CodeScene’s Post

🤖 The future of coding is here: #AI agents working behind the scenes while you build apps just by describing what you want. Dive into how AI is transforming the software development lifecycle: https://bit.ly/3WheTAL

Today's probably-obvious AI-related tip for engineers: Claude Code and other agents are good at more than just writing code - they're good at doing syseng/sysad stuff too! I save so much time with my homelab just having my AI agent do the system setup for me. I'm not just gonna let it run wild (just like I dont with my code), but as usual it is great at taking the grunt work out of building.

I'm way more excited about Claude Skills than MCP, and here's why: They're the runbook AND the implementation, readable by both humans and AI agents. We just ran an internal training at WorkOS where I packaged up a complex two-step image generation workflow: (static → animated) as a single skill. My colleagues went from "I don't know how this works" to using it in production the same afternoon. No tribal knowledge locked in DMs. No outdated docs. No rediscovery cycles when someone leaves the team. Just clone the repo, add your API key, restart Claude. The workflow you spent weeks figuring out? Now the whole team has it. Version controlled. PR-reviewable. CI-testable. The same patterns you already use for code, now applied to AI workflows. I wrote about the technical approach, the 12-factor patterns we followed, and why GitHub distribution makes perfect sense for team AI workflows. Link in comments 👇

🚫 Your AI coding assistants forget everything. Every chat starts cold. The context lives in your head—not in your tools. That kills flow. Wastes hours. Introduces bugs. ✅ Archon OS changes that. It gives your AI agents shared memory—so they remember your codebase, architecture decisions, and patterns across sessions. No more re-explaining. Just build and ship. ⚙️ What it is An open-source OS for AI agents: Microservices with React + Vite on the frontend FastAPI, Supabase, and a custom MCP server on the backend Works with both local and cloud models Fully customizable ✅ Proof Running in my stack with: Cursor Claude Code Windsurf Context handover is near-instant. Agents collaborate without memory loss. 🎥 Try it Watch the deep dive: World of AI video Codebase: GitHub Want my setup notes or a quick Loom walkthrough? Just ask. 👇 #AI #AIEngineering #SoftwareDevelopment #OpenSource #AIAgents #AgenticAI #RAG #DeveloperTools #ModelContextProtocol #MCP #KnowledgeManagement #AIAutomation #LLM #AIWorkflow #SoftwareEngineering

Run Claude Code, Codex, and Gemini in parallel on one canvas. This tool manages your AI coding sessions with spatial boards, git worktrees, and real-time team collaboration. 100% open-source. More such AI tools and projects in theunwindai.com: Get access to 100+ AI Agent, RAG, LLM, and MCP tutorials with open-source code - All for FREE. Check it out: https://lnkd.in/dHcR73QK

Dev teams: stop hardcoding every AI call like it’s 2022. cascadeflow is here: Lemony’s open-source model router that figures out the cheapest, fastest path for every query (so you don’t have to). ✅ Dynamic model selection ✅ Real-time cost tracking ✅ Multi-provider flexibility ✅ Full transparency baked in Tell your devs. Star it. Fork it. Run with it. 🍋 Now on GitHub → https://lnkd.in/eYm5b2P7

You can now generate a complete visual walkthrough of any codebase and save it to README in less than 5 minutes. Why? You will be able to understand complex software logic up to 3x faster than asking AI agents, relying on documentation, or tracing lines of code. Step 1: Paste any GitHub repo link (public or private) at code-canvas.com Step 2: Get a complete visual map of the codebase's business use-cases to their underlying source code that implements them. Step 3: Save the interactive diagram to your README #AI #DeveloperTools #Codex #Cursor

We've been exploring what it looks like when AI fits naturally into the tools you already use, like GitHub, without changing your workflow. In this demo, Jules runs entirely inside GitHub. Add a "jules" label to any issue, and it handles the rest: understanding context, writing and reviewing code, validating results, and handing you a PR. In the demo I add an issue in the jules-invaders repo to update the easter egg on julesinvaders.com. Check it out: https://lnkd.in/gwaCkDm4

Jules (from Google) is a “Yolo-mode”* assistant that doesn’t just code but acts as an extra pair of hands and brain to go do the harder part of maintaining collaborative code bases for apps etc known as repositories. Now summoning Jules is easier than ever by just sticking a label on stuff in the repository. *YOLO-mode, or “you-only-live-once” is a term Google Gemini CLI uses to describe going fully-automated mode. That is, these code assistants operate without asking you at every step a yes/yes this once/no pause. Have you ever asked someone if you can leave the meeting to go to the restroom? Not since grade school right? YOLO mode is sort of like graduating from grade school when you believe the Agent is responsible enough and has safeguards to go do their business and come back and/or you’ve got enough hall monitors to make sure they’re observable and doing ok.

🚀 Elevate your projects with Machine Learning! Our introductory guide makes it accessible for all professionals—no coding required. Discover how to seamlessly integrate ML into your business operations and drive innovation. Embrace the future with ease. #MachineLearning #NoCode #Innovation