What are the recommended resource limits for multi-agent workflows?

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

Can I use this with tools other than Claude Code?

Yes, this skill supports various harnesses including Claude Code, Codex, OpenCode, Cline, Gemini, and Qwen, allowing for cross-harness workflows.

How does this skill prevent file conflicts during parallel work?

The skill utilizes Git worktree integration to isolate file changes into separate directory structures, allowing multiple agents to work on different branches of the same repository simultaneously without interference.

dmux is a tmux-based orchestration tool that manages multiple AI agent panes, allowing you to run and coordinate several agent sessions in parallel within a single terminal window.

dmux Multi-Agent Orchestration

Name: dmux Multi-Agent Orchestration
Author: saar120

bysaar120

0•

생산성 및 워크플로

Orchestrates parallel AI agent sessions using tmux to handle complex multi-track development workflows and task parallelization.

This skill enables advanced multi-agent workflows by leveraging dmux, a tmux-based pane manager designed specifically for AI agent harnesses. It allows developers to coordinate multiple Claude Code, Codex, or OpenCode sessions simultaneously, facilitating complex 'divide-and-conquer' strategies like parallel feature implementation, research-to-code pipelines, and automated test-fix loops. By integrating with Git worktrees, it ensures file isolation during parallel operations, making it an essential tool for scaling AI-driven development productivity across large and complex codebases.

주요 기능

01Automatic context merging and result aggregation

02Cross-harness support for Claude Code, Codex, and OpenCode

03Predefined patterns for research, implementation, and code review

04Native Git worktree integration for file conflict prevention

05Parallel agent session management via tmux panes

060 GitHub stars

사용 사례

01Setting up a 'Test + Fix' loop where one agent monitors failures while another applies fixes

02Executing large-scale refactors across multiple independent modules simultaneously

03Splitting research-heavy tasks from implementation to maximize speed and token efficiency

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add saar120/skillgate-registry dmux-workflows

For use in Claude.ai and ChatGPT

주요 기능

01Automatic context merging and result aggregation

02Cross-harness support for Claude Code, Codex, and OpenCode

03Predefined patterns for research, implementation, and code review

04Native Git worktree integration for file conflict prevention

05Parallel agent session management via tmux panes

060 GitHub stars

사용 사례

01Setting up a 'Test + Fix' loop where one agent monitors failures while another applies fixes

02Executing large-scale refactors across multiple independent modules simultaneously

03Splitting research-heavy tasks from implementation to maximize speed and token efficiency

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add saar120/skillgate-registry dmux-workflows

For use in Claude.ai and ChatGPT