devpod-gh

Quality of life plugin.

I like vibe coding and yolo mode, this is to make using local devcontainers easier and more seamless for "safer" yolo.

I get GitHub Copilot through work, so getting "automatic" sign in to me is important. I have been learning neovim (and am use to vscode), so auto port forwarding.

This is a simple function wrapper around devpod which checks for devpod and github cli.

Dependencies

• devpod - Development container management
• github cli - GitHub authentication and token generation
• gum - Interactive workspace selection
• jq - JSON processing for port monitoring
• ssh/scp - Remote connection and file transfer
• stdbuf - Unbuffered output handling for port monitoring

Installation

Oh My Zsh

Clone this repository into your Oh My Zsh custom plugins directory:

git clone https://github.com/scaryrawr/devpod-gh.zsh ${ZSH_CUSTOM:-~/.oh-my-zsh/custom}/plugins/devpod-gh

Then add devpod-gh to your plugins array in ~/.zshrc:

plugins=(... devpod-gh)

Antidote

antidote install scaryrawr/devpod-gh.zsh

Configuration

SSH ControlMaster (Required for optimal performance)

Configure SSH ControlMaster in your ~/.ssh/config for optimal port forwarding performance:

Host *
  ControlMaster auto
  ControlPath ~/.ssh/cm-%C
  ControlPersist 10m

This enables connection multiplexing, which allows multiple SSH connections to share a single network connection. This significantly speeds up establishing new port forwards by avoiding repeated SSH handshakes and authentication.

Features

Automatic GitHub Token Injection

When you ssh into a devpod using devpod ssh:

devpod ssh

The wrapper automatically injects environment variables:

• GH_TOKEN: Using gh auth token
• GH_COPILOT_TOKEN: Automatically extracted from ~/.config/github-copilot/apps.json if available

devpod ssh --set-env GH_TOKEN=`gh auth token` --set-env GH_COPILOT_TOKEN=<token>

This enables things like github copilot cli and github cli to just work... automagically, in devpods.

Automatic Port Forwarding

The plugin automatically monitors and forwards ports that are bound inside your devpod workspace. When an application starts listening on a port inside the devpod, it will be automatically forwarded to your local machine on the same port.

This uses a background port monitoring process that watches for port binding events and establishes SSH tunnels as needed. The port forwarding is cleaned up automatically when you disconnect.

Automatic Reverse Port Forwarding (Self-hosted LLMs)

The plugin automatically detects and reverse-forwards local LLM services to your devpod:

• Port 1234: LM Studio
• Port 11434: Ollama

When these services are running locally, they become accessible inside your devpod, allowing you to use self-hosted AI models in your remote development environment.