Alexander Shulman’s Post

Most portfolio projects could be built in a weekend. I wanted to build something that couldn't. So I reverse-engineered how GitHub works and built RepoDoc — a full-stack version control platform with a custom Git-like CLI, completely from scratch. Here's what it actually does: 🗂️ Repos → Full CRUD, in-browser file editing, commit history, S3-backed storage 👤 Users → Profile management, image uploads, JWT auth 🐛 Issues → Create, update, close, delete per repository ⭐ Social → Star / Unstar repos 💻 Custom CLI: init → spins up a local .repoGit directory add → stages a file commit → commits everything staged push → ships files to AWS S3 pull → pulls latest from S3 revert → rolls back to any commit by ID The CLI and integrating AWS with it alone took longer than most of my previous projects combined. Because when you build the tool instead of just using it, you finally understand it. The real takeaway wasn't the tech stack. It was finally understanding why the tools we use every day are designed the way they are. You don't get that from tutorials. You get it from breaking things and fixing them yourself. 🎥 Short walkthrough below. Drop your thoughts! 🔗 Live link in the comments. If you've ever built dev tooling from scratch, what broke first? #softwareengineering #buildinpublic #fullstackdevelopment #reactjs #nodejs #javascript #aws

🚨 WE JUST CHANGED THE GAME. AND THERE'S NO GOING BACK. 🚨 Forget everything you knew about MCP setup. Forget the hours wasted on config files. Forget the "it works on my machine" nightmares. We just shipped something that makes all of that IRRELEVANT. 👇 ━━━━━━━━━━━━━━━━━━━━━━ ⚡ Introducing @global-packages/mcp-kit The LAST MCP tool your team will ever need. ━━━━━━━━━━━━━━━━━━━━━━ One command. Every MCP. Zero friction. Every OS. ``` npm install -g @global-packages/mcp-kit mcp-kit install mcp-kit init --dev ``` That's it. You're done. Go touch grass. 🌿 🔥 What we built: ✅ Auto-detects your .vscode folder across ANY project, ANY OS ✅ Configures Azure, AWS, GCP, Confluence, Postgres, Playwright, Slack, Figma, Git & MORE — out of the box ✅ Secure credential storage via system keychain (no more .env nightmares in shared repos) ✅ Tailored setups for BOTH devs AND non-devs — finally, a tool for the whole team ✅ Works flawlessly on Windows, macOS & Linux ✅ `mcp-kit doctor` — because your setup deserves a health check too 🌍 16+ MCPs supported. Cloud. Databases. DevOps. Design. Testing. Productivity. Azure DevOps. SonarQube. Figma. Playwright. Slack. Confluence. PostgreSQL. MSSQL. This isn't just a tool. This is INFRASTRUCTURE for how modern teams work with AI in VS Code. We didn't just onboard to the MCP ecosystem — 💥 WE REDEFINED WHAT ONBOARDING LOOKS LIKE. The old world of painful MCP configuration? Gone. Deleted. It doesn't exist anymore. And the best part? 🤝 It's OPEN SOURCE. Community-driven. MIT licensed. PRs welcome. If you want to add new MCPs, suggest improvements, or make it even more insane — come build with us. ━━━━━━━━━━━━━━━━━━━━━━ 📦 NPM Package: https://lnkd.in/gqpNXQGU 🐙 GitHub Repository: https://lnkd.in/gt5eY5eX ━━━━━━━━━━━━━━━━━━━━━━ Drop a 🚀 in the comments if you're done tolerating broken MCP setups. Share this with the one developer on your team who's STILL configuring things manually. You know who they are. 👀 #MCP #VSCode #DeveloperTools #OpenSource #DevTools #AI #ArtificialIntelligence #SoftwareEngineering #CloudComputing #Azure #AWS #GCP #DevOps #Automation #TypeScript #NodeJS #GitHub #Productivity #TechInnovation #Programming #Coding #SoftwareDevelopment #Tech #Innovation #JavaScript #BuildInPublic #100DaysOfCode #MachineLearning #AITools #FutureOfWork

Trangly v0.1.2 is out. Here's what shipped: — Web terminal: a full shell inside your running container, directly from the dashboard. No SSH client. No docker exec. Just click and you're in — right next to your live logs and deploy history. — Queue and internal improvements: the rate limiter cleanup goroutine now shuts down cleanly instead of leaking on server stop. Errors that were previously swallowed silently in the deploy pipeline and queue are now properly logged, making debugging a lot less of a guessing game. — v0.1.2 also adds actual screenshots to the README (finally 😅) — Dashboard, Deployments, Live Logs, Terminal, and Settings. The terminal felt overdue. Debugging still meant SSH-ing into your VPS, hunting down the container ID, and running docker exec just to get a shell inside it. Every single time. So we closed that loop. Trangly is a self-hosted, single-binary CI/CD tool for developers running Docker apps on a VPS. No YAML. No cloud accounts. No DevOps setup. Push → webhook → build → health check → swap. The old container stays live until the new one is healthy. If you're tired of Coolify, CapRover, or wrangling GitHub Actions YAML just to deploy a side project — give it a look. https://lnkd.in/d5-4RB4v #OpenSource #DevOps #SelfHosted #Go #CICD #DeveloperTools

🚀 What is Docker? (Simple Developer Explanation) Let’s understand Docker with a real-world developer scenario 👇 Imagine you built a simple Node.js application. To run it on your system, you: • Install Node.js • Use the correct Node version • Run npm install to install dependencies • Start the app Everything works perfectly on your machine 👍 Now you share this project with your friend via GitHub. Here’s the problem ❌ Your friend: • Doesn’t have Node.js installed • Might install the wrong version • Might face dependency issues • Ends up saying: “It’s not working on my machine” 😅 💡 This is where Docker comes in. If one friend — it’s fine, you can help. But if it’s 10 friends… or 100 developers in a company? How will you manage? 🤯 Instead of asking your friend to install Node.js and set everything up manually, you can: 👉 Package your application along with: • Node.js • Dependencies • Required environment into a Docker Image. Now your friend just needs to: • Install Docker • Run your Docker container That’s it. No manual setup. No version issues. No dependency conflicts. 🏢 How big companies handle this In real-world companies: • Hundreds of developers work on the same project • Code runs across multiple environments (Dev, QA, Prod) • Applications are deployed on cloud servers Without Docker ❌ 👉 “Works on my machine” becomes a daily problem With Docker ✅ 👉 Every developer, tester, and server runs the same container Companies use Docker to: • Standardize environments across teams • Simplify onboarding (new devs can start in minutes) • Deploy applications reliably using CI/CD pipelines • Scale applications easily using tools like Kubernetes ✨ In simple terms: Docker allows you to bundle your application + environment together and run it anywhere consistently. 📦 “Build once, run anywhere” — that’s the power of Docker. If you’re a developer, learning Docker is not optional anymore — it’s a must-have skill for modern development and DevOps. Check this repo to handson: https://lnkd.in/gxB-fBVZ #Docker #DevOps #NodeJS #SoftwareEngineering #BackendDevelopment #CloudComputing

Every deploy used to be a ritual. SSH into the server. Pull the code. Rebuild the frontend. Restart the backend. Watch the site throw 502s for 30 seconds while everything came back up. Hope nothing broke. Today, I pushed a commit to main and walked away. 90 seconds later, the new version was live. Frontend and backend both. No downtime. No manual steps. No anxiety. Here's what I wired up this weekend: 👉 GitHub Actions triggers on push to main 👉 Docker builds fresh production images for the frontend and the backend 👉 Server pulls the images and starts new containers alongside the old ones 👉 Health checks run on both before any traffic hits them 👉 Nginx cuts traffic over once they pass 👉 Old containers shut down gracefully 👉 One command rollback if something breaks No more maintenance windows. No more "please don't refresh" messages in Teams. No more scheduling deploys at 2am to dodge users. Two months ago I thought zero downtime deployment was something only teams with dedicated SREs needed. Turns out it's a few containers, a health check, and a proxy that knows how to wait. What deploy process did you outgrow this year? 👇 Tech: Docker · GitHub Actions · Nginx · Ubuntu · PostgreSQL · NestJS · React · TypeScript #DevOps #CICD #Docker #GitHubActions #SoloDeveloper #BuildInPublic #FullStackEngineer #ZeroDowntime

GitHub Actions observability is a fucking joke. You get a wall of YAML logs and a prayer. Something breaks on step 8 of 10? Good luck figuring out what happened in the seven steps before it. Most teams are flying blind on their own CI pipelines. They can't answer basic questions: - How much CPU is my build actually consuming? - How much memory? - What's my average queue time? - What's my average build time? - Are my builds flaky, or is something actually broken? As you scale code velocity, these questions aren't nice-to-haves. They're the difference between shipping fast and shipping into a wall. Here's what we're building with Depot CI: Observability: Track the metrics that actually matter. CPU consumption. Memory usage. Queue time. Build time. Not buried in logs, exposed through an API so you can feed them to your own tools, your agents, your dashboards. You need to see how scaling your velocity impacts your delivery pipeline before it becomes a bottleneck. Debugability: When something breaks, reproduce it in the exact environment where it failed. Not from the beginning. Not from scratch. Your build has 10 steps. It fails on step 8. You've accumulated state across the first seven steps before that failure. With Depot, you drop into step 8 with all of that state intact. Same CPU. Same memory. Same filesystem. Same everything. Integration: Logs behind an API so you can feed errors directly to an agent or your own CLI tools. CI environments are finicky, you've customized them in a very specific way tailored to your build. Something slightly off and you're staring at a failure you can't reproduce locally. You need the ability to jump into any step of your build, not guess your way through six re-runs. Observability tells you what's happening. Debugability lets you fix it. Most CI platforms give you observability in the form of a wall of YAML logs. That's not observability. That's a prayer. Because re-running a build three times and squinting at logs isn't debugging. It's wasting time.

🚀 Built NGINX Web Server Using Docker Today! 🐳 Today was a great hands-on DevOps learning experience where I created a web server (NGINX) using Docker and served a custom HTML page 🔥 Here’s how I did it step-by-step 👇 🔹 Step 1: Pull & Run NGINX Container Pulled the NGINX image from Docker Hub and exposed it on port 80 docker run --name docker-nginx -p 80:80 nginx 🔹 Step 2: Verify & Manage Container Checked running containers and practiced stopping/removing it docker ps docker stop docker-nginx docker rm docker-nginx 🔹 Step 3: Create Custom Web Page Created a directory and added an HTML file mkdir -p ~/docker-nginx/html cd ~/docker-nginx/html vi index.html 🔹 Step 4: Run with Volume Mapping Mapped local HTML directory to NGINX container docker run --name docker-nginx -p 80:80 -d -v ~/docker-nginx/html:/usr/share/nginx/html nginx 🔹 Step 5: HTML File Created a simple HTML page which is now served via NGINX 🚀 This small project gave me a clear understanding of how Docker simplifies deployment and avoids “it works on my machine” issues 😄 #Docker #NGINX #DevOps #CloudComputing #AWS #LearninginPublic

Claude Code routines just launched as a research preview – and they solve a problem we've seen across dozens of automation builds. The pattern was always the same: – Wire up a GitHub Action – Shell out to claude -p with a prompt – Hope the token budget holds – Debug YAML at midnight Routines replace that entire stack. You define a prompt, connect your repos and MCP tools, pick a trigger, and Claude runs it on Anthropic's cloud. No local machine required. Three trigger types: 1. Scheduled – hourly, daily, weekly, or custom cron 2. API – monitoring tools POST to a dedicated endpoint 3. GitHub – reacts to PR and release events The key shift: GitHub Actions runs commands you wrote. A routine receives a prompt and decides how to reach the outcome. We wrote up a full breakdown with six practical use cases:

Agree with this. Ghostty leaving GitHub isn’t just about outages - it’s a signal about how fragile the entire delivery layer has become. We’ve quietly turned GitHub into more than Git hosting: * PRs * CI/CD * reviews * releases * collaboration itself So when it’s unstable, it’s not “inconvenient” - it’s stoppage. And Mitchell’s point in the announcement is key: it’s not Git that’s the issue, it’s everything built around it now. This isn’t a “leave GitHub” take. It’s a reminder that we’ve centralized a lot more than we admit - and that deserves a second look.

GitHub losing Ghostty is not just open source drama. It is a warning shot for developer infrastructure. Mitchell Hashimoto says Ghostty is leaving GitHub after months of reliability issues, including frequent outages affecting PR review, GitHub Actions, and day-to-day maintainer work. He also makes the key point: the problem is not Git itself. It is everything around Git that modern teams depend on now: issues, pull requests, CI, releases, identity, and community workflow. That's the real signal. For years, GitHub has been treated like a default utility. -> Not a vendor. -> Not a dependency. -> Not a platform risk. Just “where the code lives.” But modern software teams do not merely store code on GitHub anymore. They operate through it. When GitHub is down, the repo might still exist, but the factory floor stops: -> PRs wait. -> CI stalls. -> Reviews block. -> Releases slip. -> Maintainers lose momentum. The contrast is interesting. We spent years preaching distributed systems, multi-cloud, backups, failover, and durability. Then we centralized the entire software collaboration layer into one platform and called it convenience. This is not a “leave GitHub” post. GitHub is still one of the most important developer platforms ever built. But Ghostty leaving should make engineering leaders ask a serious question: -> What parts of our software delivery process are only “distributed” in theory?