an attack since that is what we’re usually attempting to detect. We can break it down like this: <br> <br>Attack (X) was performed against a Service (Y) creating a disturbance (Z)</p><p>Identifying and prioritizing detection research leads is out of scope for this post but for now let’s assume we have an attack we would like to detect.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*MMPiWxMXpMvRunEG" /></figure><h3><strong>Understand the attack</strong></h3><p>If we do not know how and why the attack works we cannot build a reliable detection. Also, understanding the attack is fun!</p><p><em>1. What is the goal of the attack technique?</em></p><p>This is where you identify the specific reasons for an adversary to perform the attack. There could be many depending on the attack.</p><p><em>2. Why does this technique work?</em></p><p>This is where you identify specifically why this attack works. It could be due to a specific configuration setting, vulnerability or software feature. This analysis can shed light into how to prevent this attack too.</p><p><em>3. At what phase of the attack would this technique be leveraged?</em></p><p>This is excellent background so we know when an adversary would attempt to use this technique from their toolkit.</p><p><em>4. What artifacts would an adversary require to perform this attack?</em></p><p>An adversary requires information to perform attacks. That could be user credentials, computer certificates or even just a target IP address. Knowing this is critical when understanding how to detect their attacks but also knowing what they are capable of at certain points during an incident. This research on the attack and our own environment will become the basis of the documentation that will help educate others quickly if this alert were to trigger.</p><p><em>5. Does this attack work in our organization?</em></p><p>It’s important to ask this since if the attack doesn&#39;t work within our organization then we can possibly be done here and focus on higher priorities.</p><p>For example if this attack requires MacOS and we don’t have any then well there are probably better things we can do with your time.</p><p>However, in some cases it may still be valuable to build a detection. In the cases where it is a no, would an adversary realize the attack will not work and still attempt it? An example could be <a href="https://attack.mitre.org/techniques/T1558/004/">AS-REP Roasting</a>. Your organization may not have any AS-REP roastable accounts but if an attacker kicks off a noisy scan you may still be able to detect them. The ideal situation is that an attack fails and the attacker is detected. They don’t always have to be caught trying something that actually works!</p><h3>Understand normal system and user behavior</h3><p><strong>Understanding normal is the most overlooked component of detection engineering</strong>. In fact I believe it’s the key to being a competent Detection Engineer. Life is chaotic keeping up with every attack ever and not all attack chains are publicly shared (think well funded nation states). But if you know your own environment and what normal looks like (in telemetry) then you can recognize when something is out of place. Your spidey sense will be like <em>I don’t know what this is</em> and that’s a concern. Then you will dig in to determine what you’re looking at. In some cases if it turns out to be nothing then you’ve just learned even more about your own environment which is always a positive.</p><p>When building detections in an isolated lab it’s not too difficult to find attacker behavior. What’s tricky is isolating that behavior from normal, or even unusual but benign user behavior in a functioning enterprise environment. To add to the complexity there will generally be data gaps within the enterprise versus a lab. To successfully build a robust detection that’s difficult to evade we must understand what normal looks like.</p><p><em>1. How is this service used normally? What is it used for?</em></p><p>Its always important to know the “boring” stuff. Which is why does this service exist, how is it intended to be used and how is it used within my environment since that can be different too.</p><p><em>2. Is this “exploited” feature needed in the enterprise? And why?</em></p><p>In some cases if a feature isn’t needed at all we can disable it. Sadly, that’s not too common but it’s a good question to ask. If at this point you realize this feature isn’t needed at all then that is covered in the “<strong>Is prevention possible”</strong> section.</p><p><em>3. Is this behavior outside of a standard process?</em></p><p>When learning what is normal it’s important to know when this activity occurs within the enterprise. A good example would be user account creation. Is there a standard process for all account creation in the enterprise? Generally, its never this simple but the goal would be to know all of the ways some action occurs so you can say anything outside of this process is either not someone following enterprise standards or it is in fact malicious. <strong>This creates process choke points attackers have to go through to potentially remain undetected.</strong> This is also a great opportunity to make pals with the Identity and Access teams. Understanding those systems and process will be very important since their mission overlaps with detection engineering!</p><p><em>4. Do administrators or power users perform this operation?</em></p><p>It’s important to think about who is authorized to do this. If it’s a small subset of the environment those are very useful details when building a detection or preventing this activity outright.</p><h3>Visibility into the attack</h3><p>To detect an attack we’d require visibility from a logging source. In some cases an attack will only be visible in a lab environment and not at scale in an enterprise. <strong>We do want to focus on telemetry that is accessible within an enterprise but identifying artifacts left behind (that we cannot feasibly collect at scale) is still valuable for learning or during an incident where forensic artifacts can be grabbed manually.</strong></p><p><em>1. Which steps from the attack can we see within telemetry?</em></p><p>From part 1 (Understanding the attack) and part 2 (Understanding normal) we should have a good high level understanding on what occurs during the attack so next we can make a list of each step to attempt to identify that within the telemetry. The process diagram shows arrows going between each of the 3 parts (See picture below).<strong> During complex attacks you’ll be going back and forth learning more about the attack, then whats normal, then back to visibility and then back again until you fully understand all parts.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rJ97ueMXeDy-uMEf4OMYHQ.png" /><figcaption>Keep revisiting each step to fully understand all parts</figcaption></figure><p>This step may also depend on what testing environments you have access to. Ideally, you can perform it in a lab environment with all logging (even more than within your organization) enabled. Then you’ll have a full picture and understanding of what occurred. At some point it would be valuable to perform the attack in the enterprise but understanding what happens in a lab would allow for you to learn the attack prior to jumping into the deep end. As we had discussed before the enterprise will have data gaps and more un-explainables to deal with. It’ll be important to understand all of this but it’s nice to not have to drink from the fire hose immediately. In addition to that it generally requires approvals and assets to test attacks in an enterprise environment which adds to the preparation time.</p><p><em>2. Does this telemetry exist in your organization?</em></p><p>If you are aware of what telemetry occurs when the attack happens you can go looking in your log analytics platform to see if it’s there and see whats missing.</p><ul><li><em>Can it be acquired at scale?</em></li><li><em>Can it be acquired locally?</em></li></ul><p>If we cannot query the logs for a detection then we cannot build a detection which is pretty straight forward. If the logs are not available it’s important to consider how they could be acquired. <strong>This research to detect an attack could be the evidence needed to get these logs ingested into a log analytics platform.</strong></p><p>Let’s take an extreme example which would be Operating System API calls. It is highly unlikely to be able to acquire these at scale. Even <a href="http://www.rohitab.com/apimonitor">monitoring</a> them locally can cause a system to grind to a halt. Active Directory events for example are noisy but could still be justified to collect with how valuable they are.</p><p>Powershell Events would be expensive to log centrally for large organizations however if your research produces some good artifacts that would appear in powershell events there could be a push to get them enabled across the organization so they could be used if there was an incident. They can be logged and sit on an endpoint if needed. They would not be retained forever but still its better than not logging them at all.</p><h3>Is prevention possible?</h3><p>Detections are wonderful but stopping the attack path is better! <strong>Detections cannot be the solution for everything or we will drown in alerts.</strong> Pushing for prevention can be a long term plan but it’s important to dig into this angle as well. Enterprises have many prevention tools at their disposal so it’s very important to understand how they work and how to leverage them.</p><p>A good example of this is running a browser in debug mode. The general user population would not need to do this ever and this feature is required to steal session cookies from the <a href="https://www.fortisec.co.uk/blog/extracting-cookies-chromium-remote-debugger-using-ttpforge/">browser</a>. In this case we could check endpoint telemetry to determine if this is used within the enterprise (or Enterprise Security Standards 😊 ). It’s unlikely that its needed for the vast majority of the enterprise so this could possibly be a good candidate to disable via <a href="https://learn.microsoft.com/en-us/deployedge/microsoft-edge-browser-policies/remotedebuggingallowed">GPO</a>. Anyone who needs it for testing can request it and then you’ve significantly reduced the attack surface.</p><p>This is an oversimplification of this attack but for more details refer to this awesome <a href="https://specterops.io/blog/2025/08/27/dough-no-revisiting-cookie-theft/">SpecterOps blog post</a>.</p><h3>Isolate malicious behavior</h3><p>Now that we have all the pieces we can attempt to isolate the malicious behavior.</p><p><em>1. Is this a rare event itself?</em></p><p>If this is rare then this is a simpler one, we can just develop the alert looking for the rare events.</p><p><em>2. Is there a rare sequence of events that occurs?</em></p><p>This is more advanced and how we can really reduce benign events. Looking at the telemetry identify multiple rare-ish steps that’d need to occur for this attack to be successful. This would require a way to track the sequence of events. Some log analytics platforms have methods to do so. This is how you would do that with <a href="https://www.splunk.com/en_us/blog/security/risk-based-alerting-splunk-sequenced-events.html">Splunk</a>.</p><p><em>3. Are there multiple suspicious events that could be correlated regardless of order?</em></p><p>My <a href="https://medium.com/detect-fyi/detection-pipeline-maturity-model-076984779651">previous post</a> about the detection pipeline is relevant here since a risk engine would be a method to correlate and implement a risk score for multiple suspicious events. These would be events that are suspicious but the sequence does not matter.</p><p><em>3. Can statistics be used?</em></p><p>Let’s take a simple brute force password spray. We can look for a high failure ratio from a specific IP. The failure ratio is important since aggregators or morning logins by users can create spikes in fails but when looking over the activity window it really isn’t anything spectacular.</p><ul><li>300 login failures, 800 login successes over a 10min time span which equates to a 27.3% failure rate— nothing special, I forget my passwords all the time too</li><li>300 login failures, 1 login success over 10min time span which equates to a 99.6% failure rate — this is more suspicious</li></ul><p><em>4. How resilient is this detection to evasion?</em></p><p>This is important to understand so we can communicate how effective a detection is. As long as everyone is on the same page for the fidelity of the detection (and even better it’s tracked within Detection Metrics) then we are good. The goal is to be at the top of the <a href="https://center-for-threat-informed-defense.github.io/summiting-the-pyramid/">pyramid</a> but depending on the situation that isn’t always the best approach.</p><p>Let’s say we have an incident with malware in our environment and the malware uses a unique filename/location combination. We could create an alert temporarily for that pretty quickly. We would not want to tell leadership we’ll be able to Reverse Engineer the binary in 6 months then build the best detection ever. We’d first implement the basic detection to aid in response then work on the more permanent detection.</p><p>Even if we build the best detection feasible it&#39;ll be important to understand evasion. When a detection is evade-able it may still be valuable and would still raise the bar for attackers.</p><h3>Document</h3><p>After all this experimenting and simulating it&#39;ll be critical to document all of this for others to learn from and build upon. At the end of this research journey it may not even be feasible to isolate the malicious activity so a detection would not be built. In those scenarios especially it’s important to document all our research and the insights gained. A detection opportunity may come to mind later or by another teammate as the environment matures. For documentation this format can be used as well.</p><ol><li>Attack Details</li><li>Attack Telemetry</li><li>Service Background and Usage</li><li>Prevention opportunities (explored)</li><li>Malicious Activity</li><li>Detection Gaps</li></ol><h3>Development</h3><p>The final step if a detection concept is identified would be to proceed to development!</p><p>If you’ve made it this far I have walked through these steps in the companion post developing a Remote DCSync detection <a href="https://medium.com/@plantainstan/a-detection-researcher-mindset-dcsync-t1003-006-2532bddefbf5">here</a>.</p><h3>Sources &amp; Additional Reading</h3><p><a href="https://www.splunk.com/en_us/blog/security/risk-based-alerting-splunk-sequenced-events.html">https://www.splunk.com/en_us/blog/security/risk-based-alerting-splunk-sequenced-events.html</a></p><p><a href="https://specterops.io/blog/2025/08/27/dough-no-revisiting-cookie-theft/">https://specterops.io/blog/2025/08/27/dough-no-revisiting-cookie-theft/</a></p><p><a href="https://attack.mitre.org/techniques/T1558/004/">https://attack.mitre.org/techniques/T1558/004/</a></p><p><a href="http://www.rohitab.com/apimonitor">http://www.rohitab.com/apimonitor</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f2ed045480c5" width="1" height="1" alt=""><hr><p><a href="https://detect.fyi/a-detection-researcher-mindset-f2ed045480c5">A Detection Researcher Mindset</a> was originally published in <a href="https://detect.fyi/">Detect FYI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Detection Pipeline Maturity Model]]></title>
            <link>https://detect.fyi/detection-pipeline-maturity-model-076984779651?source=rss-7b5ee9155ea7------2</link>
            <guid isPermaLink="false">https://medium.com/p/076984779651</guid>
            <category><![CDATA[blue-team]]></category>
            <category><![CDATA[threat-detection]]></category>
            <category><![CDATA[threat-hunting]]></category>
            <category><![CDATA[detection-engineering]]></category>
            <dc:creator><![CDATA[Scott Plastine]]></dc:creator>
            <pubDate>Thu, 05 Mar 2026 01:01:40 GMT</pubDate>
            <atom:updated>2026-03-24T02:49:01.495Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_5E1PYFQ-BEOosJsPIftRg.png" /></figure><p>High fidelity detections are critical but if they get lost in the noise of other alerts they may get missed or not actioned appropriately. A Detection engineering team is required to build a system to support the monitoring of “everything” so its important to have a robust pipeline. When there is a logical and intuitive place for each type of detection or data source rolling out detections will be quicker.</p><p>There should be a place for your custom built high fidelity rules and the noisy closed source signature tool that was recently acquired without your knowledge or input. We want the data sources and analytics we have the most confidence to rise to the top while the other noise machines can continue to be monitored without clogging the pipeline.</p><p>The two types of data sources for a detection pipeline are security tools and telemetry. Closed source signature security tools have built-in detections provided by the vendor and in some cases they also offer the ability to write complex custom detections within them. There are open source signature security tools but sadly they are becoming more rare. Security tools look at the telemetry they are designed to collect to discover malicious activity within it.</p><p>Some examples of popular security tools are below:</p><ul><li>Endpoint — Elastic, CrowdStrike, Sentinel One, Microsoft Defender, etc</li><li>Application — F5, Fortinet, Akamai, etc</li><li>Email — O365 Sec &amp; Compliance, Proofpoint, Sophos, etc</li><li>Cloud — Azure Identity Protection, AWS Guard Duty, Google SCC, etc</li><li>Network — Cisco Firepower, Palo Alto, ZScaler, Netskope, etc</li></ul><p>Telemetry are events from technology indicating some activity occurred. Custom analytics are required to discover malicious activity within these logs.</p><p>Some examples of core telemetry sources are below:</p><ul><li>Endpoint — Windows Event Log, Powershell Events, Windows Registry, Linux Syslog, etc</li><li>Network —Web Proxy events, Netflow, DNS events, etc</li><li>Cloud — AWS Cloudtrail, Azure EventHub, etc</li><li>Access — Active Directory, Entra ID, etc</li><li>Email — O365, Email Flow, etc</li></ul><p>Here are some excellent telemetry frameworks to show the vast amount of event types to build detections from. These are great for mapping your data sources to better understand your current logging visibility.</p><ul><li><a href="https://www.osquery.io/schema/5.21.0/">OSQuery</a></li><li><a href="https://schema.ocsf.io/">OCSF</a> — Open Cybersecurity Schema Framework</li><li><a href="https://attack.mitre.org/datacomponents/">Mitre Data Components</a></li></ul><p>A large enterprise will have data sources from different environments. In general that is some variation of Production, Staging/QA, and Development environments. From a security perspective having malware or a threat actor in any environment is an issue. It would unwise to only monitor Production thinking that’s the most important data. An attacker could hang out in the Staging environment indefinitely finding ways to sneak into the production environment. If a staging environment is ransomware’d that would be a major problem for any enterprise as well. At a minimum production and staging environments should be monitored by security tools and custom analytics. Since there is generally more freedom in non-prod environments the sensitivity of the detections could be toned back a bit.</p><p><strong>Maturity Level — None</strong></p><p>Before walking through the full maturity model we will start with the first level which is no detection pipeline at all. This is when analysts are bouncing between security consoles looking at alerts. This is not sustainable and this method is highly unlikely to catch a sophisticated adversary.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*zK301mPx7-wWlgaD" /><figcaption>Very hectic, look at that unhealthy heart rate for the analyst</figcaption></figure><p>Pros</p><ul><li>Cheap and quick from a technology perspective</li></ul><p>Cons</p><ul><li>Not Measurable — Security consoles retain alert logs for a limited amount of time so if an alert rolls over its gone. It would not be feasible to prove the appropriate alerts were actioned when working directly from consoles.</li><li>Manual Correlation — Sophisticated adversaries may trip some low severity alerts on different systems and in this scenario you would rely on an analyst to remember what they saw before. If there are multiple analysts across different regions and time zones then there are even less chances of manual correlation occurring.</li><li>Inconsistent prioritization — The analyst will need to prioritize each console based on their experience leading to inconsistency and favoritism of specific security tools.</li><li>Mistake prone — Alerts from security consoles are manually exported and placed within a case management system which will be done differently by every analyst and mistakes and data loss are likely to occur.</li><li>Manual Tuning — Any alert tuning or more likely suppression would have to occur in each console manually. This does not give much flexibility to tune and the way to tune would rely on the console itself which makes it all even more complex.</li><li>Manual data enrichment — Adding context from other log sources is not feasible in this scenario increasing investigation times.</li><li>Limited custom detection capabilities — It would only be possible to write custom rules within the security tool itself which would be useful during an incident but not sustainable long term.</li><li>No Log analytics platform — At this stage it’s unlikely there is a central log analytics platform.</li></ul><p><strong>Maturity Level — Basic</strong></p><p>At the basic level the security tools are integrated with a central case management which is better than using the consoles separately to look at alerts. Some case management systems are able to perform alert aggregation and field normalization to at least provide a consistent alert structure for review and prioritization based on the vendors configuration. Investigation would likely need to be performed separately within another analytics platform and the security console itself.</p><p>An alternative basic setup would include sending alerts through an analytics platform but it would mostly be a pass through. This would still be an improvement since going beyond basic requires all data sources going through an analytics platform.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/953/0*psCaiAZrBaU5EUdz" /><figcaption>Improved but still mostly manual and hectic</figcaption></figure><p>Pros</p><ul><li>Data Retention — Alerts are retained for a longer period within the auditable case management system</li><li>Measurable — Alerts that were reviewed are auditable since they will always be sent to case management for investigation.</li><li>Single Pane of Glass — Analysts are not jumping around in many consoles looking for new triggered alerts.</li><li>Basic prioritization — Some prioritization can occur based on the data source.</li></ul><p>Cons</p><ul><li>Minimal data correlation — Limited to basic correlation across multiple security tools which would not scale for custom analytics.</li><li>Manual data enrichment — Adding context from other log sources is not feasible in this scenario adding to longer investigation times.</li><li>Limited custom detection capabilities — It would only be possible to write custom rules within the security tool itself which would be useful during an incident but not sustainable long term.</li><li>Manual Tuning — Any alert tuning or more likely suppression would have to occur in each console manually. This does not give much flexibility to tune and tuning procedures and capabilities would be different in each console.</li></ul><p><strong>Maturity Level — Standard+ Architecture</strong></p><p>At this level you will have the core building blocks for a scalable pipeline that can continue to grow with more sophisticated detections. Let’s start with a high level view of the detection pipeline.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*XXNGdNsPQj4dIFqZ" /></figure><p><em>Data Sources</em></p><p>Everything no matter where it’s coming from is a data source. Microsoft defender for endpoint is a data source just like email flow logs are. They would both need to be on-boarded to a central data analytics platform with the available fields extracted and normalized.</p><p><em>Analytics</em></p><p>Analytics is the logic to identify malicious or suspicious activity within an environment. Analytics are both security tools or custom searches against telemetry. We will need to break it down one step further since validation will be different depending on the type of analytic we are using. These could also be called signatures or signals but to avoid the marketing mess of behavior vs signatures I’ll call them all analytics.</p><p><em>Security Tools — Closed Source Analytics</em></p><p>Closed source analytics are from security tools that are purchased by the enterprise and must be incorporated into the pipeline. Closed source security tool analytics are commercial security tools where the rule sets have not been vetted or tested within your environment. These are common in every environment which do catch malicious activity however validating detection coverage is cumbersome since the detection logic is hidden.</p><p><em>Security Tools — Open Source Analytics</em></p><p>Some security tools do share the detection logic that is running within them. This is wonderful since this gives the consumer the ability to validate detection coverage, modify these rules for your own environment and add new rules. With a complete understanding of how the rule will perform the detection engineering team is able to determine the appropriate priority if the rule ever fires.</p><p>Examples:</p><ul><li>Malicious file discovered on an endpoint</li><li>Log4j exploit attempt against an external web application</li><li>Custom Snort rule scanning network traffic</li><li>Script writing MZ header to a file</li></ul><p><em>Telemetry — Custom Analytics</em></p><p>Custom analytics are the rules to identify malicious or suspicious activity using the logs within your environment. These rules would typically run within a log analytics platform such as Splunk or Kibana.</p><p>Examples:</p><ul><li>Kerberoasting against Active Directory</li><li>High login attempt failure ratios against External Web Applications</li></ul><p><em>Enrichment</em></p><p>Enrichment adds useful context to the output of an analytic. For example if a malicious file is discovered on an endpoint the log may have the hostname with the username. Enrichment could add the department the user works in, the region, email address and domain of the endpoint. This context would be valuable for correlation, alerting thresholds and investigations.</p><p><em>Risk Engine</em></p><p>All detection analytics whether from telemetry or security tools pass through the risk engine. Within the risk engine many factors are considered to determine the risky events that will be bubbled up and forwarded to case management to action. Building a risk engine is out of scope for this post but at a high level it takes the output from every analytic and sums the risk scores for each by (at least) asset, user or threat object correlating all data sources. When those reach a threshold over a specific period of time it will trigger an alert to send to case management.</p><p>Below is an example table on how this could work. In an enterprise there would be more factors to consider such as attack technique and confidence in the data source.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*xopL8p8H33OwlSjG" /></figure><p>To continue with the same theme from the other diagrams the pipeline architecture would look like this.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*aZ1YPTm4J65QaLrH" /></figure><p><strong>Maturity Level — Standard</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ZCQdmLG4U4741Czr" /></figure><p>The standard level takes security tool analytics and feeds them through the risk engine. Its important to focus on the biggest wins first. While custom signatures are more fun (and in some cases more likely to detect an adversary) the security tools need to be addressed first. The enterprise made an investment and monitoring will be required.</p><p>Security tools have hundreds or more analytics built in so implementing monitoring for them does enable so much visibility. It’s critical that all security tools are filtered through a risk engine especially if the security tool has closed source analytics. This allows for aggressive correlation and scoring depending on the fidelity of the security tool.</p><p>In the maturity model this pipeline would replace the None and Basic levels. As an organization matures it’s rarely a quick lift and shift of everything but the end goal would be to ensure all detection data sources flow through the analytics platform first.</p><p><strong>Maturity Level — Advanced</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*j3jamYaVqQqq0L3I" /></figure><p>Advanced is where organizations will have the majority of their rules. The majority of the triggered alerts may still however originate from commercial rules within security tools. The advanced rules will be custom built and validated against attack simulations within your own environments so they will be higher fidelity. Advanced does not replace standard; instead it’s how custom rules are integrated into the pipeline along with commercial rules.</p><p><em>Atomic detections with telemetry</em></p><p>All rules go through the detection engine however some high fidelity signatures will be set to alert every time. They will still be correlated with lower severity signatures within the risk engine to collect other signatures if anything else triggered as well.</p><p>Examples:</p><ul><li>Script writing MZ header to a file</li><li>High password attempt failure ratios against External Web Applications</li></ul><p><em>Risk based detections with security tools (custom rules)</em></p><p>These are similar to other rules however diving into building rules within security appliances generally requires collaboration with security engineering teams and justifiably more scrutiny since a bad rule could take down an entire appliance or even worse entire network zones or halt critical business processes.</p><p>Examples:</p><ul><li>Snort rule looking for double SSL encrypted traffic</li><li>Email quarantine rules scanning email headers and bodies for targeted phishing</li></ul><p><em>Risk based detections with telemetry</em></p><p>Adversaries are attempting to blend in with normal users and perform actions that power users or administrators use day-to-day. In these situations the risk engine is critical for clustering many more common events within an enterprise. Alone these types of events are business as usual but as more are performed then the risk scores would rise. When they reach the alerting threshold they would trigger and be actioned.</p><p>Examples:</p><ul><li>Many discovery commands (ipconfig, dsquery, route, hostname, etc) performed on an endpoint</li><li>Many requests for RC4 encrypted service tickets against Active Directory from the same user</li><li>Many downloads of the same image (.gif, .jpg, etc) file name with different hashes</li><li>RDP tunneling with the same username across many systems</li></ul><p><strong>Maturity Level — Leading</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*vXFTaRWIytVRErgo" /></figure><p><em>Data science backed detections with telemetry</em></p><p>Some detections will require more horse power and special analytics to find outliers. These generally cannot be performed in a standard analytics platform which is OK! Not every platform can do everything. This is where some of the data crunching can be offloaded to another platform such as Databricks or Dataiku then results can be sent back.</p><p>Example:</p><ul><li>DLL Sideloading detection looking for outliers</li></ul><p><em>Deception</em></p><p>Finally, lets have some fun and set traps for adversaries. Having detections is great but adding deception into the mix will make it even more likely to discover an attacker in your environment or at least someone snooping around where they shouldn&#39;t. But be careful, it would be devastating and embarrassing if an adversary was able to turn our deceptions against us.</p><p>Examples:</p><ul><li>Incorrect credentials for a valid account placed within endpoint scripts</li><li>A juicy sounding service account with an SPN</li></ul><p><strong>Measuring Maturity</strong></p><p>Measuring your maturity would not mean only having detections in the leading category. It will be a mixture of Standard, Advanced, and Leading with the majority of the analytics being within Advanced.</p><p>The goal is to have measurable, reliable and validated detection coverage. It is very challenging for detection engineers to measure coverage, build new (where there are gaps) and tune detection sources with closed source rules. A solution with closed source analytics is maintained by the vendor and new rules get added and removed without the clients consent. It could have a rule one day then next week that alert could be disabled making it impossible to trully understand your detection coverage.</p><p>In addition to that sophisticated threat actors test their tool sets against commercial security tools until they can evade them. Its critical to have custom analytics adversaries are unaware of to have the best chance to discover them in your environment. Wonderful <a href="https://x.com/MalwareJake/status/1508512289623875592?t=Zgn7_-OyCP3S5RKU1cvKqQ&amp;s=19">post by malwarejake</a> making the same point.</p><p>As more advanced and leading type detections are rolled out we can reduce the reliance on unmeasurable analytics thus increasing the detection fidelity. The standard level analytics would not be removed since they do provide value but their risk thresholds can continue to be reduced if they are noisy which will result in an overall higher detection fidelity.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*LsoPAMoUlV9Ak3xz" /><figcaption>Reduce the reliance on closed source analytics!</figcaption></figure><p>Here’s a new Risk engine example table on what it could look like to reduce the risk thresholds for analytics that have not or cannot be validated.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*rEGcRBc4NZm8fwUK" /></figure><p>Within the table any analytic that has not been validated (default rating) gets their risk score adjusted. This could be done for many more factors as well to increase or decrease risk scores.</p><p>Asset1 — This asset did generate an alert but only after the rule was correlated with another event. There is only so much this example table can show but logic could also be applied to require that the second event comes from a different data source entirely. If two different data sources see suspicious activity on a system that generally should result in alert. This is a great method to bring down the noise when modifying the analytic is not feasible.</p><p>Event2/Event3 — These events are both rated Medium however have different risk scores. There isn’t enough detail here to explain the why nor is that the point. This is to show that there is freedom when applying risk scores and just because the severity is medium does not mean it will always have to be the same risk score.</p><p><strong>Bonus Hunting Opportunities</strong></p><p>To further validate your detection pipeline time should be allocated to validate risk thresholds are appropriate. The risk engine is a place where many rules with data enriched results are flowing and would be an excellent place for hunting to occur to discover anything that may be flying under the radar. Detection engineers should also be regularly reviewing and validating thresholds as well.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Eat74NaoIYRtbz0C" /></figure><p><strong>Sources and Additional Reading</strong></p><p><a href="https://medium.com/@nburns9922/building-a-detection-engine-part-1-what-is-a-detection-engine-e223119fad7e">https://medium.com/@nburns9922/building-a-detection-engine-part-1-what-is-a-detection-engine-e223119fad7e</a></p><p><a href="https://redcanary.com/blog/testing-and-validation/detection-validation/">https://redcanary.com/blog/testing-and-validation/detection-validation/</a></p><p><a href="https://lantern.splunk.com/Security_Use_Cases/Threat_Investigation/Implementing_risk-based_alerting">https://lantern.splunk.com/Security_Use_Cases/Threat_Investigation/Implementing_risk-based_alerting</a></p><p><a href="https://www.splunk.com/en_us/form/the-essential-guide-to-risk-based-alerting.html">https://www.splunk.com/en_us/form/the-essential-guide-to-risk-based-alerting.html</a></p><p><strong>Updates</strong></p><p>2026–03–23: Added a more appealing and fun AI (Gemini) picture of a detection pipeline for the featured image</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=076984779651" width="1" height="1" alt=""><hr><p><a href="https://detect.fyi/detection-pipeline-maturity-model-076984779651">Detection Pipeline Maturity Model</a> was originally published in <a href="https://detect.fyi/">Detect FYI</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>