For Those Who Know the Future of Engineering is Off Your Laptop

That's why I'm building this. Downloading, sharing, managing, editing, and updating skills across your repositories, your Obsidian vault, and all those used to manage your fleet of local machines and laptops is a pain in the ass.

But then, you still need to manage plugins, your AGENTS.md, and other permissions and .toml setup across your machines. But they don't all have the same config! Some must have these secrets, some may not. The same skill might need to say something different on a different machine.

Git was designed for a pre-AI era. Worktrees are cumbersome. And in practice, your agents should all be working as if they were on the same machine.

Context Nine's suite of tools allows your code folder to stay in sync while you work, with conflicts duplicated and made easily greppable. Your task as an AI engineer is to know what machines your task needs to run on, and have enough of an understanding so as not to cause conflicts while working on the same branch across multiple machines. However, your machines should still be able to talk, communicate with each other, and use computer use when needed. I've found that having a spare Mac for computer use and a Linux machine for large-scale changes and a primary Mac that you orchestrate this all from is the best setup, although the more machines the merrier!

I built all of these tools and context systems because there simply wasn't a way to do all this. The way we work as engineers has fundamentally changed. And I believe the solution to that is a well-defined system based on your own file system with lots of Markdown files.

I can't believe I'm typing this, but try it for yourself, and there'll be no going back.

I'm a huge K3s fan, and so a lot of the included skills relate to how I interact with Kubernetes. I haven't published everything yet publicly, but as things get more polished, I will make them available. This repo allows me to send a voice note on my phone and have my Linux worker build and deploy an entire tested application on Kubernetes live behind Cloudflare on my existing cheap cluster hosted on Hetzner while using Computer Use on my Mac Air.

To get things done without me being there. If it can be done on a computer, it can be done with a voice note. I use VoiceInk (I know, not WhisperFlow, ooohhh) with ElevenLabs, btw. V2/V3 is just so good at capturing technical terms.

Since I'm a Kubernetes fan, my context setup involves Helm charts and deployments. It can triage Grafana alerts into governed incidents, publish PR files and readable HTML or Markdown through Publisher, run evidence-driven agents, and manage your secrets per repository with global secrets propagated into your working repositories.