mihazs

Lazysh — AI skill for Claude Code

AI community

Cross-platform SSH terminal and AI coding-agent workspace for Android, iOS, macOS, Linux and Windows.

How to install Lazysh

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

What Lazysh does

Cross-platform SSH terminal and AI coding-agent workspace for Android, iOS, macOS, Linux and Windows. Built with Rust, Tauri 2, React and xterm.js. In development.

Alternatives in AI

Open Science — Open Science Desktop — local-first, model-agnostic AI research workbench for macOS, Windows & Linux 1.5k ★
CLI Manager — A multi-project AI CLI workspace for local terminals, SSH hosts, and mobile-assisted workflows 749 ★
Memory Forge Rs — Stop resetting satisfying AI chats — edit the memory instead 474 ★

README

lazysh — Cross-platform SSH terminal and AI agent workspace

**lazysh** is a cross-platform SSH client and terminal workspace in development for **Android, iOS, macOS, Linux, and Windows**. It is designed for developers who want to connect directly to their own servers, work in remote shells, and manage long-running AI coding agents from desktop or mobile.

The project is built with **Rust, Tauri 2, React, and xterm.js**, using **Bun and Turborepo** for the application workspace. Direct SSH works with the user's own hosts; the companion and external services are optional.

**Project status:** Work in progress. The native implementation is available as a development preview. There is no stable release or verified cross-platform Moshi parity. See [implementation PR #22](https://github.com/mihazs/lazysh/pull/22), the [GitHub Project](https://github.com/users/mihazs/projects/2), and the [parity epic](https://github.com/mihazs/lazysh/issues/1) for test evidence and remaining work.

What is lazysh for?

SSH terminal on desktop and mobile: connect directly to remote SSH hosts using a shared native client.
Remote development and persistent sessions: use tmux, Herdr, and Zellij to keep long-running programs on the host when the client disconnects.
AI coding-agent workflows: view supported events and transcripts from host-side tools such as Claude Code, Codex, and OpenCode. Integration depth varies by agent.
Local-first connections: use SSH independently, with optional companion tools for agent and session workflows.

SSH and developer tools

The implementation includes password/private-key SSH, explicit server fingerprint verification, concurrent terminal tabs, output backpressure, and an optional [local agent companion](docs/companion.md). The workspace adds transcript readers, tmux/Herdr/Zellij controls, historical Git browsing and review comments, SFTP transfers, image attachments, SSH-forwarded previews, [terminal preferences and