hbasheer

Yardbell — AI skill for Claude Code

AI community

One self-hosted page for AI coding agent sessions across all your machines: Claude Code, Codex, Gemini CLI, opencode, Aider and more.

How to install Yardbell

This entry records only its repository, not the path inside it, so there is no exact command to give. Open hbasheer/yardbell and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Yardbell does

One self-hosted page for AI coding agent sessions across all your machines: Claude Code, Codex, Gemini CLI, opencode, Aider and more. Health, sessions, tunnelled app ports, and a terminal from any device. Nothing installed on the machines it manages.

Alternatives in AI

  • Claude Squad — by smtg-ai - Claude Squad is a terminal app that manages multiple Claude Code, Codex (and other local agents i 6.5k ★
  • Chat Ollama — ChatOllama is an open source agentic app for running AI agents across local and hosted models 3.5k ★
  • Omnara — by Ishaan Sehgal - A command center for AI agents that syncs Claude Code sessions across terminal, web, and mo 2.6k ★

README

Yardbell

**One page for every AI coding agent session across all your machines.**

Yardbell runs on one machine and watches the others over ssh. It shows which machines are up and how much memory they have left, which agent session (Claude Code, Codex, Gemini CLI, opencode, Aider and more) runs where and for which ticket, and which local port opens the app each session is working on. It opens a full terminal into any session from your desktop or your phone, creates a git worktree per branch, and adds new machines without hand-editing ssh config.

Nothing is installed on the machines it manages. It is one Rust binary with a plain HTML UI, and a [desktop app](#apps) for macOS and Linux adds a tray icon and a notification when an agent needs you.

![Yardbell dashboard](docs/screenshots/dashboard-dark.png)

Why

If you spread work across a desktop, a laptop and a Linux box (usually because Docker stacks and several agent sessions eat all the RAM on one machine), you lose track quickly:

  • Is that machine even connected right now, and does it have memory to spare?
  • Which terminal tab is the session for ENG-142, and on which machine?
  • Which port do I open to test that branch?
  • How do I check on a long-running session from my phone?

Yardbell answers all of these on one page.

Features

  • Machine health first. Connection state (mesh, ssh, snapshot) with the reason when something fails, memory drawn to scale with the agents' own share, CPU load, and mesh VPN status.
  • Every agent session. Claude Code, Codex, Gemini CLI, Qwen Code, opencode, Copilot CLI, Cursor Agent, Crush, Goose and Aider out of the box, and any other CLI agent you add in the config. Machine, agent, folder, ticket (read from the branch name), branch, working or idle, and whether it can be opened. Sessions on unreachable machines stay visible as "last known".
  • Ports, already tunnelled. Docker Compose service ports from each session's folder are forwarded to the host over