<?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" xmlns:media="http://search.yahoo.com/mrss/" version="2.0">
    <channel>
        
        <title>
            <![CDATA[ Malware - freeCodeCamp.org ]]>
        </title>
        <description>
            <![CDATA[ Browse thousands of programming tutorials written by experts. Learn Web Development, Data Science, DevOps, Security, and get developer career advice. ]]>
        </description>
        <link>https://www.freecodecamp.org/news/</link>
        <image>
            <url>https://cdn.freecodecamp.org/universal/favicons/favicon.png</url>
            <title>
                <![CDATA[ Malware - freeCodeCamp.org ]]>
            </title>
            <link>https://www.freecodecamp.org/news/</link>
        </image>
        <generator>Eleventy</generator>
        <lastBuildDate>Sun, 22 Nov 2026 15:34:08 +0000</lastBuildDate>
        <atom:link href="https://www.freecodecamp.org/news/tag/malware/rss.xml" rel="self" type="application/rss+xml" />
        <ttl>60</ttl>
        
            <item>
                <title>
                    <![CDATA[ How AI Is Changing Malware Detection: From Traditional Antivirus to Next-Gen Protection ]]>
                </title>
                <description>
                    <![CDATA[ Malware used to be simple to describe. A virus attached itself to a file, and antivirus software removed it. That world is gone. Today, a single attack can steal your passwords, lock up your photos, w ]]>
                </description>
                <link>https://www.freecodecamp.org/news/how-ai-is-changing-malware-detection/</link>
                <guid isPermaLink="false">6aa41cc6739dc5dd502bad21</guid>
                
                    <category>
                        <![CDATA[ Security ]]>
                    </category>
                
                    <category>
                        <![CDATA[ Malware ]]>
                    </category>
                
                    <category>
                        <![CDATA[ AI ]]>
                    </category>
                
                    <category>
                        <![CDATA[ ai agents ]]>
                    </category>
                
                <dc:creator>
                    <![CDATA[ Manish Shivanandhan ]]>
                </dc:creator>
                <pubDate>Fri, 11 Sep 2026 15:22:46 +0000</pubDate>
                <media:content url="https://cdn.hashnode.com/uploads/covers/5e1e335a7a1d3fcc59028c64/891dc84b-49d3-4823-bf82-fc5acac84c43.png" medium="image" />
                <content:encoded>
                    <![CDATA[ <p>Malware used to be simple to describe. A virus attached itself to a file, and antivirus software removed it.</p>
<p>That world is gone. Today, a single attack can steal your passwords, lock up your photos, watch what you type, and hide inside software you trust.</p>
<p>The bigger problem is volume. The <a href="https://www.av-test.org/en/statistics/malware/">AV-TEST Institute records over 450,000 new malicious programs</a> every single day. No security team can review that many files by hand. So the job has moved to machines, and antivirus software's decision-making has changed with it.</p>
<p>In this article, we'll look at how signature scanning worked and why it started to fail. We'll also cover what machine learning adds, how behaviour tracking catches ransomware while it runs, how cloud threat data turns every device into a sensor, and where AI still gets things wrong.</p>
<h3 id="heading-what-well-cover">What We'll Cover:</h3>
<ul>
<li><p><a href="#heading-signature-scanning-worked-until-malware-learned-to-change">Signature Scanning Worked Until Malware Learned to Change</a></p>
</li>
<li><p><a href="#heading-why-todays-malware-is-so-hard-to-spot">Why Today's Malware is So Hard to Spot</a></p>
</li>
<li><p><a href="#heading-what-machine-learning-adds-to-the-picture">What Machine Learning Adds to the Picture</a></p>
</li>
<li><p><a href="#heading-watching-what-software-does-not-just-what-it-looks-like">Watching What Software Does, Not Just What it Looks Like</a></p>
</li>
<li><p><a href="#heading-stopping-ransomware-while-its-still-running">Stopping Ransomware While it's Still Running</a></p>
</li>
<li><p><a href="#heading-the-cloud-turns-every-device-into-a-sensor">The Cloud Turns Every Device into a Sensor</a></p>
</li>
<li><p><a href="#heading-catching-the-attack-before-malware-ever-lands">Catching the Attack Before Malware Ever Lands</a></p>
</li>
<li><p><a href="#heading-where-ai-still-gets-it-wrong">Where AI Still Gets it Wrong</a></p>
</li>
<li><p><a href="#heading-layers-beat-any-single-trick">Layers Beat Any Single Trick</a></p>
</li>
<li><p><a href="#heading-what-this-means-for-you">What This Means For You</a></p>
</li>
</ul>
<h2 id="heading-signature-scanning-worked-until-malware-learned-to-change">Signature Scanning Worked Until Malware Learned to Change</h2>
<p>Early antivirus software used signatures. A signature is a small pattern taken from a known bad file, a bit like a fingerprint.</p>
<p>Researchers found a piece of malware, pulled out its pattern, and added it to a database. Your antivirus downloaded that database and compared it against every file on your disk. A match meant the file was blocked.</p>
<p>This was fast, cheap, and easy to trust. It also had one big weakness.</p>
<p>Attackers figured out that they only had to change the file a little. A new name, some padding, a different way of packing the code, and the fingerprint no longer matched. The old signature was suddenly useless.</p>
<p>That created a gap. A new threat would spread for hours or days before anyone wrote a signature for it. Attackers now build thousands of tiny variations of the same program on purpose, making it a losing race to write one signature per version.</p>
<p>Signatures are still worth keeping. They catch known threats in milliseconds. They just can't be the only thing standing between you and an attack.</p>
<h2 id="heading-why-todays-malware-is-so-hard-to-spot">Why Today's Malware is So Hard to Spot</h2>
<p>Modern malware tries hard to look boring.</p>
<p>It may sit quiet for days before doing anything. It may arrive as a harmless-looking script and download the real payload later. Some of it never writes a file to disk at all, which is why Microsoft groups these as <a href="https://learn.microsoft.com/en-us/defender-endpoint/malware/fileless-threats">fileless threats</a>.</p>
<p>Worse, plenty of attacks use tools that are already on your computer. An attacker who gets access can run commands through <a href="https://attack.mitre.org/techniques/T1059/001/">PowerShell</a>, a normal Windows administration tool. The <a href="https://lolbas-project.github.io/">LOLBAS project</a> catalogues hundreds of trusted Windows programs that can be abused this way.</p>
<p>Nothing malicious is being installed here. A trusted tool is simply being used for the wrong reason. A file scanner has almost nothing to grab onto.</p>
<p>So the question security software asks has changed. It's no longer just "have I seen this file before?" It's "what is this program actually doing?"</p>
<h2 id="heading-what-machine-learning-adds-to-the-picture">What Machine Learning Adds to the Picture</h2>
<p>Machine learning doesn't give software a sixth sense. It gives it a way to make a judgment call from evidence.</p>
<p>A model is trained on huge sets of files, both safe and harmful. Over time, it learns which traits tend to show up in each group. It learns about file structure, how the code is packed, which system calls it makes, what it talks to over the network, and how it interacts with other programs.</p>
<p>When something new arrives, the model weighs those traits and estimates the risk. It never saw this exact file, but it has seen the shape of the problem before.</p>
<p>This matters most for variants. Attackers often rewrite the surface of their code and keep the guts the same. Signatures miss the family resemblance. A model trained on behaviour and structure often catches it.</p>
<h2 id="heading-watching-what-software-does-not-just-what-it-looks-like">Watching What Software Does, Not Just What it Looks Like</h2>
<p>The biggest shift in antivirus protection is the move from scanning files to watching actions.</p>
<p>Picture an unknown program starting up on a laptop. Within seconds, it opens hundreds of documents, rewrites each one, changes their file extensions, deletes the recovery copies, and calls out to a server nobody recognises.</p>
<p>There may be no signature for it anywhere. But that pattern is ransomware, and it's unmistakable.</p>
<p>Microsoft describes this approach in its documentation on <a href="https://learn.microsoft.com/en-us/defender-endpoint/behavioral-blocking-containment">behavioral blocking and containment</a>, where machine learning models score a chain of actions rather than a single file. Individual steps can look innocent. Read together, they tell a story.</p>
<p>The advantage is timing. The software doesn't need to have met this exact malware before. It only needs to notice the shape of the attack early enough to cut it off.</p>
<h2 id="heading-stopping-ransomware-while-its-still-running">Stopping Ransomware While it's Still Running</h2>
<p>Ransomware shows why this matters more than any other threat type.</p>
<p>Known ransomware families get caught by signatures without trouble. But a brand new variant will often slip straight past them. That's the whole point of building a new variant.</p>
<p>Behavior monitoring gives you a second chance. The software watches for rapid file changes across many folders, attempts to delete backups or shadow copies, and processes trying to shut off security tools. Federal guidance on the <a href="https://www.cisa.gov/stopransomware">CISA StopRansomware hub</a> leans on the same signals, alongside offline backups you can actually restore from.</p>
<p>Once enough warning signs stack up, the software can kill the process or pull the device off the network. Even a partial save matters here. Stopping an attack after fifty encrypted files is a very different day than stopping it after fifty thousand.</p>
<p>Products aimed at everyday users are moving the same way. The <a href="https://nordvpn.com/next-gen-antivirus/">next-gen antivirus from NordVPN</a> blocks malicious downloads and scam pages before they reach the device, which shows how consumer tools have widened past plain file matching.</p>
<h2 id="heading-the-cloud-turns-every-device-into-a-sensor">The Cloud Turns Every Device into a Sensor</h2>
<p>Everything so far has been about <em>what</em> security software examines: files first, then behaviour. The other big change is <em>where</em> that examination happens.</p>
<p>Signature scanning ran start to finish on your machine. Your antivirus pulled down a database, compared files against it locally, and that was the entire decision.</p>
<p>Modern protection splits the work instead. Cheap checks stay on the device so they're instant, and the harder calls get handed to the vendor's cloud, which can see far more than any one laptop ever will.</p>
<p>Here's the actual sequence: First, the agent installed on your device records security-relevant events: process launches, parent-child process relationships, registry edits, outbound connections, file hashes. When it meets something it can't classify on its own, it sends metadata about it (typically the hash, the file's structural traits, and the surrounding activity, rather than the whole document) to the vendor's backend over an encrypted channel.</p>
<p>Microsoft documents this handoff for Defender in its notes on <a href="https://learn.microsoft.com/en-us/defender-endpoint/cloud-protection-microsoft-defender-antivirus">cloud protection</a>, where the local client queries the cloud service for a verdict and briefly holds the file while it waits for an answer.</p>
<p>The backend does two jobs at once. Automated systems compare the submission against what millions of other devices have reported and score it with models far too large to ship to a laptop.</p>
<p>If the picture is clear, a verdict comes back in under a second with no human involved. If it isn't, the case escalates to the vendor's threat research and security operations teams, the analysts employed specifically to hunt for clusters like this.</p>
<p>That's what makes the "few thousand devices" signal worth something. If the same unfamiliar binary, or the same odd process chain, appears across thousands of unrelated organisations inside an hour, no single one of them would notice. The backend sees the cluster immediately and flags it for a human to open up. Analysts pull samples, detonate them in a sandbox, confirm what the code does, and write a detection for it.</p>
<p>Then the loop closes. Confirmed cases become labelled training data, which is exactly what the next generation of models needs. The output goes out in two speeds: new indicators like hashes, domains and behavioural rules reach every protected device within minutes, while retrained models follow on a slower cycle of days or weeks.</p>
<p>Either way, the person targeted next is protected by what your device reported, and nobody had to wait for the next big database download.</p>
<h2 id="heading-catching-the-attack-before-malware-ever-lands">Catching the Attack Before Malware Ever Lands</h2>
<p>Not every threat arrives as a program. Most start with a message.</p>
<p>Phishing is still the front door. The APWG counted <a href="https://apwg.org/trendsreports/">more than one million phishing attacks in a single quarter</a>, and the fake pages are getting harder to eyeball. Attackers copy bank branding, login screens, and delivery notices closely enough to fool careful people.</p>
<p>AI helps by checking the things humans skip: how old the domain is, whether the link redirects somewhere odd, and whether the page matches a known scam kit.</p>
<p>Blocking a page like that stops the attack one step earlier. No download, no file to scan, and no cleanup.</p>
<h2 id="heading-where-ai-still-gets-it-wrong">Where AI Still Gets it Wrong</h2>
<p>AI brings real problems along with the benefits, and it's worth being clear about them.</p>
<p>False positives are the everyday one. Unusual isn't the same as malicious, and software that blocks a legitimate app because it looked odd trains people to switch protection off.</p>
<p>Speed is another. All this analysis has to happen without making the machine feel slow.</p>
<p>Then there's the arms race. Attackers study these models too. NIST's report on <a href="https://csrc.nist.gov/pubs/ai/100/2/e2025/final">adversarial machine learning</a> lays out how models get poisoned during training or fooled at the moment of decision. A model is a target, not a fortress.</p>
<p>Privacy deserves a mention as well. Cloud analysis means some information about your files and connections leaves your device. It's fair to ask any vendor what they collect and how long they keep it.</p>
<h2 id="heading-layers-beat-any-single-trick">Layers Beat Any Single Trick</h2>
<p>None of this replaces what came before. The strongest setups stack methods on purpose.</p>
<p>Signatures handle known malware instantly. Reputation checks block bad sites and untrusted programs. Behaviour monitoring catches suspicious activity as it happens. Machine learning fills the gap for threats nobody has named yet.</p>
<p>The <a href="https://attack.mitre.org/">MITRE ATT&amp;CK framework</a> is useful here because it maps out attacker techniques so teams can see which layer covers which step.</p>
<p>Layers also give the software context. A file with a clean history that starts acting strangely deserves a closer look. A file already known to be malware doesn't need any analysis at all.</p>
<h2 id="heading-what-this-means-for-you">What This Means For You</h2>
<p>Antivirus software has moved a long way from matching files against a list.</p>
<p>Signatures still earn their place, but they only answer one question. AI and behaviour monitoring answer a better one: what is this software doing right now, and does it make sense?</p>
<p>The practical takeaway is short. Good protection today is less about recognizing bad files and more about noticing bad behavior quickly. If you're choosing security software, ask whether it watches activity or only scans files, and check whether it blocks dangerous sites and downloads before they arrive.</p>
<p>Attacks keep getting faster and more automated. The defenses have to work the same way, and that's the real reason AI ended up at the centre of malware detection.</p>
<p>Hope you enjoyed this article. You can <a href="https://linkedin.com/in/manishmshiva">connect with me on LinkedIn</a>.</p>
 ]]>
                </content:encoded>
            </item>
        
            <item>
                <title>
                    <![CDATA[ How to Protect Your GitHub Repos Against Malicious Clones ]]>
                </title>
                <description>
                    <![CDATA[ The world of open-source development comes with various cyber threats. GitHub is still facing a type of attack that is ongoing since last year where attackers mirrored a huge number of repositories. S ]]>
                </description>
                <link>https://www.freecodecamp.org/news/protect-github-repos-from-malicious-clones/</link>
                <guid isPermaLink="false">68781639d48174ae283f8a5b</guid>
                
                    <category>
                        <![CDATA[ repo confusion ]]>
                    </category>
                
                    <category>
                        <![CDATA[ repository confusion ]]>
                    </category>
                
                    <category>
                        <![CDATA[ fake repos ]]>
                    </category>
                
                    <category>
                        <![CDATA[ GitHub ]]>
                    </category>
                
                    <category>
                        <![CDATA[ clone ]]>
                    </category>
                
                    <category>
                        <![CDATA[ Security ]]>
                    </category>
                
                    <category>
                        <![CDATA[ #cybersecurity ]]>
                    </category>
                
                    <category>
                        <![CDATA[ Malware ]]>
                    </category>
                
                    <category>
                        <![CDATA[ Supply Chain Attack ]]>
                    </category>
                
                    <category>
                        <![CDATA[ malicious ]]>
                    </category>
                
                    <category>
                        <![CDATA[ checklist ]]>
                    </category>
                
                    <category>
                        <![CDATA[ phishing ]]>
                    </category>
                
                    <category>
                        <![CDATA[ infostealer ]]>
                    </category>
                
                    <category>
                        <![CDATA[ Prevention ]]>
                    </category>
                
                    <category>
                        <![CDATA[ best practices ]]>
                    </category>
                
                <dc:creator>
                    <![CDATA[ brooklyn ]]>
                </dc:creator>
                <pubDate>Wed, 16 Jul 2025 21:14:33 +0000</pubDate>
                <media:content url="https://cdn.hashnode.com/res/hashnode/image/upload/v1752700407765/5fe06816-3d3a-40e4-8a4e-5cfe96a22368.png" medium="image" />
                <content:encoded>
                    <![CDATA[ <p>The world of open-source development comes with various cyber threats. GitHub is still facing a type of attack that is ongoing since last year where attackers mirrored a huge number of repositories. So as it turns out…the clone wars are not over!</p>
<p>If you haven’t heard about what’s going on:</p>
<blockquote>
<p>GitHub is struggling to contain an ongoing attack that’s flooding the site with with millions of code repositories. These repositories contain obfuscated malware that steals passwords and cryptocurrency from developer devices. … The result is millions of forks with names identical to the original one.</p>
<p>–&nbsp;Dan Goodin, <a href="https://arstechnica.com/security/2024/02/github-besieged-by-millions-of-malicious-repositories-in-ongoing-attack/">Ars technica</a></p>
</blockquote>
<p>Because search engines and GitHub’s own search rankings favor recent activity, these cloned repositories often float to the top – then they lure unsuspecting developers into pulling code that may contain malware.</p>
<p>One of my <a href="http://github.com/hyperphantasia/miniature-fortnight">repositories</a> has been targeted by such an attack, prompting me to monitor it closely. This guide offers tips to spot malicious repository clones before they catch you off guard.</p>
<h2 id="heading-table-of-contents"><strong>Table of Contents</strong></h2>
<ol>
<li><p><a href="#heading-what-is-a-repository-confusion-attack">What is a Repository Confusion Attack?</a></p>
<ul>
<li><a href="#heading-supply-chain-attacks">Supply Chain Attacks</a></li>
</ul>
</li>
<li><p><a href="#heading-basic-mitigation-strategies">🛡️ Basic Mitigation Strategies</a></p>
<ul>
<li><p><a href="#heading-verify-the-contributors-profiles">Verify the contributors profiles</a></p>
</li>
<li><p><a href="#heading-search-for-clone-repositories">Search for clone repositories</a></p>
</li>
<li><p><a href="#heading-examine-the-commit-pattern">Examine the commit pattern</a></p>
</li>
<li><p><a href="#heading-examine-the-commit-history">Examine the commit history</a></p>
</li>
<li><p><a href="#heading-examine-the-commit-contents">Examine the commit contents</a></p>
</li>
<li><p><a href="#heading-compare-the-concerned-files">Compare the concerned files</a></p>
</li>
<li><p><a href="#heading-some-information-about-the-malware">Some information about the malware</a></p>
</li>
</ul>
</li>
<li><p><a href="#heading-action-time">Action Time</a></p>
</li>
<li><p><a href="#heading-conclusion">Conclusion</a></p>
</li>
<li><p><a href="#heading-more-resources">More Resources</a></p>
</li>
</ol>
<h2 id="heading-what-is-a-repository-confusion-attack"><strong>What is a Repository Confusion Attack?</strong></h2>
<p>A repository confusion attack involves:</p>
<ul>
<li><p>Cloning legitimate repositories.</p>
</li>
<li><p>Injecting malicious code into the clone.</p>
</li>
<li><p>Uploading the clone.</p>
</li>
<li><p>Spreading through various unaware actors.</p>
</li>
</ul>
<h3 id="heading-supply-chain-attacks"><strong>Supply Chain Attacks</strong></h3>
<p>If you search for repository confusion on the internet, you'll find out it's a type of <em>supply chain attack</em>.</p>
<p>A supply chain attack is an&nbsp;<em>indirect</em>&nbsp;threat where hackers try infiltrating a system by targeting a trusted third-party or software component, rather than attacking the primary target directly.</p>
<p>It's not the first time this has happened. Before GitHub was targeted, PyPI was attacked in 2023 with&nbsp;<a href="https://arstechnica.com/information-technology/2023/01/more-malicious-packages-posted-to-online-repository-this-time-its-pypi/">fake packages</a>&nbsp;posing as legitimate. These packages lured negligent pip users into downloading malicious payloads (containing in most cases&nbsp;<a href="https://en.wikipedia.org/wiki/Infostealer">infostealer malware</a>).</p>
<h2 id="heading-basic-mitigation-strategies"><strong>🛡️ Basic Mitigation Strategies</strong></h2>
<p><strong>Before</strong>&nbsp;using any repository, make sure you follow these steps and take these precautions.</p>
<h3 id="heading-verify-the-contributors-profiles"><strong>Verify the contributors profiles</strong></h3>
<p>That's a first check: if you see a rather empty GitHub profile&nbsp;– one without reputation that contains just one repository but with a lot of daily commits to it – well, that's a bit suspicious.</p>
<p>In the fake repository, the original author will be listed as a contributor, too. Check that profile. You should be able to find the legitimate repository and do some comparisons.</p>
<img src="https://cdn.hashnode.com/res/hashnode/image/upload/v1752335573817/c39aca11-2605-47a2-8a6b-aded16547783.png" alt="GitHub screenshot of a repository contributors" width="316" height="156" loading="lazy">

<p>In the above screenshot you can see solotech143, my evil doppelgänger <em>(he’s been taken down since)</em>.</p>
<h3 id="heading-search-for-clone-repositories"><strong>Search for clone repositories</strong></h3>
<p>You can do a GitHub search by repository name and sort the results by most recent first. Malicious repositories tend to appear at the top of the search results because they are updated more frequently. The original repository might be hidden deeper in the search results.</p>
<img src="https://cdn.hashnode.com/res/hashnode/image/upload/v1752335785307/943c6dd3-aa28-4d72-b63a-65736de06dcf.png" alt="GitHub clone search results." width="1172" height="314" loading="lazy">

<p>It’s like clone wars.</p>
<p><strong>This is where it’s dangerous:</strong> users generally click on the first few search results, and in that type of attack, you’re almost guaranteed to see the attacker’s fake repository at the top of the results. The attacker achieves that by giving the fake repository regular fresh commits (and sometimes even a few stars!).</p>
<p>In my case, the original repository is a submission for the HackaViz 2025 competition. Hackathons offer a good attack surface because, beyond the fact they draw niche communities, they are also time sensitive.</p>
<p>Now, let’s move forward a year and imagine Hackaviz 2026 is starting soon. The attacker has easily outranked the untouched original submission. Which repository is most likely to be visited when future competitors – unaware of the scam – will look for the previous submissions?</p>
<h3 id="heading-examine-the-commit-pattern"><strong>Examine the commit pattern</strong></h3>
<p>Here’s when things take a weird turn. Malicious clones are run by automated agents, so the commit history fits a pattern that is rather unusual for a human. Of course, you can automate for many legitimate reasons but… this will always follow a clear goal and there will always be a human-touch at some point. In this case, commits are not adding up.</p>
<p>Let's see how that looks in the screenshots below:</p>
<img src="https://cdn.hashnode.com/res/hashnode/image/upload/v1752335872381/1238dee9-3568-4d2b-88bb-f63258ffb045.png" alt="1238dee9-3568-4d2b-88bb-f63258ffb045" width="390" height="197" loading="lazy">

<p>Regular like a clock...</p>
<img src="https://cdn.hashnode.com/res/hashnode/image/upload/v1752335891381/77f835fe-cccf-409f-85d7-789f918d4aa3.png" alt="A GitHub screenshot of a very active contribution activity.." width="764" height="602" loading="lazy">

<p>... and hyperactive!</p>
<h3 id="heading-examine-the-commit-history"><strong>Examine the commit history</strong></h3>
<p>You can’t! And that's the weird part. You're just able to see the last and the initial commit. So why is it hiding all of them? Do you like it when someone hide things from you?</p>
<img src="https://cdn.hashnode.com/res/hashnode/image/upload/v1752336385127/6274dd87-0a97-4c38-8849-9d547b9edb22.png" alt="A github commits history screenshot for one day." width="1304" height="225" loading="lazy">

<p>For July 10th, we should be able to see 11 commits, where are the ten others?</p>
<img src="https://cdn.hashnode.com/res/hashnode/image/upload/v1752336355084/4c7c343d-2000-4359-ae98-dcc98fb08732.png" alt="A github commits history screenshot for a whole period." width="1307" height="369" loading="lazy">

<p>Well, you can only check the first and last commit. That is not a lot for a repository that has more than 2000 commits registered.</p>
<h3 id="heading-examine-the-commit-contents"><strong>Examine the commit contents</strong></h3>
<p>Well, since I can always check the last commit, I checked some of them. They share the same pattern: the bot is constantly looping over the README file doing the same modifications. As you can see in the screenshot below, it’s updating the file with links to an infected release.</p>
<img src="https://cdn.hashnode.com/res/hashnode/image/upload/v1752336493881/e8b57b4c-4d13-44f8-bca6-bd8fbad2738c.png" alt="A github screenshot of commits to a malicious repository." width="1382" height="800" loading="lazy">

<p>Above you can see an AI agent stuck in the Readme loop of change.</p>
<p>Human edits are more varied. In a human-driven project, you will see a large mix of commits: feature commits, exploratory experiments, bug fixes, styling tweaks, and sometimes reverts. A bot clone will often just overwrite files, bump versions, or re-inject the same malicious payload repeatedly with no real contribution to the codebase.</p>
<h3 id="heading-compare-the-concerned-files"><strong>Compare the concerned files</strong></h3>
<p>This is where common sense comes handy. So, you have two README's:</p>
<ol>
<li><p>The&nbsp;<a href="https://web.archive.org/web/20250711182419/https://github.com/solotech143/miniature-fortnight/blob/main/README.md">first</a>&nbsp;consists of AI-generated content that is cluttered with emojis and low-value information. It is designed solely to entice you into clicking the download link of the release.</p>
</li>
<li><p>The <a href="https://github.com/hyperphantasia/miniature-fortnight/blob/main/README.md">other</a> follows <a href="https://www.freecodecamp.org/news/how-to-write-a-good-readme-file/">best practices</a> for creating a good README file. It is accurate and well-structured and functions as a valuable helper and explainer to the code. It also goes deep into the most important aspects of the project. This is usually a good sign that a repository is organic and genuine.</p>
</li>
</ol>
<h3 id="heading-some-information-about-the-malware"><strong>Some information about the malware</strong></h3>
<p>What do we have so far? Well, a suspicious link in a phishy, AI-generated README file that is consistent with a very suspicious pattern in the commit history.</p>
<p>Now, let’s have a closer look at that dubious release and let’s see what an online antivirus scanner might reveal about it.</p>
<img src="https://cdn.hashnode.com/res/hashnode/image/upload/v1752336589656/124aecf2-39e9-4158-9a06-f0fd59cbf8c1.png" alt="A  github screenshot of commits to a malicious repository." width="1216" height="404" loading="lazy">

<p>The malware is packed only in the miniature-fortnight-v1.7.6.zip release.</p>
<img src="https://cdn.hashnode.com/res/hashnode/image/upload/v1752336609780/7eaf7fc8-73f2-4d9d-b169-1d5c50ce84f2.png" alt="A malware analysis result." width="1317" height="757" loading="lazy">

<p>Above you can see the result of a scan with an online scanner.</p>
<p>The .zip file contains <strong>only</strong> four files:</p>
<ul>
<li><p>config.txt</p>
</li>
<li><p>launch.bat</p>
</li>
<li><p>lua51.dll</p>
</li>
<li><p>luajit.exe</p>
</li>
</ul>
<p>These files are <strong>totally unrelated</strong> to the source project (a Python data science project with Jupyter notebooks combined to a React app using three.js).</p>
<p>I will not go into the detail in this article. But for the curious ones, it's an infostealer malware (a malware that will exfiltrate your credentials and other precious information about your configuration) similar to the one described in&nbsp;<a href="https://www.trendmicro.com/en_us/research/25/c/ai-assisted-fake-github-repositories.html#">detail here</a>.</p>
<h2 id="heading-action-time"><strong>Action Time</strong></h2>
<p>If you discover a potentially malicious repository, here are some steps you can take:</p>
<ol>
<li><p>Document some evidence.</p>
</li>
<li><p>Notify the original repository maintainers.</p>
</li>
<li><p>Report the malicious clone to GitHub.</p>
</li>
</ol>
<p>Reporting a repository or a profile on GitHub is easy and fast. Go to the user’s profile page, click “Block or report” in the left sidebar and choose “Report abuse” in the pop-up. You will have to complete a short contact form with some details about the behavior before submitting. If needed, you can find more information on <a href="https://docs.github.com/en/communities/maintaining-your-safety-on-github/reporting-abuse-or-spam#reporting-a-user">GitHub</a>.</p>
<h2 id="heading-conclusion"><strong>Conclusion</strong></h2>
<p>This is a description of just one attack, from the perspective of someone who found out that one of his repository had been targeted. There are likely cases of more sophisticated attacks. But the clone repository flood we can see on GitHuB is definitely massive low quality automation. Quantity over quality.<br>To be honest, I'm quite surprised algorithms crafted at GitHub didn't manage to spot this one.</p>
<p>This also raises questions related to AI.</p>
<ul>
<li><p>What happens when LLMs are trained on malicious content? That’s a more general question about&nbsp;<a href="https://arstechnica.com/information-technology/2024/01/ai-poisoning-could-turn-open-models-into-destructive-sleeper-agents-says-anthropic/">AI poisoning</a>.</p>
</li>
<li><p>A human might easily spot the patterns and the low quality content&nbsp;<em>for now</em>. But..</p>
<ul>
<li><p>Imagine you are using coding agents, many of them. Will the agents pick-up the malicious clone instead of the original one? How to distinguish the repositories from an automaton's perspective?</p>
</li>
<li><p>The attackers <strong>will</strong> refine their tactics, making the clones more human-like and therefore luring us more easily into their traps.</p>
</li>
</ul>
</li>
<li><p>This is really a situation that makes me wonder about the early days of Google. Back then, the company had to fight huge amounts of spam due to keyword stuffing and manipulative SEO tactics. Will big tech companies have to go through a&nbsp;<a href="https://en.wikipedia.org/wiki/Timeline_of_Google_Search#Full_timeline">Florida update</a>&nbsp;moment to face the rise of AI generated spam ?</p>
</li>
</ul>
<h2 id="heading-more-resources"><strong>More Resources</strong></h2>
<ul>
<li><p><a href="https://www.trendmicro.com/en_be/research/23/j/infection-techniques-across-supply-chains-and-codebases.html">A detailed description of the attack</a></p>
</li>
<li><p><a href="https://www.cisa.gov/sites/default/files/publications/ESF_SECURING_THE_SOFTWARE_SUPPLY_CHAIN_DEVELOPERS.PDF">Complete safety recommendations</a></p>
</li>
</ul>
<p><strong>Stay Informed, Stay Secure!</strong></p>
<p>A <strong>cheat-sheet</strong> is also available on my&nbsp;<a href="https://github.com/hyperphantasia/repo-confusion-guard">GitHub</a>. Feel free to contribute to it!</p>
 ]]>
                </content:encoded>
            </item>
        
    </channel>
</rss>
