team

Orchestrates parallel Codex and Claude CLI workers in tmux panes with shared task state.

Updated Feb 1, 2024
One-click install
npx skills add https://github.com/ogiboy/portfolio-app --skill team-ogiboy
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: team
Source: https://github.com/ogiboy/portfolio-app/tree/main/.codex/skills/team
Command: npx skills add https://github.com/ogiboy/portfolio-app --skill team-ogiboy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Coordinating multiple AI coding agents on a shared task list is error-prone when done through ad-hoc terminal typing. This Skill provides a structured operator workflow for launching, monitoring, and shutting down N parallel Codex or Claude CLI workers in tmux panes, coordinated through durable state files and mailbox-based messaging instead of fragile keystroke injection. ## Core Features & Use Cases - Parallel worker orchestration: Launch N workers with role prompts via omx team N:agent-type "<task>", with per-worker CLI selection (Codex or Claude) and optional git worktree isolation. - State-first coordination: Tasks, mailboxes, heartbeats, and dispatch queues live under .omx/state/team/<team>/, mutated through the omx team api ... --json CLI rather than direct tmux send-keys. - Lifecycle contract: Enforces startup verification, active leader monitoring loops, terminal-state gating before shutdown, and clean-slate recovery for stale panes. - Use Case: A developer needs three workers to analyze a feature, fix bugs, and run verification in parallel. They run omx team 3:executor "analyze feature X and report flaws", monitor with omx team status, and shut down only after all tasks reach terminal state. ## Quick Start Ask the agent to launch a coordinated team by running omx team with a worker count, role, and task description, then monitor progress with omx team status until all tasks complete.

Frequently Asked Questions about team

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

FAQPage Schema
How do I run multiple AI coding agents in parallel with tmux?

Use the omx team command with a worker count, role, and task, such as omx team 3:executor "analyze feature X". It splits the current tmux window into worker panes running independent Codex or Claude CLI sessions coordinated through shared state files.

When should I use omx team instead of native subagents?

Use native subagents for bounded in-session parallelism where one leader fans out a few independent subtasks. Use omx team when you need durable tmux workers, shared task state, mailbox coordination, worktrees, or long-running execution that survives beyond one reasoning burst.

Can I mix Codex and Claude workers in the same team?

Yes, set OMX_TEAM_WORKER_CLI_MAP with comma-separated values like codex,claude to assign a CLI per worker index. OMX_TEAM_WORKER_CLI=claude forces all workers to Claude, and auto mode selects Claude when the worker model argument contains claude.

Why does omx team api return ENOENT errors from workers?

ENOENT means the team state directory was removed while workers were still running, typically because shutdown ran before tasks reached terminal state. Enforce the completion gate of zero pending and in-progress tasks before running omx team shutdown.

What are the prerequisites for launching an omx team?

You need tmux installed, the leader session running inside tmux, the omx command resolving to the intended build, and a context snapshot in .omx/context before launch. Duplicate hud --watch panes should also be removed before splitting worker panes.