orchestrate

Launch AI agents in tmux sessions and produce merge-ready reports.

28|5|Updated Feb 3, 2026
One-click install
npx skills add https://github.com/leonletto/thrum --skill orchestrate-leonletto
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: orchestrate
Source: https://github.com/leonletto/thrum/tree/main/codex-plugin/skills/orchestrate
Command: npx skills add https://github.com/leonletto/thrum --skill orchestrate-leonletto

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Coordinated execution of complex plans by launching and supervising multiple AI agents across worktrees, enabling gated approvals and consolidated reports.

Core Features & Use Cases

  • Phase-driven execution: validate handoffs, configure autonomy, plan worktrees, launch agents, and execute epics with review gates.
  • Parallel epics and gated reviews: run independent epics concurrently where possible and require human gates before progression.
  • Finalize and merge: compile a cross-branch merge report and guide safe integration.

Quick Start

Initiate an orchestration by loading the handoff plan and letting the orchestrator launch and manage agents across worktrees.

Frequently Asked Questions about orchestrate

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

FAQPage Schema
How do I coordinate multiple AI agents for plan execution?

You can coordinate multiple AI agents for plan execution by loading a multi-epic handoff plan into an orchestrator that launches agents in separate tmux sessions. This enables supervised execution across worktrees with gated reviews before progression.

How do I manage parallel epic execution across different git worktrees?

Parallel epic execution across git worktrees is managed by launching independent AI agents concurrently in separate tmux sessions. The orchestrator supervises these agents, validating handoffs and requiring human review gates before allowing further progression.

Can I use tmux to run autonomous AI agents for software engineering tasks?

Yes, you can use tmux to run autonomous AI agents by launching them in dedicated sessions configured with negotiated autonomy levels. The orchestrator manages these tmux sessions to execute software engineering epics and validate phase handoffs.

What is the best way to finalize and merge cross-branch reports from AI agents?

The best way to finalize and merge cross-branch reports is to use an orchestrator that compiles a consolidated merge report after all agents complete their epics. This guides safe integration of work across multiple branches.

How do I implement gated reviews for an orchestrated AI project workflow?

Gated reviews for an orchestrated AI project workflow are implemented by configuring the orchestrator to require human approval before progressing between phases. This ensures phase-by-phase validation is met before executing subsequent epics.

What are the limitations of automating plan execution with AI agents across worktrees?

A key limitation of automating plan execution across worktrees is the necessity of human intervention at configured review gates. The orchestrator cannot bypass these gates, requiring manual validation of handoffs before finalizing cross-branch merges.