flywheel-swarm

Coordinate multi-agent coding swarms across Claude Code, Codex, and Gemini.

Updated Mar 18, 2026
One-click install
npx skills add https://github.com/quangdang46/flywheel-skills --skill flywheel-swarm
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: flywheel-swarm
Source: https://github.com/quangdang46/flywheel-skills/tree/main/flywheel-swarm
Command: npx skills add https://github.com/quangdang46/flywheel-skills --skill flywheel-swarm

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Launch and tend a coordinated swarm of coding agents to enable parallel exploration, coordination, and governance over complex AI-driven tasks.

Core Features & Use Cases

  • Orchestrate multiple agents (e.g., Claude Code, Codex, Gemini) in parallel using br, bv, and Agent Mail to ensure deterministic progress.
  • Enforce pre-launch checks, staggered starts, phase-based model selection, and AGENTS.md compliance to reduce coordination failures.
  • Manage stuck agents, run review and hardening rounds, and maintain an auditable bead/Agent Mail workflow.

Quick Start

Run the pre-launch checklist, stagger agent starts, and monitor Agent Mail to coordinate and sustain the swarm.

Frequently Asked Questions about flywheel-swarm

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

FAQPage Schema
How do I coordinate multiple AI coding agents to work in parallel?

You coordinate a multi-agent coding swarm by launching agents like Claude Code, Codex, and Gemini in parallel, using Agent Mail and bead workflows to ensure deterministic task routing and synchronized progress across all active agents.

What is Agent Mail and how does it help with multi-agent orchestration?

Agent Mail is a coordination protocol used to manage communication across a parallel coding swarm, enabling deterministic task routing, monitoring for stuck agents, and facilitating review and hardening rounds to maintain auditable progress.

Can I run Claude Code, Codex, and Gemini simultaneously without coordination failures?

Yes, you can run these models simultaneously by enforcing pre-launch checks, staggered starts, and phase-based model selection, alongside AGENTS.md compliance to significantly reduce coordination failures across the swarm.

How do I handle stuck agents during a parallel coding task?

To handle stuck agents during parallel coding tasks, you monitor the Agent Mail workflow to detect blockages, then apply swarm management techniques to resolve dependencies and run review or hardening rounds to restore progress.

What's the best way to enforce governance and pre-launch checks for AI agents?

The best way to enforce governance is by applying AGENTS.md compliance requirements and running a pre-launch checklist that verifies phase-based model selection and staggered starts before initiating the parallel agent swarm.

Do I need AGENTS.md to manage a multi-agent coding swarm?

Yes, AGENTS.md is required to establish governance over the swarm, ensuring deterministic task routing, phase-based model selection, and compliance checks are applied uniformly across Claude Code, Codex, and Gemini agents.