# Team

> Orchestrate parallel tmux workers for a multi-agent team task.

- Skill: `yeachan-heo-oh-my-gemini/team` (Agent Skill)
- Install (CLI): `npx skillmds@latest add yeachan-heo-oh-my-gemini/team`
- Raw SKILL.md: https://api.skillmd.com/api/skills/yeachan-heo-oh-my-gemini/team/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: Yeachan-Heo (https://skillmd.com/u/yeachan-heo-oh-my-gemini)
- Updated: 2026-09-10
- Page: https://skillmd.com/skills/yeachan-heo-oh-my-gemini/team

---


# Team Skill (oh-my-gemini)

Use this skill when the user wants to run a multi-worker team orchestration task.

## What it does

Launches `omg team run` with the specified task and worker count, creating a tmux session with N parallel workers. Each worker receives the team context via GEMINI.md and executes the task, then writes a completion signal.

## Usage

```
omg team run --task "<task description>" --workers <N>
```

## Examples

- `omg team run --task "Review all PRs" --workers 3`
- `omg team run --task "Run smoke tests" --workers 2`
- `omg team run --task "Audit dependencies" --workers 4`

## Orchestration steps

1. Hook system writes team context to `.gemini/GEMINI.md`
2. tmux session is created with N worker panes
3. Each worker runs `omg worker run --team <name> --worker <id>`
4. Workers read context, execute task, write done signal to state
5. Orchestrator polls until all workers complete
6. Results aggregated and displayed

## Acceptance criteria

- [ ] tmux session created with N panes
- [ ] Each pane executes the worker run command
- [ ] Each worker writes a done signal (`.omg/state/team/<name>/workers/<id>/done.json`)
- [ ] `omg team status` reflects worker completion