Verified:GitHub

dmux-workflows

Orchestrate parallel AI agent sessions with dmux and merge pane outputs.

Updated Sep 13, 2025

Search Tags:#tmux #git worktree #multi-agent orchestration #agent harness #dmux #parallel workflows #code review pipeline

Ready to run

npx skills add https://github.com/llmh333/employee_management_spring --skill dmux-workflows-llmh333

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: dmux-workflows
Source: https://github.com/llmh333/employee_management_spring/tree/main/.gemini/skills/dmux-workflows
Command: npx skills add https://github.com/llmh333/employee_management_spring --skill dmux-workflows-llmh333

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

dmux-workflows helps you run multiple AI agent tasks in parallel without losing track of pane outputs, reduces coordination overhead, and minimizes merge conflicts when working across overlapping files.

Core Features & Use Cases

Parallel multi-agent orchestration: Split work into independent panes and merge results back into a main session for faster completion.
Cross-harness execution: Coordinate agent sessions across Gemini CLI, Codex, OpenCode, Cline, Gemini, and Qwen to match each task to the best harness.
Conflict-safe workflows: Use git worktrees to isolate pane changes when different agents touch the same or overlapping paths.
Workflow patterns: Apply repeatable strategies like research + implement, multi-file feature development, and test + fix loops.

Quick Start

Start a dmux session and create multiple agent panes for separate prompts, then merge pane outputs back into the main session as each pane completes.

Frequently Asked Questions about dmux-workflows

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

FAQPage Schema

How do I run parallel AI agent sessions in tmux without losing track of pane outputs?▼

You can orchestrate parallel AI agent sessions using tmux-based pane management to split independent workloads, then merge pane outputs back into a main session for consolidated results.

Can I coordinate multiple AI coding agents across different harnesses like Gemini CLI and Codex?▼

Yes, cross-harness execution coordinates agent sessions across Gemini CLI, Codex, OpenCode, Cline, Gemini, and Qwen to match specific task requirements with the best available harness.

How do I prevent merge conflicts when multiple AI agents touch overlapping files?▼

You can prevent merge conflicts by using git worktrees to isolate pane changes, ensuring that different agents working on overlapping paths remain separated until merged.

What is the best way to structure a multi-agent workflow for research and implementation?▼

The best approach applies repeatable workflow patterns like research plus implement, multi-file feature development, and test plus fix loops to split and consolidate independent tasks.

Do I need tmux to manage parallel development workflows with multiple agents?▼

Yes, this parallel multi-agent orchestration requires tmux-based pane management to create separate agent panes and execute merge workflows for consolidating completed outputs.

Why does coordinating multiple AI agents manually lead to merge conflicts and lost outputs?▼

Manual coordination lacks isolated pane management and consolidated merge workflows, causing overlapping file edits and lost terminal outputs when running multiple agents simultaneously.