<?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 Masha Vaverova on Medium]]></title>
        <description><![CDATA[Stories by Masha Vaverova on Medium]]></description>
        <link>https://medium.com/@masha.vaverova?source=rss-317ee55b8432------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*aiyWgdw_eN4W6Xk6</url>
            <title>Stories by Masha Vaverova on Medium</title>
            <link>https://medium.com/@masha.vaverova?source=rss-317ee55b8432------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 23:07:46 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@masha.vaverova/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[A Verifiable Equal Pay Registry on Ethereum]]></title>
            <link>https://coinsbench.com/a-verifiable-equal-pay-registry-on-ethereum-1f946080f508?source=rss-317ee55b8432------2</link>
            <guid isPermaLink="false">https://medium.com/p/1f946080f508</guid>
            <category><![CDATA[ethereum]]></category>
            <category><![CDATA[smart-contracts]]></category>
            <category><![CDATA[equal-pay]]></category>
            <category><![CDATA[transparency]]></category>
            <category><![CDATA[blockchain-technology]]></category>
            <dc:creator><![CDATA[Masha Vaverova]]></dc:creator>
            <pubDate>Mon, 16 Mar 2026 19:45:18 GMT</pubDate>
            <atom:updated>2026-03-27T15:21:16.969Z</atom:updated>
            <content:encoded><![CDATA[<h3>Replacing “trust us” with cryptographic proof.</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ma2wi_GX72oUMXNnLQI-Fg.png" /></figure><h3>The Problem With Pay Transparency</h3><p>Apparently the standard way to prove equal pay is still… a PDF report.</p><p>Companies publish documents with aggregated salary numbers, sometimes once per year. If you’re lucky, you get a spreadsheet. Most of the time it’s a PDF.</p><p>And then everyone is expected to trust the numbers.</p><p>There are obvious problems with this approach:</p><ul><li>reports can be changed later</li><li>the calculation methods are often unclear</li><li>there is no cryptographic proof that the data hasn’t been modified</li><li>independent verification is difficult</li></ul><p>In other words, the current system relies heavily on <strong>institutional trust</strong>.</p><p>I wanted to explore a simple question:</p><blockquote><strong><em>What would pay transparency look like if it were verifiable instead of trusted?</em></strong></blockquote><p>So I built this prototype during the <strong>Chainlink Convergence Hackathon</strong>, mostly out of curiosity.</p><p>I wanted to see whether pay transparency could be turned into something <strong>cryptographically verifiable instead of institutionally trusted</strong>.</p><p>Turns out… it actually works.</p><p>The result is a prototype I ended up calling:</p><h3>The Verifiable Equal Pay Registry</h3><p>The idea is actually pretty simple — at least conceptually.</p><p>Instead of publishing a static report, companies submit <strong>structured aggregated pay data</strong> that can be cryptographically verified.</p><p>The process looks like this:</p><ol><li>A company submits aggregated pay statistics →</li><li>The data is canonicalized and hashed →</li><li>The hash is committed on-chain →</li><li>A score is computed using a transparent methodology →</li><li>Independent attestors verify the report →</li><li>If everything checks out, the company receives a <strong>non-transferable verification badge</strong></li></ol><p>The key concept is simple:</p><p><strong>“Trust us” → replaced by cryptographic proof.</strong></p><p>Importantly, the system never stores individual salaries.<br> Only aggregated statistics are used.</p><h3>Privacy by Design</h3><p>A critical requirement was avoiding personal data.</p><p>Each report only contains <strong>aggregated metrics</strong>, such as:</p><ul><li>headcount per role</li><li>total base pay</li><li>total bonus pay</li><li>total working hours</li></ul><p>Example structure:</p><pre>{<br>  gender: &quot;F&quot;,<br>  level: &quot;Senior&quot;,<br>  nHeadcount: 5,<br>  sumBasePay: &quot;3000000&quot;,<br>  sumHours: 9400<br>}</pre><p>From this information the system calculates <strong>mean pay per FTE</strong>, which allows comparison between groups without exposing individual salaries.</p><h3>Making Reports Cryptographically Verifiable</h3><p>The key technical step is converting the report into a <strong>deterministic canonical form</strong>.</p><p>JSON objects normally allow arbitrary key ordering. That means these two objects are technically identical but produce different hashes:</p><pre>{a:1,b:2}<br>{b:2,a:1}</pre><p>To prevent this, the system uses <strong>canonical JSON formatting</strong> before hashing.</p><p>The pipeline looks like this:</p><pre>Report JSON<br>   ↓<br>Canonicalization<br>   ↓<br>keccak256 hash<br>   ↓<br>reportHash</pre><p>The resulting hash becomes the <strong>commitment stored on-chain</strong>.</p><p>Once committed, the report cannot be altered without producing a different hash.</p><h3>Smart Contract Architecture</h3><p>The on-chain system consists of two contracts.</p><h3>EqualPayRegistry</h3><p>This contract stores report commitments and verification state.</p><p>Key information stored on-chain:</p><ul><li>companyId</li><li>reportHash</li><li>reporting period</li><li>scoreBps</li><li>methodologyId</li><li>attestation count</li><li>verification status</li></ul><p>Roles control the process:</p><ul><li><strong>Admin</strong> — manages system parameters</li><li><strong>Oracle</strong> — publishes computed scores</li><li><strong>Attestors</strong> — verify reports</li></ul><p>Once a report receives enough attestations, it becomes <strong>Verified</strong>.</p><h3>EqualPayBadge</h3><p>Verified companies receive a <strong>soulbound badge</strong>.</p><p>This badge:</p><ul><li>is implemented as an ERC-721 token</li><li>cannot be transferred</li><li>represents a verified report</li></ul><p>The token ID corresponds directly to the report ID, making verification straightforward.</p><p>Anyone can check:</p><ul><li>the registry state</li><li>the badge ownership</li><li>the verification status</li></ul><p>All on-chain.</p><h3>Backend Architecture</h3><p>The off-chain system is built using <strong>NestJS</strong>.</p><p>It provides endpoints for:</p><pre>POST /reports<br>GET /reports/:reportId<br>GET /reports/:reportId/meta<br>GET /reports/:reportId/score<br>POST /oracle/publish/:reportId</pre><p>Responsibilities include:</p><ul><li>report validation (JSON schema)</li><li>canonicalization</li><li>hash computation</li><li>score calculation</li><li>oracle publishing</li></ul><p>Reports are stored in a database together with their hashes and metadata.</p><h3>Computing the Pay Gap Score</h3><p>The current prototype uses a simple methodology:</p><p><strong>Mean base pay per FTE</strong></p><p>The score is expressed in <strong>basis points</strong> to avoid floating-point issues.</p><p>Example output:</p><pre>scoreBps: 742<br>methodology: mean_gap_v1_fte<br>meanBasePayPerFte_F: 464285.71<br>meanBasePayPerFte_M: 501479.29</pre><p>The methodology itself is also hashed and stored as:</p><pre>methodologyId = keccak256(&quot;mean_gap_v1_fte&quot;)</pre><p>This ensures that scoring methods remain transparent and verifiable.</p><h3>Why Use Blockchain?</h3><p>This system could technically be built using a traditional database.</p><p>But blockchain introduces several important properties:</p><ul><li><strong>immutability</strong></li><li><strong>public verification</strong></li><li><strong>transparent governance</strong></li><li><strong>tamper resistance</strong></li></ul><p>Once a report hash is committed on-chain, it cannot be quietly replaced or edited.</p><p>Anyone can independently verify the integrity of the data.</p><h3>What Could This Become?</h3><p>A system like this could potentially support:</p><ul><li>regulatory reporting</li><li>independent audits</li><li>NGO verification</li><li>academic research datasets</li><li>public transparency dashboards</li></ul><p>The prototype is just a starting point.</p><h3>Final Thoughts</h3><p>Blockchain obviously doesn’t fix social problems by itself.</p><p>But it can change <strong>how trust is built</strong>.</p><p>And sometimes improving the infrastructure of trust is exactly where technology can make a real difference.</p><h3>Tech Stack</h3><ul><li>Solidity (Smart Contracts)</li><li>Hardhat (Development + Testing)</li><li>NestJS (API &amp; report processing)</li><li>Viem (on-chain interaction)</li></ul><p>Feedback is very welcome !</p><p>All the best!</p><p>Link to the repo: <a href="https://github.com/mashavaverova/EqualPay_Chainlink">https://github.com/mashavaverova/EqualPay_Chainlink</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1f946080f508" width="1" height="1" alt=""><hr><p><a href="https://coinsbench.com/a-verifiable-equal-pay-registry-on-ethereum-1f946080f508">A Verifiable Equal Pay Registry on Ethereum</a> was originally published in <a href="https://coinsbench.com/">CoinsBench</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I Thought I Understood Uniswap V2]]></title>
            <link>https://medium.com/@masha.vaverova/i-thought-i-understood-uniswap-v2-4cfce3ba9240?source=rss-317ee55b8432------2</link>
            <guid isPermaLink="false">https://medium.com/p/4cfce3ba9240</guid>
            <category><![CDATA[uniswap-v2]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[defi]]></category>
            <dc:creator><![CDATA[Masha Vaverova]]></dc:creator>
            <pubDate>Wed, 25 Feb 2026 19:49:43 GMT</pubDate>
            <atom:updated>2026-02-25T19:49:43.097Z</atom:updated>
            <content:encoded><![CDATA[<p>Turns Out I Mostly Understood My Imagination.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*klIxdfPlQ41IGDu2ixd1Dw.png" /></figure><p>There’s a very specific kind of confidence you get after reading about x*y = k.</p><p>You nod.<br> You get it.<br> Constant product. Reserves. Simple.</p><p>Uniswap V2 feels elegant. Minimal. Almost obvious.</p><p>So I built a small local lab around it — not to hack it, not to write a dramatic vulnerability thread — just to test whether my understanding survived contact with reality.</p><p>(It mostly didn’t)</p><h3>Assumption #1: If I Send 100 Tokens, 100 Tokens Arrive</h3><p>This feels morally correct.</p><p>If transfer(100) succeeds, then 100 tokens moved.<br> Right?</p><p>Not always.</p><p>I deployed a fee-on-transfer token that takes 2% on every move. If you send 1000 tokens, 980 actually arrive. Nothing malicious. Just a haircut. I plugged that token into a local Uniswap V2 pair and ran swaps.</p><p>The router gave me a confident quote. Everything looked normal.</p><p>But the pair received less than the intended amount. And since Uniswap calculates output based on what actually arrives, not what you meant to send, the output was smaller than the quote suggested. Nothing broke.The protocol did exactly what it promises.</p><p>The only thing that broke was my mental shortcut:</p><blockquote><em>amountIn equals amount received.</em></blockquote><p>It doesn’t — not as a guarantee.</p><h3>Assumption #2: A Router Quote Is Basically Destiny</h3><p>There’s something deeply satisfying about getAmountsOut(amountIn) returning a clean number. It feels official. Like a commitment. But that quote assumes the pair actually receives amountIn. With fee-on-transfer tokens, it doesn’t.</p><p>So now you have two outputs:</p><ul><li>Naive output (based on intended input)</li><li>Safe output (based on actual received input)</li></ul><p>And mathematically, it behaves predictably:</p><p>naiveOut ≥ safeOut</p><p>Usually strictly greater.</p><p>The math isn’t wrong. The system isn’t inconsistent. It’s just operating on the true input, not your hopeful one. If you treat quotes like guarantees, you’re quietly assuming uniform ERC20 behavior.</p><p>That assumption is not part of Uniswap’s contract.</p><h3>Assumption #3: More Tokens in the Pair Means More Liquidity</h3><p>Then I tried something even simpler.</p><p>I transferred tokens directly to the pair contract. Not through mint, not through swap. Just a raw ERC20 transfer. Surely something meaningful changes? Yes. But not what I expected.</p><p>The ERC20 balance increased. The stored reserves did not. Because Uniswap V2 keeps its own accounting. It updates reserves only through specific functions. It doesn’t constantly sync with token balances.</p><p>So now there’s a mismatch:</p><ul><li>Actual token balances are higher.</li><li>Internal reserves are unchanged.</li></ul><p>And what happens to that excess? There’s a function called skim()! Anyone can call it and extract the difference between balances and reserves.</p><p>Again: not a bug. Not an exploit. Just not intuitive if you believe:</p><blockquote><em>Sending tokens to the pair increases liquidity.</em></blockquote><p>It doesn’t. Liquidity changes only through defined protocol actions. The protocol is very strict about how state evolves. My intuition was… looser.</p><h3>The Math Was Never the Problem</h3><p>What’s interesting is that none of the mathematical properties failed.</p><p>Monotonicity held.<br> The invariant held.<br> Fuzz tests behaved as expected.<br> Everything was internally coherent.</p><p>The weirdness didn’t come from broken math. It came from unexamined assumptions. Uniswap’s formula operates on reserves. Reserves are updated in specific contexts. Token transfers are external events.</p><p>Blending those layers in your head produces nice mental models — and subtle mistakes.</p><h3>Why This Was New to Me</h3><p>None of this is secret knowledge. People who’ve lived inside AMMs for years will say, “Of course.” But it wasn’t of course for me.</p><p>I understood the formula.<br> I read the docs.<br> I knew reserves mattered.</p><p>And I still assumed:</p><ul><li>Transfers are exact.</li><li>Quotes are commitments.</li><li>Balances and reserves are interchangeable.</li></ul><p>They aren’t.</p><p>And the only reason I noticed is because I stopped trusting my intuition and started writing small, boring tests.</p><h3>Why I Built the Lab</h3><p>The lab has around eight small tests:</p><ul><li>Fee-on-transfer input mismatch</li><li>Naive quote vs actual output</li><li>Donation-induced balance/reserve drift</li><li>Skim extraction</li><li>Some simple fuzz properties</li></ul><p>Nothing groundbreaking.</p><p>But enough to ask a simple question:</p><blockquote><em>What is actually guaranteed here?</em></blockquote><p>Not what feels obvious.<br> Not what usually works.<br> What is explicitly guaranteed.</p><p>That shift changes how you approach integrations. Not just with Uniswap. With anything.</p><h3>The Real Lesson</h3><p>Uniswap V2 isn’t broken. Fee-on-transfer tokens aren’t evil.</p><p>skim() isn’t a vulnerability. Everything behaves exactly as specified.</p><p>The only thing that failed was the gap between specification and assumption. And that gap is where most integration mistakes live.</p><p>Sometimes the safest thing to audit isn’t the protocol.</p><p>It’s the sentence in your head that starts with:</p><p>“Well, obviously…”</p><p>— — — —</p><p>I turned all of this into a tiny Foundry lab.<br> Nothing fancy — just a few tests that break common assumptions.</p><p>If you’re learning Uniswap V2 or integrating with it, feel free to poke around, reuse it, or extend it:<br> 👉 <a href="https://github.com/mashavaverova/DEX-Assumption-Breaker">https://github.com/mashavaverova/DEX-Assumption-Breaker</a></p><p>It helped me. Maybe it’ll help you too.</p><p>All the best</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4cfce3ba9240" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[One more text about Web3 Security]]></title>
            <link>https://medium.com/@masha.vaverova/one-more-text-about-web3-security-af293f49f817?source=rss-317ee55b8432------2</link>
            <guid isPermaLink="false">https://medium.com/p/af293f49f817</guid>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[web3]]></category>
            <category><![CDATA[online-safety]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[security]]></category>
            <dc:creator><![CDATA[Masha Vaverova]]></dc:creator>
            <pubDate>Mon, 02 Feb 2026 18:12:22 GMT</pubDate>
            <atom:updated>2026-02-02T18:14:22.202Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>But with some practical links and extra reading tips (for those who want to go all in and read even more)</strong></p><p>As you probably already understand, Web3 gives you freedom, ownership, and control. This is something you can see on every corner and hear from every microphone. And to be honest, it <em>is</em> really good and really powerful.</p><p>But at the same time, it also gives you something exciting called <strong>“personal responsibility with zero refunds.”</strong><br>Yes — it means there’s no more “forgot password”, no “chargeback”, and not even a “please undo”.”</p><p>So let’s talk about how <strong>not</strong> to accidentally donate your crypto to strangers on the internet.</p><ol><li><strong>The One Rule That Saves People Money:</strong></li></ol><blockquote>If you don’t understand <strong>why</strong> you’re being asked to sign something — don’t sign it. That’s it. Yeah, so simple.</blockquote><p>You can close the article now.<br> <em>(But plz don’t — the rest makes this rule easier to follow.)</em></p><p><strong>2. “Signing” Is NOT THE SAME AS “Logging In”</strong></p><p>In Web3, every time you see a wallet popup, it can mean:</p><ul><li>giving permission to move your tokens</li><li>approving future spending (sometimes unlimited)</li><li>transferring assets immediately</li></ul><p>In <strong>EVERY</strong> case, your wallet — like <strong>MetaMask</strong> — is trying to tell you what’s really happening. The website? Sometimes… not so honest.</p><blockquote><strong>Rule:</strong><br> Trust the wallet popup more than the website text.<br> Websites can lie. Wallets are brutally honest.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Uni4h0_57AOWne-a7kPLEQ.png" /></figure><p><strong>3. Please Stop Using One Wallet for Everything. </strong><em>(And finally buy a hardware one)</em></p><p>Using one wallet for: long-term savings and random mints and experimental dApps is like using the same password for your bank, email, and that sketchy site you signed up for in 2014.</p><p>Instead, do this (it’s free):</p><blockquote><strong><em>Cold wallet</em></strong><em> → savings, NFTs, governance<br> Long-term storage. Rarely used. Maximum security.</em></blockquote><blockquote><strong><em>Hot wallet</em></strong><em> → daily use<br> Your everyday wallet for apps and DeFi. Keep limited funds.</em></blockquote><blockquote><strong><em>Burner wallet</em></strong><em> → new sites, mints, experiments<br> High-risk wallet. Assume it can break. Use and discard.</em></blockquote><p>If (when) you mess up, this limits the damage.</p><p><strong>Also! USE Hardware Wallets!</strong></p><p>Yes, this can cost money.<br> But it’s not sexy — it’s <strong>VERY, VERY effective</strong>.</p><p>If your keys live only in your browser, your security depends on: your laptop, your browser extensions, your typing accuracy at 2 a.m. That’s… ambitious.</p><p>Hardware wallets keep your keys off your computer.<br> If you’ve never heard about them — check these:</p><ul><li><strong>Ledger (</strong><a href="https://www.ledger.com/">https://www.ledger.com/</a>)</li><li><strong>Trezor (</strong><a href="https://trezor.io/">https://trezor.io/</a>)</li></ul><blockquote><strong>Rule:</strong><br> If losing the funds would hurt — use a hardware wallet.</blockquote><p><strong>One more rule:</strong> buy <strong>only original devices</strong> from the official websites.<br> Not on marketplaces.<br> Not “somewhere cheaper.”<br> Not second-hand.<br> Only original. Only hardcore.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*C-sDqwBh6zTLjBNDMABT2w.png" /></figure><p><strong>4. The Silent Killer: Forgotten Token Approvals</strong></p><p>Many people don’t get hacked immediately.<br> They get drained later, because at some point (probably when they were green and young) they clicked:</p><p>“Approve unlimited”</p><p>…and then forgot about it. Totally understandable — but unforgivable 😌</p><p>If that website from your early Web3 days gets hacked and still has your approval — congrats, you have no funds anymore.</p><blockquote>So please check and clean this up here:<br> <strong>Revoke.cash</strong><br> <a href="https://revoke.cash/">https://revoke.cash</a></blockquote><blockquote>Make it a habit: monthly check and extra check after trying new dApps</blockquote><p>Think of it as <strong>brushing your crypto teeth</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JlGy-lXJCo0-2ugSQRKhfg.png" /></figure><p><strong>5. Links, DMs, and “Support”: A Horror Story</strong></p><p>Most scams start with a Discord DM, a Twitter reply, a fake “support” message, you name it… So here is not much to say, except:</p><blockquote>Navigate to sites yourself. Don’t click links sent to you.</blockquote><p><strong>Confusion Is Your Built-In Alarm System</strong></p><p>If you’re thinking:</p><ul><li>“Wait… what am I approving?”</li><li>“I don’t really get this but okay”</li><li>“Everyone else is doing it”</li></ul><p>Stop!</p><p>That uncomfortable feeling is not stupidity! <br> It’s your brain saving you money.</p><blockquote><strong>Rule:</strong><br> If you can’t explain the transaction in one sentence, don’t sign it.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5DGvLPGdBBtWJc93wg5wxw.png" /></figure><p><strong>You Don’t Need to Be Paranoid — Just Consistent</strong></p><h3><strong>For Curious Humans Who Want to Go Deeper</strong></h3><p><em>(No hype. No grindset. No bullshit.)</em></p><p>If you actually want to understand <strong>why</strong> things break in Web3 — and not just what to click — start here:</p><ul><li><strong>MetaMask — Signature phishing explained</strong><br><a href="https://support.metamask.io/stay-safe/protect-yourself/wallet-and-hardware/signature-phishing"> https://support.metamask.io/security/privacy-and-security/signature-phishing/<br></a> → Explains why “just signing” is where most disasters begin.</li><li><strong>Ethereum Foundation — Staying safe in crypto</strong><br><a href="https://ethereum.org/security/"> https://ethereum.org/en/security/<br></a> → Calm, neutral, grown-up security basics. No drama.</li><li><strong>Revoke.cash — Why approvals are dangerous</strong><br> <a href="https://revoke.cash/learn">https://revoke.cash/learn</a><br> → The silent killer explained clearly (and painfully).</li></ul><p>No bullshit.<br> Just reality.</p><h3><strong>Useful Links (For People Who Scrolled Straight Here )</strong></h3><p><strong>1. Revoke token approvals</strong></p><p><a href="https://revoke.cash/">https://revoke.cash</a></p><p><strong>What it does:</strong></p><p>Shows who can move your tokens &amp; Lets you revoke old or <em>unlimited</em> approvals</p><p><strong>2. Transaction understanding &amp; simulation</strong></p><p><a href="https://tenderly.co/">https://tenderly.co</a></p><p>Many wallets use Tenderly under the hood and If your wallet <strong>warns you → listen! </strong>If a dApp says <em>“ignore the warning”</em> → <strong>run!</strong></p><p>Seriously. That’s not confidence. That’s <strong>a red flag.</strong></p><p><strong>3. Scam &amp; phishing detection</strong></p><p><a href="https://www.scamsniffer.io/">https://www.scamsniffer.io</a></p><p>Browser extension + website scanner, Flags known scam domains and warns you before malicious signatures. Not perfect — but way better than vibes.</p><p>Stay safe.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=af293f49f817" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>