<?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 Joyz on Medium]]></title>
        <description><![CDATA[Stories by Joyz on Medium]]></description>
        <link>https://medium.com/@joyzoursky?source=rss-57d6cffeab2a------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*ERvLnhqXPVVgTCEGxcS2Tg.png</url>
            <title>Stories by Joyz on Medium</title>
            <link>https://medium.com/@joyzoursky?source=rss-57d6cffeab2a------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 15:19:20 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@joyzoursky/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[From Prompt to Video: Why Remotion Skills Suddenly Matter]]></title>
            <link>https://medium.com/@joyzoursky/from-prompt-to-video-why-remotion-skills-suddenly-matter-1927307dcf1f?source=rss-57d6cffeab2a------2</link>
            <guid isPermaLink="false">https://medium.com/p/1927307dcf1f</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[video-production]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[product-marketing]]></category>
            <dc:creator><![CDATA[Joyz]]></dc:creator>
            <pubDate>Sat, 24 Jan 2026 19:00:28 GMT</pubDate>
            <atom:updated>2026-01-24T19:05:35.515Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ICVtEj3ECQ4JI30uvcZ1ug.png" /><figcaption>Remotion</figcaption></figure><p>Last week, <strong>Remotion Skills</strong> started showing up everywhere.</p><p>That alone was interesting. But what really caught my attention was <em>why</em>.</p><p><a href="https://www.remotion.dev/">Remotion</a> itself isn’t new. It has existed for years as a code based video library, mostly used by developers who were comfortable treating video like React. Powerful, but niche.</p><p>The <strong>Skills</strong> release changes that.</p><p>The official Remotion team recently introduced Remotion Skills for AI agents, and that shifts Remotion from “code to video” into something more approachable.</p><p><a href="https://www.remotion.dev/docs/ai/skills">https://www.remotion.dev/docs/ai/skills</a></p><p>Agents can now reason about scenes, timing, and structure, and turn prompts directly into working videos.</p><p>That is what makes this moment interesting. Not as a developer toy, but as a practical way to create animation videos faster, cheaper, and with far less friction than traditional production workflows.</p><h3>What Remotion Skills Actually Unlock</h3><p>Before Skills, using Remotion meant you needed to think in components, frames, and composition from day one. That was fine if you already lived in code, but it created a high starting cost for anyone else.</p><p>With Remotion Skills, the starting point shifts.</p><p>Instead of asking “how do I build this animation,” you can start with “what should this video communicate.” An AI agent can draft a storyboard, define scenes, and generate a working baseline video that you can run immediately.</p><p>This doesn’t replace design or motion thinking. It removes the blank canvas problem.</p><h3>The Prompt to Video Workflow</h3><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fplayer.vimeo.com%2Fvideo%2F1157963591%3Fapp_id%3D122963&amp;dntp=1&amp;display_name=Vimeo&amp;url=http%3A%2F%2Fvimeo.com%2F1157963591&amp;image=https%3A%2F%2Fi.vimeocdn.com%2Fvideo%2F2111839777-ccb9ed7574325489139a6dc515ab78709b8b669ff09c07ca3de26e2b351165ad-d_960%3Fregion%3Dus&amp;type=text%2Fhtml&amp;schema=vimeo" width="1080" height="1080" frameborder="0" scrolling="no"><a href="https://medium.com/media/b642e61c32fee8ad3466aef3040991d5/href">https://medium.com/media/b642e61c32fee8ad3466aef3040991d5/href</a></iframe><p>Here’s the workflow I ended up using, and the one that made everything click.</p><h3>1. Choose an AI coding agent</h3><p>I used a mix of Claude Code, Codex, and Antigravity. The exact agent matters less than the approach. Treat the agent as a collaborator that proposes structure, not a button that magically produces perfection.</p><h3>2. Add Remotion Skills</h3><p>This is the key piece. Remotion Skills gives the agent a mental model of how Remotion works. Scenes, timing, and composition become things the agent can reason about instead of guess.</p><h3>3. First prompt: storyboard and structure</h3><p>The first prompt is not about visuals. It is about structure.</p><p>I asked the agent to draft:</p><blockquote>Scenes and their order<br>Approximate duration per scene<br>Key messages or visuals in each scene</blockquote><p>This step is cheap and fast, and it sets direction without locking anything in.</p><h3>4. Second prompt: generate the video</h3><p>Next, I asked the agent to generate an actual Remotion video using Skills, with clear constraints.</p><blockquote>Brand colors<br>Fonts<br>Design tone</blockquote><p>At this point, the video is no longer an idea. It runs.</p><h3>5. Iterate and commit</h3><p>This is where the workflow starts to feel very different from traditional video production.</p><p>I tweaked values, re-ran generations, and committed good states to Git. No fear of breaking things. No endless manual edits. Just iteration.</p><h3>6. Add music at the end</h3><p>I added music separately using <a href="https://suno.com/">Suno</a>. Keeping audio out of the early iterations made visual changes much faster.</p><p>From zero to a solid first version took around two hours. The first generation took the longest, which was perfect time to grab a cup of coffee and let it run☕️</p><h3>Why This Matters for Product, Branding, and Marketing Teams</h3><p>Animation videos are incredibly useful. Landing pages, feature launches, onboarding, social clips.</p><p>The problem has never been value. The problem is cost and speed.</p><p>Traditional workflows make small changes expensive. A wording tweak or timing adjustment often means another round of coordination or re rendering.</p><p>With Remotion Skills, iteration becomes cheap.</p><p>That changes behavior. Teams are more willing to try animation when it is no longer a one way door.</p><h3>The Small Tweaks That Make the Difference</h3><p>Once a first version exists, quality comes from small, deliberate changes. These are the adjustments I reach for most often.</p><h4>Adjusting scene length</h4><pre>const fps = 30;<br><br>&lt;Sequence from={0} durationInFrames={fps * 3}&gt;<br>  &lt;IntroScene /&gt;<br>&lt;/Sequence&gt;<br>&lt;Sequence from={fps * 3} durationInFrames={fps * 5}&gt;<br>  &lt;FeatureScene /&gt;<br>&lt;/Sequence&gt;</pre><p>Thinking in seconds and frames makes timing explicit and easy to adjust.</p><h4>Keeping brand colors consistent</h4><pre>export const COLORS = {<br>  primary: &#39;#1E90FF&#39;,<br>  background: &#39;#0F172A&#39;,<br>  text: &#39;#FFFFFF&#39;,<br>};</pre><p>Once colors live in code, consistency stops being a manual task.</p><h4>Typography control</h4><pre>&lt;div<br>  style={{<br>    fontFamily: &#39;Inter, sans-serif&#39;,<br>    fontSize: 64,<br>    fontWeight: 700,<br>    color: COLORS.text,<br>  }}<br>&gt;<br>  Prompt to Video<br>&lt;/div&gt;</pre><p>Everything is explicit and easy to change.</p><h3>Final Thoughts</h3><p>Remotion Skills does not magically remove the need for taste or storytelling.</p><p>What it removes is friction.</p><p>For teams that need animation videos but cannot afford slow cycles or high costs, that matters. A lot.</p><p>This feels like the moment Remotion stops being a niche developer tool and starts becoming a practical option for product, branding, and marketing teams who want to move faster.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1927307dcf1f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Modern Approach to AI-Assisted End-to-End Validation]]></title>
            <link>https://medium.com/@joyzoursky/the-modern-approach-to-ai-assisted-end-to-end-validation-a9cde59699f7?source=rss-57d6cffeab2a------2</link>
            <guid isPermaLink="false">https://medium.com/p/a9cde59699f7</guid>
            <category><![CDATA[software-testing]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[test-automation]]></category>
            <category><![CDATA[product-development]]></category>
            <dc:creator><![CDATA[Joyz]]></dc:creator>
            <pubDate>Mon, 19 Jan 2026 02:49:33 GMT</pubDate>
            <atom:updated>2026-01-19T02:49:33.144Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TFjZiT5p26PkLL4OHBHjqQ.png" /><figcaption>3 Practical AI Testing Approaches for 2026</figcaption></figure><p>Software teams today don’t struggle because they lack tools.<br>They struggle because <strong>systems are moving faster than certainty</strong>.</p><p>New features ship daily. UI changes constantly. Dependencies update silently. And somewhere between <em>“it works on my machine”</em> and <em>“users are complaining”</em>, reality breaks.</p><p>At <a href="https://www.rconnect.tech/blog/ai-devtools-taipei"><strong>AI DevTools Taipei</strong></a>, I shared how teams are starting to use AI — not to <em>test more cases</em>, but to <strong>validate real user flows faster and more realistically</strong>.</p><blockquote>This is not about QA as a function.<br>This is about confidence in real user behavior.</blockquote><h3>From “Testing” to Validating Reality</h3><p>Traditionally, end-to-end work has been framed as <em>testing</em>:</p><ul><li>Write scripts</li><li>Assert expected outputs</li><li>Maintain selectors</li><li>Rerun endlessly</li></ul><p>That still matters — but it no longer covers the full problem space.</p><p>What teams increasingly need is the ability to ask:</p><ul><li><em>Can a real user still complete this flow?</em></li><li><em>Does the UI still communicate the right information?</em></li><li><em>Did a change silently break something obvious?</em></li></ul><p>AI changes <strong>how</strong> we answer these questions with the help of Tools.</p><h3>Three Practical AI Testing Approaches</h3><p>Based on real usage (not demos), AI-assisted validation now falls into <strong>three complementary approaches</strong>.</p><h3>1. AI Testing Frameworks</h3><p><em>(Structured, repeatable validation)</em></p><p><strong>Examples: </strong><a href="https://midscenejs.com/">Midscene</a>, <a href="https://www.stagehand.dev/">Stagehand</a>, <a href="https://github.com/VibiumDev/vibium">Vibium</a></p><p>These tools sit closest to traditional automation, but replace brittle selectors with <strong>visual and language-based understanding</strong>.</p><p>Instead of scripting <em>how</em> to click something, you describe <em>what should happen</em>:</p><ul><li>Log in</li><li>Add item to cart</li><li>Verify total price</li></ul><p>The AI agent interprets screenshots and UI context to execute and validate the flow.</p><p><strong>Best for:</strong></p><ul><li>Repeatable user journeys</li><li>Regression confidence</li><li>Cross-functional visibility (Non-tech PMs can read results)</li></ul><p><strong>Trade-off:</strong></p><ul><li>Slower than pure code-based tests</li><li>Still requires intent and structure</li></ul><p>This approach works well when <strong>certainty matters more than speed</strong>.</p><h3>2. MCP Tools (Model Context Protocol)</h3><p><em>(Exploration, planning, understanding)</em></p><p><strong>Examples: </strong><a href="https://developer.chrome.com/blog/chrome-devtools-mcp">Chrome DevTools MCP</a>, <a href="https://github.com/microsoft/playwright-mcp">Playwright MCP</a>, <a href="https://docs.maestro.dev/getting-started/maestro-mcp">Maestro MCP</a></p><p>MCP tools allow AI to <em>interact</em> with browsers and applications directly, using models as an execution layer.</p><p>In practice, teams use MCP tools to:</p><ul><li>Explore unfamiliar systems</li><li>Generate draft user flows</li><li>Assist with test planning</li><li>Validate assumptions quickly</li></ul><p>They are powerful, but beware of your <em>context tokens usage</em>!</p><p><strong>Best for:</strong></p><ul><li>Discovery and exploration</li><li>Early-stage validation</li><li>Understanding “what should be tested”</li></ul><blockquote><strong>Important limitation:</strong><br>MCP tools are <strong>not reliable enough for final verification</strong>.<br>They may report success where subtle failures exist.</blockquote><p>Think of MCP as <strong>a thinking assistant</strong>, not a judge.</p><h3>3. Browser Agents</h3><p><em>(Action-oriented, goal-driven validation)</em></p><p><strong>Examples: </strong><a href="https://chromewebstore.google.com/detail/antigravity-browser-exten/eeijfnjmjelapkebgockoeaadonbchdd?pli=1">Antigravity with Chrome</a>, <a href="https://cursor.com/docs/agent/browser">Cursor Browser</a>, <a href="https://claude.com/chrome">Claude in Chrome</a></p><p>A newer category focuses on AI systems that control browsers end-to-end to accomplish goals.</p><p>This includes:</p><ul><li>Agent-based browser control</li><li>Goal-driven execution</li><li>Tools built around AI-controlled browser interaction</li></ul><p>These agents don’t follow scripts.<br>They <strong>reason</strong>, adapt, retry, and navigate like a human would.</p><p>Example tasks:</p><ul><li>“Find the cheapest suitable product on this site”</li><li>“Complete onboarding and report where it fails”</li><li>“Check if checkout still works after deployment”</li></ul><p><strong>Best for:</strong></p><ul><li>One-off validation</li><li>Operational checks</li><li>Complex, non-linear flows</li></ul><p><strong>Trade-off:</strong></p><ul><li>Higher cost</li><li>Lower repeatability</li><li>Harder to debug</li></ul><p>Browser agents are most useful when <strong>the path is unclear</strong>, but the outcome matters.</p><blockquote><strong>The important mindset shift:</strong><br>MCP tools and agent browsers are powerful, but they are <strong>not “free clicks.”</strong><br>Treat them like you would treat expensive integration tests: use them intentionally.</blockquote><h3>What This Means for Product Teams</h3><p>For business teams, this shift matters because:</p><ul><li>Confidence no longer depends solely on scripts</li><li>Validation can happen earlier, even when requirements are still evolving</li><li>Results are easier to understand across roles</li><li>Real user behavior becomes the reference point</li></ul><p>For developers, it means:</p><ul><li>Fewer brittle assumptions about UI and structure</li><li>Better support for exploratory and edge-case validation</li><li>Less time spent maintaining glue code for scenarios that rarely repeat</li></ul><p>The goal is not to automate everything.<br>The goal is to validate what actually matters, with less friction.</p><h4>A Note on Responsibility</h4><p>AI does not remove responsibility.</p><ul><li>Remember your goals to ensure quality</li><li>Avoid over-automation in the hype</li><li>Adopt for delegation, not to add more work</li><li>Test as a Real User</li></ul><p>Your product users are humans, not AI. Manual testing remains essential.</p><h3>Watch the Full Talk</h3><p>This article is based on my talk at <strong>AI DevTools Taipei</strong>, where I shared real demos, trade-offs, and lessons from production use.</p><p>🎥 <strong>Full talk on YouTube:</strong><br><a href="https://www.youtube.com/watch?v=vJlp8X1tzeY">https://www.youtube.com/watch?v=vJlp8X1tzeY</a></p><p>This is also the foundation for some internal tools and experiments we will be sharing publicly soon, built to support this exact philosophy.</p><p>Stay tuned!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a9cde59699f7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[From “IDK How” to Pro: Riding the 7 Levels of AI Vibe Coding]]></title>
            <link>https://medium.com/@joyzoursky/from-idk-how-to-pro-riding-the-7-levels-of-ai-vibe-coding-684a9e3465e8?source=rss-57d6cffeab2a------2</link>
            <guid isPermaLink="false">https://medium.com/p/684a9e3465e8</guid>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[code-generation]]></category>
            <category><![CDATA[vibe-coding]]></category>
            <category><![CDATA[ai-coding]]></category>
            <dc:creator><![CDATA[Joyz]]></dc:creator>
            <pubDate>Sat, 12 Apr 2025 06:48:17 GMT</pubDate>
            <atom:updated>2025-04-12T06:51:37.217Z</atom:updated>
            <content:encoded><![CDATA[<p>As someone who spends their days thinking about software quality (but also loves sleek design and getting things done <em>fast</em>), I’ve been watching the rise of what we’re calling <strong>“Vibe Coding”</strong>.</p><p>You know the scene: tools like <a href="https://windsurf.com/">Windsurf</a> (which I personally like the most), <a href="https://www.cursor.com/">Cursor</a>, <a href="https://code.visualstudio.com/">VS Code</a> with <a href="https://github.com/features/copilot">GitHub Copilot</a>, <a href="https://www.trae.ai/">Trae</a>, <a href="https://firebase.studio/">Firebase Studio</a>, connecting all kinds of <a href="https://docs.anthropic.com/en/docs/agents-and-tools/mcp">MCP</a> servers to further power-up. Or maybe even prompt-based generators like <a href="https://v0.dev/">v0</a>, <a href="https://bolt.new/">Bolt</a>, <a href="https://lovable.dev/">Lovable</a>, <a href="https://replit.com/">Replit</a>, or <a href="https://www.canva.com/ai-code-generator/">Canva AI</a>, cranking out code and even whole app skeletons. It’s fast, it’s flashy, and honestly, it feels pretty cool.</p><p>But let’s be real, especially for our junior devs (and maybe some seniors learning new tricks!), this vibe coding journey often follows a familiar pattern. I call it the <strong>7 Levels of the Vibe Coding Grind</strong>:</p><h3><strong>Lv 1: The Quick Change Artist</strong></h3><p>“Need this button blue? Copilot, make it so!” Code changes, PR ships. How does it <em>actually</em> work under the hood? ¯\_(ツ)_/¯</p><p>But hey, it worked! Speed: 10. Understanding: 1.</p><p>So what did it do? IDK.</p><h3><strong>Lv 2: The Instant Website Wizard</strong></h3><p>Prompt: “Build me a landing page for artisanal dog biscuits.” <em>Poof!</em> A website appears via v0 or similar. Looks decent.</p><p>How does the routing work? State management? CSS framework? Who knows! It’s live, right?</p><p>So how to update the website? IDK.</p><h3><strong>Lv 3: The Doc Feeder</strong></h3><p>Stuck on a complex API integration? Feed the docs to the AI.</p><p>Doc in, Code out. It spits out code and it feels like you wire it up.</p><p>It… works? Mostly? How? The AI absorbed the ancient scrolls, performed some digital ritual, and blessed you with functioning code. Don’t question the magic…for now.</p><h3><strong>Lv 4: The Guided Builder</strong></h3><p>Okay, now you’re using AI step-by-step. “Design and plan the implementation with the requirements below. Breakdown your plan into step-by-step tasks.”</p><p>“Generate a user model,” “Create the login function,” “Add tests for this…”</p><p>You start seeing <em>some</em> connections. But as the project grows, generated bits bolted onto other generated bits start looking… funky. Maintainability? We’ll worry about that later.</p><h3><strong>Lv 5: The Modular Architect (Almost)</strong></h3><p>You’re getting smarter! Yes, YOU, not the AI.</p><p>You ask the AI to build things in modules. Looks clean! Components here, services there.</p><p>But when it’s time to refactor, debug a deep issue, or <em>really</em> scale it? That “IDK how it works” feeling creeps back in. Understanding the <em>glue</em> between the modules is still hazy.</p><h3><strong>Lv 6: The Rewriter</strong></h3><p>This is the turning point. You hit a high wall.</p><p>The AI’s code, while functional, is hard to maintain or has hidden issues.</p><p>You sigh, grab coffee, <em>actually read the framework/library docs</em>, understand the core concepts, grab the huge chunks of AI code part-by-part, script-by-script, and rewrite (or copy &amp; paste, reorganize) everything by your own hands, applying proper structure and patterns.</p><p>It feels… good, real, not a toy anymore, not a university FYP, not a one-day prototype that only works for one day.</p><p>Slower, but <em>solid</em>.</p><h3><strong>Lv 7: The Enlightened Vibecoder</strong></h3><p>You’re back at Lv 1, needing a quick change.</p><p>But <em>this time</em>, when the AI suggests code, you instantly know <em>why</em> it’s suggesting that. You can evaluate it, tweak it, or reject it knowledgeably. You use the tools as accelerators, not crutches.</p><p>You <em>know</em> how it works now.</p><h3><strong>A QA Manager’s Quick Reality Check on Vibe Coding</strong></h3><p>Look, these AI tools are awesome productivity boosters. Embrace the vibe! But keep these in mind:</p><ol><li><strong>Security Blindspots:</strong> AI doesn’t always write secure code. It learns from vast datasets, including potentially vulnerable patterns. <em>Always</em> review generated code, especially for inputs, auth, and data handling. Check out the <em>OWASP Top 10 For GenAI </em>👉 <a href="https://genai.owasp.org/">https://genai.owasp.org/</a></li><li><strong>Tech Debt Traps:</strong> Fast code isn’t always good code. AI might generate repetitive, inefficient, or hard-to-read code. This builds tech debt faster than you can say “refactor” (and also burns your AI credits or tokens as fast as the ugly code grows!!). Be mindful of maintainability from the start.</li><li><strong>The “IDK” Danger Zone:</strong> Relying too heavily on AI without understanding the underlying principles hinders your growth. Debugging becomes a nightmare if you don’t grasp the fundamentals.</li><li><strong>Ownership:</strong> Remember, <em>you</em> are ultimately responsible for the code you commit, not the AI. Test thoroughly, understand the implications.</li></ol><h4><strong>The Takeaway</strong></h4><p>Vibe coding is fun, fast, and here to stay. Use these amazing tools to accelerate your workflow and learn quicker. <strong>But don’t skip Level 6.</strong> The real level-up happens when you combine the AI’s speed with your own hard-earned understanding.</p><p>Keep coding, keep learning, and keep that quality bar high! 😉</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2lTXuLo1wAsOlUuXmF77bA.jpeg" /><figcaption>Let’s Grind!</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=684a9e3465e8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[4 Essential AI Testing Prompts to Streamline Your QA Process]]></title>
            <link>https://medium.com/oursky-team/4-essential-ai-testing-prompts-to-streamline-your-qa-process-b185b7763717?source=rss-57d6cffeab2a------2</link>
            <guid isPermaLink="false">https://medium.com/p/b185b7763717</guid>
            <category><![CDATA[prompt]]></category>
            <category><![CDATA[qa]]></category>
            <category><![CDATA[software-testing]]></category>
            <category><![CDATA[project-management]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Joyz]]></dc:creator>
            <pubDate>Fri, 08 Nov 2024 08:12:24 GMT</pubDate>
            <atom:updated>2024-11-08T08:12:24.299Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*akZhaFU1fj8ldhIz9PvEtw.jpeg" /></figure><p>In today’s fast-paced software development landscape, efficient testing isn’t just a technical necessity — it’s a business imperative. Whether you’re a business owner looking to cut costs, a project manager seeking higher efficiency, or a QA professional aiming to streamline your workflow, AI-powered testing prompts can transform your quality assurance process.</p><h3>Why This Matters to Your Bottom Line</h3><ul><li>🎯 <strong>Cost Reduction</strong>: Automate test case generation and reduce manual planning time</li><li>⚡ <strong>Increased Efficiency</strong>: Generate comprehensive test scenarios in minutes, not hours</li><li>📊 <strong>Better Coverage</strong>: Ensure thorough testing with AI-generated edge cases</li><li>🔄 <strong>Consistency</strong>: Standardize testing approaches across teams and projects</li></ul><h3>⚠️ Security Advisory: Protecting Your Test Information</h3><p>Before diving into the prompts, an important note on security:</p><ul><li><strong>Keep Information Confidential</strong>: Testing prompts often contain sensitive information about your systems, business logic, and potential vulnerabilities</li><li><strong>Use Trusted Sources</strong>: Only share testing information with trusted, secure LLM providers or use self-hosted models</li><li><strong>Sanitize Data</strong>: Remove sensitive data, credentials, and internal identifiers from prompts</li><li><strong>Consider Self-Hosting</strong>: For maximum security, consider using self-hosted models for test case generation, such as <a href="https://ollama.com/">Ollama</a>.</li></ul><p>Remember: Your test cases are a blueprint of your system. Treat them with appropriate security measures.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2Ctu0Uvt0NHZW9SpTAzfLg.png" /><figcaption>Self-hosted Local AI Chatbot</figcaption></figure><p>Now let’s dive into 4 essential testing scenarios and see how AI prompts can assist your approach.</p><h3>1: Post-Sprint Exploratory Testing</h3><p>For finding bugs on new features and changes after a sprint.</p><h4>The Prompt:</h4><blockquote><em>This is the updates done for the project [Project Name] in the past sprint, please summarize the updates, and list the corresponding test cases to test the updated functions manually. Print it in a document.</em></blockquote><p>Use the prompt with uploading a CSV of issues that you have completed in the sprint, including the title and description of each issue.</p><h4>Quick Tips:</h4><ul><li>Provide all updated features as context.</li><li>Compare results from different LLMs.</li></ul><h3>2: Pre-Release Smoke Testing</h3><p>For quick validation of critical functionalities before release.</p><h4>The Prompt:</h4><blockquote><em>You are a product owner going to release a new version of [Project Name]. Generate a set of Smoke Test cases for manual QA to ensure these main flows work without any critical issues. Categorize the test cases and divide into sections. Make a checklist for each section with columns ID, Scenario, Expected behaviour, Pass/Fail and Remarks. Keep it short and clean.</em></blockquote><blockquote><em>Core functionalities:<br>[List your core features here]</em></blockquote><p><strong>Quick Tips:</strong></p><ul><li>List all core features.</li><li>Keep the list in 20 items, dont’ be too long.</li></ul><h3>3: Complex Feature Testing</h3><p>For thorough testing of features with complex business rules, requires a high coverage and deepness.</p><h4>The Prompt:</h4><blockquote><em>You are a manual QA tester. Given the user flow and rules of the system, create a set of comprehensive functional test cases to test each step in the user flow, covering positive test scenarios, negative test scenarios, edge cases and additional cases.</em></blockquote><blockquote><em>Categorize the test cases and divide into sections. Each section shows a table format with a meaningful “ID”, “Test Case”, “Sample Test Data” and “Expected Behaviour” columns. Keep the sentences short and clean.</em></blockquote><blockquote><em>Happy flow:<br>[List your user flow steps]</em></blockquote><blockquote><em>Rules:<br>[List your business rules]</em></blockquote><p><strong>Quick Tips:</strong></p><ul><li>Include a complete happy flow, but start from the feature’s entry point (not from system login).</li><li>List all validation requirements and business rules.</li><li>Break down large features into smaller, focused test suites for better depth.</li><li>Try multiple prompt attempts and combine the best test cases for comprehensive coverage.</li></ul><h3>4: Post-Bug-Fix Regression Testing</h3><p>For ensuring bug fixes don’t break related functionalities.</p><h4>The Prompt:</h4><blockquote><em>You are a manual QA tester. Below is a bug reported. The bug is already fixed by developer. Please suggest a list of regression test cases to check if any related area is affected after the fix. Return your results in a checklist table with columns “ID”, “Aspect”, “Description”, “Expected Result”, “Priority”.</em></blockquote><blockquote><em>Original bug report:<br>[Include full bug report with environment details and steps to reproduce]</em></blockquote><p><strong>Quick Tips:</strong></p><ul><li>Include the complete bug report with title, environment details and testing platforms, description, steps to reproduce and expected behaviours.</li><li>Focus on the issues one by one, don’t retest multiple issues at a time.</li><li>Use <a href="https://issuesnap.oursky.app/">IssueSnap</a> to help you create a clear and informative bug report.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*BZ4kFvt3q0vXT3MebSteww.png" /><figcaption>IssueSnap — Your Bug Report Helper</figcaption></figure><h3>Common Pitfalls to Avoid</h3><ol><li><strong>Too Vague</strong>: “Test the login feature” vs. “Test the login feature including password requirements, error handling, and session management”</li><li><strong>Missing Context</strong>: Always include relevant business rules and technical constraints</li><li><strong>Unclear Priorities</strong>: Specify which scenarios are most critical</li><li><strong>Incomplete Information</strong>: Include all necessary details about the feature or bug</li></ol><p>Remember: The quality of your test cases directly depends on the quality of your prompts. Take time to craft detailed, specific prompts that clearly communicate your testing needs and constraints.</p><p><em>At the heart of Oursky is the art of software engineering: translating ideas into opportunities. We work with startups in bringing their ideas to life with the right blend of people, process, and technology — from MVP development to scalable digital products powered by the latest technologies. </em><a href="https://oursky.com/contact"><em>Schedule a no-commitment consultation with us</em></a><em> and we can explore how to creatively harness the digital space to solve real-world problems.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b185b7763717" width="1" height="1" alt=""><hr><p><a href="https://medium.com/oursky-team/4-essential-ai-testing-prompts-to-streamline-your-qa-process-b185b7763717">4 Essential AI Testing Prompts to Streamline Your QA Process</a> was originally published in <a href="https://medium.com/oursky-team">Oursky Team</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The New Era of E2E Test Automation: Faster, Cheaper, and More Accessible]]></title>
            <link>https://medium.com/oursky-team/the-new-era-of-e2e-test-automation-faster-cheaper-and-more-accessible-90b0e8595171?source=rss-57d6cffeab2a------2</link>
            <guid isPermaLink="false">https://medium.com/p/90b0e8595171</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[quality-assurance]]></category>
            <category><![CDATA[test-automation]]></category>
            <category><![CDATA[ai-tools]]></category>
            <category><![CDATA[software-testing]]></category>
            <dc:creator><![CDATA[Joyz]]></dc:creator>
            <pubDate>Fri, 27 Sep 2024 08:56:18 GMT</pubDate>
            <atom:updated>2024-09-27T08:56:18.104Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Vgozzm5TpqNgrMwKaGrhaA.png" /></figure><p>End-to-end (E2E) test automation has long been a critical component of software quality assurance. Traditionally, when comparing automated testing with manual testing, factors such as coverage, limitations, and cost have been the primary focus. However, the landscape is changing rapidly, and it’s time to reassess the value proposition of E2E test automation.</p><h3>Overcoming the Cost Hurdle</h3><p>One of the main hesitations in adopting E2E test automation has been the perceived high cost. Many organizations believe that implementing robust automated testing requires significant investment in tools, infrastructure, and skilled personnel. This perception has led to a misconception that E2E test automation is only for <strong>“rich product owners”</strong>. But what if we told you that this is no longer the case?</p><h3>Enter Cursor: A Game-Changer in AI Code Editing</h3><p><a href="https://www.cursor.com/">Cursor</a> is one of the tools that changes the way we approach coding and test automation. As an AI-powered code editor, it offers significant advantages over traditional tools like <a href="https://code.visualstudio.com/">Visual Studio Code</a> and even surpasses the capabilities of <a href="https://github.com/features/copilot">GitHub Copilot</a> in some areas like its powerful autocomplete. Cursor’s intelligent code suggestions and automated refactoring capabilities make it a competitive handy asset for test automation engineers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tkNRh989QIPx5N9jY8LD5A.png" /><figcaption>Cursor: The AI Code Editor</figcaption></figure><h3>Choosing Your Test Framework and Language</h3><p>Before diving into the practical aspects, it’s essential to select the right test framework and language for your project. For web applications, popular choices include:</p><ul><li><a href="https://playwright.dev/">Playwright</a></li><li><a href="https://www.selenium.dev/">Selenium</a></li><li><a href="https://www.cypress.io/">Cypress</a></li></ul><p>These frameworks support various programming languages, with Python and Node.js being common choices. For mobile applications, <a href="https://appium.io/docs/en/latest/">Appium</a> is a popular option that supports multiple languages as well.</p><h3>Speedrun a Test Script with Cursor</h3><p>Let’s walk through a practical example of how Cursor can help you create an E2E test case using Playwright, <a href="https://docs.pytest.org/en/stable/">Pytest</a>, and <a href="https://allurereport.org/">Allure Report</a> in just 5 minutes:</p><p><strong>Step 1: </strong>Create a virtual environment in Python</p><pre>python3 -m venv myenv<br>source myenv/bin/activate</pre><p><strong>Step 2: </strong>Use Cursor to guide you through installing required packages</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bZMq1A6Uk-eiG4Vrwr2s-A.png" /><figcaption>Ask AI: How to install required packages for playwright pytest with allure report?</figcaption></figure><p><strong>Step 3: </strong>Generate raw Playwright script by Playwright Inspector using playwright codegen (different test frameworks have their own code generator or inspector)</p><p><strong>Step 4: </strong>Click through the test flow once and copy the generated code</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rBgBnlKZ4zaZ8vLRMWKmIA.png" /><figcaption>Use playwright codegen to generate the initial script</figcaption></figure><p><strong>Step 5:</strong> Create a test script file and paste the generated code</p><p><strong>Step 6: </strong>Ask Cursor to refactor the code to Playwright Pytest with best practices</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AOufZhgNIokUqcYXld5cAA.png" /><figcaption>Ask AI: Refactor this script to Playwright Pytest with good coding practice</figcaption></figure><p><strong>Step 7: </strong>Run the test using pytest --headed to verify the updated script</p><p><strong>Step 8: </strong>Debug with Cursor or edit any steps as needed</p><p><strong>Step 9: </strong>Use Cursor to add Allure report steps</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*I3-S2WR4AuFqYI8SLOiHOA.png" /><figcaption>Ask AI: Add Allure Report steps</figcaption></figure><p><strong>Step 10: </strong>Run the test script and generate the HTML report using Allure functions</p><pre>pytest test_script.py --alluredir=./allure-results<br>allure generate --clean --single-file allure-results</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hhkp20_Vo4nbmUG6V8cYcg.png" /><figcaption>Or simply ask AI: How to run the script and generate the report?</figcaption></figure><h3>The New Economics of Test Automation</h3><p>With the development of LLM and AI-powered tools, the efficiency of creating and maintaining E2E tests is dramatically improved. This shift in productivity translates to significantly lower costs, making test automation accessible to a broader range of organizations and projects. No longer is E2E test automation the exclusive domain of well-funded enterprises — it’s now within reach for startups, small teams, and budget-conscious projects.</p><h3>Looking Ahead: The Future of E2E Testing</h3><p>As we embrace these new tools and methodologies, it’s important to focus on the following areas to maximize the benefits of E2E test automation:</p><ol><li><strong>Test Stability and Maintenance</strong>: Develop strategies to create more robust tests that require less maintenance over time.</li><li><strong>Test Script Quality</strong>: Leverage AI tools to improve the overall quality and readability of test scripts.</li><li><strong>Enhanced Test Case Design</strong>: Use the time saved on script creation to focus on designing more comprehensive and effective test cases.</li><li><strong>Manual Testing Complementarity</strong>: Allocate more resources to exploratory and user experience-focused manual testing, complementing the automated suite.</li></ol><p>By focusing on these areas, teams can create a more balanced and effective testing strategy that combines the strengths of both automated and manual approaches.</p><p>In conclusion, the landscape of E2E test automation is evolving rapidly. With tools like Cursor, we’re entering an era where high-quality test automation is accessible to all, regardless of budget constraints. It’s time to embrace these new possibilities and elevate your testing practices to new heights.</p><p><em>At the heart of Oursky is the art of software engineering: translating ideas into opportunities. We work with startups in bringing their ideas to life with the right blend of people, process, and technology — from MVP development to scalable digital products powered by the latest technologies. </em><a href="https://oursky.com/contact"><em>Schedule a no-commitment consultation with us</em></a><em> and we can explore how to creatively harness the digital space to solve real-world problems.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=90b0e8595171" width="1" height="1" alt=""><hr><p><a href="https://medium.com/oursky-team/the-new-era-of-e2e-test-automation-faster-cheaper-and-more-accessible-90b0e8595171">The New Era of E2E Test Automation: Faster, Cheaper, and More Accessible</a> was originally published in <a href="https://medium.com/oursky-team">Oursky Team</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Building High-Performing Software Teams: The Hive Mind Strategy in Action]]></title>
            <link>https://medium.com/@joyzoursky/building-high-performing-software-teams-the-hive-mind-strategy-in-action-4587a5020566?source=rss-57d6cffeab2a------2</link>
            <guid isPermaLink="false">https://medium.com/p/4587a5020566</guid>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[project-management]]></category>
            <category><![CDATA[teamwork]]></category>
            <dc:creator><![CDATA[Joyz]]></dc:creator>
            <pubDate>Tue, 09 Apr 2024 06:50:56 GMT</pubDate>
            <atom:updated>2024-04-19T04:54:04.193Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Note: This is an experimental content written with assist of GenAI.</em></p><p>Software project managers often struggle to transform a group of skilled developers into a truly <strong>cohesive</strong>, high-performing team. Traditional hiring and team-building methods can identify individual talent, but fall short in cultivating the seamless collaboration and shared purpose needed for remarkable results.</p><p>The <strong>Hive Mind Strategy</strong> offers a compelling solution, drawing inspiration from the remarkable efficiency of <strong>Eusociality</strong> in beehives. This approach focuses on building a team-oriented environment. It unites diverse skills within the team and directs them towards a common goal, leveraging the team’s full potential.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*efhUE-rTbpmcuMNDKo4eXg.jpeg" /><figcaption>Photo by David Hablützel: <a href="https://www.pexels.com/photo/circular-honeycombs-in-beehive-6043084/">https://www.pexels.com/photo/circular-honeycombs-in-beehive-6043084/</a></figcaption></figure><h3><strong>Eusociality: The Blueprint for a Beehive Mind</strong></h3><p>Eusociality is the pinnacle of social organization in the animal kingdom. Here’s what makes it special:</p><ul><li><strong>Unified Goal:</strong> The entire colony works tirelessly for the <strong>prosperity of the hive</strong>.</li><li><strong>Division of Labor:</strong> Bees have specialized roles: queen for reproduction, workers for foraging and construction. This specialization optimizes the hive’s efficiency.</li><li><strong>Seamless Knowledge Transfer:</strong> Older bees nurture and train younger generations, ensuring the continuity of essential skills.</li><li><strong>Collective Ownership:</strong> Each bee takes pride in the hive’s success, fostering a sense of shared responsibility.</li></ul><h3><strong>The Hive Mind Strategy in Action: Building Your A-Team</strong></h3><p>By applying the principles of Eusociality, you can build your own high-performing software development team:</p><h4><strong>1. Shared Vision: From Project Roadmap to Individual Purpose</strong></h4><p>Just like a bee colony works for the hive’s prosperity, your team needs a <strong>shared vision for the project</strong>. Translate your project’s high-level goal into a clear roadmap with specific user stories, ensuring everyone understands <strong>the “why” behind their tasks</strong> and feels invested in the bigger picture.</p><h4><strong>2. Cross-Functional Collaboration: Building Your Dream Team</strong></h4><p>Build <strong>cross-functional teams</strong> during project planning, including roles like UI/UX designers and testers from the start, not just developers. Encourage team members to demonstrate <strong>discipline in their own specialties</strong> while also learning from and collaborating with one another. Just like bees with specialized roles, your team will be a versatile unit that can tackle any challenge.</p><h4><strong>3. Knowledge Sharing: Empowering Your Hive Mind</strong></h4><p>Similar to how older bees train younger generations, create a central repository for <strong>knowledge sharing</strong> and organize regular sharing sessions where team members can share their expertise on specific technologies, tools, or project-related topics. By creating a collaborative learning environment akin to<strong> an internal community</strong> of practice, your team will develop a collective understanding of the projects and the ability to solve problems together more effectively.</p><h4><strong>4. Collective Ownership: Building Pride in the Project</strong></h4><p>Imagine each bee taking pride in the hive’s success! Instill a sense of <strong>collective ownership</strong> within your team. Assign clear ownership of user stories or functionalities to team members or sub-teams, and create a culture of accountability to encourage individuals to take initiative. <strong>Celebrate team wins and milestone achievements</strong> to reinforce the sense of shared accomplishment.</p><h4><strong>5. Effective Communication: The Hive Buzz That Keeps Everyone Informed</strong></h4><p>Just like a beehive constantly adjusts based on environmental changes, your team needs <strong>effective communication</strong>. Regular communication practices are crucial in project management.</p><ul><li>Implement daily stand-up meetings to keep everyone informed on progress and roadblocks.</li><li>Hold sprint planning sessions to collaboratively define goals and tasks.</li><li>Conduct regular project retrospectives and individual one-on-one sessions to discuss what’s working well and identify areas for improvement.</li></ul><p>By keeping everyone on the same page, you ensure your team functions as a cohesive unit.</p><h3><strong>The Hive Mind Strategy: More Than Just a Buzzword</strong></h3><p>Although the beehive comparison isn’t perfect (given the complexity of human motivations), the fundamental aspects of Eusociality provide useful takeaways. Encouraging a shared vision, cross-functional collaboration, knowledge sharing, collective ownership, and effective communication can help build a high-performing software team that functions as a true <strong>“hive mind”</strong>, achieving remarkable results together.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4587a5020566" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI vs Human: Test Case Generation]]></title>
            <link>https://medium.com/oursky-team/ai-vs-human-test-case-generation-9d6699b3a584?source=rss-57d6cffeab2a------2</link>
            <guid isPermaLink="false">https://medium.com/p/9d6699b3a584</guid>
            <category><![CDATA[software-testing]]></category>
            <category><![CDATA[test-case]]></category>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[quality-assurance]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Joyz]]></dc:creator>
            <pubDate>Mon, 17 Apr 2023 03:26:48 GMT</pubDate>
            <atom:updated>2023-04-17T04:32:56.987Z</atom:updated>
            <content:encoded><![CDATA[<p>Over the past few months, people have seen a surge of AI generated contents from articles to illustrations and more. ChatGPT is currently the most powerful AI model with the ability to chat with humans like a real person, and has a huge professional knowledge base to perform even techincal tasks.</p><p>As experts in software testing, we have a question here:</p><blockquote><em>Can AI help human generate test cases?</em></blockquote><p>Below we have done a small experiment, to ask ChatGPT to draft some test scenarios for a sample app, and compare the results with those written by a real human tester. The app will have authentication functions such as Sign Up, Sign In, Email Verification, Forgot Password, and Sign Out. Let’s see how it turns out, no cheating!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Z_HlkckwxcDd66zKz34nyw.jpeg" /><figcaption>AI vs Human</figcaption></figure><p>In the experiment, we will only give 1 short question to both ChatGPT and the human tester, without explanation or examples ahead, let’s see what they will show us from scratch:</p><p><strong><em>“I have a mobile app which authenticate users with email and password. It has Sign Up, Sign In, Email verification, Forgot password and Sign Out functions. Name some test scenarios for each of these functions.”</em></strong></p><blockquote>As a quick experiment, we are only testing the “default” answers from both sides by not providing extra information. However, for practical ChatGPT usage, we should include much more specific information in the prompt to generate the ideal result. More techniques about it can be found on the <a href="https://github.com/openai/openai-cookbook/blob/main/techniques_to_improve_reliability.md">OpenAI Cookbook</a>.</blockquote><p>Alright Let’s Start!</p><h3>Round 1 — Sign Up</h3><h4>ChatGPT’s Test Cases</h4><ul><li>Verify that the user is able to enter valid email and password information</li><li>Verify that the user is able to submit the form and is <strong>redirected to a success page</strong></li><li>Verify that the user is not able to submit the form if any required fields are missing or invalid</li><li>Verify that the user is not able to sign up with an already registered email address</li><li>Verify that the user is <strong>prompted with helpful error messages</strong> if any errors occur during the sign up process</li></ul><h4>Human’s Test Cases</h4><ul><li>Sign up with an email and password that meet the requirements</li><li>Sign up with missing mandatory fields</li><li>Sign up with an email with <strong>incorrect format</strong>, e.g. missing “@” sign</li><li>Sign up with a password that <strong>does not meet the complexity requirements</strong></li><li>Sign up with an email that is already registered</li></ul><h4>Result — Human wins a bit💪</h4><p>Both ChatGPT and the human tester are able to cover the most common scenarios for successful and invalid sign-up attempts. A slight difference we could see here is that ChatGPT prioritizes the test scenarios for <strong>UI behaviors</strong> more, while the human tester focuses more on <strong>conditional cases</strong> of user input.</p><p>We would say both are important to the app, but in real-life executions, the human’s test cases actually have higher test coverage as the UI behaviors could be considered as some sub-steps of the conditional test scenarios.</p><h3>Round 2 — Sign In</h3><h4>ChatGPT’s Test Cases</h4><ul><li>Verify that the user is able to enter valid email and password information</li><li>Verify that the user is able to sign in and is <strong>redirected to the correct page</strong></li><li>Verify that the user is not able to sign in with invalid email or password information</li><li>Verify that the user is redirected to the login page if they <strong>try to access a protected page while not authenticated</strong></li><li>Verify that the user is <strong>prompted with helpful error messages</strong> if any errors occur during the sign in process</li></ul><h4>Human’s Test Cases</h4><ul><li>Sign in with the correct email and password of a verified account</li><li>Sign in with the correct email and password of <strong>an unverified account</strong></li><li>Sign in with missing mandatory fields</li><li>Sign in with an <strong>unregistered email</strong></li><li>Sign in with a wrong password</li></ul><h4>Result — 50/50⚔️</h4><p>Similar to Sign Up case, both ChatGPT and human tester can cover successful and unsuccessful attempts. We can see that the human tester focus on conditional user inputs as in the last round, while ChatGPT also includes an important case <strong>related to authentication security</strong>, although it is not directly related to clicking the Sign-In button.</p><p>On the other hand, the human tester is able to relate test cases with other features such as email verification, while for ChatGPT, it seems that it needs to be given with more guidance to include additional cases related to cross-feature user flows.</p><h3>Round 3 — Email Verification</h3><h4>ChatGPT’s Test Cases</h4><ul><li>Verify that the user is able to successfully sign up for the app with a <strong>valid</strong> email address and password</li><li>Verify that the user receives the verification email promptly and that it contains the <strong>correct</strong> verification link</li><li>Verify that the user is able to click on the verification link and that it takes them to the <strong>correct</strong> page to verify their email address</li><li>Verify that the user is able to <strong>successfully</strong> verify their email address using the verification link</li><li>Verify that the user is redirected to the <strong>correct</strong> page after their email address has been verified</li><li>Verify that the user is <strong>able to log in</strong> with their email and password after their email address has been verified</li><li>Verify that the user is not able to log in with an incorrect email or password</li><li>Verify that the user is not able to access any protected pages until their email address has been verified</li><li>Verify that the user is <strong>able to request</strong> a new verification email if they did not receive the first one</li><li>Verify that the user is not able to request a new verification email if their email address <strong>has already been verified</strong></li><li>Verify that the user is <strong>able to receive</strong> and use a new verification email if they accidentally delete the first one or the link expires.</li><li>Verify that the user is able to <strong>successfully</strong> verify their email address using the new verification email</li></ul><h4>Human’s Test Cases</h4><p>Assume the user cannot sign in before verifying their email. Assume the email verification link is a deep link to open the app and finish the process in the app.</p><ul><li>Click a valid email verification link to complete the verification</li><li>Click an email verification link after it has expired</li><li>Click an email verification link with an invalid URL</li><li>Request to resend the email verification link through the app</li><li>Click an old verification link after requesting to send a new link</li><li>Click an email verification link when the app is signing in with another account</li><li>Click an email verification link when the app is in the background</li><li>Click an email verification link when the app has been force quit</li><li>Click an email verification link when the app has been uninstalled</li></ul><h4>Result — Human wins by a mile🏃🏃🏃🏃</h4><p>ChatGPT’s default answer is more like a set of steps to ensure that the Email Verification function works properly in <strong>happy flows</strong>. It generates a lot of points, but we can actually group some of them like point 1–6 into one test flow, while some are duplicated with previous sets like point 7–8, or too similar as separated cases like point 9 &amp; 11.</p><p>The human side style is very different on this part. We can see that it actually covers a wider variety of conditions and real-life situations, such as clicking the verification link of different statuses.</p><p>As a test on a complex user function without more information given to ChatGPT, it will be difficult for it to create good enough result.</p><p>💡 Frankly speaking email verification is a complicated feature, especially when handling on a mobile app, as there are a lot of factors that could affect the behaviours depending on the app and business logic, which should be clarified before starting to write the test cases. For example,</p><p><strong>Regarding the verification status:</strong></p><ul><li>Can user still sign in when the email is not verified?</li><li>If it verifies with a link, how many error states there are? Expiry, invalid link, old link, etc.</li><li>If it verifies with a code instead of clicking a link, how many error states there are? Expiry, wrong code, exceed no. of attempts, etc.</li><li>Are there other user statuses that will fail the verification? Suspened user, deleted account, etc.</li></ul><p><strong>Regarding the user flow:</strong></p><ul><li>How many entry points to trigger a email verification?</li><li>How many places will have status change after a successful verification, including other web portal or app platforms of the system?</li><li>Will email or push notification be sent to user after a verification?</li><li>How should the link route when the app is under different conditions? Staying on specific screens, app in background, app closed, etc.</li></ul><h3>Round 4 — Forgot Password</h3><h4>ChatGPT’s Test Cases</h4><ul><li>Verify that the user is able to enter their email address and submit the forgot password form</li><li>Verify that the user receives an email with a password reset link</li><li>Verify that the user is able to click the password reset link and is redirected to a success page</li><li>Verify that the user is prompted with helpful error messages if <strong>any errors occur</strong> during the forgot password process</li><li>Verify that the user is able to reset their password and sign in using their new password</li></ul><h4>Human’s Test Cases</h4><ul><li>Submit forgot password form with a valid and registered email</li><li>Submit forgot password form with <strong>empty email field</strong></li><li>Submit forgot password form with <strong>an unregistered email</strong></li><li>Receive reset password email, follow the instructions and set a new password</li><li>Reset password with a password the <strong>same as the old one</strong></li><li>Reset password with a password that <strong>does not meet the complexity requirements</strong></li><li>Sign in with new password after reset</li><li>Try to <strong>sign in with old password</strong> after reset</li></ul><h4>Result — Human wins by accuracy and sustainability🌱</h4><p>The “Forgot Password” feature is another example where ChatGPT generates something more like a test script that goes through the positive path of the feature with multiple steps, and includes only one unclear error scenario that tries to cover “any errors”, which is not specific enough for test execution.</p><p>In contrast, the human tester states the exact error conditions and divides the points by different test flows instead of different steps in a flow. This is also more recommended for higher efficiency in reading and maintaining the test scenarios, especially in high-speed agile projects.</p><h3>Last Round — Sign Out</h3><h4>ChatGPT’s Test Cases</h4><ul><li>Verify that the user is able to click the sign out button and is <strong>redirected to the login page</strong></li><li>Verify that the user is not able to access any protected pages after signing out</li><li>Verify that the user is <strong>redirected to the login page</strong> if they try to access a protected page while not authenticated</li><li>Verify that the user is able to sign back in using their original email and password after signing out</li><li>Verify that the user’s session <strong>is completely destroyed</strong> after signing out</li></ul><h4>Human’s Test Cases</h4><ul><li>Show <strong>confirmation dialogue</strong> and confirm sign out</li><li>Cancel logout in confirmation dialogue</li><li>After sign out, user should not be able to access pages that requires login</li><li>After sign out, user <strong>should not recceive push notifications</strong> sending to the last logged in user</li><li>Sign in as the same user again after sign out, check if user info is correct</li><li>Sign in as <strong>a different user</strong> after sign out, check if user info is correct</li></ul><h4>Result — Woohoo! Another 50/50 🤜🤛</h4><p>Similar to the Sign In round, ChatGPT and human tester are both able to list the common scenarios, including successful sign in, restricted content after sign out, and signing back in after sign out.</p><p>One interesting point is that since the given description is not clear, both of them tried to add some common UX behaviours based on their understanding of a sign out feature too, i.e.</p><ul><li>Redirecting to sign in page after sign out and accessing protected page</li><li>Showing confirmation dialogue before user confirms to sign out</li></ul><p>We can also see that they are slightly more precise in their “strengths” respectively, where ChatGPT describes more details about UI transition, and the human tester thinks of more checkpoints to test for a destroyed session.</p><h3>What’s Next?</h3><p>It is significant that AI could act as a supporting role in writing test cases for human, and it takes just a second to generate the content while the prompt is sent. Even without specific information and examples given, ChatGPT could already generate some results as good as human testers, especially in straightforward questions, e.g. common scenarios or happy flows, user input validation, CRUD features, etc.</p><p>In the future, instead of spending time writing step-by-step test cases, we could foresee that QA testers will transform their skills to develop techniques related to the usage of ChatGPT, such as <a href="https://github.com/openai/openai-cookbook/blob/main/techniques_to_improve_reliability.md#few-shot-examples">few-shot chain of thought prompting</a>, and widely adopt it in the field of test case generation.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9d6699b3a584" width="1" height="1" alt=""><hr><p><a href="https://medium.com/oursky-team/ai-vs-human-test-case-generation-9d6699b3a584">AI vs Human: Test Case Generation</a> was originally published in <a href="https://medium.com/oursky-team">Oursky Team</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[User Acceptance Testing (UAT): The Guide for End-user and Business Analyst]]></title>
            <link>https://medium.com/oursky-team/user-acceptance-testing-uat-the-guide-for-end-user-and-business-analyst-415a3d3983c7?source=rss-57d6cffeab2a------2</link>
            <guid isPermaLink="false">https://medium.com/p/415a3d3983c7</guid>
            <category><![CDATA[user-acceptance-testing]]></category>
            <category><![CDATA[software-testing]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[functional-testing]]></category>
            <category><![CDATA[exploratory-testing]]></category>
            <dc:creator><![CDATA[Joyz]]></dc:creator>
            <pubDate>Thu, 02 Mar 2023 07:55:24 GMT</pubDate>
            <atom:updated>2023-03-02T07:55:24.188Z</atom:updated>
            <content:encoded><![CDATA[<p>User Acceptance Testing (UAT) is a gatekeeping process in the software development lifecycle, to ensure a new product or version update is ready for public release. The application needs to be fully tested and accepted by the real End-Users and/or the Business Analysts (BA), to make sure the business requirements are met and the features are worked as intended from the users’ point of view.</p><p>Depending on the scale, complexity and nature of products, there are many different approaches to perform the UAT. For general user-facing applications, a common method is to run through a combination of <strong>Functional Testing</strong> and <strong>Exploratory Testing</strong>, to evaluate whether the functionality, UI/UX, performance, stability and overall quality match with the business needs.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uwduQTiyygLEttrajX4Aew.jpeg" /><figcaption>User Acceptance Test (UAT)</figcaption></figure><h3>Functional Testing</h3><p>Functional Testing validates whether the application works as expected. Through comparing the product behaviours against the functional requirements in specification, the test result acts as a green light for release if all acceptance criteria are passed.</p><p>To get started with Functional Testing, testers design and write <strong>Test Cases</strong> based on the business requirements and use cases. They then execute the test cases step-by-step to ensure that all procedures result in the expected behaviors.</p><p>A <strong>Test Suite</strong> is a group of test cases that includes <strong>happy paths (positive cases)</strong>, where the user uses the product as expected, e.g. <em>filling in all fields correctly and submitting a form</em>, and <strong>unhappy paths (negative cases)</strong>, where the user performs unexpected actions, e.g. <em>filling in a form with invalid input and trying to submit it</em>. A good functional test suite should include both types of cases to ensure that the product can handle both success and failure cases properly.</p><h4>Functional Test Cases Example</h4><p>Imagine that you are working on an e-commerce mobile app. Recently, the app has added a credit card payment option via <a href="https://stripe.com/">Stripe</a>. To design the test cases for this feature, you could check the product specification with support from the <a href="https://stripe.com/docs/testing">documentation of Stripe</a>, and add various cases to your functional test suite, for example, payment with credit cards issued in different countries, card brands, invalid cards, 3D secured cards, etc.</p><p>Here is a simple format of Functional Test Cases:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*udhCrVfFHg1MP3cdkbQxGQ.png" /><figcaption>Functional Test Cases Example</figcaption></figure><h3>Exploratory Testing</h3><p>Exploratory Testing is a context-driven methodology of testing, which relies on the subject knowledge of testers to simulate how the product will be used in real-life situations. As a supplement to functional testing, testers explore the application creatively instead of following any scripted test cases or specifications, to discover defects, usability problems, or even business concerns from the perspective of an end-user.</p><h3>Session-based Exploratory Testing</h3><p>To execute Exploratory Testing, there are <strong>Session-based</strong> and <strong>Scenario-based</strong> methodologies to control the scope and resources of test. The idea of Session-based Exploratory Testing is to set a timebox, like a 60–90 minutes session, and assign a group of testers to explore the app within this time, given a <strong>Test Charter</strong> that describes the test focus and objectives of this session.</p><h4>Exploratory Test Charter Template</h4><p>Below is a template of an Exploratory Test Charter, proposed by Elisabeth Hendrickson in the book <em>Explore It!, Chapter 2, page 67 of 502 (ebook) — “A simple charter template”</em>.</p><blockquote><em>Explore (target) With (resources) To discover (information)</em></blockquote><ul><li><strong>Target</strong>: What are you exploring? It could be a feature, a requirement, or a module.</li><li><strong>Resources</strong>: What resources will you bring with you? Resources can be anything: a tool, a data set, a technique, a configuration, or perhaps an interdependent feature.</li><li><strong>Information</strong>: What kind of information are you hoping to find? Are you characterizing the security, performance, reliability, capability, usability or some other aspect of the system? Are you looking for consistency of design or violations of a standard?</li></ul><h3>Scenario-based Exploratory Testing</h3><p>Scenario-based Exploratory Testing is another direction to explore the app based on the story about how the product is used, including information about the motivations of the users involved.</p><p>Let’s use a social app with a login function as an example.</p><p>The <strong>User Story</strong> is simple:</p><blockquote><em>As a user, I want to securely log in to the system, so that I can access the Home page content that is exclusive to me.</em></blockquote><p>Based on the story, testers can brainstorm different scenarios that may occur in real-life situations and execute the flows to discover misbehaviors, defects in edge cases, UIUX problems, or even business concerns.</p><p>For example,</p><ul><li>What if the user double clicks on the login button?</li><li>What if the internet is disconnected while loading?</li><li>What if the user cannot receive the one-time password (OTP)?</li><li>What if the user has already logged in on another device?</li></ul><p>The expected behaviours of these scenarios may not be fully stated in the functional requirements, but should also be addressed if they go against a good user experience.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Mm0QgsBeq5LGL8sTOcKprA.jpeg" /><figcaption>Be creative, and timebox your Exploratory Test session.</figcaption></figure><h3>Test Preparation</h3><p>Before running the UAT, testers should have a basic understanding of the test run workflow and information, product scope and features, and any necessary resources to ensure a smooth and efficient test execution. Typically, the test manager will create a test document that includes all required references to support the test run, such as:</p><ul><li>Testing environment</li><li>Testing build version and Installation guide</li><li>Supporting browsers, devices and platforms</li><li>Type of test and related documents</li><li>Feature scope</li><li>Specifications</li><li>Design reference</li><li>Test accounts and Test data</li><li>Issue reporting instructions</li><li>Known issues</li></ul><p>Test documentation serves as a knowledge base for testers and helps new testers onboard when the test run is distributed. Without sufficient preparation, testers may easily get blocked during testing or need to spend a lot of time gathering information from different parties to resolve obstacles. This could cause delays in the process and, in the worst case, it could also affect the release schedule.</p><h3>After the UAT</h3><p>After completing all the test runs for UAT, a list of defects and feedback will be collected. The product owner and project team will need to understand and discuss the issues in order to prioritize and arrange for further enhancements.</p><p>Once the bugs have been fixed and passed regression testing, Congratulations🎉 — the project team can finally proceed to prepare for public release! The testing period to perform a full UAT can sometimes be long, maybe up to 1–2 months, but the high quality of the final product and the satisfaction of end-users will definitely make it worthwhile.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=415a3d3983c7" width="1" height="1" alt=""><hr><p><a href="https://medium.com/oursky-team/user-acceptance-testing-uat-the-guide-for-end-user-and-business-analyst-415a3d3983c7">User Acceptance Testing (UAT): The Guide for End-user and Business Analyst</a> was originally published in <a href="https://medium.com/oursky-team">Oursky Team</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Building an Agile Testing Workflow in Startup]]></title>
            <link>https://medium.com/oursky-team/building-a-one-man-testing-team-in-agile-startup-4d5459170885?source=rss-57d6cffeab2a------2</link>
            <guid isPermaLink="false">https://medium.com/p/4d5459170885</guid>
            <category><![CDATA[qa]]></category>
            <category><![CDATA[agile]]></category>
            <category><![CDATA[scrum]]></category>
            <category><![CDATA[project-management]]></category>
            <category><![CDATA[software-testing]]></category>
            <dc:creator><![CDATA[Joyz]]></dc:creator>
            <pubDate>Wed, 23 Nov 2022 08:46:06 GMT</pubDate>
            <atom:updated>2022-11-23T09:04:27.584Z</atom:updated>
            <cc:license>https://creativecommons.org/licenses/by-sa/4.0/</cc:license>
            <content:encoded><![CDATA[<p><strong>Agile</strong> methodology has been a trend in software industry in the recent years. Publishing speedy new releases with high quality is the goal of all product teams.</p><p><strong>Scrum</strong> is a framework of agile project development, which breaks the big scope of features development into small cycles called <strong>“Sprints”</strong>, and deliver product increment by iteration called <strong>“Milestones”</strong>. There are various ways of works, terms and principals in actual, but in general it refers to a mechanism that can provide fast and continuous delivery.</p><p>Keeping up the quality as its speed is one of the key challenges. Teams could easily rush for new features and miss out <strong>testing and maintenance</strong> for the app quality, breaking the product with accumulating defects and growing tech debt.</p><p>Below we would like to introduce a model of <strong>QA Workflow</strong> for specifically small-scale scrum teams, where all one-man testing team gets started from.</p><h3>Team Structure</h3><p>Let’s say you have a startup scrum team with the following members:</p><ul><li>Project Manager (PM) x 1</li><li>UI/UX Designer x 1</li><li>Senior Developer x 1</li><li>Junior Developer x 1–2</li><li>QA Tester (QA) x 1</li></ul><p>Regarding to testing, PM, Developer and QA are the 3 main roles that contribute to the quality control of products. We could have a simple division of work for them according to their skill set.</p><p><strong>Project Manager</strong></p><ul><li>Responsible person of the project, including product quality</li><li>Consolidate project scope and use cases</li><li>Break down user stories and acceptance criteria</li><li>Manage development, test and release schedule</li><li>Collect feedback, prioritise bug fix and enhancements</li></ul><p><strong>Developer</strong></p><ul><li>Complete assigned tasks with basic testing</li><li>Code quality control</li><li>Fix bugs by assignment</li><li>Continuous delivery</li></ul><p><strong>QA Tester</strong></p><ul><li>Continuous Exploratory Test based on project scope</li><li>Run different types of test according to project need</li><li>Find, Verify and Report issues</li><li>Regression test after new changes and bug fix</li></ul><h3>Sprint &amp; Milestone</h3><p>In a scrum lifecycles, tasks to do are scheduled by recurring “Sprints”, while the batch features to release is planned by “Milestone”.</p><p>For example, say if each sprint is around 2 weeks of work, and the product owner will like to release a version of the app every 3 months, then the cycle will be like:</p><p>Sprint 1 → Sprint 2 → Sprint 3 → Sprint 4 → Sprint 5 → Sprint 6 (Milestone 1)</p><h3>QA Workflow</h3><p>To suit the nature of “Sprint” and “Milestone”, the testing worflow will also be slightly different in different stages of development, while the testing work does not actually vary a lot.</p><p>A suggested workflow with PM, Developer, Tester involved:</p><p>When the milestone starts, PM hosts a <strong>Kickoff meeting</strong> with the project team, to brief and clarify the background and requirements of the whole project</p><p>Within each sprint,</p><ol><li>PM prepares user stories and host a <strong>Sprint planning meeting</strong> with the team, to debrief the work done in the previous sprint, and go through the tasks scheduled in this sprint;</li><li>Developers do a <strong>Technical planning</strong> to break down the todos;</li><li>Developers code and deploy changes to <strong>Testing environment;</strong></li><li>PM runs an <strong>Acceptance Test</strong>, to check the acceptance criteria and testability of app;</li><li>When it is ready for test, QA runs an <strong>Exploratory Test</strong>, to find bugs in any possible scenarios, related to the features scope of new updates, and also the overall app stability;</li><li>After QA reports bugs to the team, PM and QA may discuss to <strong>prioritize the issues</strong>, and arrange whether they should be fixed within this sprint, the next sprint, other slots or put to a backlog;</li><li>Looping the bug fix and test run within the sprint until it is stable to ship, PM can summarize a <strong>Progress Update</strong> as the work done in this sprint</li></ol><p>Coming to the end of a milestone after several sprints of work, a final check on whole project should also be arranged to make sure the it is good for a <strong>Production release</strong>:</p><ol><li>Developers prepare a <strong>Alpha or Beta Production environment</strong> for testing</li><li>Apart from the regular sprint testing flow, PM and QA also runs a <strong>Smoke Test</strong> for the whole app, as a health check for all main user flows, no matter the feature is related to the milestone update or not</li><li>When everything looks good and submitted to app store, it’s time for a release party 🎉</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JUzfmNE3xPNTez8BKVqlww.png" /><figcaption>A sample workflow chart for small-scale agile team</figcaption></figure><p>Actually, when working out a testing flow in a scrum team, there is also a lot of project-specific tests runs that we may consider add-ons, e.g. Usability Test, Functional Test, Performance Test, Load Test, Security Test, etc.</p><p>These could be more technical or time-consuming work which also depend on the needs of the project. But as a recommendation for Startup Scrum Team, <strong>Exploratory Test</strong> is good to be one of the main testing methods, as it suits the nature of Agile — Fast, Flexible and User-Driven.</p><p>Last but not least, we would also like to share you a sample of an Issue Report format that our QA team used to write, hope it will also be useful for anyone getting started to software testing :)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*R6TDwBNmZI6K_1Zk6kJbzQ.png" /><figcaption><a href="https://github.com/authgear/authgear-server/issues/2537">https://github.com/authgear/authgear-server/issues/2537</a></figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4d5459170885" width="1" height="1" alt=""><hr><p><a href="https://medium.com/oursky-team/building-a-one-man-testing-team-in-agile-startup-4d5459170885">Building an Agile Testing Workflow in Startup</a> was originally published in <a href="https://medium.com/oursky-team">Oursky Team</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Bug Report Template]]></title>
            <link>https://medium.com/oursky-team/the-bug-report-template-fbab2ebe99b8?source=rss-57d6cffeab2a------2</link>
            <guid isPermaLink="false">https://medium.com/p/fbab2ebe99b8</guid>
            <category><![CDATA[software-testing]]></category>
            <category><![CDATA[testing]]></category>
            <category><![CDATA[reporting]]></category>
            <category><![CDATA[bugs]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Joyz]]></dc:creator>
            <pubDate>Mon, 24 Feb 2020 07:40:06 GMT</pubDate>
            <atom:updated>2020-02-24T07:52:16.313Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*IUGqfhuddbIg4Dqe" /><figcaption>Photo by <a href="https://unsplash.com/@fempreneurstyledstock?utm_source=medium&amp;utm_medium=referral">Leone Venter</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>There are many articles on the net teaching testers how to write a good bug report. But most of them are too long to read.</p><p>Here is a simple one in markdown format that you can just copy &amp; paste to start reporting your bug. We have been using it for years in our own team. Feel free to take reference and modify for your needs 🧙‍♂️</p><h3>Bug Report Title:</h3><pre>[Page / Section] One-line description</pre><h3>Bug Report Body (for mobile apps):</h3><pre>## Environment:<br>- Testing environment: Nightly / Staging / Production<br>- Testing platform: Android / iOS<br>- Testing device (OS version):<br>- App version:</pre><pre>## Description:<br>Further explanation / Any special conditions required</pre><pre>## Steps to reproduce:<br>1. Go to page xxx<br>2. Click xxx<br>3. Click xxx<br>4. ...</pre><pre>## Expected result:<br>What you expect to see</pre><pre>## Video:<br>Any video demonstrating how to reproduce the bug</pre><pre>## Screenshot:<br>Any screenshot(s) of the current behaviour</pre><pre>## Crash log:<br>Any crash log captured</pre><h3>Bug Report Body (for web applications):</h3><pre>## Environment:<br>- Testing environment: Nightly / Staging / Production<br>- Testing device:<br>- Testing device OS version:<br>- Testing browser (version):<br>- Build version:</pre><pre>## Description:<br>Further explanation / Any special conditions required</pre><pre>## Steps to reproduce:<br>1. Go to page xxx<br>2. Click xxx<br>3. Click xxx<br>4. ...</pre><pre>## Expected result:<br>What you expect to see</pre><pre>## Video:<br>Any video demonstrating how to reproduce the bug</pre><pre>## Screenshot:<br>Any screenshot(s) of the current behaviour</pre><pre>## Crash log:<br>Any crash log captured</pre><h3>Example:</h3><pre>[My Apps] Last updated time of app is shown as “invalid date”</pre><pre>## Environment:<br>- Testing environment: Staging<br>- Testing device: MacBook Pro (Retina, 13-inch, Early 2015)<br>- Testing device OS version: macOS Catalina 10.15.3<br>- Testing browser (version): Google Chrome (79.0.3945.130)<br>- Build version: v1.0.1</pre><pre>## Description:<br>When a new app is created, the last updated time showing on My Apps page is wrongly displaying as &quot;invalid date&quot;.</pre><pre>## Steps to reproduce:<br>1. Login<br>2. Create new app<br>3. See Get Started page<br>4. Click back button<br>5. See My Apps page<br>6. See “Last updated at invalid date” below app name</pre><pre>## Expected result:<br>It should show last updated date correctly. If it is a newly created app, it should show the app created date as its last updated date.</pre><pre>## Screenshot:</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/589/1*P2uHjSTH6BJ-Ewy2wsDnvw.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fbab2ebe99b8" width="1" height="1" alt=""><hr><p><a href="https://medium.com/oursky-team/the-bug-report-template-fbab2ebe99b8">The Bug Report Template</a> was originally published in <a href="https://medium.com/oursky-team">Oursky Team</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>