🔥 Hot Repo: oh-my-claudecode Brings 32-Agent Teams to Claude Code
oh-my-claudecode added 1,400 GitHub stars in a day. The TypeScript plugin turns Claude Code into a multi-agent pipeline with 32 specialized agents, smart model routing, and tmux-based Codex/Gemini workers — installed in two commands.
By OMC Editorial on 2026-03-27
oh-my-claudecode OMC added 1,400 stars in a single day on March 27, 2026 — a notable surge for a repo already sitting at 13,500 stars — as developers discover its v4.9.1 release from three days prior. The TypeScript framework adds multi-agent orchestration to Claude Code with zero configuration, and has become the most-starred third-party plugin for Claude Code since its January 2026 launch.
What it does
OMC installs directly from the Claude Code plugin marketplace:
bash
/plugin marketplace add https://github.com/Yeachan-Heo/oh-my-claudecode
/plugin install oh-my-claudecode
Once installed, it replaces single-agent Claude Code with a staged pipeline spanning planning, spec writing, parallel execution, verification, and fix loops. The canonical surface is its team command:
bash
/team 3:executor "fix all TypeScript errors"
That fires a five-stage pipeline — team-plan → team-prd → team-exec → team-verify → team-fix — with three Claude agents running in parallel during the execution phase. The pipeline loops on failures until the task passes verification.
Multi-provider workers via tmux
Version 4.4.0 February 2026 added support for spawning Codex and Gemini CLI processes as first-class team members inside tmux panes:
bash
omc team 2:codex "review auth module for security issues"
omc team 2:gemini "redesign UI components for accessibility"
omc team 1:claude "implement the payment flow"
Workers spawn on demand and terminate when their task completes. The /ccg skill routes a single prompt to both Codex and Gemini and has Claude synthesize both responses — useful for tasks that benefit from two code-review perspectives at once.
32 specialized agents and smart model routing
OMC ships with 32 specialized agents covering architecture, research, design, testing, and data science. A built-in model router sends simple tasks to Haiku and complex reasoning to Opus, which the project claims saves 30–50% on token costs compared to routing everything to the most capab