netstack

SSH + Kiro CLI Workflow

Connect to a federation dev node via SSH and use Kiro CLI for AI-assisted development. No GUI needed — pure terminal, zero lag.

Why

RDP/VNC adds video encoding latency (especially over ZeroTier)
SSH is text-only — instant response, works on any connection quality
Kiro CLI provides the same AI assistant capabilities as the full IDE
Works from any device with an SSH client (laptop, tablet, phone)

Prerequisites

ZeroTier installed and joined to federation network on your client device
SSH client (built into Windows 10+, macOS, Linux)
Target node has OpenSSH server + Kiro CLI installed

Quick Connect

# Connect to nsdockerhv (primary dev node)
ssh nsadmin@10.147.17.176

# Start Kiro CLI chat (from the SSH session)
cd ~/code/wip
kiro-cli chat

One-liner (direct to Kiro chat)

ssh nsadmin@10.147.17.176 -t "cd ~/code/wip && kiro-cli chat"

Setup: SSH Key Auth (passwordless)

From Windows (PowerShell)

# Generate key (if you don't have one)
ssh-keygen -t ed25519

# Copy public key to target node
type $env:USERPROFILE\.ssh\id_ed25519.pub | ssh nsadmin@10.147.17.176 "cat >> ~/.ssh/authorized_keys"

# Test (should connect without password)
ssh nsadmin@10.147.17.176

From WSL / Linux / macOS

# Generate key (if you don't have one)
ssh-keygen -t ed25519

# Copy public key to target node
ssh-copy-id nsadmin@10.147.17.176

# Test
ssh nsadmin@10.147.17.176

Setup: Windows Terminal Profile (optional)

Add to Windows Terminal settings.json for one-click access:

{
    "name": "Wip (nsdockerhv)",
    "commandline": "ssh nsadmin@10.147.17.176 -t \"cd ~/code/wip && kiro-cli chat\"",
    "icon": "🤖"
}

Kiro CLI Commands

Once in kiro-cli chat:

Command	What
(just type)	Chat with Kiro (same as IDE chat)
`/quit`	Exit chat
`/clear`	Clear conversation

Other kiro-cli subcommands:

kiro-cli chat          # Start chat session
kiro-cli agent         # Manage AI agents
kiro-cli doctor        # Debug installation issues
kiro-cli settings      # Customize behavior

Federation Node SSH Addresses

Node	ZeroTier IP	User	Purpose
nsdockerhv	10.147.17.176	nsadmin	Primary dev (Wip, Docker, Gitea)
CyberTruck	10.147.17.219	ghadmin	Host (Hyper-V, Plex, backups)
cat9fin	10.147.17.218	nsadmin	HWPC production (hwpc-rp, QB)
slwin11ops	10.147.17.94	ghadmin	SL ops (backup target)

Troubleshooting

Problem	Fix
Connection refused	Verify ZeroTier is connected: `zerotier-cli status`
Permission denied	Copy your SSH public key (see Setup above)
kiro-cli not found	Add to PATH: `export PATH="$HOME/.local/bin:$PATH"`
Session drops	Use `tmux` or `screen` for persistent sessions

Persistent Sessions with tmux

# Start a named session
ssh nsadmin@10.147.17.176 -t "tmux new-session -A -s wip"

# Inside tmux, start Kiro
cd ~/code/wip && kiro-cli chat

# Detach: Ctrl+B then D
# Reconnect later — session persists even if SSH drops
ssh nsadmin@10.147.17.176 -t "tmux attach -t wip"

remote-desktop.md — GUI remote access (when you need it)
../backup/windows-openssh-setup.md — enabling SSH on Windows nodes
contributor-setup.md — full dev environment setup

Mobile Access: Termius (iOS/Android)

Termius is an SSH client for mobile devices that supports:

Saved host profiles (one-tap connect)
Snippets (saved commands)
SFTP file browsing
Siri Shortcuts integration (voice-triggered SSH sessions)

Setup

Install Termius from App Store / Play Store
Add host:
- Label: nsdockerhv (Wip)
- Hostname: 10.147.17.165 (requires ZeroTier on phone)
- Port: 22
- Username: nsadmin
- Auth: SSH key (import or generate in Termius)
Save and test connection

Siri Integration (iOS)

Termius supports Siri Shortcuts. You can create a shortcut that:

Opens Termius
Connects to nsdockerhv
Runs a command (e.g., starts tmux session)

Setup:

Open iOS Shortcuts app
Create new shortcut: “Hey Siri, open Wip”
Action: Open Termius → Connect to saved host “nsdockerhv (Wip)”

Voice paste pattern:

Dictate to Siri → text goes to clipboard
Paste into Termius terminal (SSH session)
This enables voice-to-terminal input from phone

Alternative: SSH from iOS without Termius

iOS has no built-in SSH client, but options include:

Termius (recommended — best UX, Siri support)
Blink Shell (power user, mosh support)
a]Shell (free, basic)

tmux Best Practices for Wip Sessions

Named Sessions

# Create/attach to named session
tmux new-session -A -s wip

# Multiple sessions for different contexts
tmux new-session -A -s wip      # Wip coordination
tmux new-session -A -s dev      # Development work
tmux new-session -A -s ops      # Ops/monitoring

Useful tmux Commands

Key	Action
`Ctrl+B d`	Detach (session keeps running)
`Ctrl+B c`	New window
`Ctrl+B n`	Next window
`Ctrl+B p`	Previous window
`Ctrl+B "`	Split horizontal
`Ctrl+B %`	Split vertical
`Ctrl+B [`	Scroll mode (q to exit)

Reconnect After Disconnect

# From any device (phone, laptop, tablet):
ssh nsadmin@10.147.17.176 -t "tmux attach -t wip"

# If session doesn't exist, create it:
ssh nsadmin@10.147.17.176 -t "tmux new-session -A -s wip"

Why tmux + SSH > RDP

	RDP	SSH + tmux
Bandwidth	High (video stream)	Minimal (text only)
Latency feel	Laggy on slow links	Instant
Session persistence	Lost on disconnect	Survives disconnect
Mobile	Clunky (small screen + desktop UI)	Natural (terminal fits any screen)
Voice input	No	Yes (Siri → paste into terminal)
Multiple sessions	One desktop	Many tmux windows