A powerful terminal multiplexer controller that enables programmatic management of workspaces, surfaces, and notifications via a Unix socket API.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install cmux
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
Copy this prompt to OpenClaw to install it automatically.
Help me install cmux using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
The cmux skill provides a bridge for AI agents and developers to interact with the cmux terminal multiplexer programmatically. By utilizing a Unix socket or the dedicated CLI, this tool allows for the dynamic creation of workspaces, complex window splitting, and precise input injection. It is an essential component for users building sophisticated environments within Openclaw Skills to manage multi-terminal workflows without manual intervention.
This skill excels at bridging the gap between automated scripts and the user's visual terminal environment. Whether you are managing complex microservices or simple command-line tasks, cmux ensures that your workspace remains organized and actionable through high-level abstractions for terminal surfaces and metadata.
Ensure the cmux CLI is installed and the daemon is running. The default socket is typically located at /tmp/cmux.sock. You can verify your environment using the following commands:
# Check if cmux is available
command -v cmux &> /dev/null && echo "CLI ready"
# Check for the active socket
[ -S "/tmp/cmux.sock" ] && echo "Socket ready"
To use a custom socket path, export the CMUX_SOCKET_PATH environment variable before initiating your Openclaw Skills session.
The skill communicates using a structured JSON-RPC format. Below is the metadata taxonomy used for organizing terminal states:
| Component | Description | Primary Fields |
|---|---|---|
| Workspace | A high-level container for terminal sessions. | id, name, visibility |
| Surface | An individual terminal pane or split. | id, workspace_id, focus_state |
| Notification | Ephemeral messages displayed in the UI. | title, body |
| Sidebar | Metadata for tracking process health. | status_key, progress_value, log_level |
Loading
A protocol for AI agents to automatically trigger and execute installed skills based on conversation context and pre-defined rules.

A professional tool for analyzing Xiaohongshu and Douyin content and profiles to extract engagement metrics and viral strategies.

An intelligent development assistant that automates code analysis, detects vulnerabilities, and generates comprehensive Pull Request documentation.

A cross-platform social media distribution engine that automates content adaptation and publishing for major social networks.

A high-speed deployment utility for scaffolding and shipping static websites to production platforms in under five minutes.

A professional-grade framework for engineering high-performance AI prompts using proven structural methodologies like RACE and Chain-of-Thought.








































