<?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 Shreyash Tambe on Medium]]></title>
        <description><![CDATA[Stories by Shreyash Tambe on Medium]]></description>
        <link>https://medium.com/@shreyash_tambe?source=rss-83f6d1b096d5------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*QETvxItQQwisZ0XHpFYyaw.jpeg</url>
            <title>Stories by Shreyash Tambe on Medium</title>
            <link>https://medium.com/@shreyash_tambe?source=rss-83f6d1b096d5------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 01:31:13 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@shreyash_tambe/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[RedSun : A Clever Windows Defender Zero Day Exploit]]></title>
            <link>https://medium.com/@shreyash_tambe/redsun-a-clever-windows-defender-zero-day-exploit-0892e08c513b?source=rss-83f6d1b096d5------2</link>
            <guid isPermaLink="false">https://medium.com/p/0892e08c513b</guid>
            <dc:creator><![CDATA[Shreyash Tambe]]></dc:creator>
            <pubDate>Thu, 16 Apr 2026 17:35:10 GMT</pubDate>
            <atom:updated>2026-04-16T17:46:50.549Z</atom:updated>
            <content:encoded><![CDATA[<p>Yo folks back again with new zero-day explaination!</p><p>Researcher Nightmare-Eclipse discovered this another variant of zero-day exploit to gain SYSTEM privilege on 14th April 2026.</p><p>POC : <a href="https://github.com/Nightmare-Eclipse/RedSun">https://github.com/Nightmare-Eclipse/RedSun</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jOEDuJRhg91aRRKAIWqqOg.png" /></figure><p>When Defender detects any THREAT it follows a multi-step cleanup process called remediation:</p><p><strong>Step A :</strong> Create a Backup Snapshot (Volume Shadow Copy) VSS. <br>Before quarantine/deleting, Defender creates a backup snapshot of your disk. This is a safety measure of windows if something goes wrong during cleanup, Windows can roll back using this snapshot.<br>e.g. \Device\HarddiskVolumeShadowCopy1</p><p><strong>Step B : </strong>Remove the Threat<br>Defender quarantines or deletes the malicious file. Standard operating procedure.</p><p><strong>Step C :</strong> “Cloud Tag” Based file Remediation.<br>This is where things get weirdoo. If the detected file has something called a “cloud tag” i.e. CfApi (a special metadata that says “This file belongs to a cloud service like OneDrive”), Defender’s logic says “This file came from a cloud sync provider. Maybe it got corrupted during sync. I should restore a clean version from my backup snapshot to the original location.”<br> <br> — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —<br> Let’s explore the exploit and its TTP</p><h4>STAGE 1 :: DROPPING EICAR INTO BAIT</h4><p>Exploit named <strong>Redsun.exe </strong>is dropped in any working directory. <br> what this exploit does is it will drop tampered <strong>TieringEngineService.exe </strong>(LOLBIN) into %<strong>TEMP</strong>% directory. this TieringEngineService.exe will work as bait to fool Defender.<br> <br> Exploit redsun.exe will already contain <strong>reversed</strong> EICAR text.</p><pre>EICAR is a fake virus string <br>i.e. a harmless 68-character text string used by security professionals <br>to test antivirus behaviour. <br>It&#39;s not malicious, but every AV detects it as a threat.<br><br>X50IP%@AP[4\PZX54(P^)7CC)7)SEICAR-STANDARD-ANTIVIRUS-TEST-FILEISH+H*</pre><p>Why reversed? <br>If the attacker stored the EICAR string normally in their exploit code, Defender would flag the exploit at compile time itself before it even runs, So they store it backwards and flip it in correct order at runtime into a bait LOLBIN file i.e. TieringEngineService.exe.</p><h4>STAGE 2 :: EXECUTING BAIT</h4><p>Initially Redsun.exe <strong>will take snapshot of all VSS copies</strong> present and it will constantly watches for new shadow copies created using “<strong>ShadowCopyFinderThread</strong>”.</p><p>Afterwards Redsun Exploit will <strong>trigger Defender</strong> by executing TieringEngineService.exe using <strong>FILE_EXECUTE </strong>flag. <br> The moment Defenders real time protection sees file being executed that contains EICAR text it start its remediation procedure explained at beginning. starting with creation of VSS copy i.e. STEP 1 of remediation. <br> <br> As soon as new VSS is created : “ShadowCopyFinderThread” from Redsun.exe <strong>triggered</strong> an alarm and it will then create “<strong>batch oplock</strong>” on that VSS. From Exploits POV the appearance of a new VSS is proof that Defender has started remediation process.</p><pre>What the hack is this batch oplock? <br>Think of It as a trigger. <br>It gives the exploit right timing to execute the next steps.</pre><h4>STAGE 3 :: FOOLING DEFENDER BY CLOUD TAG</h4><p>As alarm is triggered what TA have to do now is create the new bait file from scratch as “cloud tag” in same %TEMP% folder but this is not possible with same name in same directory.</p><p>For that we first need to delete existing bait but but but hook/handle to existing bait is occupied by Defender so we cannot delete it. you’ll get “file is in use” errors.<br> <br> TA use clever technique to evade that i.e. POSIX delete</p><pre>POSIX delete (using special flags): <br>The file&#39;s name immediately vanishes from the directory, <br>even while programs still have it open. <br><br>The actual data still exists on disk (open handles can still read/write it), <br>but the filename is gone from the directory listing.</pre><p>so Exploit will write new bait file with cloud tag in same %TEMP% directory. [why new file from scratch ? Is because we cannot make existing file as cloud tagged it needs structural + metadata changes]<br> <br> <strong>All this happen in fraction of seconds After STEP A of remediation</strong>. [Also knows as <strong>TOCTOU </strong>time-of-check to time-of-use in Defender]</p><p>so now Defender will start STEP B for that it will again go to path in %TEMP% directory to delete the bait file but whatttt it sees “cloud tag” and it goes like dayumn I have to revert this lol</p><h4>STAGE 4 :: DEFENDER FOOLED AND STARTED ROLL BACK</h4><p>Now bait file has “cloud tag” Instead of deleting the threat, Defender tries to rewrite back to where it found the file</p><p>So, now Defender will assume TieringEngineService.exe as cloud based and will try to roll it back to it’s original location. To roll back Defender will access previously created VSS copy of it. And this triggers the <strong>“batch oplock” </strong>of exploit<strong> indicating Defender is trying to roll back our bait.</strong></p><h4>STAGE 5 :: REDIRECTING NTFS MOUNT POINT JUNCTION</h4><p>As “batch oplock” triggered an alarm, exploit will change the <strong>Junction </strong>from %TEMP% to \system32\</p><pre>A junction is a directory redirect. <br>It makes one folder silently point to another folder. <br>No special permissions needed!</pre><p>so, at <strong>STEP C of Remediation </strong>Defender have started rolling back bait file into %TEMP% but in reality it is being written into \system32\ and as we all know Defender runs with SYSTEM context hence, it can easily write into \system32\ protected directory.</p><h4>STAGE 6 :: YOU’RE COOKED!</h4><p>So file TieringEngineService.exe lands into \system32\ but still Defender was not aware about this scam and hence it did not applied any System32’s restrictive ACL’s. Holy shitt!<br> <br>Now TieringEngineService.exe is in protected directory with permissive ACL’s that a normal user can write and here after Exploit redsun will copy itself entirely into TieringEngineService and gain SYSTEM level previlege. <br> <br> BOOM GAME OVER !</p><p>Thnks ; )<br> SHREYASH TAMBE<br> Threat Detection Developer / Malware Analyst</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0892e08c513b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[️ MongoBleed Detection: Hunting CVE-2025–14847]]></title>
            <link>https://medium.com/@shreyash_tambe/%EF%B8%8F-mongobleed-detection-hunting-cve-2025-14847-709c2cd68b26?source=rss-83f6d1b096d5------2</link>
            <guid isPermaLink="false">https://medium.com/p/709c2cd68b26</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[cve]]></category>
            <category><![CDATA[threat-hunting]]></category>
            <category><![CDATA[mongodb]]></category>
            <category><![CDATA[threat-detection]]></category>
            <dc:creator><![CDATA[Shreyash Tambe]]></dc:creator>
            <pubDate>Sat, 03 Jan 2026 06:33:10 GMT</pubDate>
            <atom:updated>2026-01-03T06:33:10.023Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*o6hBorOkx5AYaNUfEQ8PqQ.png" /></figure><p>olá pessoal 👋</p><p>In the first part, we went through a detailed exploration of how <strong>MongoBleed (CVE-2025–14847)</strong> leaks heap memory at the protocol level.<br>Now comes the fun part : <strong>detecting it in the real world</strong>.</p><p>Instead of just talking theory, I went ahead and <strong>built an actual hunt and detection logic</strong> from a <strong>threat detection/threat hunting perspective</strong>, validated it against real exploit traffic in a lab setup, and stress-tested it against logs.</p><h3>The Core Detection Insight</h3><p>The public MongoBleed PoC behaves very differently from legitimate MongoDB clients.</p><p>Two signals stand out immediately:</p><h4>1️. Extreme Connection Bursts</h4><p>The exploit does <strong>not reuse connections</strong>. Instead, it rapidly opens and closes connections to repeatedly trigger memory disclosure,<strong> </strong>hundreds of connections in <strong>fractions of a second, </strong>which is wildly abnormal for real applications.</p><h4>2️. Missing Client Metadata</h4><p>Every legitimate MongoDB driver: PyMongo / Node.js / mongosh / Java all send <strong>client metadata immediately after connecting</strong>. Metadata is not optional for real drivers; it’s automatic. MongoDB logs this as <strong>event ID 51800</strong>.</p><p>The exploit Connects (22943) -&gt; Extracts memory → Disconnects (22944) -&gt; <strong>Never sends metadata</strong></p><p>This creates a powerful detection signal:</p><blockquote><em>High connection volume + zero metadata = very suspicious traffic</em></blockquote><h3>🧱 MongoDB Log Events We Correlate</h3><p>MongoDB’s JSON logs give us exactly what we need.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/830/1*PdBcL_PXynD4huqVwfmjaA.png" /></figure><p><strong>Key observation:</strong><br>Legitimate traffic shows <strong>22943 → 51800 → 22944</strong><br>MongoBleed traffic shows <strong>22943 → 22944</strong> only.</p><h3>🧠 Detection Strategy (High-Level)</h3><p>We aggregate MongoDB logs <strong>by source IP</strong> and calculate:</p><ul><li>Total number of connections</li><li>Metadata presence rate</li><li>Connection velocity (connections per minute)</li><li>Then we apply <strong>risk scoring</strong>.</li></ul><blockquote><strong>Risk Classification Logic .:<br>HIGH : </strong>connections ≥100 AND metadata rate &lt;10% AND velocity ≥500 <br><strong>MEDIUM : </strong>connections ≥100 AND metadata rate &lt;10%<br><strong>LOW : </strong>connections ≥100</blockquote><h3>🛠 The Splunk SPL Detection Query</h3><p>Here’s the <strong>exact SPL</strong> I built to hunt MongoBleed exploitation:</p><pre>index=* sourcetype=&quot;*mongo*&quot; id IN (22943,22944,51800)<br>| eval source_ip = &#39;attr.remote&#39;<br>| rex field=source_ip &quot;^(?&lt;src_ip&gt;[^:]+):\d+$&quot;<br>| eval event_time=strptime(&#39;t.$date&#39;, &quot;%Y-%m-%dT%H:%M:%S.%f%z&quot;)<br>| eval event_type=case(<br>    id=&quot;22943&quot;, &quot;connection&quot;,<br>    id=&quot;22944&quot;, &quot;disconnection&quot;,<br>    id=&quot;51800&quot;, &quot;metadata&quot;,<br>    1=1, &quot;other&quot;<br>)<br>| stats <br>    min(event_time) as first_seen<br>    max(event_time) as last_seen<br>    count(eval(event_type=&quot;connection&quot;)) as connection_count<br>    count(eval(event_type=&quot;disconnection&quot;)) as disconnection_count<br>    count(eval(event_type=&quot;metadata&quot;)) as metadata_count<br>    by src_ip<br>| eval total_seconds = max(last_seen - first_seen, 1)<br>| eval total_minutes = total_seconds / 60<br>| eval total_minutes = if(total_minutes &lt; 0.0167, 0.0167, total_minutes)<br>| eval total_connection_velocity_pm = round(connection_count / total_minutes, 2)<br>| eval total_metadata_rate = round(if(connection_count &gt; 0, (metadata_count / connection_count) * 100, 0), 2)<br>| eval risk_factor = case(<br>    connection_count &gt;= 100 AND total_metadata_rate &lt; 10 AND total_connection_velocity_pm &gt;= 500, &quot;HIGH: ALERT_MongoBleed_detected&quot;,<br>    connection_count &gt;= 100 AND total_metadata_rate &lt; 10, &quot;MEDIUM: ALERT_MongoBleed_pattern&quot;,<br>    connection_count &gt;= 100, &quot;LOW: ALERT_Excessive_connection_rate&quot;,<br>    1=1, &quot;INFO&quot;<br>)<br>| where risk_factor != &quot;INFO&quot;<br>| eval first_seen = strftime(first_seen, &quot;%Y-%m-%d %H:%M:%S&quot;)<br>| eval last_seen = strftime(last_seen, &quot;%Y-%m-%d %H:%M:%S&quot;)<br>| table src_ip, first_seen, last_seen, risk_factor, connection_count, total_connection_velocity_pm, metadata_count, total_metadata_rate, total_minutes</pre><h3>🧪 Validation Against Real Exploitation</h3><p><strong>Lab Setup</strong></p><ul><li>Vulnerable MongoDB with zlib enabled</li><li>Ran the public MongoBleed PoC <a href="https://github.com/joe-desimone/mongobleed">LINK</a></li><li>Collected logs and ran the detection</li><li>Detection triggered correctly. (FP might come due to a misconfigured MongoDB setup)</li></ul><h3>⚠️ Important Caveat</h3><p>This detection is based on <strong>current public PoC behavior</strong>. A skilled attacker could: Add fake client metadata &amp; may slow down the exploit to look more “normal”.</p><p>But that introduces a trade-off:</p><ol><li>More stealth → much slower data extraction</li><li>Faster extraction → obvious detection</li></ol><p><strong>Defense in depth applies. Detection helps, but</strong> <strong>patching is mandatory</strong>.</p><p>Until then, patch your MongoDB, restrict network access, and keep hunting</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=709c2cd68b26" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ MongoBleed: A Granular Breakdown of MongoDB’s Pre-Auth Memory Leak]]></title>
            <link>https://medium.com/@shreyash_tambe/mongobleed-a-granular-breakdown-of-mongodbs-pre-auth-memory-leak-cbaeaaa8dc43?source=rss-83f6d1b096d5------2</link>
            <guid isPermaLink="false">https://medium.com/p/cbaeaaa8dc43</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[cve]]></category>
            <category><![CDATA[mongodb-atlas]]></category>
            <category><![CDATA[mongodb]]></category>
            <category><![CDATA[mongobleed]]></category>
            <dc:creator><![CDATA[Shreyash Tambe]]></dc:creator>
            <pubDate>Thu, 01 Jan 2026 17:12:53 GMT</pubDate>
            <atom:updated>2026-01-01T17:12:53.320Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RijhHQZTOUHfXZmNuCv7UA.png" /></figure><p><strong>Hola folks!</strong><br> Been busy building detections, reversing things, and living the security life, but it’s time to get back to blogs. Your friendly <strong>malware researcher and threat detection engineer</strong> is here to explain a fresh CVE, not just what it is, but how it really works under the hood, with examples and attacker-level clarity.</p><ul><li><strong>CVE ID:</strong> <em>CVE-2025–14847 </em>nicknamed <strong>MongoBleed</strong> by security folks due to its similarity in impact to the infamous Heartbleed.</li><li><strong>Severity:</strong> <strong>CVSS Score: 8.7/10 (High)</strong></li><li><strong>Affected Software:</strong> MongoDB Server versions, including all 3.6, 4.0, 4.2, and specific ranges within 4.4, 5.0, 6.0, 7.0, 8.0, and 8.2</li><li><strong>Introduced:</strong> The vulnerable code was merged way back around <strong>June 2017</strong>, meaning it had been sitting unnoticed for <em>years</em> and probably being exploited for 8+ years : )</li><li><strong>Disclosed:</strong> MongoDB publicly warned about it in <strong>late December 2025</strong>.</li><li><strong>PoC Released:</strong> Proof-of-concept exploit went public <strong>Dec 26, 2025.</strong></li><li><strong>Who’s Affected:</strong> Any <strong>public-facing MongoDB instance with zlib compression enabled (18000+),</strong> whether or not it has authentication configured.</li></ul><h4><strong>🌐 Why So Many Mongo Servers Are Exposed Online</strong></h4><p>Here’s a reality check: MongoDB is wildly popular as a NoSQL database in cloud environments and microservices. But security? Not always a priority for most startup-scale companies.</p><p>Many developers deploy MongoDB:</p><ul><li>For quick prototyping without tight firewall rules</li><li>With default configurations and no authentication</li><li>On cloud VMs where the database port (27017) is Internet-reachable</li></ul><p>🔍 Scans show <strong>tens of thousands</strong> of MongoDB instances reachable from the public internet, a giant attack surface just waiting. So combine that with a pre-auth vulnerability and boom🔥 (luckily there’s no RCE)</p><h4><strong>☁️ Quick Memory 101 | Heap, C++, and Why This Matters</strong></h4><p>Before getting to how MongoBleed works under the hood, let’s keep memory basics simple:</p><ul><li>In languages like <strong>C++</strong> (which MongoDB’s core is written in), memory is manually managed. There’s <strong>no garbage collector</strong> like Python or Java has, which means leftover memory can stick around and not freed.</li><li>When a function allocates a buffer but doesn’t zero out its contents, there’s a chance <strong>previous data stays in that chunk of RAM</strong>.</li><li>If you accidentally (or intentionally) read beyond what you should, you can see that random memory, which might contain secrets, APIs, tokens, and whatnot. 👀</li></ul><h4><strong>🧠 How MongoDB Authenticates</strong></h4><p>1. <strong>Initial TCP handshake:</strong> Client connects to the MongoDB server on port 27017 : no authentication is required.</p><p>2. <strong>Compression negotiation:</strong> Client and server agree on using <strong>zlib compression</strong> for the wire protocol, and here lies the BUG / CVE. This happens before any login/authentication</p><p>3. <strong>Authentication</strong>: After receiving the command server decided whether that command required authentication or not</p><h4><strong>🧱 MongoDB Wire Protocol: OPCODES and BSON</strong></h4><p>MongoDB communicates using a <strong>binary wire protocol</strong>. Instead of plain JSON, it uses <strong>BSON (Binary JSON)</strong> compact and efficient for machines.</p><p>Some key terms:</p><ul><li><strong>OP_MSG:</strong> The main opcode used to send commands like ping, find, etc.</li><li><strong>OP_COMPRESSED:</strong> When compression is enabled, messages get wrapped in this structure and then sent over to the server, and then the server decompresses the original message wrapped and then takes actions accordingly. <a href="https://www.mongodb.com/docs/manual/reference/mongodb-wire-protocol/#mongodb-wire-protocol">LINK</a></li></ul><p>Here’s the structure of an OP_COMPRESSED message:</p><pre>struct {<br> MsgHeader header; // standard message header<br> int32 originalOpcode; // value of the wrapped opcode<br> int32 uncompressedSize; // size after decompression, excluding header<br> uint8 compressorId; // which compressor is used<br> char *compressedMessage; // the compressed payload (excluding header)<br> }</pre><p>Explanation:</p><ul><li><strong>originalOpcode</strong> tells the server what the real command is (ex: OP_MSG).</li><li><strong>uncompressedSize</strong> claims how big the decompressed data <em>should</em> be.</li><li><strong>compressorId</strong> is an ID for zlib, snappy, zstd, etc.</li><li><strong>compressedMessage</strong> is the actual payload to decompress.<br> (<em>The flaw lies in how uncompressedSize is blindly trusted by server</em>)</li></ul><h4><strong>🩸 The MongoBleed Explained:</strong></h4><p>Threat actors craft a simple <strong>message (OP_MSG), e.g. ping command, and use </strong>OP_COMPRESSED<strong> opcode to compress the original message/command</strong></p><p><strong>As you can see from the above structure of OP_COMPRESSED opcode, there’s one field “uncompressedSize” which tells the server the amount of memory required for this message to get decompressed</strong></p><p><strong>At the server side </strong><a href="https://github.com/mongodb/mongo/blob/master/src/mongo/transport/message_compressor_zlib.cpp"><strong>message_compressor_zlib</strong></a><strong>.cpp contains code for decompression, which blindly trusts the claimed size and returns the entire allocated memory after decompression instead of returning the actual decompressed memory.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/772/1*Yu4XF4E-9nstxrtswhX0Vg.png" /><figcaption>Implemented fix</figcaption></figure><p>Let’s take an example:</p><ul><li>TA sends a regular ping command, compressed it’s ~3KB and uncompressed ~5KB normally</li><li>But in the crafted exploit, they <strong>set uncompressedSize to 500KB intentionally</strong></li><li>The server dutifully allocates <strong>500KB of heap memory</strong>, decompresses the actual small message into that memory, and then sends back the entire 500KB buffer instead of 5KB in response.</li><li><strong>Result:</strong> That “extra” memory (455KB) likely contains sensitive bits like credentials, APIs , secrets, keys, session tokens got leaked. Not every byte is useful, but <strong>repeated probing</strong> can leak high-value secrets over time.</li></ul><h4><strong>🛠 Recommendations &amp; Mitigations</strong></h4><p>If you manage MongoDB systems, here’s your patch-and-protect checklist:</p><ol><li><strong>Patch Immediately : </strong>Upgrade to <strong>patched versions</strong> e.g. 8.2.3+, 8.0.17+, 7.0.28+, 6.0.27+, 5.0.32+, 4.4.30+ as provided by MongoDB since <strong>Dec 19, 2025</strong>.</li><li><strong>Network Segmentation : </strong>Make sure your MongoDB servers aren’t directly exposed to the Internet.<br> Only trusted networks should access database ports.</li><li><strong>Disable zlib Compression : </strong>If you can’t patch <em>right now</em>, turn off <strong>zlib</strong> in your network message compressors.</li><li><strong>Filter &amp; Firewall : </strong>Block port <strong>27017</strong> from public access &amp; Use VPN or private peering for environments</li><li><strong>Rotate credentials </strong>: your creds might have been leaked in the past, as this CVE has been exploited for a long so better to rotate all the creds</li><li><strong>Monitor &amp; Audit : </strong>Check the MongoDB logs, transport them to SIEM &amp; look for unusual connection attempts or malformed compressed traffic.</li></ol><p><strong>✨ In the next part, I’m going to explain how to hunt/detect this from your mongo logs | gonna write a query based on risk scoring factor, so stay connected!</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cbaeaaa8dc43" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ CrowdStrike’s Automated Leads: Signal AI Fatigue is Real — Here’s How to Tackle It]]></title>
            <link>https://medium.com/@shreyash_tambe/crowdstrikes-automated-leads-signal-ai-fatigue-is-real-here-s-how-to-tackle-it-f4d7a7306dc8?source=rss-83f6d1b096d5------2</link>
            <guid isPermaLink="false">https://medium.com/p/f4d7a7306dc8</guid>
            <category><![CDATA[crowdstrike]]></category>
            <category><![CDATA[crowdstrike-outage]]></category>
            <category><![CDATA[edr]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Shreyash Tambe]]></dc:creator>
            <pubDate>Thu, 07 Aug 2025 16:59:24 GMT</pubDate>
            <atom:updated>2025-08-07T16:59:24.806Z</atom:updated>
            <content:encoded><![CDATA[<h3>🚨 CrowdStrike’s Automated Leads: Signal AI Fatigue is Real — Here’s How to Tackle It</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Fnq2hIK8Fb7wJ1n9EDXZyQ.png" /></figure><h3>📌 What’s Going On?</h3><p>CrowdStrike recently rolled out a feature called <strong>Automated Leads</strong>, powered by their new <strong>Signal AI engine</strong>. Unlike traditional global detection models that rely on static signatures or rules, Signal operates with <strong>local context</strong> , learning behaviors unique to your endpoints, users, and tenants.</p><p>In theory? Smart.<br>In practice (at least right now)? <strong>Painful.</strong></p><h3>💥 The Pain Point</h3><p>Since the launch, many users are reporting serious alert fatigue. Here’s what’s causing frustration:</p><ul><li><strong>Flood of alerts</strong>: Every change in detection confidence creates a new event in SIEM.</li><li><strong>Vanishing detections</strong>: Clicking into alerts sometimes leads nowhere — some detections disappear (alert_url).</li><li><strong>Low signal, high noise</strong>: Most leads don’t result in anything worth investigating.</li><li><strong>RTR (Remote Response) confusion</strong>: Even CrowdStrike’s own automated RTR activity gets flagged.</li><li><strong>Lack of clarity</strong>: Events tied to EPP (Endpoint Protection Platform) show up as detections, but with no visibility in the actual EPP dashboard.</li><li><strong>Hard to correlate</strong>: The link between Indicators and Detections inside a lead isn’t always clear.</li></ul><h3>What Are “Automated Leads” Anyway?</h3><p>Think of this as a <strong>Local Intelligence, Not Global Rules:</strong> Unlike global detection logic that applies across all Falcon customers, <strong>Signal AI evaluates what’s “normal” specifically in <em>your</em> environment</strong> .: per host, per user, per tenant.</p><p>For example:</p><ul><li>PowerShell being run by IT_Admin_01 on a server? Probably fine.</li><li>Same activity by user_temp12 on a finance endpoint? Suspicious.</li></ul><p><strong>❌ Not Traditional Detections</strong><br> These aren’t your usual rule-based detections like malware signatures or known IOCs. <strong>Automated Leads</strong> are surfaced even when no formal detection is triggered.</p><p>Leads can contain:</p><ul><li><strong>Known detections</strong> (visible in the Detections UI)</li><li><strong>Indicators of Behavior (IOBs</strong>) — subtle signals from telemetry like process lineage, script execution, rare network activity, etc.</li><li>Or a<strong> combinatio</strong>n of both</li></ul><p>These Indicators are not typically shown in the EPP console, but are accessible in:</p><blockquote><em>NG SIEM → Dashboards → Indicator Activity</em></blockquote><p><strong>Dynamic and Adaptive</strong><br>Signal constantly <strong>learns behavioral baselines</strong>. As patterns evolve, so does its judgment — which means:</p><ul><li><strong>False positives can go down over time</strong></li><li>But <strong>alerts may feel random at first</strong> (especially in early rollout)</li></ul><h3>🔄 Will This Replace CrowdScore?</h3><p>Yes. Eventually, Automated Leads + Signal AI is expected to replace<strong> CrowdScore</strong>, but only after it’s matured and battle-tested with customer feedback.</p><blockquote>CrowdScore, in the context of <a href="https://www.google.com/search?rlz=1C1RXQR_enIN1023IN1023&amp;cs=1&amp;sca_esv=84a972cf9bb6250a&amp;sxsrf=AE3TifMoDPd1LSCpsulMm105UfaGh0_HTg%3A1754584382909&amp;q=CrowdStrike&amp;sa=X&amp;ved=2ahUKEwjMgeWWkPmOAxVCRWcHHfqwDRAQxccNegQIAxAB&amp;mstk=AUtExfAjqFGK6hkRr3EthwyaL8znRYGXuB7SkiOzl5_svnQkd99bQygf8AA3iRDNNxaDv3T2hdgMqtBEd7ZPOAehuC2Tn60rmyGvCJP6ATZKWZ9tYG33lhjArPeOIfNL8Gf5P7TUykO2eZVSmdZHpFOyiSnFV_7lAfcEe9ciFs8s1m-ucuw&amp;csui=3">CrowdStrike</a>, is a real-time threat-level metric that assesses the potential risk of hostile activity targeting an organization</blockquote><h3>🛠️ How to Tame the Noise (for Now)</h3><p>While there’s <strong>no official toggle to turn it off</strong>, here’s what you <em>can</em> do:</p><p><strong>Filter out low-confidence alerts in your SIEM</strong> (if possible).</p><p>Work with your <strong>CrowdStrike TAM or SE</strong> — raise a ticket, share logs, report false positives.</p><p>Keep tabs on updates — the team is iterating fast and will likely add:</p><ul><li>Ability to close detections bundled inside Leads</li><li>API improvements to avoid alert duplication</li></ul><h3>Final Thoughts</h3><p>New AI-powered detections sound great on paper, but when they trigger constant noise without actionability : it burns out the SOC. Signal AI and Automated Leads have potential, but right now? It feels like the tool needs more tuning before it can truly help analysts, not overwhelm them.</p><p>If you’re feeling the heat — <strong>you’re not alone</strong>. Engage your TAM, filter wisely, and give feedback. Let’s hope this powerful engine gets some better brakes soon.</p><p><em>Seen similar behavior or figured out a better workaround? Drop it in the comments — let’s enlighten each other.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f4d7a7306dc8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[️The SharePoint Heist :]]></title>
            <link>https://medium.com/@shreyash_tambe/%EF%B8%8Fthe-sharepoint-heist-0eca0b32880b?source=rss-83f6d1b096d5------2</link>
            <guid isPermaLink="false">https://medium.com/p/0eca0b32880b</guid>
            <category><![CDATA[detection]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[webshell]]></category>
            <category><![CDATA[vulnerability]]></category>
            <category><![CDATA[microsoft]]></category>
            <dc:creator><![CDATA[Shreyash Tambe]]></dc:creator>
            <pubDate>Wed, 23 Jul 2025 19:16:54 GMT</pubDate>
            <atom:updated>2025-07-23T19:16:54.827Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>🛡️</strong>The SharePoint Heist :</h3><h3>Ft. ToolShell, Machine Keys &amp; ViewState Mischief</h3><p><strong>About the vulnerability ::</strong></p><p>Meet <strong>CVE-2025‑53770</strong> <br> It is a <strong>CRITICAL</strong> (CVSS 9.8) deserialization vulnerability that affected exclusively<strong> on-prem</strong> Microsoft <strong>SharePoint</strong> servers that allows unauthorized Remote Code Execution <strong>RCE</strong>.</p><blockquote>Which leads Threat Actors to exploit two flaws:<br><strong>Authentication bypass</strong> to sneak in via header spoofing <strong>CVE-2025–53771</strong><br><strong>Insecure .NET deserialization</strong> of untrusted input <strong>CVE-2025‑53770</strong></blockquote><p>These two vulnerabilities are chained together in an exploit known as <strong>ToolShell</strong> :<br><br> Let’s dive deeper inside <strong>ToolShell :</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fJkTCYlfG5QYCNrD4m4_sw.jpeg" /></figure><blockquote>Initial Entry (CVE-2025–53771):</blockquote><p>TA sends a malicious POST request to</p><p>/_layouts/15/ToolPane.aspx?DisplayMode=Edit&amp;a=/ToolPane.aspx HTTP/1.1</p><p>using a crafted <strong>Referer header</strong> : (/_layouts/SignOut.aspx) to bypass authentication. This exploits a header spoofing vulnerability that causes SharePoint to treat the request as authenticated.</p><pre>WHY ToolPane.aspx ?<br>ToolPane.aspx is used when editing web parts on a SharePoint page. <br>It accepts dynamic parameters, including paths to other files, and <br>blindly deserializes ViewState object on backend.<br><br>Why SignOut.aspx ?<br>Server just checks:<br> &quot;Hey, did this request come from SignOut.aspx?&quot;<br>If yes, Server assumes:<br> &quot;This must be a user logging out or in the process of re-authentication. Let&#39;s allow it.&quot;</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/721/1*DXdJI2nnzLjG06tgaTU_Bw.png" /></figure><p>Initial access: unlocked. Now let’s see what chaos they cooked up next</p><blockquote>Shell Delivery (CVE-2025–53770):</blockquote><p>With authenticated access to the vulnerable <strong>ToolPane.aspx</strong> within the server, the attacker exploits an insecure deserialization vulnerability by submitting a malicious PowerShell payload at the bottom in the POST header [highlighted in above screenshot]. <br> <br>SharePoint deserializes attacker-controlled data without validation, leading to the execution of embedded commands. This results in dropping a stealthy ASPX <strong>web shell</strong> (<strong>spinstall0.aspx</strong>) into a directory such as</p><p>C:\Program Files\Common Files\Microsoft Shared\Web Server Extensions\16\TEMPLATE\LAYOUTS\.</p><p>This web shell acts as a <strong>backdoor</strong>, giving attackers a VIP pass for persistent remote access. More details about PowerShell payload are in the screenshot below .:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Yb1RSGer6HDsq3565DGZzA.png" /></figure><p>This Base64 decoded PowerShell payload will extract the<strong> server’s machine keys</strong> and write them to the spinstall0.aspx file and drop it in the SharePoint layouts directory.</p><blockquote>Key Extraction:</blockquote><p>The dropped web shell ( spinstall0.aspx) contains stolen machine keys.<br> Further, this file is used to extract sensitive cryptographic material, including the <strong>ValidationKey</strong> and <strong>DecryptionKey</strong> from the server’s <strong>machineKey configuration</strong>. These keys are critical for crafting valid &amp; signed <strong>ViewState payloads</strong> that SharePoint will trust.</p><pre> This GET request will return stolen machine keys as a response : <br><br>2025-07-18 18:xx:05 &lt;proxy masked&gt; GET /_layouts/15/spinstall0.aspx - 443 - &lt;proxy masked&gt; Mozilla/5.0+(Windows+NT+10.0;+Win64;+x64;+rv:120.0)+Gecko/20100101+Firefox/120.0 /_layouts/SignOut.aspx 200 0 0 31</pre><blockquote>Remote Code Execution:</blockquote><p>Using tools like <a href="https://github.com/frohoff/ysoserial"><strong>ysoserial</strong></a>, the attacker crafts malicious <strong>__VIEWSTATE</strong> payloads <strong>signed</strong> with the stolen keys. Because SharePoint trusts the signature, it deserializes and executes the further embedded payloads. This enables unauthenticated remote code execution <strong>RCE</strong>, even after the initial foothold — completing the attack chain.</p><p><strong>Observed IOCs</strong></p><ul><li>Shell name: spinstall0.aspx</li><li>SHA256 hash: 92bb4ddb98eeaf11fc15bb32e71d0a63256a0ed826a03ba293ce3a8bf057a514</li><li>Referer: /_layouts/SignOut.aspx</li><li>POST to /_layouts/15/ToolPane.aspx?DisplayMode=Edit</li><li>GET request to /_layouts/15/spinstall0.aspx</li><li>Process name: <strong>w3wp.exe</strong> spawning encoded PowerShell in server</li><li>IP addresses: 107.191.58[.]76, 104.238.159[.]149, 96.9.125[.]147, 103.186.30[.]186</li><li>File path:<br> C:\PROGRA~1\COMMON~1\MICROS~1\WEBSER~1\16\TEMPLATE\LAYOUTS\spinstall0.aspx</li></ul><p><strong>Mitigations .:</strong></p><ul><li>Apply the latest SharePoint <strong>security</strong> <strong>patch</strong></li><li>Enable <strong>AMSI</strong> in Full Mode</li><li><strong>Rotate</strong> ASP.NET <strong>Machine Keys</strong> after patching or enabling AMSI:</li><li>Assume compromise if the server was exposed to the internet:</li><li><strong>Isolate</strong>/shut down the server</li><li>Revoke credentials and rotate secrets</li><li>Engage <strong>CSIRT</strong> teams</li><li>Monitor for Indicators of Compromise (IOCs).</li><li>Enable <strong>WAF</strong> rules</li></ul><p>Stay safe, folks — patch and rotate those keys!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0eca0b32880b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[DNS Tunneling Att&ck]]></title>
            <link>https://medium.com/@shreyash_tambe/dns-tunneling-att-ck-53b5584eccc1?source=rss-83f6d1b096d5------2</link>
            <guid isPermaLink="false">https://medium.com/p/53b5584eccc1</guid>
            <category><![CDATA[hacking]]></category>
            <category><![CDATA[data-exfiltration]]></category>
            <category><![CDATA[dns-servers]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[dns]]></category>
            <dc:creator><![CDATA[Shreyash Tambe]]></dc:creator>
            <pubDate>Sat, 28 Dec 2024 19:53:53 GMT</pubDate>
            <atom:updated>2024-12-28T19:53:53.115Z</atom:updated>
            <content:encoded><![CDATA[<h4>Stealth in the Shadows</h4><p>In this blog, we will Unlock the Secrets of DNS Tunneling, explore how attackers use it, and outline ways to detect and prevent it — all in simple terms.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nyzhLyvRQP-MTO2ujXSY9A.jpeg" /></figure><h4><strong>What is DNS?</strong></h4><p>Let’s briefly revise what DNS (Domain Name System) is. DNS acts as the internet’s phonebook, translating human-readable domain names like example.com into IP addresses that computers use to communicate.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/850/0*RHpfRijBIHs0dwnJ.png" /><figcaption>DNS Resolution</figcaption></figure><p><strong>DNS Tunneling</strong> manipulates this legitimate process of DNS resolution. It’s a method where attackers use DNS queries and responses, to create a covert communication channel between a compromised system and a remote server (C2). This tunnel can then be used to send commands to malware or exfiltrate sensitive data, often bypassing firewalls and intrusion detection systems (IDS).</p><h4>How Does DNS Tunneling Work?</h4><p>DNS Tunneling typically involves the following steps:</p><ol><li><strong>Infecting the Victim</strong>: Attackers first compromise a device within the target network using phishing emails, malicious downloads, or exploiting vulnerabilities.</li><li><strong>Domain Registration:</strong> The attacker registers a domain name that they control.</li><li><strong>Establishing a Covert Channel</strong>: The malware on the victim’s machine communicates with a command-and-control (C2) server by embedding data inside DNS queries. These queries look like regular DNS traffic, making them difficult to detect.</li><li><strong>Exfiltrating Data or Sending Commands</strong>: The attacker can use this tunnel to:</li></ol><ul><li>Exfiltrate sensitive information like passwords or proprietary data.</li><li>Send commands to the malware to carry out further attacks.</li></ul><h4>Example Scenario .:</h4><p>Imagine an attacker embeds sensitive information inside a DNS query, such as confidentialdata.attacker.com. The attacker’s Authoritative DNS server decodes the query, extracts the information, and sends a response back through the same channel.</p><pre>Suppose the attacker wants to exfiltrate the following information:<br><br>Password: pass456<br><br>Base64 Encoding:<br>Password (pass456) encoded in Base64: cGFzczQ1Ng==<br><br>DNS Query: The attacker could send a DNS query like this <br>to the domain example.com:<br><br><br>cGFzczQ1Ng==.example.com</pre><h4>Why is DNS Tunneling Effective?</h4><ol><li><strong>The Global presence of DNS</strong>: DNS is a foundational internet protocol rarely blocked or scrutinized by firewalls. Most firewalls and security systems are configured to allow DNS traffic without thorough inspection.</li><li><strong>Bypassing Security Tools</strong>: Since DNS traffic often goes unchecked, it’s an attractive medium for covert operations.</li><li><strong>Stealthiness</strong>: Since DNS queries are common, high in volume and expected, unusual patterns can be difficult to detect amidst regular network noise.</li><li><strong>Limited network traffic monitoring in EDR</strong>: EDR solutions primarily focus on endpoint-level monitoring and behavior analysis. They may not always have full visibility into network-level activities such as DNS tunneling.</li><li><strong>Lack of DNS-specific detection rules in SIEM</strong>: SIEM systems rely heavily on predefined rules. Many SIEMs are not optimized for detecting DNS tunneling unless they are specifically configured with DNS-specific rules and thresholds, which might not always be in place.</li></ol><h3>How to Detect DNS Tunneling</h3><ol><li><strong>Monitor DNS Traffic</strong>: Use tools to analyze DNS traffic for anomalies such as <strong><em>unusually long domain names</em></strong>, excessive query rates, or <strong><em>encrypted DNS Traffic</em></strong>.</li><li><strong>Machine Learning Models</strong>: Use AI-Based Solutions, Some advanced security solutions utilize artificial intelligence to detect patterns indicative of DNS tunneling attacks</li><li><strong>Regular Audits for port 53:</strong> Conduct periodic reviews of your network’s DNS logs to identify any anomalies.</li></ol><pre>Wireshark:<br><br>Apply the filter:<br>udp.port == 53 || tcp.port == 53<br>Look for packets labeled DNS.<br><br><br>tcpdump:<br><br>Use the command:<br>tcpdump -n port 53<br></pre><p>Have you encountered DNS Tunneling in your organization, or are you looking for more ways to secure your DNS traffic? Share your thoughts below.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=53b5584eccc1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[DLL File Analysis: A Practical Guide for Malware Analysts]]></title>
            <link>https://medium.com/@shreyash_tambe/dll-file-analysis-a-practical-guide-for-malware-analysts-9e0a8cf82d1d?source=rss-83f6d1b096d5------2</link>
            <guid isPermaLink="false">https://medium.com/p/9e0a8cf82d1d</guid>
            <category><![CDATA[dll-hijacking]]></category>
            <category><![CDATA[dll-injection]]></category>
            <category><![CDATA[windows]]></category>
            <category><![CDATA[dll-file]]></category>
            <category><![CDATA[malware-analysis]]></category>
            <dc:creator><![CDATA[Shreyash Tambe]]></dc:creator>
            <pubDate>Sat, 26 Oct 2024 14:42:55 GMT</pubDate>
            <atom:updated>2024-10-26T14:42:55.260Z</atom:updated>
            <content:encoded><![CDATA[<p>Dynamic Link Libraries (DLLs) are essential components in Windows operating systems, enabling <strong>code reuse</strong> and modular programming. However, they can also be misused for malicious purposes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dt0JcbnqLhXVtVg3sx6F7g.png" /></figure><h4>What are DLLs?</h4><p>DLLs are <strong>shared libraries</strong> that contain <strong>code and data</strong> that can be<strong> used by multiple programs simultaneously. </strong>They allow developers to create modular applications and reduce the overall memory footprint by sharing common functions (<strong>Data Redundancy</strong>).</p><p><strong>How are DLLs loaded into memory?</strong><br> — <strong>Load-Time Dynamic Linking:</strong> DLLs are loaded when the executable is initially run. [prone to DLL sideloading / DLL hijacking vulnerabilities] <br> — <strong>Run-Time Dynamic Linking: </strong>DLLs are loaded at a specific point in the program’s execution, allowing for dynamic and flexible function calls. [prone to vulnerabilities such as DLL injection]</p><h4>Analyzing COM vs Non-COM DLLs :</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/675/1*NKQiNwNz_vPrW571ky59HQ.png" /></figure><blockquote><strong>COM DLLs</strong></blockquote><p>The Component Object Model (COM) architecture is a Windows framework that enables <strong>reusable software components </strong>to interact with each other, even if they’re written in different programming languages, by using <strong>Windows Registry</strong> to manage component <strong>registration</strong> and accessibility.</p><p><strong>COM DLLs</strong> are designed to create and manage these objects within the COM architecture by <strong>manipulation through registered components in the Windows Registry. That’s why COM DLLs need to be registered.</strong></p><p><strong>Example-:</strong> A typical use case might be a <strong>shared spell-check component [COM DLL] in Microsoft Office that both Word and Excel can use without duplicating the code for spell-checking in each program. </strong>When Word or Excel needs spell-checking capabilities, they query the <strong>Windows Registry</strong> to locate and load the registered spell-check component by referencing its unique Class ID (<strong>CLSID</strong>) or Programmatic Identifier (<strong>ProgID</strong>).</p><p><strong>How to identify COM DLL-:</strong></p><p>Use tools like <strong>Dependency Walker</strong> or <strong>PE Explorer</strong> to examine the exports in the DLL:</p><ul><li>Look in the “Export” section for DllRegisterServer &amp; DllUnregisterServer. If these exports are there it indicates that it’s a COM DLL</li></ul><h4>Analysis Techniques Of COM DLLs-:</h4><p>1. Registering COM DLL:<br>The regsvr32.exe utility is primarily designed to register and unregister only <strong>COM DLLs</strong>, not non-COM DLLs.</p><p><strong>What Does Registering a DLL Mean?</strong></p><p>When a COM DLL is registered, information about the DLL and its interfaces (such as CLSIDs and ProgIDs) is added to the <strong>Windows Registry</strong>. The registry entries act as a map for Windows, allowing other applications to locate and use the COM object. For instance, Microsoft Word or any program that needs a shared component (like spell-checking or a graphics filter) can look up the registry, find the CLSID, and use that COM object.</p><pre>[Open cmd as administrator and type below commands]</pre><pre>1. Registering a COM DLL</pre><pre>      regsvr32 path\to\your\comdll.dll<br></pre><pre>2. Unregistering a COM DLL</pre><pre>      regsvr32 /u path\to\your\comdll.dll<br></pre><pre>When you run regsvr32.exe with a DLL file, <br>it attempts to call below functions within the DLL:</pre><pre><strong>DllRegisterServer</strong>: Adds the necessary registry entries for the COM object.<br><strong>DllUnregisterServer</strong>: Removes the COM object&#39;s registry entries.</pre><p>2. Registry Analysis:<br>After running the above 1st command Check the Windows registry for entries related to the DLL, such as <strong>CLSID</strong> and <strong>ProgID</strong> keys. Use tools like <strong>RegScanner</strong>, <strong>Regedit, ProcMon </strong>or <strong>Autoruns</strong> to i<strong>nspect the registry for suspicious entries.</strong></p><p>If the DLL is registered, you can check the Windows Registry for entries associated with it. Look for the <strong>CLSID</strong> or <strong>ProgID</strong> entries, typically found under:</p><ul><li>HKEY_CLASSES_ROOT\CLSID\ – contains Class IDs for registered COM components.</li><li>HKEY_CLASSES_ROOT\Interface\ – lists interfaces exposed by COM components.</li><li>Also, look for any suspicious network activity in the background.</li><li><strong>Legitimate Use:</strong> Register or unregister COM DLLs, such as ActiveX controls.</li><li><strong>Malicious Use:</strong> Executing hidden code during DLL registration.</li></ul><blockquote><strong>Non-COM DLLs</strong></blockquote><p>Non-COM DLLs are standard traditional DLLs that do not use the COM architecture. They are typically loaded and used through conventional function calls. In Non-COM DLLs <strong>no registries are involved instead directly functions are exported to running programs.</strong></p><p><strong>Non-COM DLLs</strong>, don’t need this registration because they are typically used only by a specific application that knows where they’re located on the file system. They are loaded directly by the application without any need to “announce” their availability through the registry.</p><h4>Analysis Techniques Of NON-COM :</h4><p><strong>1. Simply check the Export Function Name / Ordinal :</strong><br> — Use a tool like CFF Explorer and check for the Export Name / respective ordinal for export.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/817/1*Z7052cenfm56nYbb8Wo9Pg.png" /></figure><p><strong>2. Execute Exports one by one:</strong></p><pre>1. Executing an Exported Function by Name<br>      Syntax:        <br>            <strong>rundll32.exe path\to\dll,FunctionName</strong><br>2. Executing an Exported Function by Ordinal<br>      Syntax:<br>            <strong>rundll32.exe path\to\dll,#OrdinalNumber<br>**NOTE**</strong><br>In case of too many exports use the tool <strong>DLLRunner</strong> to execute all exports at once.<br><strong>Ref: </strong><a href="https://github.com/Neo23x0/DLLRunner"><strong>https://github.com/Neo23x0/DLLRunner</strong></a></pre><p><strong>3. Memory Analysis:</strong> <br> — Use all Dynamic Analysis tools like ProcMon, RegShot, NetMon, Fiddler, ProcessExplorer, etc., and check the changes made by malicious DLL in a system like file dropping, network activity, etc.</p><h4>**NOTE**</h4><p>Remember we are using rundll32.exe as a dummy application to load the DLL under analysis.<br>Sometimes rundll32.exe doesn’t support the architecture of a few DLLs.</p><p>At that time what you can do is just go to <strong>VirusTotal</strong> and try to find out the <strong>DLLs executable parent in the relation tab.</strong><br>Now you have found out parent executable to confirm it’s the right one go to the Import section of that executable and try to find out if the name of our dll is there if it’s there it means that the executable is going to load our dll during execution.<br>So now <strong>download that executable and our dll in the same folder &amp; run that executable.</strong> here <strong>we simply use the parent executable rather than rundll32.exe. </strong><br>This tip will be more useful for analysis. :)</p><h4>Analyzing DLLs with Export vs Without Exports</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*CabREpYEMcFd9eI9ozhDNg.png" /></figure><blockquote><strong>DLLs with Exports</strong></blockquote><p>Exports are functions that a DLL makes available for use by other executables. These exports are listed in the DLL’s export table and can be called directly by other modules. <strong>Each export</strong> has a <strong>Name</strong> or an <strong>Ordinal</strong>, <strong>(A Numeric identifier often used by malware to evade detection.)</strong></p><pre>Analysis :</pre><pre>If Exports contains <strong>DllRegisterServer</strong> &amp; <strong>DllUnregisterServer</strong> <br>then use <strong>regsvr32</strong> commands</pre><pre>Otherwise, <br>use <strong>rundll32.exe</strong> command mentioned above for the rest of the exports.</pre><blockquote><strong>DLLs without Exports</strong></blockquote><p>DLLs without exports do not expose any functions or variables for external use and are typically used only by the application that loads them. <strong>DLLs without exports can be used as resources, allowing applications to access images, icons, and other data without needing direct function calls.</strong></p><pre>Analysis :<br><br>If a DLL doesn&#39;t have any exports, it can still contain malicious code, <br>especially within its DllMain function. <br><br>The <strong>DllMain function</strong> is a special entry point that the system calls when the DLL is loaded or unloaded, <br>and it can perform various actions, such as initializing global variables, allocating resources, or executing malicious payloads.<br><br>To Execute DllMain we need a dummy command :<br><br><strong>Syntax:        <br>            rundll32.exe path\to\dll,test<br><br></strong>Here, <strong>test is a dummy name</strong> There is no such export in our DLL so obviously it&#39;ll give an error popup,<br>But this trick will help you to execute DllMain function.<br>Most of the malicious DLLs write <strong>C2 connection code</strong> in DllMain, so better to use network monitoring tools while analyzing :)</pre><p>This expanded guide provides more depth and actionable insights, making it a valuable resource for those analyzing DLL files in cybersecurity. Let me know if you’d like to add further details!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9e0a8cf82d1d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cracking AgentTesla: Revealing Steganographic Payload]]></title>
            <link>https://medium.com/@shreyash_tambe/cracking-agenttesla-revealing-steganographic-payload-37f5c0c54a8d?source=rss-83f6d1b096d5------2</link>
            <guid isPermaLink="false">https://medium.com/p/37f5c0c54a8d</guid>
            <category><![CDATA[reverse-engineering]]></category>
            <category><![CDATA[malware]]></category>
            <category><![CDATA[debugging]]></category>
            <category><![CDATA[malware-analysis]]></category>
            <category><![CDATA[agenttesla]]></category>
            <dc:creator><![CDATA[Shreyash Tambe]]></dc:creator>
            <pubDate>Thu, 05 Sep 2024 18:05:37 GMT</pubDate>
            <atom:updated>2024-09-05T18:05:37.022Z</atom:updated>
            <content:encoded><![CDATA[<p>Reversing .NET Executables and DLL Payloads</p><p>AgentTesla, a notorious malware that has evolved into one of the most prevalent info-stealers, continues challenging security professionals. Named after the famed inventor Nikola Tesla, the malware is just as ingenious as its namesake, employing sophisticated techniques like steganography to hide its payloads in plain sight. But what makes it even more elusive is its integration into .NET applications, using Microsoft Intermediate Language (MSIL) to obfuscate its code.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*59qVPicRqHz58Bf8g7oQxQ.png" /></figure><p>In this blog, we’ll explore AgentTesla’s origins and how it has become a favorite tool for cybercriminals. I’ll also walk you through practical steps for <strong>reverse-engineering and debugging MSIL code in .NET samples for Exe’s and Dll’s</strong>. Whether you’re new to malware analysis or a seasoned pro, this tutorial will equip you with the tools to dismantle AgentTesla’s tricks and expose its malicious payloads.</p><p>For this blog, I have selected .net based sample from <strong>MalwareBazaar</strong> named <strong>cOyVy.exe</strong><br>SHA256: <strong>c4e8c058d4f4a9976ca89a275963e6c4e2b3a2a7e90ad2dd095e173c2f5859bf</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*83mITucO4mq3T9dQR36I0A.png" /><figcaption>Attack Flow</figcaption></figure><h4>Initial Loader :</h4><p>The initial loader, a .NET executable cleverly disguised as <strong>cOyVy.exe</strong>. It further loads 3 different DLLs hidden with 3 different techniques.<br>Let’s drag it quickly to <strong>DnSpy</strong> and get to the Entry Point of the code.</p><p>Inside cOyVy.exe it contains an encrypted payload in the form of a <strong>byte array named “re” in the resource section </strong>of the executable. At runtime, this payload is decrypted using an <strong>XOR</strong> function.</p><p>The XOR function relies on a list of numbers, where each element is paired with a corresponding byte of the encrypted data. This <strong>decryption loop</strong> resets every 22 bytes, creating a sophisticated cycle that adds an extra layer of complexity to the loader’s tactics.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/253/1*naTeSGMJJjkguLHTDDre5Q.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/643/1*1MMrf4eQdJo128oaMl9ycg.png" /><figcaption>Image-I</figcaption></figure><p>Finally, we get the Decrypted payload in List named <strong>Data_ <br>0x4D 0X5A</strong> indicates the ‘<strong>MZ</strong>’ header in <strong>PE</strong> files.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/476/1*RvXtJTc3-kOt1BVDtTy9DQ.png" /><figcaption>Image-II</figcaption></figure><p>but it’s not a complete payload, to get that complete Decrypted Payload we further need to put a breakpoint on <strong>Wr_99. <br></strong>Also, the Malware Author has split a String ‘<strong>MN0</strong>’ based on <strong>‘$’</strong> and saved it in an Array named <strong>‘JJ’.</strong> Further, they have created another Array named <strong>‘array’</strong> that has 3 different values as per below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fqQCrl0WJvJ78ol6B_VyEA.png" /><figcaption>Image-III</figcaption></figure><p>Now we have a DLL payload named ‘<strong>pathologist.dll</strong>’ inside ‘<strong>Data_</strong>’ list as per ‘<strong>Wr_99</strong>&#39;. Also, an Array named ‘<strong>JJ</strong>’ contains 2 values i.e. secret keys that we’ll find out later in this blog.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/719/1*iOmDKw1_DPySSa-6d0Jihw.png" /><figcaption>Image-IV</figcaption></figure><p>so now as per Image-I <strong>CreateInstance</strong>() method will load <strong>‘pathologist.dll’</strong> in memory along with ‘array’ that contains <strong>3 secret values.</strong><br>Further <strong>‘pathologist.dll’ </strong>is packed with <strong>SmartAssembly</strong> so I’ll suggest to use de4Dot to make things easier.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/346/1*2YU05jtzobrK4_aVh0a0QA.png" /><figcaption>Image-V</figcaption></figure><p>The running executable will go to sleep for 10sec then our second payload is hidden in ‘<strong>Ar.reg’</strong> as compressed <strong>GZIP</strong> that further gets decompressed and creates its instance with the name ‘<strong>obj</strong>’. Further, a function named ‘<strong>CausalitySource</strong>’ of obj payload {Gamma.dll} is called along with 2 secret values one by one.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1015/1*gBnArj8i6ZHcyAc2nf7jfg.png" /><figcaption>Image-VI</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/545/1*ngn4Ce2grkM9HFN0yUfung.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/762/1*LJOytGRhNU6WX4Qus2NBHQ.png" /><figcaption>Image-VII</figcaption></figure><p>A function named ‘CausalitySource’ returned the decoded secret keys as it’s simply hex to Ascii conversion: <br>Now we have decoded 3 Secret keys passed by cOyVy.exe</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/472/1*tdVQchzoq8Jkj_zCl9BnJw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/466/1*MxkNca262i4hMGCxCBJ6pw.png" /></figure><p>Further in Image-VI, below 2 secret keys are passed to get the bitmap image.</p><pre>{<br>StringTypeInfo=&#39;UJARP&#39; #Name of steganographic image<br>EscapedIRrmotingFormatter=&#39;FlightBookingApp&#39; #Location of steganographic image<br>}</pre><p>If you look closely, Image-I of <strong>cOyVy.exe</strong> it contains a PNG File named “<strong>UJARP</strong>“ at the location “<strong>FlightBookingApp”</strong> in the resource section.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/793/1*kKudSyTCCgkYt47g0apB0Q.png" /></figure><p>Here, the malware author cleverly employs <strong>steganography</strong>, a technique used to hide malicious code or data within seemingly harmless images. In this case, the <strong>PNG file acts as a vessel</strong>, concealing further payloads or instructions, making detection by traditional security tools even more challenging. This hidden data can be extracted and executed at runtime, further obfuscating the malware’s true purpose.</p><p>Further, in Image-VI bitmap image is changed by <strong>150 pixels</strong> so that malware can load its 3rd payload named ‘<strong>Tyrone.dll</strong>’ in <strong>array2</strong> from <strong>steganographic </strong>image.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/532/1*zsNo7BFBYv_uj9pFH3lWZA.png" /></figure><p>This ‘Tyrone.dll’ is also SmartAssembly packed and highly obfuscated so we need to use de4dot as well as renaming of few functions and variables to understand things clearly.</p><p>It also can download a file into %TEMP% file.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/837/1*pDHFtfp_A0BvIawO3eVCpA.png" /></figure><p>It will load the initial file into %APPDATA% folder. The name is different but it is actually the same as the initial file {same SHA256}.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/861/1*qwZecLcnvJF75BTvTs1NDw.png" /></figure><p>As you continue debugging, you’ll eventually come across the core of <strong>AgentTesla</strong>. At this point, it’s best to dump the malware for code analysis, as further debugging won’t yield useful results. Static analysis will allow you to dissect its inner workings without relying on dynamic execution.</p><p>The malware collects and <strong>exfiltrates sensitive browser information</strong> from the victim’s system.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/774/1*5gxspaBPI2eb2KRBX9Bv0A.png" /></figure><p>AgentTesla actively scans for database files and extracts any stored <strong>usernames and passwords</strong> it discovers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/670/1*TMYCDo1WhHGxeJv2AQIDBA.png" /></figure><p>By understanding how these payloads are hidden and executed, we can develop more effective defenses and enhance our malware analysis skills. The key takeaway here is that no matter how well malware conceals itself, a methodical approach to reversing and static analysis can reveal its secrets. <strong>Enjoy Reversing ! ! !</strong></p><p><strong>Regards,<br>Shreyash Tambe</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=37f5c0c54a8d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[CrowdStrike’s update led to chaos]]></title>
            <link>https://medium.com/@shreyash_tambe/crowdstrikes-update-led-to-chaos-be3255dcf85f?source=rss-83f6d1b096d5------2</link>
            <guid isPermaLink="false">https://medium.com/p/be3255dcf85f</guid>
            <category><![CDATA[microsoft]]></category>
            <category><![CDATA[learning-opportunities]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[cloudstrike]]></category>
            <category><![CDATA[incident-response]]></category>
            <dc:creator><![CDATA[Shreyash Tambe]]></dc:creator>
            <pubDate>Sun, 11 Aug 2024 22:31:43 GMT</pubDate>
            <atom:updated>2024-08-11T22:31:43.160Z</atom:updated>
            <content:encoded><![CDATA[<p>Key Takeaways from this incident</p><p>🚨CrowdStrike’s Falcon Update : A Critical Lesson in Cybersecurity🚨</p><p>Recently, a major update to CrowdStrike Falcon led to unexpected blue screen errors (BSOD) for numerous Microsoft🪟 users worldwide. A critical mismatch in input parameters within the Falcon sensor caused severe disruptions across vital sectors like healthcare🏥 and transportation🚚.</p><p>I’ve created a crisp report on this incident, highlighting the key takeaways.</p><p>While this incident was challenging, it also offers valuable lessons for all security professionals. Let’s learn, improve, and prevent such disruptions in the future.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LsW0t5RZX0sO4Ui5W_LtiQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3be6quXp-df7b-MnxSWNww.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vMGDVWrKJ_p1KLRYtg6FCw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kJRLr5SRdYxcqF0DhxprBw.png" /></figure><p>~ Happy Learning : )<br> Shreyash Tambe</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=be3255dcf85f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Detect Malicious Strings at Lightning Speed with Python]]></title>
            <link>https://medium.com/@shreyash_tambe/detect-malicious-strings-at-lightning-speed-with-python-4d00137eed17?source=rss-83f6d1b096d5------2</link>
            <guid isPermaLink="false">https://medium.com/p/4d00137eed17</guid>
            <category><![CDATA[malware]]></category>
            <category><![CDATA[cyber-security-awareness]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[windows]]></category>
            <category><![CDATA[malware-analysis]]></category>
            <dc:creator><![CDATA[Shreyash Tambe]]></dc:creator>
            <pubDate>Mon, 22 Jul 2024 10:52:47 GMT</pubDate>
            <atom:updated>2024-07-22T10:52:47.784Z</atom:updated>
            <content:encoded><![CDATA[<h4>Streamline Your Malware Detection Process!</h4><h4>Introduction:</h4><p>Malware analysis is a critical task for cybersecurity professionals, requiring meticulous attention to detail and a thorough understanding of potential threats. To aid malware analysts in their work, I’ve developed a Python script that extracts and separates suspicious strings and IP addresses from an application’s output file generated by “<strong>bintext</strong>”. This script will save time and enhance the accuracy of your analysis.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/992/1*Toait-1k7Mq4aWfkDWkJqA.png" /></figure><h4>How Script Works? .:</h4><blockquote>Steps to use this script are very straight forward .:</blockquote><ol><li>Simply drag the application to bintext and save it as (&quot;a.txt”)</li><li>Now execute the below python script (&quot;shreyash.py”) at same location to “a.txt”</li><li>It will automatically drop a new file (&quot;b.txt”) which we have to go through, making it easier for analysts to pinpoint potential malware indicators.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/389/1*ZTBxTCq9sUfLSAQytaiiPA.png" /></figure><h4>Key Features For Malware Detection .:</h4><p><strong>Comprehensive Substrings List</strong>:</p><ul><li>The script includes extensive list of substrings {more than 100 keywords} associated with malware including :</li></ul><pre>* File extensions (.exe, .dll, .bat, .ps1, etc.)<br>* Suspicious Activities (encrypt, bypass, hook, etc.)<br>* System related terms (svchost, regedit, rundll, etc.)<br>* Networking related terms (URL&#39;s, FTP, download, socket, etc.)<br>* AntiVM &amp; AntiDebug (vmcheck, vmware, sandbox, ida, ollydbg, etc.)<br>* Clipboard, Keylogging, Credential theft, stealers, etc.</pre><p><strong>IP Address Extraction</strong>:</p><ul><li>In addition to substrings, the script also extracts any IP addresses found within the input file. {Exception : Version number may also give FP}</li></ul><h3>Python Script:</h3><pre>#made with curiosity by Shreyash_Tambe_010<br>def extract_and_separate(input_file_path, output_file_path, substrings):<br>    try:<br>        with open(input_file_path, &quot;r&quot;) as input_file:<br>            lines = input_file.readlines()<br><br>        # Initialize a dictionary to store matching lines for each substring<br>        matching_lines_dict = {substring: [] for substring in substrings}<br><br>        # Iterate through each line in the input file<br>        for line in lines:<br>            # Check if any of the specified substrings are present in the line<br>            for substring in substrings:<br>                if substring.lower() in line.lower():<br>                    matching_lines_dict[substring].append(line.strip())  # Remove leading/trailing whitespace<br><br>        # Write the matching lines to the output file with separator lines<br>        with open(output_file_path, &quot;w&quot;) as output_file:<br>            for substring, matching_lines in matching_lines_dict.items():<br>                output_file.write(f&quot;---|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*| {substring} |*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|---\n&quot;)<br>                for matching_line in matching_lines:<br>                    output_file.write(matching_line + &quot;\n&quot;)<br><br>            # Append the IP address extraction part<br>            import re<br>            ip_addresses = []<br>            for line in lines:<br>                ip_pattern = r&#39;\b\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}\b&#39;<br>                matches = re.findall(ip_pattern, line)<br>                ip_addresses.extend(matches)<br><br>            output_file.write(f&quot;---|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*| IP Addresses |*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|*|---\n&quot;)<br>            for ip_address in set(ip_addresses):<br>                output_file.write(ip_address + &quot;\n&quot;)<br><br>        print(f&quot;Extracted lines containing the specified substrings and IP addresses.&quot;)<br>        print(f&quot;Results saved to {output_file_path}&quot;)<br>    except FileNotFoundError:<br>        print(&quot;Error: Input file not found.&quot;)<br><br># Specify the input file path, output file path, and substrings to search for<br>input_file_path = &quot;a.txt&quot;  <br>output_file_path = &quot;b.txt&quot;  <br>substrings = [<br>    &quot;Http&quot;, &quot;download&quot;, &quot;www.&quot;, &quot;.com&quot;, &quot;Browser&quot;, &quot;Edge&quot;, &quot;Chrome&quot;, &quot;Firefox&quot;,<br>    &quot;.exe&quot;, &quot;.dll&quot;, &quot;Crypt&quot;, &quot;Encry&quot;, &quot;Key&quot;, &quot;Passw&quot;, &quot;username&quot;, &quot;Login&quot;,<br>    &quot;Credential&quot;, &quot;load&quot;, &quot;Rundll&quot;, &quot;Sql&quot;, &quot;select&quot;, &quot;Run&quot;, &quot;.pdb&quot;, &quot;.cmd&quot;,<br>    &quot;encode&quot;, &quot;base64&quot;, &quot;Powershell&quot;, &quot;mine&quot;, &quot;game&quot;, &quot;hack&quot;, &quot;clipboard&quot;,<br>    &quot;GetAsyncKeyState&quot;, &quot;mouse&quot;, &quot;hook&quot;, &quot;bypass&quot;, &quot;monitor&quot;, &quot;Firewall&quot;,<br>    &quot;ransom&quot;, &quot;payload&quot;, &quot;HKEY&quot;, &quot;VMcheck&quot;, &quot;Virus&quot;, &quot;DOS&quot;, &quot;task&quot;, &quot;rat&quot;,<br>    &quot;ftp&quot;, &quot;smtp&quot;, &quot;socket&quot;, &quot;connect&quot;, &quot;send&quot;, &quot;recv&quot;, &quot;autorun&quot;, &quot;startup&quot;,<br>    &quot;services.msc&quot;, &quot;svchost&quot;, &quot;regedit&quot;, &quot;regsvr32&quot;, &quot;vmware&quot;, &quot;vbox&quot;, &quot;qemu&quot;,<br>    &quot;xen&quot;, &quot;sandbox&quot;, &quot;ollydbg&quot;, &quot;windbg&quot;, &quot;ida&quot;, &quot;trojan&quot;, &quot;worm&quot;, &quot;backdoor&quot;,<br>    &quot;rootkit&quot;, &quot;keylogger&quot;, &quot;stealer&quot;, &quot;exploit&quot;, &quot;shellcode&quot;, &quot;xor&quot;, &quot;base64&quot;,<br>    &quot;createremotethread&quot;, &quot;virtualallocex&quot;, &quot;writeprocmemory&quot;, &quot;loadlibrary&quot;,<br>    &quot;getprocaddress&quot;, &quot;.bat&quot;, &quot;.vbs&quot;, &quot;.js&quot;, &quot;.ps1&quot;, &quot;.hta&quot;, &quot;.wsf&quot;, &quot;.scr&quot;,<br>    &quot;.pif&quot;, &quot;bitcoin&quot;, &quot;monero&quot;, &quot;ethereum&quot;, &quot;wallet&quot;, &quot;miner&quot;, &quot;pool&quot;,<br>    &quot;stratum&quot;, &quot;overwrite&quot;, &quot;killprocess&quot;<br>]<br><br># Call the function to extract matching lines<br>extract_and_separate(input_file_path, output_file_path, substrings)</pre><h4>Benefits in Malware Analysis .:</h4><ol><li><strong>Time Efficiency : </strong><br>Rapidly Process Larger files, saving analysts valuable time. Automatically extracts relevant information, reducing manual search efforts.</li><li><strong>Pattern recognition :</strong> <br>Helps identifying common malware patterns, also useful for initial stage &amp; prioritization of suspicious files.</li><li><strong>Adaptability : </strong><br>The Substring list can be easily updated as per your preferences to include new malware indicators.</li></ol><p><strong>Regards, <br>Shreyash Tambe</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4d00137eed17" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>