Can I use different AI models simultaneously with this skill?

Yes, dmux-workflows allows you to run different harnesses like Claude Code, Codex, and OpenCode in separate panes at once to leverage their specific strengths.

What is the 'Merge' feature in dmux-workflows?

The merge feature allows you to capture the output and findings from a specific agent's pane and bring that context back into your main development session.

Is there a recommended limit to parallel agent sessions?

To manage API token usage and system resources effectively, it is recommended to keep the total number of parallel panes between 5 and 6.

What is dmux in the context of AI agents?

dmux is a tmux-based pane manager designed specifically for orchestrating multiple AI agent sessions, such as Claude Code and Codex, within a single terminal environment.

How do I handle file conflicts when running multiple agents?

The skill recommends using Git worktrees to isolate file changes per pane, ensuring that agents working on overlapping files do not overwrite each other's work.

dmux Multi-Agent Workflows

Name: dmux Multi-Agent Workflows
Author: affaan-m

byaffaan-m

•

156,032

•

Productivity & Workflow

Orchestrates parallel AI agent sessions using dmux to streamline complex, multi-track development tasks across different harnesses.

dmux-workflows provides a structured framework for running multiple AI agent sessions simultaneously within a tmux-managed environment. It enables developers to divide complex tasks—such as simultaneous frontend and backend development or parallel research and implementation—across different harnesses like Claude Code and Codex. By utilizing patterns like the 'Test + Fix Loop' and 'Cross-Harness' execution, this skill optimizes developer throughput and prevents context switching during intensive multi-file operations.

Key Features

01Integrated output merging to consolidate results from multiple panes

02156,032 GitHub stars

03Support for Git worktree isolation to prevent file conflicts during parallel work

04Cross-harness orchestration for leveraging different AI models for specific tasks

05Standardized patterns for research, implementation, and code review

06Parallel execution of Claude Code, Codex, and other agent harnesses

Use Cases

01Executing a 'Research + Implement' loop where one agent gathers documentation while another builds

02Building full-stack features by running frontend, backend, and database agents in parallel panes

03Running a continuous test watcher in one pane while a fixing agent iterates in another

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add affaan-m/everything-claude-code dmux-workflows

For use in Claude.ai and ChatGPT

Key Features

01Integrated output merging to consolidate results from multiple panes

02156,032 GitHub stars

03Support for Git worktree isolation to prevent file conflicts during parallel work

04Cross-harness orchestration for leveraging different AI models for specific tasks

05Standardized patterns for research, implementation, and code review

06Parallel execution of Claude Code, Codex, and other agent harnesses

Use Cases

01Executing a 'Research + Implement' loop where one agent gathers documentation while another builds

02Building full-stack features by running frontend, backend, and database agents in parallel panes

03Running a continuous test watcher in one pane while a fixing agent iterates in another

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add affaan-m/everything-claude-code dmux-workflows

For use in Claude.ai and ChatGPT