cmux-agent-observer

Orchestrate codex, opencode, or shell workers in cmux and tmux panes.

Updated Mar 19, 2026
One-click install
npx skills add https://github.com/meengi07/cmux-agent-observer-skill --skill cmux-agent-observer
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cmux-agent-observer
Source: https://github.com/meengi07/cmux-agent-observer-skill/tree/main/cmux-agent-observer
Command: npx skills add https://github.com/meengi07/cmux-agent-observer-skill --skill cmux-agent-observer

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

Orchestrates Codex/OpenCode agent workflows by creating, observing, and merging results across multiple cmux panes, reducing manual session management and context switching.

Core Features & Use Cases

  • Automatically creates and arranges worker panes from a single coordinator pane using cmux and tmux.
  • Launches codex, opencode, or shell workers in dedicated panes with per-worker state and handoff notes.
  • Supports long-running sessions with tmux persistence and reattach capability, plus per-worker progress tracking.

Quick Start

Run the launcher script to create worker panes for codex or opencode.

Frequently Asked Questions about cmux-agent-observer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I coordinate multiple codex and opencode workers in tmux?

You can coordinate multiple codex and opencode workers in tmux by using a launcher script to automatically create, arrange, and observe dedicated worker panes from a single coordinator pane. This reduces manual session management and context switching.

Can I launch shell workers alongside codex or opencode in the same cmux workspace?

Yes, you can launch shell workers alongside codex or opencode within the same cmux workspace. The orchestrator supports launching codex, opencode, or shell workers in dedicated panes, each with per-worker state and handoff notes.

How does cmux handle reattaching to long-running AI agent sessions?

cmux handles reattaching to long-running AI agent sessions by leveraging tmux persistence and loading per-worker configuration from surface state and handoff artifacts. This supports state restoration and reattachment when you return.

What is the best way to merge outputs from multiple AI agent panes into a main workspace?

The best way to merge outputs from multiple AI agent panes is by using handoff notes. The orchestrator merges results across cmux panes into the main workspace by loading these handoff artifacts to track per-worker progress.

Do I need to manually configure tmux panes for opencode workers?

No, you do not need to manually configure tmux panes for opencode workers. The orchestrator automatically creates and arranges worker panes from a single coordinator pane using cmux and tmux, reducing manual configuration effort.