Hi PH! I built a K8s collaboration tool for debugging — and it actually works

Hey Product Hunt! 👋

I'm an independent developer from China. Before I quit my job, I worked with Kubernetes every single day. Every time an incident happened, our internal communication channels were flooded with links and screenshots. Online meetings turned into chaos — everyone fighting for the mic, trying to explain what they were seeing. It was painful.

So I built Kikimora — a Kubernetes cluster management tool that enables in-subnet collaboration for troubleshooting. It solves one specific problem: sharing debugging context instantly.

Here's the demo:

What it does:

1. Select any resource in your cluster — logs, status, or any critical debugging info. Pick the teammate you want to notify, and click "Share".

2. Everyone in the same subnet receives the notification instantly. One click, and they jump exactly to the resource you highlighted.

No more "which namespace?" or "let me share my screen". Just point, click, and share.

Why I built it:
I'm a big fan of Lens — it's an incredible tool. But every time my team needed to collaborate on an incident, we were back to copy-pasting and screenshots. I wanted something lightweight, focused, and built specifically for the "collaborative troubleshooting" scenario — a tool that makes information sync effortless.

Right now a (rough) MVP is live on the Mac App Store [1]. I'd really love to get feedback from this community — especially from SREs and platform engineers. What's your biggest pain point when troubleshooting with your team?

Here's what I've built and what's coming next:

1. ✅ Done — Sync kubeconfig from Infisical / Vault

   

2. ✅ Done — Local kubeconfig version management (One thing I realized: currently the version tracking is tightly coupled with Kikimora's own operations. If we decouple it into a separate file watcher that continuously monitors the kubeconfig, we could automatically version changes made by any tool — not just Kikimora.)

   

3. 🔄 Under testing — AWS EKS cluster support

4. 🔄 Under testing — MCP (Model Context Protocol) support, allowing custom AI agents to assist with troubleshooting (read‑only)

5. 📅 Planned — Open‑source reference implementation of a centralized server, with an Agent that automatically archives troubleshooting sessions into a knowledge base that grows with your project

Would love to hear your thoughts!