<?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 Michael Wills on Medium]]></title>
        <description><![CDATA[Stories by Michael Wills on Medium]]></description>
        <link>https://medium.com/@showprove?source=rss-9c38d3a41cbb------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*dVUgMq8dU6VitWKJ3W40Iw.jpeg</url>
            <title>Stories by Michael Wills on Medium</title>
            <link>https://medium.com/@showprove?source=rss-9c38d3a41cbb------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 22:47:51 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@showprove/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[️ Part 5: FIPS 140–3 — The Final Boss]]></title>
            <link>https://medium.com/@showprove/%EF%B8%8F-part-5-fips-140-3-the-final-boss-1f0f2fde104b?source=rss-9c38d3a41cbb------2</link>
            <guid isPermaLink="false">https://medium.com/p/1f0f2fde104b</guid>
            <dc:creator><![CDATA[Michael Wills]]></dc:creator>
            <pubDate>Thu, 19 Feb 2026 21:49:04 GMT</pubDate>
            <atom:updated>2026-09-24T01:39:23.575Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3VPVO6YmUVeu7Mc14z8DGA.png" /></figure><p>At the top of the pagoda in <em>Game of Death</em>, Bruce Lee faces his final opponent.</p><p>This is not a contest of strength or speed.<br> It is a test of <strong>truth</strong> — what is real, what can be trusted, and what will fail under pressure.</p><p>In hardened Linux environments, this final test is <strong>cryptography</strong>.</p><p>And the master guarding this floor is <strong>FIPS 140–3</strong>.</p><p><strong>Why Cryptography Is the Final Layer</strong></p><p>Every other security control ultimately depends on cryptography:</p><ul><li>SSH relies on it</li><li>TLS relies on it</li><li>Disk encryption relies on it</li><li>Authentication relies on it</li><li>Integrity relies on it</li></ul><p>If cryptography is weak, misconfigured, or unvalidated, <strong>every layer below it is compromised</strong>, regardless of how well it is implemented.</p><p>FIPS 140–3 exists to answer a single critical question:</p><p><em>Can this cryptography be trusted?</em></p><p><strong>What FIPS 140–3 Actually Enforces</strong></p><p>FIPS 140–3 is not “strong encryption mode.”</p><p>It is <strong>cryptographic assurance</strong>.</p><p>It enforces:</p><ul><li>Approved algorithms only</li><li>Approved key lengths</li><li>Approved operating modes</li><li>Validated cryptographic modules</li></ul><p>When FIPS mode is enabled:</p><ul><li>Weak or deprecated algorithms are disabled</li><li>Non-validated crypto paths are blocked</li><li>Applications must use approved crypto providers</li></ul><p>This aligns directly with:</p><ul><li><strong>NIST SP 800–53 — IA, SC, CP families</strong></li><li>DoD encryption and key management requirements</li><li>Compliance mandates for classified and controlled systems</li></ul><p><strong>Validation vs Assumption</strong></p><p>A critical distinction in regulated environments:</p><p><strong>Encryption that is not validated is not trusted.</strong></p><p>Modern systems often support strong cryptography — <br> but unless it is <strong>FIPS-validated</strong>, it cannot be assumed compliant.</p><p>FIPS shifts security from:</p><ul><li>“We believe this is secure”<br> to:</li><li>“This has been tested, validated, and certified”</li></ul><p>That difference matters during:</p><ul><li>Audits</li><li>Incident response</li><li>Legal scrutiny</li><li>Mission failure analysis</li></ul><p><strong>Operational Impact</strong></p><p>Enabling FIPS mode:</p><ul><li>Is intentional</li><li>Is irreversible without reinstallation</li><li>Must be planned</li><li>Must be tested</li></ul><p>This is not a tuning option — it is a <strong>security posture decision</strong>.</p><p>In DoD environments, FIPS is not optional.<br> It is foundational.</p><p><strong>Defense-in-Depth Comes Full Circle</strong></p><p>At the top of the pagoda, FIPS reinforces every layer below it:</p><ul><li>Firewalld protects encrypted services</li><li>SELinux confines crypto-capable processes</li><li>Fapolicyd ensures crypto libraries are trusted</li><li>STIGs enforce compliant crypto configuration</li></ul><p>Without FIPS, the system may be secure — <br> but it cannot be <strong>certified as trustworthy</strong>.</p><p><strong>Closing Reflection</strong></p><p>In <em>Game of Death</em>, the final opponent is not the strongest.</p><p>They are the most deceptive.</p><p>FIPS ensures that your system’s trust mechanisms are <strong>provable</strong>, not assumed.</p><p><strong><em>Security that cannot be validated cannot be trusted.</em></strong></p><h3>Why This Series Exists</h3><p>Most cybersecurity failures don’t happen because one control was missing.</p><p>They happen because defense-in-depth quietly eroded.</p><ul><li>A layer wasn’t implemented</li><li>A layer was misconfigured</li><li>A layer was disabled “temporarily”</li><li>A layer was misunderstood</li></ul><p>Security rarely collapses in one dramatic moment.<br> It degrades — slowly, quietly, and predictably.</p><p>This series uses <em>Bruce Lee’s Game of Death</em> as a metaphor for something very real: attackers don’t win because they’re unstoppable. They win because resistance weakens at each step.</p><p>Defense-in-depth is about making every step harder than the last.</p><p>Each layer is a discipline.<br> Each discipline addresses a different failure mode.<br> Together, they form a system that is resilient, auditable, and ready for pressure.</p><p>Because in cybersecurity, the goal isn’t perfection.</p><p>It’s layered resistance.</p><p><em>This is Level 4 of a complete five-level compliance framework. Read the full book: </em><a href="https://www.amazon.com/dp/B0HK2GYZ35"><em>The 5 Levels of Defense-in-Depth</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1f0f2fde104b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ Part 4: STIGs — The Strategist Game of Death for Cybersecurity (Defense-in-Depth Series)]]></title>
            <link>https://medium.com/@showprove/part-4-stigs-the-strategist-game-of-death-for-cybersecurity-defense-in-depth-series-91cd553ead27?source=rss-9c38d3a41cbb------2</link>
            <guid isPermaLink="false">https://medium.com/p/91cd553ead27</guid>
            <category><![CDATA[govtech]]></category>
            <category><![CDATA[security]]></category>
            <category><![CDATA[sysadmin]]></category>
            <dc:creator><![CDATA[Michael Wills]]></dc:creator>
            <pubDate>Tue, 03 Feb 2026 22:17:50 GMT</pubDate>
            <atom:updated>2026-09-24T01:38:27.200Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3VPVO6YmUVeu7Mc14z8DGA.png" /></figure><h3>🧠 Part 4: STIGs — The Strategist Game of Death for Cybersecurity (Defense-in-Depth Series)</h3><p><strong>Introduction</strong></p><p>By the fourth floor of the pagoda, raw skill is no longer enough.</p><p>In <em>Game of Death</em>, this is where discipline, preparation, and strategy determine the outcome.<br> Every movement is calculated. Every mistake is punished.</p><p>In cybersecurity, this level belongs to STIGs. Not a tool. Not a daemon. <br> A doctrine.</p><h3>What STIGs Actually Are</h3><p>Security Technical Implementation Guides (STIGs) are often misunderstood as:</p><ul><li><strong>Checklists</strong></li><li><strong>Compliance paperwork</strong></li><li><strong>“Audit requirements”</strong></li></ul><p>STIGs represent codified operational lessons derived from:</p><ul><li><strong>Real incidents</strong></li><li><strong>Known attack patterns</strong></li><li><strong>Repeated system failures</strong></li></ul><p>They define how systems must be configured to behave predictably under stress.</p><h3>Configuration Is a Security Control</h3><p>Misconfiguration remains one of the leading causes of compromise.</p><p>STIGs address this by enforcing:</p><ul><li><strong>Secure service configurations</strong></li><li><strong>Strong authentication policies</strong></li><li><strong>Proper file permissions</strong></li><li><strong>Required auditing and logging</strong></li><li><strong>Disabled unnecessary functionality</strong></li></ul><p>From a control standpoint, STIGs align with:</p><ul><li><strong>NIST SP 800–53 — CM family</strong></li><li><strong>IA, AC, AU, and SI controls</strong></li><li><strong>POA&amp;M reduction initiatives</strong></li></ul><h3>Discipline at Scale</h3><p>STIGs are not about perfection — they are about consistency.</p><p>In DoD environments:</p><ul><li><strong>Systems are built by different teams</strong></li><li><strong>Maintained over long lifecycles</strong></li><li><strong>Operated under mission pressure</strong></li></ul><p>STIGs ensure that:</p><ul><li><strong>Security posture does not drift</strong></li><li><strong>Systems remain auditable</strong></li><li><strong>Risk decisions are intentional and documented</strong></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3VPVO6YmUVeu7Mc14z8DGA.png" /></figure><p>Automation (Ansible, OpenSCAP) transforms STIGs from static documents into living controls.</p><p>STIGs and Defense-in-Depth</p><p>STIGs reinforce every other layer:</p><ul><li><strong>Firewalld rules are standardized</strong></li><li><strong>SELinux remains enforcing</strong></li><li><strong>Fapolicyd policies are consistent</strong></li><li><strong>FIPS settings are preserved</strong></li></ul><p>Without STIGs, layers erode over time.</p><p>With STIGs, discipline is maintained even as systems change.</p><h3>Operational Reality</h3><p>STIG compliance does not eliminate risk — <br> it reduces unknown risk.</p><p>It creates an environment where:</p><ul><li><strong>Deviations are visible</strong></li><li><strong>Exceptions are documented</strong></li><li><strong>Audits validate posture, not scramble for evidence</strong></li></ul><h3>Closing Reflection</h3><p>In <em>Game of Death</em>, the strategist doesn’t fight harder — <br> they ensure the opponent never gains advantage.</p><p>STIGs do the same for your systems.</p><p>Discipline before deployment prevents chaos after compromise.</p><p><em>This is Level 4 of a complete five-level compliance framework. Read the full book: </em><a href="https://www.amazon.com/dp/B0HK2GYZ35"><em>The 5 Levels of Defense-in-Depth</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=91cd553ead27" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[⚔️ Part 3: Fapolicyd — The Weapons Master]]></title>
            <link>https://medium.com/@showprove/%EF%B8%8F-part-3-fapolicyd-the-weapons-master-0456f2f5160f?source=rss-9c38d3a41cbb------2</link>
            <guid isPermaLink="false">https://medium.com/p/0456f2f5160f</guid>
            <dc:creator><![CDATA[Michael Wills]]></dc:creator>
            <pubDate>Wed, 21 Jan 2026 22:35:48 GMT</pubDate>
            <atom:updated>2026-09-24T01:36:21.727Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*pX7z3jWWbV2flVq92u4z_Q.png" /></figure><h3>⚔️ Part 3: Fapolicyd — The Weapons Master</h3><p><em>In Game of Death, Bruce Lee eventually encounters an opponent who doesn’t rely on strength, speed, or technique.</em></p><p><em>Instead, this master controls the </em><strong><em>weapons</em></strong><em>.</em></p><p>The fight is no longer about <em>how well</em> Bruce can strike it. It’s about <strong>what he is allowed to use</strong>.</p><p>In hardened Linux environments, that role belongs to <strong>fapolicyd</strong>.</p><p>If Firewalld controls entry and SELinux controls movement, <strong>fapolicyd controls execution</strong>.</p><p><strong>The Problem Fapolicyd Solves</strong></p><p>Many successful compromises do not rely on exploiting complex vulnerabilities.<br> They rely on something far simpler:</p><p><em>Getting unauthorized code to run.</em></p><p>Scripts dropped in writable directories<br> Binaries copied into /tmp or /var/tmp<br> Malicious libraries injected into trusted paths</p><p>If the code executes, the attacker gains momentum.</p><p>Fapolicyd exists to stop that moment.</p><p><strong>Runtime Execution Control</strong></p><p>Fapolicyd (File Access Policy Daemon) enforces <strong>execution whitelisting</strong>.</p><p>It determines:</p><ul><li>What binaries may execute</li><li>What shared libraries may load</li><li>What scripts may run</li></ul><p>Authorization is typically based on:</p><ul><li>RPM package trust</li><li>File integrity</li><li>Explicit policy rules</li></ul><p>If a binary is not trusted, it <strong>does not execute</strong> — regardless of permissions.</p><p>This aligns directly with:</p><ul><li><strong>NIST SP 800–53 — CM-7 (Least Functionality)</strong></li><li><strong>SI-7 (Integrity Controls)</strong></li><li>ACAS findings related to unauthorized executables</li></ul><p><strong>Execution vs. Existence</strong></p><p>A critical distinction in regulated environments:</p><p><strong>A file existing on disk is not the same as a file executing.</strong></p><p>Fapolicyd enforces this distinction.</p><p>An attacker may succeed in copying a file to disk — <br> but without execution rights, the attack stalls.</p><p>This dramatically reduces:</p><ul><li>Malware effectiveness</li><li>Living-off-the-land techniques</li><li>Accidental execution of unauthorized tools</li></ul><p><strong>Fapolicyd and Defense-in-Depth</strong></p><p>Fapolicyd does not replace SELinux.<br> It complements it.</p><ul><li><strong>SELinux</strong> controls <em>what a process can do</em></li><li><strong>Fapolicyd</strong> controls <em>whether a process can start</em></li></ul><p>Together, they enforce:</p><ul><li>Pre-execution control</li><li>Post-execution containment</li></ul><p>This is layered defense — not redundancy.</p><p><strong>Operational Discipline</strong></p><p>In DoD environments, fapolicyd should be:</p><ul><li>Enabled by default</li><li>Integrated with RPM-based software delivery</li><li>Updated after authorized installations</li><li>Tuned deliberately, not disabled reactively</li></ul><p>Blocking unexpected execution is not a failure condition — <br> it is a <strong>successful security outcome</strong>.</p><p><strong>Closing Reflection</strong></p><p>In <em>Game of Death</em>, Bruce Lee adapts when his weapons are restricted.</p><p>Attackers do the same — unless the system refuses to arm them.</p><p><strong>Code that cannot execute cannot compromise a system.</strong></p><p><em>The full version of this story — including the Elasticsearch war story that started it — is Level 3 of </em><a href="https://www.amazon.com/dp/B0HK2GYZ35"><em>The 5 Levels of Defense-in-Depth</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0456f2f5160f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Game of Death for Cybersecurity (Defense-in-Depth Series)]]></title>
            <link>https://medium.com/@showprove/game-of-death-for-cybersecurity-defense-in-depth-series-e7ad1097cf47?source=rss-9c38d3a41cbb------2</link>
            <guid isPermaLink="false">https://medium.com/p/e7ad1097cf47</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[cyber-security-awareness]]></category>
            <category><![CDATA[govtech]]></category>
            <dc:creator><![CDATA[Michael Wills]]></dc:creator>
            <pubDate>Tue, 16 Dec 2025 18:31:23 GMT</pubDate>
            <atom:updated>2026-09-24T01:33:44.243Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7oVioCdSHR1YPODxb8hx0Q.png" /></figure><h3>🥋 Part 2: SELinux — The Grappler</h3><h3>Introduction</h3><p>In <em>Game of Death</em>, entering the pagoda is only the beginning.</p><p>Once inside, Bruce Lee faces opponents who don’t rely on brute strength — they rely on <strong>control</strong>.<br>They limit movement, dictate positioning, and turn speed against itself.</p><p>This is the role of <strong>SELinux</strong>.</p><p>If Firewalld controls <em>who enters</em>, SELinux controls <strong>what happens after entry</strong>.</p><h3>The Assumption SELinux Makes (And Why It Matters)</h3><p>SELinux operates on a critical assumption:</p><blockquote>Something will eventually break.</blockquote><p>This is not pessimism — it is realism.</p><p>SELinux enforces <strong>Mandatory Access Control (MAC)</strong>, meaning:</p><ul><li>Access decisions are policy-driven</li><li>Users and processes cannot override them</li><li>Compromise does not equal control</li></ul><p>From a compliance perspective, SELinux directly supports:</p><ul><li><strong>NIST SP 800–53 — AC-6 (Least Privilege)</strong></li><li><strong>CM-7 (Least Functionality)</strong></li><li><strong>Multiple DISA STIG findings related to privilege escalation</strong></li></ul><h3>Containment, Not Prevention</h3><p>SELinux does not prevent vulnerabilities.<br>It prevents vulnerabilities from becoming <strong>mission failures</strong>.</p><p>A compromised process:</p><ul><li>Cannot access arbitrary files</li><li>Cannot bind to unauthorized ports</li><li>Cannot interact with processes outside its domain</li></ul><p>The damage is contained — intentionally.</p><h3>Labels, Domains, and Control of the Arena</h3><p>Every object on the system:</p><ul><li>Files</li><li>Processes</li><li>Ports</li><li>Sockets</li></ul><p>…is labeled.</p><p>SELinux policy defines <strong>exactly which interactions are permitted</strong>.</p><p>This transforms the system from an open battlefield into a <strong>controlled arena</strong>, where even skilled attackers are constrained by rules they cannot bypass.</p><h3>SELinux and Auditing</h3><p>SELinux is also an <strong>audit control</strong>.</p><p>Denials generate logs that:</p><ul><li>Identify misconfigurations</li><li>Reveal unexpected behavior</li><li>Support forensic analysis</li><li>Feed continuous improvement</li></ul><p>Disabling SELinux to “fix” an issue does not remove risk — it hides it.</p><h3>Operational Discipline</h3><p>In DoD environments, SELinux should be:</p><ul><li><strong>Enabled</strong></li><li><strong>Enforcing</strong></li><li><strong>Understood</strong></li><li><strong>Integrated into change management</strong></li></ul><p>SELinux policy tuning is not a failure — it is part of operating securely at scale.</p><h3>Closing Reflection</h3><p>Bruce Lee does not defeat every opponent by striking harder.<br>He wins by <strong>controlling the space</strong>.</p><p>SELinux applies the same discipline.</p><blockquote><strong><em>Breaches happen. Uncontrolled breaches are optional.</em></strong></blockquote><p><strong>Up next:</strong><br>⚔️ <strong>Part 3 — Fapolicyd: The Weapons Master</strong></p><p><em>This piece is now Level 2 of a full five-level framework — get the complete book, </em><a href="https://www.amazon.com/dp/B0HK2GYZ35"><em>The 5 Levels of Defense-in-Depth</em></a><em>, for the full climb from boundary to cryptographic core.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e7ad1097cf47" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Game of Death for Cybersecurity (Defense-in-Depth Series)]]></title>
            <link>https://medium.com/@showprove/game-of-death-for-cybersecurity-defense-in-depth-series-04d30049129c?source=rss-9c38d3a41cbb------2</link>
            <guid isPermaLink="false">https://medium.com/p/04d30049129c</guid>
            <dc:creator><![CDATA[Michael Wills]]></dc:creator>
            <pubDate>Mon, 15 Dec 2025 16:45:22 GMT</pubDate>
            <atom:updated>2026-09-24T01:31:46.154Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3VPVO6YmUVeu7Mc14z8DGA.png" /></figure><h3>🥋 Part 1: Firewalld — The Gatekeeper</h3><h3>Introduction</h3><p>In <em>Bruce Lee’s Game of Death</em>, the protagonist cannot simply rush to the top of the pagoda.<br>Before the real fight even begins, he must first <strong>enter</strong> — and that entrance is controlled.</p><p>In hardened Linux environments, <strong>Firewalld</strong> plays that same role.</p><p>Before an adversary can exploit a service, abuse a process, or drop malware, they must first cross the <strong>network boundary</strong>. Firewalld determines whether that fight ever happens at all.</p><p>This is the first layer of defense — and one of the most underestimated.</p><h3>Firewalld as Boundary Enforcement (Not “Just a Firewall”)</h3><p>In DoD and regulated environments, security controls are not implemented for convenience — they exist to enforce <strong>intentional exposure</strong>.</p><p>Firewalld provides:</p><ul><li>Explicit network boundary control</li><li>Zone-based trust segmentation</li><li>Service-level exposure enforcement</li><li>Dynamic rule management without service disruption</li></ul><p>From a compliance standpoint, this maps directly to:</p><ul><li><strong>NIST SP 800–53 — SC-7 (Boundary Protection)</strong></li><li><strong>DISA STIG</strong> requirements for limiting network-accessible services</li></ul><p>Firewalld ensures that <strong>only approved traffic reaches the system</strong>, and nothing more.</p><h3>Why the Gate Matters</h3><p>Most attacks don’t begin with zero-day exploits.<br>They begin with <strong>reachability</strong>.</p><p>If a service is exposed:</p><ul><li>It will be scanned</li><li>It will be fingerprinted</li><li>It will eventually be tested</li></ul><p>Firewalld enforces the principle that:</p><blockquote>Services that are not required should not be reachable.</blockquote><p>This dramatically reduces:</p><ul><li>Attack surface</li><li>Alert fatigue</li><li>Downstream control load (SELinux, fapolicyd, auditing)</li></ul><h3>Zones = Trust Boundaries</h3><p>Firewalld zones are not arbitrary labels — they represent <strong>explicit trust decisions</strong>.</p><p>For example:</p><ul><li>public → untrusted networks</li><li>internal → management networks</li><li>trusted → tightly controlled interfaces</li></ul><p>In controlled environments, most systems should default to:</p><ul><li><strong>deny by default</strong></li><li><strong>allow only what is mission-required</strong></li><li><strong>document every exception</strong></li></ul><p>This is not restrictive — it is deliberate.</p><h3>Operational Reality</h3><p>Firewalld does not eliminate the need for deeper controls.<br>It exists to ensure those controls are not overwhelmed.</p><p>Firewalld:</p><ul><li>Prevents unnecessary fights</li><li>Buys time</li><li>Reduces noise</li><li>Supports audit defensibility</li></ul><h3>Closing Reflection</h3><p>In <em>Game of Death</em>, Bruce Lee never wastes energy on fights he can avoid.<br>He controls distance, entry, and engagement.</p><p>Firewalld applies the same philosophy to system defense.</p><blockquote><strong><em>A fight that never begins cannot be lost.</em></strong></blockquote><p><em>This piece is now Level 1 of a full five-level framework — get the complete book, </em><a href="https://www.amazon.com/dp/B0HK2GYZ35"><em>The 5 Levels of Defense-in-Depth</em></a><em>, for the full climb from boundary to cryptographic core.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=04d30049129c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Beyond Default: Tailoring ACAS Scan Policies for Role-Based Compliance in DoD Environments]]></title>
            <link>https://medium.com/devmap/tailor-made-scanning-why-acas-policies-must-be-customized-for-each-server-role-fb86cb79c928?source=rss-9c38d3a41cbb------2</link>
            <guid isPermaLink="false">https://medium.com/p/fb86cb79c928</guid>
            <category><![CDATA[tenable]]></category>
            <category><![CDATA[govtech]]></category>
            <category><![CDATA[elasticsearch]]></category>
            <dc:creator><![CDATA[Michael Wills]]></dc:creator>
            <pubDate>Sun, 22 Jun 2025 17:48:02 GMT</pubDate>
            <atom:updated>2026-04-14T02:35:34.355Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WJi2M0PJNJnytmvVB0BpjQ.png" /></figure><p><strong><em>Author’s Note</em></strong></p><p><em>An earlier version of this article focused primarily on the mechanics of ACAS scan configuration — the how-to of cloning policies, selecting plugin families, and scheduling scan windows. That version was not wrong. But it framed the problem as a technical one, when the real problem is a people and systems problem. Default scan policies are a symptom. The root cause is that the organizations running ACAS have not built the cross-functional collaboration — between system administrators, ISSOs, SSOs, and the Tenable team — that makes scan data mean something.</em></p><p><em>I revised this article because the original version gave readers a configuration guide. This version makes an argument: that scan tailoring is a systems design problem, and organizations that treat it as a configuration task will keep producing compliance theater. The Elasticsearch bundled JDK example in the pages that follow did not come from documentation — it came from standing in front of that cluster and knowing why the finding was technically correct and operationally irrelevant at the same time. That gap between what the scanner sees and what the engineer knows, is what this article is actually about.</em></p><p><em>That argument needed more than an engineer’s voice to land the way it should. Jakobe Wills contributed to this revision by stress-testing the language against a reader who understands cybersecurity concepts but has not yet spent years inside a DoD compliance pipeline. Where earlier drafts assumed too much, his perspective pushed the writing toward clarity without sacrificing precision. The sections most likely to reach a mixed technical and compliance audience are sharper because of that collaboration.</em></p><p><em>If you are a system administrator who has spent time explaining to a compliance officer why a finding is not what it appears to be, this article is written for you — and for the compliance officer who wants to stop having that conversation after the scan runs and start having it before.</em></p><h3>Introduction</h3><p>In Department of Defense network environments, compliance scanning is not a courtesy — it is a mandate. The Assured Compliance Assessment Solution (ACAS), built on Tenable.sc and Nessus, is the primary vulnerability scanning platform across DoD enclaves and is central to the evidence packages that drive Risk Management Framework (RMF) Authorization to Operate (ATO) decisions. Most practitioners know ACAS is required. Fewer understand that how you configure it determines whether it produces actionable intelligence or an unmanageable sea of noise.</p><p>Default scan policies are a reasonable starting point for initial network discovery, but they are poorly suited for the operational reality of a hardened, multi-role DoD infrastructure. When the same scan policy runs against a web server, a database host, and an air-gapped jump box, the result is a sprawling findings list that mixes critical vulnerabilities with irrelevant plugin checks — checks designed for Windows systems running against Red Hat Enterprise Linux (RHEL) hosts, or Server Message Block (SMB) enumeration plugins scanning a server that has never run a Windows service. The real risk is not that these irrelevant findings waste remediation time, though they do. The risk is that they bury the findings that actually matter.</p><p>This article presents a practical framework for building role-based ACAS scan policies in Tenable.sc, drawn from operational experience managing compliance pipelines in high-side DoD environments. The goal is not theoretical — it is to help security and infrastructure engineers produce cleaner findings data, defend that data during ATO reviews, and build a sustainable compliance posture that holds up under Defense Information Systems Agency (DISA) audits.</p><p>More specifically, this article treats ACAS scan tailoring as a systems-based methodology — not a technical checklist, but an interconnected set of human, organizational, and technical components whose output quality depends on how well those components work together. The scan policy is a technical artifact, but its accuracy is downstream of a human process: the collaboration between system administrators, ISSOs, SSOs, and the Tenable team that shapes it. The POA&amp;M is a compliance document, but its defensibility is downstream of that same collaboration. The ATO package is an authorization artifact, but its credibility is a property of the whole system — from the initial pre-scan conversation to the findings that ultimately land in eMASS. When any component in that chain is weak or disconnected, the degradation propagates forward. When the system works as designed, the output at every stage is more accurate, more actionable, and more defensible. That is the core argument of this article: compliance quality is not a function of individual effort — it is a function of system design.</p><h3>Why Default Scan Policies Fall Short</h3><p>Tenable.sc ships with several baseline scan policy templates, including the DoD Assured Compliance Assessment Solution (ACAS) policy derived from DISA guidance. These templates are intentionally broad — they are designed to provide coverage across heterogeneous environments where the full inventory of host types may not be known in advance. That broadness is a feature during initial assessment. It becomes a liability once your environment is characterized and operationally steady.</p><p>The problem is not that ACAS produces wrong results — it is that ACAS reports what it can observe, which is not always the same as what is operationally true. That distinction matters, and conflating the two is where compliance programs go sideways. A concrete example from the field illustrates the point. Elasticsearch ships with a bundled Java Development Kit (JDK) packaged inside its installation directory. In versions 8.15.0, 8.15.1, and 8.15.2, that bundled JDK — version 22.0.1 — carried several documented CVEs serious enough to trigger an upgrade recommendation to a later patch version. ACAS will detect that JDK binary on disk, match it against known vulnerable versions, and flag it as a finding. The finding is technically accurate: the vulnerable binary exists on the system. What the scanner cannot determine — even in a credentialed scan — is whether that bundled JDK is the runtime actually in use. In hardened deployments, Elasticsearch is routinely configured to use a separately managed, enterprise-patched system JDK rather than the bundled one. The bundled binary sits inert inside the installation directory, never invoked, while the active Java runtime is maintained under the organization’s standard patch lifecycle. Without that context, the finding enters the POA&amp;M queue as an open vulnerability requiring urgent remediation. With it, the finding is documented accurately: the vulnerable binary is present but not executed; the active runtime is current and patch-managed. That distinction is invisible to the scanner and only visible to the engineer who built and configured the system. Compliance teams pulling data into DISA’s STIG Viewer or eMASS cannot make that call without engineering input — and triage without that input produces inaccurate compliance records, inflated POA&amp;M backlogs, and wasted remediation cycles chasing findings that carry no actual operational risk.</p><p>The credentialing question compounds this. Uncredentialed scans — network-only, no host authentication — are faster and less invasive, but they miss the majority of STIG-relevant findings. Kernel version, SELinux enforcement state, Pluggable Authentication Modules (PAM) configuration, audit daemon settings, Federal Information Processing Standards (FIPS) mode status: none of these are visible to an uncredentialed scanner. Organizations that rely on uncredentialed scans for ATO evidence are producing incomplete data and may not realize it until an assessor flags it.</p><h3>The Case for Policy Tailoring</h3><p>A tailored scan policy is one that is scoped to a specific system role, uses the appropriate credential method, limits plugin execution to families relevant to that role’s OS and services, and is scheduled to avoid conflict with known maintenance or high-utilization windows. The investment in building and maintaining these policies pays dividends across the entire RMF lifecycle.</p><p>From a data quality standpoint, findings produced by a role-specific policy are immediately more actionable. A web server scan that focuses on Transport Layer Security/Secure Sockets Layer (TLS/SSL) cipher suite validation, Apache and NGINX Common Vulnerabilities and Exposures (CVE) families, Hypertext Transfer Protocol (HTTP) method exposure (particularly TRACE and PUT), and HTTP security header compliance produces a findings set that a developer or system administrator can triage and remediate without first filtering out a hundred irrelevant checks. That clarity also makes it easier to demonstrate trend lines to leadership — showing that a web server’s finding count dropped from 22 to 6 over a quarter, with specific CVEs remediated, is a far more compelling ATO narrative than presenting a noisy aggregate number.</p><p>Credentialed scanning is the other major lever. When ACAS authenticates to a host using a Secure Shell (SSH) key or a privileged service account, Nessus can perform local checks that are invisible from the network: it can read /etc/ssh/sshd_config, check PAM module configurations, enumerate installed packages against the National Vulnerability Database (NVD), verify SELinux labels, and confirm whether FIPS-validated cryptographic modules are active. For RHEL systems operating under DISA STIGs, the difference in finding depth between a credentialed and uncredentialed scan is substantial. Credentialed scans should be the default for all hosts where authentication is feasible; uncredentialed scans should be the exception, reserved for network boundary discovery or hosts where privileged access cannot be safely delegated.</p><p>Tailored policies also produce consistency that matters for enclave-level comparisons. When two web servers in different network segments are both scanned with the same role-specific web server policy, their findings are directly comparable. Anomalies — one server showing a finding the other does not — stand out immediately and warrant investigation. That kind of comparative signal is difficult or impossible to extract from a heterogeneous single-policy scan.</p><h3>Building Role-Based Policies in Tenable.sc</h3><p>The practical workflow for creating tailored policies in Tenable.sc begins in the Policy &amp; Response Workflow under Scan Policies. Rather than building from scratch, clone the DISA-provided ACAS baseline or the appropriate United States Government Configuration Baseline (USGCB) template and rename it with a clear role designation — for example, ACAS-WebServer-RHEL8 or ACAS-DBServer-RHEL8-Credentialed. This preserves the DISA-derived plugin baseline while giving you a clean copy to modify without affecting the source template.</p><p>With a cloned policy open, the primary adjustment points are plugin family selection, credential configuration, and advanced scan settings. For Linux-only roles, disable the Microsoft Bulletin, SMB, and Windows Compliance plugin families entirely. For roles that do not expose Windows file shares or Remote Procedure Call (RPC) services, the associated network enumeration plugins can also be disabled. What remains is a policy focused on the plugin families that matter: General, RHEL Local Security Checks, Patch Management, and the service-specific families relevant to the role (for example, Web Servers and SSL/TLS for web hosts).</p><p>Credential configuration deserves careful attention. For SSH-based Linux scanning, use a dedicated service account with privilege escalation configured (sudo or su, depending on your local policy) rather than scanning as root. This approach limits the blast radius of a compromised credential and produces a cleaner audit trail. For systems using System Security Services Daemon (SSSD) and Kerberos authentication — common in DoD environments integrated with Active Directory (AD) — verify that the Nessus scanner’s service account has a valid keytab and that the principal naming matches what SSSD expects. A subtle mismatch (such as a restrictedKrbhost prefix on the host principal or a lowercase Kerberos realm) can cause silent authentication failures that produce uncredentialed results without any obvious error in the scan output.</p><p>Scheduling is the final dimension of policy tailoring. Define scan windows that avoid patching cycles, backup jobs, and known peak utilization periods. For Elasticsearch clusters and other services that are sensitive to I/O and CPU contention, running a credentialed scan during a replica shard rebalance operation can introduce latency spikes that appear to operations teams as a service degradation event. Coordinate scan windows with operations staff and document them in your System Security Plan.</p><h3>Role-Based Policy Reference</h3><p>The following table summarizes recommended plugin focus areas and credential approaches by server role. This is intended as a starting framework; specific plugin families should be validated against your enclave’s approved technology stack.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/775/1*hYeBBZml8DZzzu3MEinmbg.png" /></figure><h3>Lessons from the Field</h3><p>The impact of undifferentiated scan policies is not abstract. In one operational environment, ACAS finding counts on a cluster of RHEL 8 application servers jumped from approximately 5 findings per host to over 30 in a single scan cycle, with no changes to system configurations. The initial assumption was that a new vulnerability disclosure had affected the environment. The actual cause was a Tenable plugin pack update that activated several new Windows Patch Bulletin checks — checks that were reporting as findings on Linux hosts because the plugin logic was interpreting the absence of Windows registry keys as a failed compliance state.</p><p>The resolution required auditing the active plugin list in the scan policy, identifying the newly activated families, disabling them for Linux-scoped policies, and re-running the scan to establish a clean baseline. The investigation took approximately half a day of engineering time — time that should not have been necessary if the policy had been scoped correctly from the start. More importantly, during that cycle, the spike in findings caused the team to defer triage of a legitimate finding — an unpatched kernel CVE — because it was lost in the volume. That is the real cost of policy noise: not wasted time, but degraded prioritization of actual risk.</p><p>After implementing role-based policies, the same environment saw finding counts stabilize within expected ranges, with deviations from baseline reliably indicating either a genuine new vulnerability or a configuration drift event. That predictability is operationally valuable. It means engineers can respond to finding count changes with confidence rather than spending cycles determining whether the change is real.</p><h3>The Sysadmin-ISSO Partnership: Shaping Scans Before They Run</h3><p>One of the most persistent friction points in DoD compliance programs is the gap between what a scan reports and what an operator knows about why a system is configured the way it is. This gap exists because ACAS scan policies are often owned and configured by cybersecurity staff — Information System Security Officers (ISSOs) and System Security Officers (SSOs) — who may have strong knowledge of RMF requirements and DISA STIG controls but limited hands-on experience administering the systems being scanned. The result is a compliance process that treats every finding as a deficiency to be remediated, when in reality some findings reflect deliberate, operationally necessary configurations that cannot be changed without degrading mission capability.</p><p>System administrators are the people who know why a particular port is open, why a service account has elevated privileges on a specific host, why a kernel parameter is set to a non-default value, or why a particular package that a plugin flags as end-of-life is still present because it is a dependency of a mission-critical application with no approved replacement. That operational context is not visible in a scan report. It lives in the heads of the people who built and maintain the systems — and if those people are not involved in shaping the scan policy, that context never makes it into the compliance record.</p><p>The ISSO and SSO should be working with system administrators at the lowest practical level — individual system role or enclave segment — before scan policies are finalized, not after findings land in eMASS. That conversation should address at least three questions for each system type: What services and configurations on this host are operationally non-negotiable? Which STIG controls are candidates for technical exceptions or operational requirements waivers because of those configurations? And are there findings that will appear on every scan cycle not because the system is insecure but because the compensating control for that risk lives somewhere else in the architecture?</p><p>When sysadmins and cybersecurity staff work through these questions together before the scan runs, the result is a more defensible compliance posture. Findings that require POA&amp;Ms are documented with accurate root causes — not generic “remediation in progress” language, but specific operational justifications: this port is open because it supports authenticated replication between cluster nodes; this package version is pinned because upgrading it breaks a commercial off-the-shelf (COTS) application under active vendor support; this kernel parameter is set because the default causes Transmission Control Protocol (TCP) performance degradation on high-throughput data paths. That kind of documentation holds up under assessor scrutiny in a way that a vague POA&amp;M entry does not.</p><p>It also prevents a common and costly dynamic where an ISSO, seeing a finding they do not recognize, assigns a remediation action to a system administrator who then has to explain — often repeatedly, across multiple scan cycles — that the finding is expected and intentional. That explanation cycle consumes engineering time on both sides and erodes the working relationship between operations and cybersecurity teams. Front-loading the conversation through pre-scan collaboration eliminates it.</p><p>Cybersecurity professionals without a background in systems administration are not at fault for not knowing the operational nuances of every system in their environment — that knowledge takes years to develop. But they are responsible for building the processes that capture it. Engaging sysadmins as subject matter experts in scan policy design, not just as the people who remediate findings, is how you build compliance programs that produce accurate data rather than compliance theater.</p><h3>ACAS in the RMF/ATO Lifecycle</h3><p>ACAS scan outputs do not exist in isolation — they flow directly into the RMF artifacts that support ATO packages. Tailored policies make this integration significantly cleaner. When scan results are imported into eMASS or reviewed against STIG checklists in DISA STIG Viewer, role-specific findings map more directly to the controls under review. A web server’s ATO package should be reviewed against web server STIGs; presenting a findings report that includes Windows patch failures on that host creates confusion for assessors and invites questions about scan methodology.</p><p>POA&amp;M management is where the sysadmin-ISSO collaboration pays its most visible dividend. A POA&amp;M entry should reflect one of two things: a genuine vulnerability with a documented remediation path, or a deliberate operational exception with a documented justification and accepted risk. When sysadmins are involved in scan policy design and findings review, both types of entries are written accurately from the start. Findings that carry over across scan cycles because they represent operationally necessary configurations are documented as such, with compensating controls identified. Findings that represent actual open vulnerabilities have realistic remediation timelines that account for patching windows, change management processes, and dependency constraints that only the operating team understands.</p><p>When ISSOs encounter findings they cannot explain — configurations they did not expect to see, services they were not aware were running, package versions that do not match the system’s documented baseline — that is not a compliance failure. It is a signal that the scan has surfaced something the sysadmin should be asked about before a POA&amp;M is written. The right response is a conversation, not an immediate action item. Systems in operational DoD environments are complex, and the distance between “this STIG check failed” and “this system is insecure” is often significant. The sysadmin is the person who can close that interpretive gap.</p><p>For organizations pursuing continuous ATO or cATO under recent DoD guidance, the quality of ongoing scan data is a direct input to authorization decisions. Automated compliance pipelines — whether built on Ansible, custom scripts, or platform tooling — that feed ACAS results into dashboards or eMASS need clean, consistent data to produce reliable authorization evidence. Role-based policies are a prerequisite for that consistency, and the sysadmin-ISSO working relationship is a prerequisite for role-based policies that accurately reflect operational reality.</p><h3>POA&amp;Ms as a Funding Mechanism: Turning Compliance Gaps into Resource Justification</h3><p>In most discussions of Risk Management Framework compliance, the Plan of Action and Milestones is treated purely as a liability — a record of what is broken and how long it has been broken. That framing misses one of the most practically important functions a well-maintained POA&amp;M can serve: it is one of the few formal mechanisms available to an operations team to translate infrastructure shortcomings into documented, leadership-visible resource requirements.</p><p>DoD programs operate under constrained budgets, and the competition for funding — whether for hardware refresh, software licensing, contractor support, or additional personnel — requires justification that leadership and resource managers can act on. A verbal explanation from a sysadmin that a system is running an end-of-life operating system (OS) because there is no approved budget for an upgrade rarely survives the trip up the chain. A POA&amp;M entry that documents the same situation as an open finding with an identified risk level, a compensating control, an estimated remediation cost, and a milestone tied to a budget cycle is something a program manager can attach to a funding request. It gives the resource need a formal home in the compliance record.</p><p>This is especially relevant for operational shortcomings that are not the result of negligence or inattention but of resource reality — the infrastructure inherited from a predecessor program, the COTS application that cannot be upgraded until the vendor releases a compatible version, the network segment that needs a hardware refresh but is waiting on procurement approval. These situations are common in DoD environments, and they consistently produce recurring ACAS findings that compliance teams cannot close. Without a POA&amp;M that accurately documents the operational constraint and the resource requirement, those findings simply accumulate, creating the appearance of a stagnant compliance program even when the underlying cause is a funding gap rather than an engineering failure.</p><p>Accurate POA&amp;M documentation — built with sysadmin input, as discussed earlier — gives program leadership the visibility they need to advocate for resources. An ISSO who understands why a finding exists can write a POA&amp;M that explains the operational dependency, quantifies the risk of leaving it open, identifies what it would take to close it, and attaches a realistic cost estimate. That document travels through budget reviews in a way that informal conversations do not. It creates an auditable record that the program identified the gap, assessed the risk, implemented a compensating control, and is actively pursuing a funded remediation path — which is precisely what assessors and authorizing officials need to see to make informed authorization decisions.</p><p>The inverse is also true. When POA&amp;Ms are written without operational context — when an ISSO records a STIG finding and assigns a generic remediation milestone without understanding whether that remediation is feasible under current resource constraints — the resulting POA&amp;M is not only less useful for funding purposes, it can actually undermine the program’s credibility. Milestones that slip repeatedly because they were set without accounting for procurement lead times, change management processes, or dependency constraints signal to reviewers that the program does not have an accurate picture of its own environment. Sysadmin involvement in POA&amp;M development is not just good compliance hygiene — it is what makes the document credible enough to support a funding conversation.</p><h3>Conclusion</h3><p>Default ACAS scan policies are a starting point, not a destination. In mature DoD environments where system roles are defined, networks are segmented, and ATO timelines are real, the cost of running undifferentiated scans is measured in engineer hours spent on false positive triage, in findings data that cannot support meaningful trend analysis, and in the occasional missed vulnerability buried under irrelevant results.</p><p>Building role-based scan policies is not a one-time project — it is an ongoing operational practice. Plugin packs update, new system roles are introduced, and credential configurations change as identity platforms evolve. Treating ACAS policy management with the same rigor applied to STIG remediation and patch management produces a compliance posture that is both more accurate and more defensible.</p><p>Do not accept the default. Scope your scans to the systems you are actually protecting — and let the data tell you what is real.</p><h3><strong>About the Authors</strong></h3><p><strong>Michael S. Wills</strong> is a senior Linux Systems Engineer and retired U.S. Army Chief Warrant Officer with over 20 years of experience in cybersecurity, infrastructure automation, and compliance engineering across government and defense sectors. He has supported missions for Army Cyber Command (ArCyber) — where he worked directly on cyber operations infrastructure — the Naval Criminal Investigative Service (NCIS), the National Media Exploitation Center (NMEC), and the Space-Based Infrared System (SBIRS). He has also supported Palantir platform operations in support of defense intelligence and data integration missions. He is a recipient of the Bronze Star Medal and the 2019 Black Engineer of the Year Award (BEYA) Science Spectrum Trailblazer Award, and currently specializes in RMF/ATO lifecycle management, STIG automation, and large-scale Elasticsearch platform engineering in DoD environments.</p><p><strong>Jakobe Wills</strong> is a contributing writer and graduate student in the Cybersecurity program at Old Dominion University, where he also earned his Bachelor of Arts in English. His background in technical communication and his emerging expertise in cybersecurity inform his ability to bridge the gap between operational complexity and accessible technical writing — making him a natural fit for content aimed at both engineering and compliance audiences.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fb86cb79c928" width="1" height="1" alt=""><hr><p><a href="https://medium.com/devmap/tailor-made-scanning-why-acas-policies-must-be-customized-for-each-server-role-fb86cb79c928">Beyond Default: Tailoring ACAS Scan Policies for Role-Based Compliance in DoD Environments</a> was originally published in <a href="https://medium.com/devmap">Devmap</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>