<?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 Munna🎭 on Medium]]></title>
        <description><![CDATA[Stories by Munna🎭 on Medium]]></description>
        <link>https://medium.com/@spy0x1?source=rss-4c31ad2136a3------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*MEQ8QOOSQXtc2Yx4pN9WBw.jpeg</url>
            <title>Stories by Munna🎭 on Medium</title>
            <link>https://medium.com/@spy0x1?source=rss-4c31ad2136a3------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 23:26:38 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@spy0x1/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[The Punycode Paradox: When Unicode Normalization Turns Into an ATO]]></title>
            <link>https://meetcyber.net/the-punycode-paradox-when-unicode-normalization-turns-into-an-ato-d6938d7d7e62?source=rss-4c31ad2136a3------2</link>
            <guid isPermaLink="false">https://medium.com/p/d6938d7d7e62</guid>
            <category><![CDATA[infosec]]></category>
            <category><![CDATA[bug-bounty]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[hacking]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Munna]]></dc:creator>
            <pubDate>Sun, 03 May 2026 14:04:54 GMT</pubDate>
            <atom:updated>2026-07-23T13:58:58.738Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>The Illusion: <strong>“Please Provide a Valid Email ID”</strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FbQL9hWE2PgM8-OnOFKSEw.png" /></figure><p>In the world of Bug Bounties, the <strong>“Punycode Account Takeover” </strong>is a classic. Usually, the story is predictable: you try to register with a look-alike email, the server says <em>“Email already exists,”</em> and you realize the backend is merging your fake email with a real one. <em>It’s a standard logic flaw. Old news.</em></p><p>But what happens when the front door isn’t just locked, but <strong>welded shut?</strong></p><p>What happens when the application doesn’t just reject your duplicate email, but mocks you, claiming your email doesn’t even exist in the realm of reality? That’s where the mystery begins.</p><h3>The Illusion of Security</h3><p>Imagine you’re a developer. You’ve built a robust sign-up flow. You’ve got validation, you’ve got verification codes, and you’ve even blocked “weird” characters in your UI to prevent junk data. You feel safe.</p><p>But deep in the backend, there is a ghost in the machine. It’s called <strong>Unicode Normalization</strong>. It’s a process where the server tries to be helpful by treating édny.net as edny.net.</p><p>It sounds like a great UX feature — until I realized that I could use that “helpfulness” to trick the server into handing me the keys to someone else’s kingdom.</p><h3>Bypassing the Gates</h3><p>The target application had a gatekeeper. If I tried to register with an internationalized domain (IDN) using a Punycode email like,spy0x1@édny.netthe UI would instantly snap back:</p><blockquote>“Please provide a valid email ID.”</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/229/1*SvWYk9Ua4Yj6umjm-1B-wA.png" /></figure><p>The UI was smart. But the API? It was naive.</p><ol><li>I fired up <strong>Burp Suite</strong>.</li><li>I entered a “normal” email address and hit “Sign Up.”</li><li>I caught the request mid-air and swapped my boring email for the “special” one: spy0x1@édny.net.</li></ol><p><strong>The result?</strong> A 200 OK purred back from the server. The gate was open. And A six-digit verification code that didn’t belong to me, but landed in my box.</p><h3>The Twist: The Identity Crisis</h3><p>Here is where the “magic” happens. The backend was suffering from a personality disorder.</p><ul><li><strong>When checking if the account exists:</strong> The server “normalized” the email. It saw the é and thought, <em>&quot;Oh, you mean &#39;e&#39;!&quot;</em> It matched my malicious email to the victim’s legitimate account.</li><li><strong>When routing the actual SMTP email:</strong> The server forgot to normalize. It sent the verification code to the literal address I provided — the one with the accent mark — which pointed straight to <strong>my</strong> attacker-controlled server.</li></ul><p>The UI told the victim: <em>“We sent a code to your email.”</em> <strong>The UI was lying.</strong> The code was sitting in my inbox.</p><h3>The Climax: The Zero-Interaction Takeover</h3><p>Now, for the final act. I navigated to the <strong>Forgot Password</strong> page. Again, the UI tried to block me. Again, Burp Suite was my skeleton key.</p><p>I requested a password reset for the victim’s email, but injected my Punycode domain into the request.</p><blockquote><strong><em>The Server’s Logic Flow:</em></strong></blockquote><ul><li>“Does spy0x1@édny.net exist?&quot; →<em>Normalizes to</em> spy0x1@edny.net<strong> →“YES, that&#39;s a real user!”</strong></li><li>“Okay, send the reset code to the requested address: spy0x1@édny.net.&quot;</li><li><strong>The Trap:</strong> The email leaves the server and routes to my My Email Inbox.</li></ul><p>I watched the logs. <strong>Ping.</strong> There it was. A six-digit code that didn’t belong to me, but gave me everything.</p><p>I pasted the code, set a new password, and logged in. I was no longer an attacker; I was the victim. No clicks required from them, no phishing link needed. Just one character and a bit of logic-bending.</p><blockquote><strong>Disclaimer: </strong>“Since I didn’t have a registered Punycode domain at my disposal, I utilized <strong>Interactsh</strong> (and Burp Collaborator) to simulate the attacker-controlled mail server. This allowed me to demonstrate, with surgical precision, how the backend routes sensitive verification codes to an external, accented domain.”</blockquote><blockquote><strong>Victim Email:</strong> spy0x1@edny.net.kpyqkojraafnvkgsqnftkg1oxwjpoy7bu.oast.fun<br><strong>Attacker Unicode Email: </strong>spy0x1@édny.net.kpyqkojraafnvkgsqnftkg1oxwjpoy7bu.oast.fun<br><strong>Attacker Punycode Email: <br>spy0x1@xn — dny-9la.net.jnsdxidbvzkkncqfvzhvvb9lrnh4cm879.oast.fun</strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/565/1*sVXDZUY6apWMYp5nV27TLQ.png" /></figure><blockquote><em>So, What is Punycode?<br>Punycode is an encoding that converts Unicode characters into the limited ASCII subset that DNS can handle, making internationalized domain names (IDN) possible. Every Punycode-encoded domain label starts with the </em><strong><em>xn--</em></strong><em> prefix.</em></blockquote><p><a href="https://www.punycoder.com/?source=post_page-----c0afdb74a3dc---------------------------------------">Punycode Converter - Unicode to IDN Encoder &amp; Decoder 🧰</a></p><h3>The Root Cause: A Conflict of Interest</h3><p>This wasn’t a “coding error” in the traditional sense. It was a <strong>Normalization Conflict</strong>.</p><p>The application used one library to “clean” the email for the database lookup, and a completely different logic to “route” the email via SMTP. Because these two systems didn’t agree on what it é meant, the security of the entire account fell through the cracks.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1014/1*c4mFugqtiUxU7_ZFZsOZCA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/386/1*Wv5QQTrVEUbE2MYG2RYzMQ.png" /></figure><h3>The Moral of the Story</h3><p>For developers, the lesson is simple: <strong>Be consistent.</strong> If you normalize data, do it at the very edge of your application and use that normalized version for <em>everything</em>.</p><p>For bug hunters, the lesson is even simpler: <strong>Always look for the “helpful” features.</strong> Sometimes, a server trying to be smart is the easiest way to make it look very, very foolish.</p><p>💬 You can also find me on Twitter/X: @spy0x1</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d6938d7d7e62" width="1" height="1" alt=""><hr><p><a href="https://meetcyber.net/the-punycode-paradox-when-unicode-normalization-turns-into-an-ato-d6938d7d7e62">The Punycode Paradox: When Unicode Normalization Turns Into an ATO</a> was originally published in <a href="https://meetcyber.net/">MeetCyber</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The One-Click Library Wipe: Exploiting CSRF in Saved Stories]]></title>
            <link>https://infosecwriteups.com/the-one-click-library-wipe-exploiting-csrf-in-saved-stories-28523d4bc5e2?source=rss-4c31ad2136a3------2</link>
            <guid isPermaLink="false">https://medium.com/p/28523d4bc5e2</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[bug-bounty]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[hacking]]></category>
            <category><![CDATA[programming]]></category>
            <dc:creator><![CDATA[Munna]]></dc:creator>
            <pubDate>Sun, 04 Jan 2026 06:27:55 GMT</pubDate>
            <atom:updated>2026-10-26T03:37:48.429Z</atom:updated>
            <content:encoded><![CDATA[<blockquote><em>“Security is often just an illusion — and I’m </em><strong>spy0x1</strong><em>, the guy who likes to break the spell. Come along as I show you how I chained a ‘hidden’ secret with a lazy endpoint to rewrite any user’s library in total silence.”</em></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1012/0*v7UGtJEfRo1uFRZ-.png" /></figure><p>Most people look at a URL and see a way to get to a page. As a researcher, I look at a URL and see a set of instructions. Last week, while exploring, let’s call <strong>spy0x1.com</strong>I found a set of instructions that allowed me to take control of any user&#39;s &quot;Saved Stories&quot; silently, instantly, and with a single click.</p><p>It wasn’t just a simple bug; it was a lesson in why “security by obscurity” always fails.</p><h3>A URL That Does Too Much</h3><p>I was poking around the newsletter section when I decided to save an article for later. I fired up Burp Suite to watch the traffic, expecting a standard, boring background request. Instead, I saw this staring back at me:</p><pre>GET /bookmarks/apply?zx=24723rhewuiewhdf7436r7ywe&amp;nid=346236&amp;action=add</pre><p>My heart skipped a beat. <strong>A GET request? For a state-changing action?</strong></p><p>In the world of web security, that’s like leaving your front door unlocked. But then I saw the <strong>zx</strong> parameter. At first glance, it looked like a standard CSRF token, a long, random string of characters designed to prove the request was legitimate. If that token was unique to my session, the bug was dead in the water.</p><h3>The Token Hunt</h3><p>I needed to know where that <strong>zx</strong> token came from. If I couldn&#39;t guess it or find it, I couldn&#39;t exploit the vulnerability.</p><p>I started digging into other endpoints, looking for the source. That’s when I hit the jackpot. I found a completely separate endpoint — one designed to check user emails. When I queried an email address, the server didn’t just tell me if the user existed; it handed me the keys to the kingdom.</p><p><strong>The response leaked the </strong><strong>zx token for that specific user.</strong></p><p>This changed everything. I didn’t need to steal a cookie. I didn’t need to breach a server. I just needed to look in the one place the developers thought no one was watching. By hitting the email endpoint, I could grab a victim’s zx value and &quot;seal&quot; the malicious URL.</p><h3>The Invisible Fingerprint</h3><p>To prove how dangerous this was, I crafted a simple “trap” page. I didn’t want the victim to see a form or click a “<strong>Submit</strong>” button — that’s too obvious. I wanted it to happen like magic.</p><p>I built a tiny HTML file that automatically submits the request the millisecond the page loads. To the victim, it just looks like a page that failed to load or a quick redirect. To the server, it looks like a legitimate request from a logged-in user.</p><p>CSRF.html</p><pre>&lt;html&gt;<br>&lt;body&gt;<br>&lt;form action=&quot;https://spy0x1.com/bookmarks/apply&quot;&gt;<br>&lt;input type=&quot;hidden&quot; name=&quot;zx&quot; value=&quot;[STOLEN_ZX_TOKEN]&quot; /&gt;<br>&lt;input type=&quot;hidden&quot; name=&quot;nid&quot; value=&quot;346236&quot; /&gt;<br>&lt;input type=&quot;hidden&quot; name=&quot;action&quot; value=&quot;add&quot; /&gt;<br>&lt;/form&gt;<br>&lt;script&gt;<br>document.forms[0].submit();<br>&lt;/script&gt;<br>&lt;/body&gt;<br>&lt;/html&gt;</pre><h3>The “Oh No” Moment</h3><p>I hosted the file and sent the link to my test account. I was logged in on one tab, and I opened the malicious link in another.</p><p>I went back to my “Saved Stories” and hit refresh. <strong>There it was.</strong> An article I never clicked, added to my list by a script running on a totally different domain.</p><p>Then, I changed the action from add to <strong>delete</strong>.<br><strong>Refresh: My entire library was gone.</strong></p><h3>Why “Different Endpoints” Matter</h3><p>The biggest takeaway from this find? <strong>Never trust a token that isn’t tied to a specific, secret session.</strong> The developers tried to be clever by including the zx parameter, but by leaking that same parameter in a different, unrelated endpoint, they rendered the entire defense useless. It’s a reminder to all of us: security is only as strong as your leakiest endpoint.</p><p>I reported this to the team immediately. Within a short time, they moved to secure the endpoint and properly validate origins.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/367/1*EzenPPxKJN_aUQIaKKwETg.png" /></figure><h3>Enjoyed the hunt?</h3><p>I’m always digging for the next exploit that others miss. If you want to see more deep dives into broken logic and chained vulnerabilities, <strong>hit that Follow button</strong> and stay tuned for the next report.</p><p>💬 You can also find me on Twitter/X: @spy0x1</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=28523d4bc5e2" width="1" height="1" alt=""><hr><p><a href="https://infosecwriteups.com/the-one-click-library-wipe-exploiting-csrf-in-saved-stories-28523d4bc5e2">The One-Click Library Wipe: Exploiting CSRF in Saved Stories</a> was originally published in <a href="https://infosecwriteups.com/">InfoSec Write-ups</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I Tried Filling Out a Simple Contact Form and Discovered a Stored XSS in the Wild]]></title>
            <link>https://meetcyber.net/i-tried-filling-out-a-simple-contact-form-and-discovered-a-stored-xss-in-the-wild-f777dec6369e?source=rss-4c31ad2136a3------2</link>
            <guid isPermaLink="false">https://medium.com/p/f777dec6369e</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[bug-bounty]]></category>
            <category><![CDATA[security]]></category>
            <category><![CDATA[hacking]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <dc:creator><![CDATA[Munna]]></dc:creator>
            <pubDate>Tue, 23 Dec 2025 13:00:21 GMT</pubDate>
            <atom:updated>2026-07-23T14:02:47.442Z</atom:updated>
            <content:encoded><![CDATA[<h4><em>Have you ever filled out one of those “Contact Us” forms and thought, “Eh, it’s harmless”? Yeah, I thought so too.</em></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0C2-CcZZlOy8JG7Bi1OO6A.png" /></figure><p>It was just another night — me, a cup of coffee, and my usual hobby: poking around websites to see how they handle user input. Nothing fancy, nothing evil, just curiosity.</p><blockquote>Then I landed on a page that looked totally ordinary:<br> 👉<em> https://popular.spy0x1.com/xxxxxxx/xxxxx_contact-us</em></blockquote><p>A basic<strong> “Contact Us”</strong> form. Name, Email, Message, simple stuff.</p><p>But something inside me whispered, <em>“What if it’s not so simple?”</em></p><h3>The Moment Everything Changed</h3><p>I typed in my details like everyone else would.<br>But instead of a normal name like “John Doe,” I entered something a bit… different.</p><p>Let’s just say it included a mix of <strong>HTML</strong> and <strong>JavaScript</strong> magic. (Don’t worry — I’m not sharing anything harmful here, just explaining the idea.)</p><pre>&quot;&gt;&lt;h1&gt;&lt;a href=&quot;https://evil.com&quot;&gt;click here&lt;/a&gt;&lt;/h1&gt;&lt;img src=1 Onerror=alert(&quot;spy0x1-xss&quot;)&gt;</pre><p>I hit submit. The page refreshed.<br>At first, nothing happened. I thought, “Okay, maybe this one’s safe.”</p><p>Then I checked my email inbox — and boom.<br>My code had <strong>come alive</strong>.</p><p>The script I placed in the <strong>“Name”</strong> field actually <strong>executed</strong> when the site displayed my submission.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8dWQ9XxeZFPVwF7ukCsEoQ.png" /></figure><p>For a second, I just stared at the screen thinking, <em>“Wait, did I just inject code into one of the biggest financial news sites?”</em></p><h3>What I Actually Found</h3><p>What happened was something called <strong>Stored XSS</strong> (Cross-Site Scripting) and <strong>HTML Injection</strong>.</p><p>In simple words: the site wasn’t just saving what I typed — it was <strong>trusting</strong> it.<br>When that “trusted” data was displayed later, the browser treated it as real code.</p><p>And here’s why that’s dangerous:<br> If an attacker does this, they can:</p><ul><li>Steal someone’s cookies or session.</li><li>Redirect users to fake pages.</li><li>Inject links, messages, or phishing popups.</li><li>Even control what people see on the site.</li></ul><p>Scary, right?<br> One innocent text box turned into a hacker’s playground.</p><h3>How This Could’ve Been Prevented</h3><p>Fixing it isn’t rocket science.<br> Developers just need to:</p><ul><li><strong>Sanitize user input</strong> (don’t let HTML or scripts through).</li><li><strong>Escape outputs</strong> (treat data as text, not code).</li><li><strong>Use strong content security policies</strong> (so scripts can’t freely run).</li></ul><p>Basically, don’t let the browser think the user is part of the website’s code.</p><h3>Final Thoughts</h3><p>The web is full of surprises.<br>Sometimes, the most ordinary-looking places hide the biggest lessons.</p><p>So next time you fill out a contact form, remember:<br>It’s not <em>just</em> a form — it’s a door.<br>And how securely it’s built decides who can walk through it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/295/1*-cdKj6pgaFzIkKbytOQ7lA.png" /></figure><p><em>If you enjoyed this write-up, drop a 💬 or share it with someone who still thinks “it’s just an email preference page.”</em></p><p>💬 Let’s Connect:</p><ul><li>Twitter/X: @spy0x1</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f777dec6369e" width="1" height="1" alt=""><hr><p><a href="https://meetcyber.net/i-tried-filling-out-a-simple-contact-form-and-discovered-a-stored-xss-in-the-wild-f777dec6369e">I Tried Filling Out a Simple Contact Form and Discovered a Stored XSS in the Wild</a> was originally published in <a href="https://meetcyber.net/">MeetCyber</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[“Who Sent That Email?” — How one tiny Hidden API turned into a spam machine]]></title>
            <link>https://infosecwriteups.com/who-sent-that-email-how-one-tiny-hidden-api-turned-into-a-spam-machine-a56b69a72a4c?source=rss-4c31ad2136a3------2</link>
            <guid isPermaLink="false">https://medium.com/p/a56b69a72a4c</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[information-security]]></category>
            <category><![CDATA[hacking]]></category>
            <category><![CDATA[bug-bounty]]></category>
            <dc:creator><![CDATA[Munna]]></dc:creator>
            <pubDate>Mon, 15 Dec 2025 08:23:14 GMT</pubDate>
            <atom:updated>2026-07-23T14:09:02.172Z</atom:updated>
            <content:encoded><![CDATA[<h3>“Who Sent That Email?” — How one tiny Hidden API turned into a spam machine</h3><blockquote>Ever wondered how it feels when a perfectly normal feature quietly turns into a weapon? This one started with an innocent toggle button… and ended with a mail server crying for mercy.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_7QFvOn63q2D7YNugUV67w.png" /></figure><p>Have you ever gotten an email from a service you use and thought, “I don’t remember asking for this”? I did — and the reason wasn’t a buggy UI. It was the backend trusting the client too much.</p><p>This is the story of a small, boring POST request that should’ve been boring — but wasn’t. I’ll show you how I found it, why it’s dangerous, and how the engineering team can fix it in ways that don’t wreck user experience.</p><h3>The annoying click that started it all</h3><p>I was toggling email settings in a company’s web app. The UI updated like any other app: click, save, toast message, happy face. But in Burp’s HTTP history, there was an extra request I didn’t expect:</p><pre>POST /api/v1/subscription/sendSomethingNotification</pre><p>The payload included email=<strong>… </strong>and a handful of meta fields. That’s when the little alarm bell in my head rang: why is the client telling the server, “send to this email”? Who verified that address? Where was the throttle?.</p><p>A few replayed requests later, and a quick check of the target inbox confirmed it: the server would send a preference-change notification to <em>any</em> email address provided, and it would do so repeatedly.</p><h3>The simplest Reproduction</h3><ol><li>Log in to the app with a legitimate org account.</li><li>Change a preference in the “Manage Preferences” UI. Notice two outgoing requests in Burp: the preference update and a notification send.</li><li>Send the notification request to Repeater/Intruder and replace email= with any target address.</li><li>Fire repeated requests (I used 100 repeated payloads to demonstrate volume). The server kept accepting them — no 429, no ownership checks.</li><li>Inspect the target inbox: notifications arrive.</li></ol><h3>A tidy PoC snippet</h3><pre>POST /api/v1/subscription/sendSomethingNotification<br>Content-Type: application/x-www-form-urlencoded<br>Cookie: session=...<br><br><br>email=victim@example.com&amp;product_id=xxxxxxxxxxx</pre><p>Replay that request with changed email= values and watch the server accept them repeatedly. That’s it.</p><p>No confirmation token. No proof-of-ownership step. No per-recipient rate limit. Just a replaceable email field and a mailer.</p><h3>What Makes This Dangerous</h3><p>An authenticated user can trigger <strong>unlimited notification emails</strong> to arbitrary recipients because the endpoint lacks <strong>recipient validation</strong> and <strong>rate limiting</strong> — enabling spam, phishing, and abuse.</p><h3>💭 Final Thoughts</h3><p>Sometimes, the most interesting bugs aren’t loud — they whisper.<br> A single missing rate limit, a single unverified email field, and suddenly you’ve built a self-serve spam cannon.</p><p>This one taught me (again) that <strong>“business logic” bugs can be just as impactful as technical ones</strong> — especially when trust is involved.</p><p>Next time you see an endpoint namedsendSomethingNotification, maybe pause and ask:</p><blockquote><em>“What if it sends to someone it shouldn’t?”</em></blockquote><p>Because sometimes… it actually does. 😉</p><p><a href="https://infosecwriteups.com/how-i-bypassed-random-medical-center-assignment-during-appointment-booking-16daaf7b7a6d">How I Bypassed Random Medical Center Assignment During Appointment Booking</a></p><p><em>If you enjoyed this write-up, drop a 💬 or share it with someone who still thinks “it’s just an email preference page.”</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/580/1*YQvY2is-DgZg95Zf7rEQug.png" /></figure><p>💬 Let’s Connect:<br>🐦 Twitter/X: @<em>spy0x1</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a56b69a72a4c" width="1" height="1" alt=""><hr><p><a href="https://infosecwriteups.com/who-sent-that-email-how-one-tiny-hidden-api-turned-into-a-spam-machine-a56b69a72a4c">“Who Sent That Email?” — How one tiny Hidden API turned into a spam machine</a> was originally published in <a href="https://infosecwriteups.com/">InfoSec Write-ups</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Beyond Authentication — Exploiting a Nasty IDOR in Profile Update Functionality]]></title>
            <link>https://infosecwriteups.com/beyond-authentication-exploiting-a-nasty-idor-in-profile-update-functionality-25740fb56b05?source=rss-4c31ad2136a3------2</link>
            <guid isPermaLink="false">https://medium.com/p/25740fb56b05</guid>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[infosec]]></category>
            <category><![CDATA[bug-bounty]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Munna]]></dc:creator>
            <pubDate>Sat, 13 Dec 2025 07:57:05 GMT</pubDate>
            <atom:updated>2026-07-23T14:04:48.035Z</atom:updated>
            <content:encoded><![CDATA[<h3>Beyond Authentication — <strong>Exploiting a Nasty IDOR in Profile Update Functionality</strong></h3><p><em>It wasn’t a complex SQL injection or a tricky deserialization flaw. It was a failure to ask one simple question: </em><strong><em>“Are you really authorized to modify that object?”</em></strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*P3JvjzqBCO9BB9jz8k7KAw.png" /></figure><p>Every hacker knows the feeling: that moment when a simple, almost trivial input causes a cascade of unintended, catastrophic events on a production system. I recently found myself in that sweet spot while auditing a major corporate portal, leading to one of the cleanest <strong>Insecure Direct Object Reference (IDOR)</strong> vulnerabilities I’ve ever seen.</p><p>This vulnerability didn’t just expose user data — it allowed <em>any</em> basic user to take over the identity, contact details, and designation of internal employees, including management and HR.</p><h3>The Subtle Leak That Handed Me the Keys</h3><p><strong><em>But the </em></strong><strong><em>xid not guessable so easily.</em></strong></p><p>So, my first target was the application’s user validation logic. Most sites have an API that checks if an email exists for a “Forgot Password” function. This portal’s endpoint was promisingly named: /api/v1/user/checkAccountExist.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0KWKhHGKMozbrhuo6u932A.png" /></figure><p>I started probing with public emails, and I could guess that it is so easy because of the contact us page (almost 50+ employees’ emails in my hand): [support@company.com](mailto:support@company.com), [hr@company.com](mailto:hr@company.com).</p><p><strong>And that’s where the developer made their first critical mistake.</strong></p><p>When I hit a valid email, the API didn’t just say <strong><em>“Yes, this account exists.”</em></strong> Instead, it generously included the user’s primary key — the xid. This unique identifier is the application&#39;s <strong>direct object reference</strong> to the profile.</p><p>I didn’t need to guess. I didn’t need brute force. By checking a few known corporate emails, I had instantly and silently generated a verified list of high-value internal employee xids. I now had the victims&#39; coordinates.</p><h3>The Easiest Escalation I’ve Ever Seen</h3><p>With my victim’s xid notes (let&#39;s call it VICTIM_99999), I moved to the main event: the <strong>Profile Update</strong> section.</p><ol><li>I logged in with my account.</li><li>I opened my own profile settings and changed my name to “Tester.”</li><li>I intercepted the request in Burp Suite.</li></ol><p>The request body, intended to update <em>my</em> profile, looked perfectly innocent:</p><pre>{<br>  &quot;name&quot;: &quot;Tester&quot;,<br>  &quot;mobileNumber: &quot;323232323&quot;,<br>  &quot;company&quot;: &quot;ccccccc&quot;,<br>  &quot;designation&quot;: &quot;Researcher&quot;,<br>  // ... other update fields<br>}</pre><p>Now, the chilling part. The server was clearly trusting the xid parameter provided in the POST request to decide <em>who</em> to update. The authentication token in the header proved I was logged in, but the server never checked if the token’s owner (MY_XID_12345) matched the body’s target (MY_XID_12345).</p><p><strong>The Bypass:</strong> I simply deleted my own ID and inserted the CEO/HR/CISO&#39;s xid I had just enumerated.</p><p><strong>Field Original Value (My Profile) to New Value (CEO’s Profile)”</strong></p><p><strong>xid”</strong>&quot;MY_XID_12345&quot; to <strong>&quot;VICTIM_99999&quot;</strong></p><p><strong>&quot;name&quot;</strong>&quot;Tester&quot; to <strong>&quot;Evil Impersonator&quot;</strong></p><p><strong>&quot;mobileNumber&quot;</strong>&quot;1234...&quot; to <strong>&quot;9990001112&quot;</strong></p><p><strong>“company”:</strong> “dddddddddd”,</p><p><strong>“designation”:</strong> “CEO/HR/CISO”,</p><p><strong>and so on.......</strong></p><p>I hit <strong>Forward</strong>.</p><p><strong>Result: Instant Success. HTTP 200.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*V0wTEYYc5kYSGgz7Abav9Q.png" /></figure><p>I had just leveraged my privilege session to update the profile of a CEO/HR/CISO’s. The changes were immediate and persistent. I could log out and verify that the employee’s official details (through the CEO/HR/CISO’s user name that leaked in response)— Name, Contact Number, Company Affiliation — were all now fraudulent.</p><h3>What Trusting the Client Costs You</h3><p>This is the quintessential IDOR scenario, packaged neatly with a perfect enumeration vector. The impact wasn’t just “data exposure;” it was <strong>identity corruption</strong>.</p><p>Imagine an attacker changing the phone number of the Head of HR on the official portal and then calling employees, pretending to be them. The potential for large-scale social engineering and internal fraud is immense.</p><h3>The Developer’s Golden Rule:</h3><p>If you are building an API that modifies an object, <strong>NEVER, EVER</strong> rely on the object reference (like xid) provided by the client&#39;s payload.</p><ol><li><strong>Retrieve the User ID:</strong> Get the authenticated user’s ID <em>only</em> from the secure session token (JWT payload, Session Store). This ID is your source of truth.</li><li><strong>Ignore the Client ID:</strong> Discard or ignore the xid provided in the request body. If the user is updating <em>their own</em> profile, the ID from the session is the only ID you need.</li></ol><p>This vulnerability was a masterclass in why we should <strong>trust our session state and treat all client input as hostile.</strong> It’s a vivid reminder that sometimes, the most sophisticated attacks exploit the most simplistic coding oversights.</p><p><a href="https://infosecwriteups.com/how-i-bypassed-random-medical-center-assignment-during-appointment-booking-16daaf7b7a6d">How I Bypassed Random Medical Center Assignment During Appointment Booking</a></p><p><em>Have you encountered similar “trust-the-client” IDORs? Share your war stories in the comments below!</em></p><p>💬 Let’s Connect:</p><ul><li>🐦 Twitter/X: @<em>spy0x1</em></li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*SfhcFh5-VzJL_r0-" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=25740fb56b05" width="1" height="1" alt=""><hr><p><a href="https://infosecwriteups.com/beyond-authentication-exploiting-a-nasty-idor-in-profile-update-functionality-25740fb56b05">Beyond Authentication — Exploiting a Nasty IDOR in Profile Update Functionality</a> was originally published in <a href="https://infosecwriteups.com/">InfoSec Write-ups</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I Could Change Anyone’s Email Preferences — Without Logging In ]]></title>
            <link>https://infosecwriteups.com/i-could-change-anyones-email-preferences-without-logging-in-dc228b541ef9?source=rss-4c31ad2136a3------2</link>
            <guid isPermaLink="false">https://medium.com/p/dc228b541ef9</guid>
            <category><![CDATA[application-security]]></category>
            <category><![CDATA[bug-bounty]]></category>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[hacking]]></category>
            <category><![CDATA[inspiration]]></category>
            <dc:creator><![CDATA[Munna]]></dc:creator>
            <pubDate>Thu, 13 Nov 2025 13:20:56 GMT</pubDate>
            <atom:updated>2026-10-26T04:00:11.923Z</atom:updated>
            <content:encoded><![CDATA[<h3>I Could Change Anyone’s Email Preferences — Without Logging In 😳</h3><p><em>How a single overlooked API made every user’s inbox mine to control — and how a second endpoint let me confirm it instantly. BOOM.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SaaMoQY3P2kRCHQU2DMP7Q.png" /></figure><blockquote><strong>What if I told you that with nothing more than your email address, I could stop all your newsletters, alerts, and product notifications — and you’d never even know?</strong></blockquote><p>That was the question that started it all. I wasn’t trying to break anything. I was just poking around the email preferences page of a large business website, you know, the usual <em>“Manage my subscriptions”</em> page we all ignore. But what I found turned out to be a silent and surprisingly powerful vulnerability.</p><h3><strong><em>The innocent curiosity 👀</em></strong></h3><p>It began like most bug-hunting stories, a lazy evening, coffee in one hand, Burp Suite in the other.<br>I logged into the website, visited the <strong>“Manage Preferences” </strong>section, and changed one of my email settings just to see what kind of request it would send in the background.</p><p><strong>Boom — there it was.<br></strong>A simple POST request, clean and minimal:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/685/1*6q1z2Vfjyu35TE1v4NoZTg.png" /></figure><p>Nothing fancy. No tokens. No authentication headers. Just an<strong> email parameter and a few IDs.</strong></p><p>At first, it didn’t look dangerous. But the lack of complexity was exactly what caught my eye.</p><h3>“Wait… what if I put someone else’s email?”</h3><p>That thought hit me like a caffeine shot.</p><p>So I tried it.</p><p>I replaced my email with a random one, a publicly listed company email I found on their <strong><em>Contact Us</em> page</strong>. Then, I hit <strong><em>Send</em></strong>.</p><p>Response?</p><pre>{&quot;code&quot;:200,&quot;status&quot;:&quot;SUCCESS&quot;,&quot;message&quot;:&quot;OK&quot;,&quot;data&quot;:{&quot;message&quot;:&quot;Unsubscribed Successfully.&quot;}}</pre><p>Success.<br>No error, no permission check, no authentication required.</p><p>It actually <strong>worked</strong>.</p><p>The server had just accepted my request and unsubscribed someone else’s email address as if I were them.</p><h3>The confirmation endpoint —&gt; the double punch 🥊</h3><p>If that was bad, the next discovery was the knockout. While testing, I found a <strong>second endpoint</strong> — one that returns the current email preferences when you supply an email. Guess what it checked?</p><p>You guessed it: <strong>only the email</strong>. No login, no token, no captcha — just the email.</p><p>That meant the workflow was terrifyingly simple:</p><ol><li>Send an unauthenticated request to change preferences for victim@example.com.</li><li>Query the second endpoint with victim@example.com and see the response showing all their preferences now changed.</li><li>Repeat at scale.</li></ol><p><strong>BOOM!!!</strong></p><p>I could ensure my silent tampering actually worked and automate checks across thousands of addresses. The confirmation endpoint turned a one-off annoyance into an efficiently verifiable and therefore exploitable vector.</p><h3>The silent impact</h3><p>Here’s the scary part:<br>The victim (the real owner of the email) <strong>received no notification at all.</strong></p><p><strong>No “Your preferences were changed.”<br>No “If this wasn’t you, click here.”<br>Nothing.</strong></p><p>That meant I could:</p><ul><li>Silently unsubscribe anyone from all company emails.</li><li>Disable marketing or alert emails for thousands of users if I wanted to automate it.</li><li>Even prevent users from receiving important security notifications or billing updates.</li></ul><p>And with the confirmation endpoint, I had instant proof of success. The system not only let me tamper, it let me verify the tampering — all without authentication.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*ff1D6gHwGxGmDFvnbFSERA.png" /></figure><h3>Lessons learned 🧠</h3><p>This bug wasn’t glamorous. It didn’t require fancy payloads or deep protocol bugs. It was a failure of trust assumptions: the system trusted an email string as an authority and trusted that returning preference data was safe.</p><p>The lesson for engineers: treat emails and public identifiers as attributes — never as authentication. And for product folks: if a change affects someone else, make sure that someone else knows about it.</p><p><a href="https://medium.com/@mr0x1_/how-a-small-tweak-uncovered-an-idor-vulnerability-at-unicef-5461882f33f4">How a Small Tweak Uncovered an IDOR Vulnerability at UNICEF</a></p><p><em>If you enjoyed this write-up, drop a 💬 or share it with someone who still thinks “it’s just an email preference page.”</em></p><p>💬 Let’s Connect:<br>🐦 Twitter/X: @<em>spy0x1</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/387/1*2smEAjGYmqsWrFPG22NbUQ.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dc228b541ef9" width="1" height="1" alt=""><hr><p><a href="https://infosecwriteups.com/i-could-change-anyones-email-preferences-without-logging-in-dc228b541ef9">I Could Change Anyone’s Email Preferences — Without Logging In 😳</a> was originally published in <a href="https://infosecwriteups.com/">InfoSec Write-ups</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How I Bypassed Random Medical Center Assignment During Appointment Booking]]></title>
            <link>https://infosecwriteups.com/how-i-bypassed-random-medical-center-assignment-during-appointment-booking-16daaf7b7a6d?source=rss-4c31ad2136a3------2</link>
            <guid isPermaLink="false">https://medium.com/p/16daaf7b7a6d</guid>
            <dc:creator><![CDATA[Munna]]></dc:creator>
            <pubDate>Sat, 04 Oct 2025 07:41:31 GMT</pubDate>
            <atom:updated>2026-07-23T14:06:04.714Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Sometimes, the smallest parameter can change the entire system.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*C4LEi13nvbp9S_xmwyYLMA.png" /></figure><p>Hey folks, I’m Munna, also known as <strong><em>spy0x1 </em></strong>in the security circles, a security researcher and penetration tester. I usually spend my time poking around applications, testing boundaries, and uncovering the little details that can sometimes lead to big vulnerabilities. This time, I want to share one of my findings from a medical test booking platform.</p><h4>It All Started With a $5 Appointment</h4><p>Like many of my test targets, this one looked simple at first glance: This is an internal pentest project of a healthcare platform offering users the ability to book medical test appointments.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/254/1*KQC-PcFLZkWWxk46qFnvOA.png" /></figure><p>The flow seemed straightforward: fill in personal details, pay a small fee (<strong>in this case, $5</strong>), and get assigned a medical center randomly.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/431/1*dpOBjgl5CsSJuRFb67_TKg.png" /></figure><p>So far, so good, right?</p><p>But curiosity is a researcher’s best friend.</p><h4>A “Random” Choice That Didn’t Feel Random</h4><p>While booking, I noticed something unusual:</p><p>A form appeared for user details (as expected).</p><p>The system then randomly assigned a medical center — users had no control over choosing where to go.</p><p>But here’s the twist: when intercepting the request with a proxy like BurpSuite, I discovered all available medical center information was exposed. Not only visible in the UI, but the backend also shares every center’s unique <strong><em>ID with value=””</em></strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/797/1*MzGD-KpYrpgFjsYtinUxaQ.png" /></figure><p>At this point, it became clear that the random assignment wasn’t so “random” after all, it was based on those <strong>ID</strong>s.</p><h4>The Experiment: One Tiny Change</h4><p>When I looked closely at the intercepted request, I noticed a parameter: <strong>“medical_center”: “”</strong></p><pre>POST /api/v1/appointment/book HTTP/1.1<br>Host: spy0x1.com<br>Content-Type: application/json<br>Content-Length: 189<br>{<br>  &quot;name&quot;: &quot;John Doe&quot;,<br>  &quot;age&quot;: &quot;28&quot;,<br>  &quot;phone&quot;: &quot;1234567890&quot;,<br>  &quot;medical_center&quot;: &quot;&quot;,<br>  &quot;payment&quot;: &quot;pending&quot;<br>}</pre><p>It was blank.</p><p>This raised an eyebrow. What if I filled this parameter with my desired <strong>medical center ID</strong> instead of letting the system choose for me?</p><h4>Putting the Hypothesis to the Test</h4><p>I modified the request: <strong>“medical_center”: “desired_id”</strong></p><pre>POST /api/v1/appointment/book HTTP/1.1<br>Host: spy0x1.com<br>Content-Type: application/json<br>Content-Length: 189<br>{<br>  &quot;name&quot;: &quot;John Doe&quot;,<br>  &quot;age&quot;: &quot;28&quot;,<br>  &quot;phone&quot;: &quot;1234567890&quot;,<br>  &quot;medical_center&quot;: &quot;325&quot;,<br>  &quot;payment&quot;: &quot;pending&quot;<br>}</pre><p>In this case, <strong>325</strong> was the ID of my preferred medical center (collected earlier when all centers were exposed).</p><p><strong>The server accepted it without hesitation. No validation, no restriction, no second check.</strong></p><p>Proceeding to the payment interface, I completed the $5 transaction.</p><p>And there it was, an appointment booked <strong>with the medical center of my choice</strong>, not the randomly assigned one.</p><h3>Steps to Reproduce</h3><blockquote><strong>(Prerequisites: test account on the target, an intercepting proxy like Burp/Zap, and a payment method for the $5 test transaction.)</strong></blockquote><ol><li>Open the target site and log in to a test account.</li><li>Navigate to <strong>Book an Appointment</strong> and fill in the required user details (name, age, phone, passport etc.) but do <strong>not</strong> submit yet.</li><li>Start your intercepting proxy and enable request capture.</li><li>Submit the appointment form. Intercept the outgoing POST request to the booking endpoint (example path: POST /api/v1/appointment/book).</li><li>Inspect the JSON body of the intercepted request and confirm the medical_center parameter is empty:</li></ol><pre>{<br>  &quot;name&quot;:&quot;John Doe&quot;,<br>  &quot;age&quot;:&quot;28&quot;,<br>  &quot;phone&quot;:&quot;1234567890&quot;,<br>  &quot;medical_center&quot;:&quot;&quot;,<br>  &quot;payment&quot;:&quot;pending&quot;<br>}</pre><p>6. Modify the intercepted booking request by setting medical_center to a valid center <strong>ID</strong> you collected:</p><pre>{<br>  &quot;name&quot;:&quot;John Doe&quot;,<br>  &quot;age&quot;:&quot;28&quot;,<br>  &quot;phone&quot;:&quot;1234567890&quot;,<br>  &quot;medical_center&quot;:&quot;325&quot;,<br>  &quot;payment&quot;:&quot;pending&quot;<br>}</pre><p>7. Forward the modified request from the proxy to the server.</p><p>8. Proceed through the payment UI and complete the<strong> $5 </strong>transaction for the booking.</p><p>9. Verify in the UI (or via the booking details API) that the confirmed appointment is assigned to the medical_center ID you injected (i.e., you were able to choose the center instead of the system’s random assignment).</p><h4>Why This Was Bigger Than Just Convenience</h4><p>At first glance, it might seem like I simply gave users a choice they didn’t have. But the implications run deeper:</p><ul><li><strong>Business logic broken</strong> — The “random” assignment wasn’t enforced at all.</li><li><strong>Server trust issue</strong> — The backend relied entirely on client input.</li><li><strong>Potential abuse</strong> — Imagine exploiting this to always book in-demand centers, or even manipulating other appointment-related data.</li></ul><h4>How Developers Can Prevent This</h4><p>This wasn’t a flashy bug — it was a <strong>logic flaw</strong>. But logic flaws often cause the most damage if left unchecked. Here’s how to fix it:</p><ol><li><strong>Validate parameters server-side</strong> — Never trust user input for critical decisions.</li><li><strong>Enforce randomization in the backend</strong> — Don’t let the client override it.</li><li><strong>Whitelist allowed values</strong> — Only accept server-issued tokens or IDs.</li><li><strong>Avoid oversharing</strong> — Don’t expose internal IDs in responses unless absolutely necessary.</li><li><strong>Add anomaly detection</strong> — Log and flag suspicious booking behavior for review.</li></ol><h3>The Lesson I Took Away</h3><p>This was a perfect reminder that even something as small as a <strong>blank field</strong> can change how a whole system works.</p><p><strong>For developers: </strong>trust your backend, not your users.<br><strong>For researchers:</strong> keep your eyes sharp, even the tiniest parameters can tell the biggest stories.</p><p><a href="https://medium.com/@mr0x1_/how-a-small-tweak-uncovered-an-idor-vulnerability-at-unicef-5461882f33f4">How a Small Tweak Uncovered an IDOR Vulnerability at UNICEF</a></p><p>💬 Let’s Connect:</p><ul><li>🐦 Twitter/X: @<em>spy0x1</em></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=16daaf7b7a6d" width="1" height="1" alt=""><hr><p><a href="https://infosecwriteups.com/how-i-bypassed-random-medical-center-assignment-during-appointment-booking-16daaf7b7a6d">How I Bypassed Random Medical Center Assignment During Appointment Booking</a> was originally published in <a href="https://infosecwriteups.com/">InfoSec Write-ups</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How a Small Tweak Uncovered an IDOR Vulnerability at UNICEF]]></title>
            <link>https://spy0x1.medium.com/how-a-small-tweak-uncovered-an-idor-vulnerability-at-unicef-5461882f33f4?source=rss-4c31ad2136a3------2</link>
            <guid isPermaLink="false">https://medium.com/p/5461882f33f4</guid>
            <dc:creator><![CDATA[Munna]]></dc:creator>
            <pubDate>Fri, 12 Sep 2025 04:35:13 GMT</pubDate>
            <atom:updated>2026-07-23T14:08:39.863Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fKwKwNWw9hCvXWtUzBUIpw.png" /></figure><p><strong>Hey folks!</strong></p><p>Ever stumble upon something that seems completely harmless and then realize it’s actually a ticking time bomb? That’s exactly what happened during one of my recent security tests on a UNICEF domain. What started as routine testing of an “A<strong>ssessment” feature</strong> quickly turned into a full-blown discovery: with just a small tweak in the URL, I could peek into other people’s <strong>private (unpublished) assessments</strong> and even delete them.</p><p>This sneaky vulnerability is called an <strong>Insecure Direct Object Reference (IDOR)</strong>. It’s simple, surprisingly common, and potentially devastating if left unchecked. Stick with me, I’ll take you step by step through how I found it, why it matters, and how it can be prevented.</p><h3>The Curiosity That Sparked It</h3><p>It all began like any other testing session. I logged in as a legitimate user and <strong>created an assessment</strong> on the platform (<em>https://subdomain.unicef.org/</em>). Everything seemed normal… until I noticed the URL:</p><pre>https://subdomain.unicef.org/XXXXXXXXXX/YYYYYYYYYYY/57</pre><p>The <strong><em>“57” </em></strong>instantly caught my eye. It wasn’t random, hashed, or complex — just a simple number.</p><p>And when numbers look this predictable, my instincts tell me one thing: <strong>time to experiment.</strong></p><h3>Steps to Reproduce the Flaw</h3><p>Here’s how the puzzle unfolded step by step, showing both <strong>viewing</strong> and <strong>deleting</strong> other users’ assessments:</p><h4>1. Viewing Another User’s Assessment (GET Request)</h4><p>I intercepted the request in Burp Suite while viewing my own assessment:</p><pre>GET /XXXXXXXX/YYYYYYYY/57 HTTP/1.1<br>Host: Subdomain.unicef.org</pre><p>Curiosity struck, so I modified the ID from 57 to 60:</p><pre>GET /XXXXXXXXX/YYYYYYYYY/60 HTTP/1.1<br>Host: Subdomain.unicef.org</pre><p><strong>Result:</strong> Instead of blocking me, the server returned the contents of another user’s assessment, including private assessments that had <strong>not been published yet</strong>.</p><h4>2. Deleting Another User’s Assessment (DELETE Request)</h4><p>Next, I intercepted the delete request for my own assessment:</p><pre>DELETE /XXXXXXXX/YYYYYYYY/57 HTTP/1.1<br>Host: Subdomain.unicef.org</pre><p>Again, I swapped 57 with 60:</p><pre>DELETE /XXXXXXXXXX/YYYYYYYY/60 HTTP/1.1<br>Host: Subdomain.unicef.org</pre><p><strong>Twist moment:</strong> I expected the server to return an error… but instead, the assessment <strong>vanished instantly</strong>. Public or private, it didn’t matter; the system deleted someone else’s work without any restriction. My heart skipped a beat as I realized the vulnerability wasn’t just theoretical; it was real and dangerous.</p><h3>How It Could Be Fixed</h3><p>There are two simple but powerful fixes:</p><p><strong>1. Access Control Checks</strong></p><ul><li>The server should always verify that the user requesting access owns the resource.</li><li>If you didn’t create the assessment60, you should never see it, no matter what the URL says.</li></ul><p><strong>2. Non-Sequential IDs</strong></p><ul><li>Instead of using predictable numbers, switch to randomized or UUID-based identifiers.</li></ul><h3><strong>Final Thoughts</strong></h3><p>Even the smallest details can have big consequences. For organizations: never trust a URL alone — enforce server-side authorization. For researchers: stay curious — sometimes changing just a number uncovers a major flaw.</p><p>And the cherry on top? Responsible reporting earned me a spot in the <strong>Hall of Fame,</strong> proof that curiosity, persistence, and ethical hacking do get noticed.</p><p>💬 Let’s Connect: <br>🐦 Twitter/X: @<em>spy0x1</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*d7wtkHGFWg9a0bjFKwgtbQ.png" /><figcaption>Hall of Fame</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5461882f33f4" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>