<?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 Yozhef Hisem on Medium]]></title>
        <description><![CDATA[Stories by Yozhef Hisem on Medium]]></description>
        <link>https://medium.com/@yozhef?source=rss-c2c049b6f672------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*SKcJqCNlTkRekgt9Kc48zA.jpeg</url>
            <title>Stories by Yozhef Hisem on Medium</title>
            <link>https://medium.com/@yozhef?source=rss-c2c049b6f672------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 11:33:44 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@yozhef/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[“Can You Make My Website Pretty?” — The Question That Made Me Quit Waiting and Start Building]]></title>
            <link>https://yozhef.medium.com/can-you-make-my-website-pretty-the-question-that-made-me-quit-waiting-and-start-building-f5224d963f17?source=rss-c2c049b6f672------2</link>
            <guid isPermaLink="false">https://medium.com/p/f5224d963f17</guid>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[landing-pages]]></category>
            <category><![CDATA[no-code]]></category>
            <dc:creator><![CDATA[Yozhef Hisem]]></dc:creator>
            <pubDate>Wed, 14 Oct 2026 21:24:14 GMT</pubDate>
            <atom:updated>2026-10-14T21:24:14.984Z</atom:updated>
            <content:encoded><![CDATA[<p>For years I was completely buried in my day job. I work as a software architect, I love what I do, but there was never any energy left for something of my own. I had ideas. I didn’t have the time or the push.</p><p>The push came from two very ordinary situations.</p><p>First, some friends updated their business and I offered to make them a small one-page site. It turned out nice. But I knew exactly what would happen next: they’d want to add a section, change a photo, fix a phone number. And every single time they’d need “the IT guy”. Me.</p><p>A few weeks later, the guy who was building a fireplace in our house found out I work in IT. His first question: “Oh, can you help me? I have an old WordPress site, it’s ugly. I want it to look nice.”</p><p>That’s when it clicked. These people don’t need a developer. They need a place where they can just say what they want and get it, without code, without waiting until someone finds a free evening.</p><p>I teamed up with a friend, Serhii, who wanted to build his own product as much as I did. He went deep on the hardest part for non-technical people: how to actually get a site online without touching a terminal. Neither of us is doing this for the money right now. We wanted to build something we believe in and solve a problem we kept running into ourselves.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mTrSdST3JIRxezEQNiIAQw.png" /></figure><p>The result is Ploof Studio, a Mac app where:</p><ul><li>you describe a site in one sentence (add your photos or a site you like as a reference)</li><li>you watch it write the pages in real time</li><li>you click on anything that looks wrong and say what to change</li><li>you publish to your own Vercel or GitHub Pages account</li><li>every file is yours, so you can take it anywhere</li></ul><p>Something I didn’t expect: the people who got most excited weren’t only small business owners. It was product managers and marketers. A PM can describe a feature and have a clickable HTML prototype to show the team in a few minutes, instead of a long doc. A marketer can spin up three versions of a landing page to test a hypothesis this afternoon, without waiting for a dev sprint.</p><p>I know a lot of you will say “I can already do this with Claude or Cursor”. You can, and that’s great if you’re technical. Ploof is for the person sitting next to you who isn’t, so you can hand it to your marketer or PM and stop being the bottleneck.</p><p>It just launched on Setapp. The trial includes AI credits, so there’s nothing to configure and no API keys to hunt for. It’s not perfect: we’ve already shipped a fix release after early feedback, and there’s a lot more to do.</p><p>I’d really appreciate honest feedback, especially from PMs and marketers: what would you build first, and what’s missing?</p><p>Link: <a href="https://studio.ploof.dev/">https://studio.ploof.dev/</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f5224d963f17" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Unifying AI skills across Cursor and Claude Code]]></title>
            <link>https://yozhef.medium.com/unifying-ai-skills-across-cursor-and-claude-code-3c34c44eafd2?source=rss-c2c049b6f672------2</link>
            <guid isPermaLink="false">https://medium.com/p/3c34c44eafd2</guid>
            <category><![CDATA[software-architecture]]></category>
            <category><![CDATA[cursor]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[developer-productivity]]></category>
            <category><![CDATA[claude-code]]></category>
            <dc:creator><![CDATA[Yozhef Hisem]]></dc:creator>
            <pubDate>Fri, 06 Mar 2026 06:01:05 GMT</pubDate>
            <atom:updated>2026-03-06T06:01:05.146Z</atom:updated>
            <content:encoded><![CDATA[<p>In a team environment, it’s unrealistic to standardize on a single AI coding tool. Some engineers prefer Cursor, others work in Claude Code. Forcing one choice usually reduces productivity rather than improving it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xT0cTFieI6ZQVNBggwKZmA.png" /></figure><p>The real problem appears when people switch tools.</p><p>In our case, we have a single monorepo containing architecture documentation, a Next.js application, and CLI tooling. We also rely heavily on project-specific AI instructions: conventions for ADRs, OpenAPI documentation standards, repository structure, and common commands.</p><p>Originally, this context was split across tool-specific locations:</p><ul><li>CLAUDE.md</li><li>Cursor project configuration</li><li>.claude/skills/</li><li>.cursor/skills/</li></ul><p>As long as engineers stayed within one tool, everything worked. The moment someone switched, the AI agent lost awareness of our conventions. The knowledge still existed in the repository, but not in a place visible to the new agent.</p><p>Each switch effectively reset AI productivity.</p><p>We did not want to:</p><ul><li>lock the team into a single tool,</li><li>maintain duplicated skill definitions,</li><li>or continuously reconfigure context per IDE.</li></ul><p>Instead, we introduced a single source of truth for AI skills and project context.</p><h3>Architectural principle: one canonical location</h3><p>The core decision was simple: define one canonical folder for skills and make every agent read from it.</p><p>We created a skills/ directory at the repository root. This folder contains all project-specific skills:</p><pre>skills/<br>├── adr-creator/<br>│   └── SKILL.md<br>├── openapi-spec/<br>│   └── SKILL.md</pre><p>Each skill is a subdirectory with a SKILL.md file containing frontmatter (name, description) and structured instructions.</p><p>This directory is the only place where skills are edited. It is versioned like any other part of the codebase.</p><p>We explicitly avoided using .claude/skills/ or .cursor/skills/ as primary locations. Tool-specific directories should not be the source of truth.</p><h3>Making tools converge: symlinks</h3><p>Claude Code expects skills under .claude/skills/.<br>Cursor expects .cursor/skills/.</p><p>Instead of duplicating content, we created symbolic links:</p><pre>.claude/skills  → ../skills<br>.cursor/skills  → ../skills</pre><p>Both tools now resolve to the same physical directory. Updating a skill updates behavior in both agents.</p><p>The symlinks are committed to the repository so that a standard clone on macOS, Linux, or WSL preserves the structure. For Windows environments where symlinks may not be available, we document a one-time manual copy as a fallback.</p><h3>Unifying high-level context</h3><p>Beyond skills, we also maintain high-level project context: repository structure, conventions, common commands.</p><p>We store this in AGENTS.md.</p><p>Claude Code expects a CLAUDE.md file, so instead of maintaining two copies, we created another symlink:</p><pre>CLAUDE.md → AGENTS.md</pre><p>This eliminates divergence between tool-specific documentation files. There is exactly one context document.</p><h3>Compatibility with skills.sh and .agents/skills/</h3><p>The skills.sh CLI and several agents rely on .agents/skills/ as a standard location. By default, npx skills add installs skills into ./skills/, which aligns with our canonical directory.</p><p>To ensure compatibility with agents that only read from .agents/skills/, we added:</p><pre>.agents/skills → ../skills</pre><p>As a result:</p><ul><li>Custom internal skills</li><li>Community skills installed via npx skills add</li><li>Cursor</li><li>Claude Code</li><li>Any agent that respects .agents/skills/</li></ul><p>All operate on the same directory.</p><p>We ignore the rest of .claude/, .cursor/, and .agents/ in .gitignore, but explicitly keep the skills symlinks tracked.</p><h3>Operational workflow</h3><p>Adding a new internal skill:</p><ol><li>Create skills/&lt;skill-name&gt;/SKILL.md</li><li>Commit</li></ol><p>No additional configuration is required.</p><p>If we want to import a community skill:</p><pre>npx skills add owner/repo --skill skill-name</pre><p>The skill is installed into skills/, immediately visible to all agents.</p><h3>Outcome</h3><p>This structure achieves several goals:</p><ul><li>Switching between Cursor and Claude Code does not reset AI context.</li><li>There is a single update point for skills and conventions.</li><li>Onboarding is simplified: cloning the repository is enough.</li><li>The team is not locked into a single AI tool.</li><li>Community skills can be integrated without structural changes.</li></ul><p>From an architectural perspective, the key decision was to treat AI skills as versioned project artifacts rather than tool configuration. Once that shift is made, the rest becomes a matter of directory structure and indirection.</p><p>The pattern is simple: define one canonical location, and make every tool resolve to it.</p><p>The result is stable AI productivity regardless of which coding agent an engineer chooses.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3c34c44eafd2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Early Returns: Writing Cleaner and More Maintainable Code]]></title>
            <link>https://yozhef.medium.com/early-returns-writing-cleaner-and-more-maintainable-code-0e2815880c4c?source=rss-c2c049b6f672------2</link>
            <guid isPermaLink="false">https://medium.com/p/0e2815880c4c</guid>
            <category><![CDATA[php]]></category>
            <category><![CDATA[clear-code]]></category>
            <category><![CDATA[complexity]]></category>
            <category><![CDATA[early-return]]></category>
            <category><![CDATA[junior]]></category>
            <dc:creator><![CDATA[Yozhef Hisem]]></dc:creator>
            <pubDate>Mon, 03 Mar 2025 20:47:55 GMT</pubDate>
            <atom:updated>2025-03-03T20:47:55.354Z</atom:updated>
            <content:encoded><![CDATA[<h3>What Are Early Returns?</h3><p>Early returns in functions or loops are a powerful technique that makes your code cleaner, easier to understand, and simpler to maintain. By exiting a function as soon as an invalid condition is met, you save both execution time and mental effort (and possibly your sanity 😁).</p><p>Let’s compare two approaches:</p><h3>Nested Conditional Approach</h3><pre>function processOrder($order) {<br>    if ($order-&gt;isPaid()) {<br>        if ($order-&gt;hasValidShippingAddress()) {<br>            if ($order-&gt;isInStock()) {<br>                // Order processing logic<br>                return &quot;Order processed successfully.&quot;;<br>            } else {<br>                return &quot;Order cannot be processed: Out of stock.&quot;;<br>            }<br>        } else {<br>            return &quot;Order cannot be processed: Invalid shipping address.&quot;;<br>        }<br>    } else {<br>        return &quot;Order cannot be processed: Payment pending.&quot;;<br>    }<br>}</pre><h3>Early Return Approach</h3><pre>function processOrder($order) {<br>    if (!$order-&gt;isPaid()) {<br>        return &quot;Order cannot be processed: Payment pending.&quot;;<br>    }<br>    <br>    if (!$order-&gt;hasValidShippingAddress()) {<br>        return &quot;Order cannot be processed: Invalid shipping address.&quot;;<br>    }<br>    <br>    if (!$order-&gt;isInStock()) {<br>        return &quot;Order cannot be processed: Out of stock.&quot;;<br>    }<br>    // Order processing logic<br>    return &quot;Order processed successfully.&quot;;<br>}</pre><h3>Why Use Early Returns?</h3><h3>🎯 Simplified Logic</h3><p>Avoiding multiple nested if statements significantly improves code readability and makes it easier to follow the logic.</p><h3>🛑 Reduced Nesting</h3><p>Less indentation means less mental overhead. The code is easier to read and debug.</p><h3>⏩ Faster Execution</h3><p>Early exits allow the function to terminate quickly when an invalid condition is met. This aligns with the <em>Fail Fast</em> principle — popularized by Jim Shore and Martin Fowler — which states that detecting errors early makes code more reliable. If an error occurs, the function stops immediately, preventing unnecessary operations.</p><h3>🛠️ Easier to Modify</h3><p>The fewer branches a function has, the easier it is to make changes or introduce new logic without breaking existing functionality.</p><h3>🧹 Lower Coupling</h3><p>Each piece of logic remains more isolated. This simplifies refactoring and testing since dependencies between different parts of the code are reduced.</p><h3>🔍 Focus on Critical Conditions</h3><p>Early returns emphasize the most critical conditions, making the function’s purpose clearer by eliminating unnecessary distractions.</p><h3>Conclusion</h3><p>Using early returns results in elegant and efficient code, especially in complex systems with multiple conditions and branching logic. Adopting this approach will help you write maintainable, readable, and robust software.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0e2815880c4c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Using Redis cache with Symfony]]></title>
            <link>https://yozhef.medium.com/using-redis-cache-with-symfony-fad1c074c736?source=rss-c2c049b6f672------2</link>
            <guid isPermaLink="false">https://medium.com/p/fad1c074c736</guid>
            <category><![CDATA[symfony-5]]></category>
            <category><![CDATA[symfony]]></category>
            <category><![CDATA[redis]]></category>
            <category><![CDATA[symfony4]]></category>
            <category><![CDATA[symfony-framework]]></category>
            <dc:creator><![CDATA[Yozhef Hisem]]></dc:creator>
            <pubDate>Mon, 22 Feb 2021 16:57:45 GMT</pubDate>
            <atom:updated>2021-02-23T14:55:42.934Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/958/1*AROhf77lw66l4NF5XqFLZA.jpeg" /></figure><h3>History</h3><p>Redis was created in 2009 by Salvatore Sanfilippo, who still remains a lead developer of the project. Redis is sometimes described as “Memcached on steroids,” which is hardly surprising considering that parts of Redis were built in response to lessons learned from using Memcached. Redis has more features than Memcached and is,more powerful and flexible.</p><h3>How to configure Redis</h3><p>Create a blank <a href="https://symfony.com/doc/current/setup.html">Symfony project</a>:</p><pre><em># run this if you are building a traditional web application</em><br>composer create-project symfony/website-skeleton my_project_name<br><br><em># building a microservice, console application or API</em><br>composer create-project symfony/skeleton my_project_name</pre><p>Add <a href="https://github.com/symfony-bundles/redis-bundle">Redis-Bundle</a>:</p><pre>composer require symfony-bundles/redis-bundle</pre><p>After installing you see in your AppKernel class adding the Symfony Bundle Redis Bundle class in the bundle’s array, and create a new config file</p><pre><em>#projectName/config/bundles.php</em><br>SymfonyBundles\RedisBundle\SymfonyBundlesRedisBundle::class =&gt; [&#39;all&#39; =&gt; true],</pre><pre><em>#projectName/config/packages/sb_redis.yaml:</em><br>sb_redis:<br>    clients:<br>        default:<br>            $options:    []<br>            $parameters: [&#39;tcp://127.0.0.1:6379?database=3&#39;]</pre><h3>How to use</h3><p>Let’s look at a simple example of creating a RedisStorageManager</p><pre>&lt;?php<br><br>declare(strict_types=1);<br><br>namespace App\Infrastructure\Service\Redis;<br><br>use SymfonyBundles\RedisBundle\Redis\ClientInterface;<br><br>class RedisStorageManager<br>{<br>    private ClientInterface $redis;<br><br>    public function __construct(ClientInterface $redis)<br>    {<br>        $this-&gt;redis = $redis;<br>    }<br>}</pre><p>Now, let’s imagine that we need to create a function that is looking for something</p><pre>public function find(string $key): ?string<br>{<br>    return $this-&gt;redis-&gt;get($this-&gt;getRedisKey($key));<br>}</pre><p>You may be wondering what this new getRedisKey method is. It is a key that allows us to reduce the possibility of crossing similar keys for different user flows.</p><pre>private function getRedisKey(string $key): string<br>{<br>    return sprintf(&#39;%s.%s&#39;, $this-&gt;prefix, $key);<br>}</pre><p>Next, we’ll try to save something in Redis. In this case, automatic cleaning of the so-called <a href="https://redis.io/commands/ttl">lifetime</a> can be very helpful</p><pre>public function set(string $key, string $value): void<br>    {<br>        $key = $this-&gt;getRedisKey($key);<br><br>        $this-&gt;redis-&gt;set($key, $value);<br>        $this-&gt;redis-&gt;expire($key, $this-&gt;ttl);<br>    }</pre><p>If we consider cases (search, store, delete), then our Redis Storage Manager -&gt; will increase the functionality that will easily allow us to manage data</p><pre>&lt;?php<br><br>declare(strict_types=1);<br><br>namespace App\Infrastructure\Service\Redis;<br><br>use SymfonyBundles\RedisBundle\Redis\ClientInterface;<br><br>class RedisStorageManager<br>{<br>    private string $prefix;<br>    private int $ttl;<br>    private ClientInterface $redis;<br><br>    public function __construct(string $prefix, int $ttl, ClientInterface $redis)<br>    {<br>        $this-&gt;prefix = $prefix;<br>        $this-&gt;ttl = $ttl;<br>        $this-&gt;redis = $redis;<br>    }<br><br>    public function find(string $key): ?string<br>    {<br>        return $this-&gt;redis-&gt;get($this-&gt;getRedisKey($key));<br>    }<br><br>    public function findArrayValue(string $key): ?array<br>    {<br>        return $this-&gt;redis-&gt;hgetall($this-&gt;getRedisKey($key));<br>    }<br><br>    public function set(string $key, string $value): void<br>    {<br>        $key = $this-&gt;getRedisKey($key);<br><br>        $this-&gt;redis-&gt;set($key, $value);<br>        $this-&gt;redis-&gt;expire($key, $this-&gt;ttl);<br>    }<br><br>    public function setArray(string $key, array $value): void<br>    {<br>        $key = $this-&gt;getRedisKey($key);<br><br>        $this-&gt;redis-&gt;hmset($key, $value);<br>        $this-&gt;redis-&gt;expire($key, $this-&gt;ttl);<br>    }<br><br>    public function remove(string $key): void<br>    {<br>        $this-&gt;redis-&gt;remove($this-&gt;getRedisKey($key));<br>    }<br><br>    public function has(string $key): bool<br>    {<br>        return (bool) $this-&gt;find($this-&gt;getRedisKey($key));<br>    }<br><br>    private function getRedisKey(string $key): string<br>    {<br>        return sprintf(&#39;%s.%s&#39;, $this-&gt;prefix, $key);<br>    }<br>}</pre><p>That’s it! You can use the Redis Store anywhere in your project and manipulate the data easily.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fad1c074c736" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>