swarm

Coordinate four agents to plan, build features via TDD, and harden code across isolated git worktrees.

2|Updated Feb 23, 2026
One-click install
npx skills add https://github.com/jonmumm/skills --skill swarm-jonmumm
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: swarm
Source: https://github.com/jonmumm/skills/tree/main/swarm
Command: npx skills add https://github.com/jonmumm/skills --skill swarm-jonmumm

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Coordinates a multi-agent workflow to plan, implement, test, and harden software in parallel using isolated git worktrees.

Core Features & Use Cases

  • Supports four agent roles (Feature, CRAP, Mutation, Acceptance) to manage backlog-driven development, refactoring, mutation testing, and E2E validation.
  • Uses git worktrees to run agents in isolation while sharing repository state, with a Main integration branch and automated preflight/backlog steps.
  • Provides structured progress logging and cross-run lessons to improve future iterations.

Quick Start

Launch the swarm to plan, spawn parallel agents, and orchestrate feature work with isolated git worktrees.

Frequently Asked Questions about swarm

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

FAQPage Schema
How do I coordinate parallel AI agents for software development using git worktrees?

A multi-agent swarm coordinates four specialized roles—Feature, CRAP, Mutation, and Acceptance—to plan, build via TDD, and harden code across isolated git worktrees. Agents share repository state while executing backlog-driven tasks independently in parallel.

What is backlog-driven development in a multi-agent coding workflow?

Backlog-driven development in a multi-agent workflow uses structured progress logging and cross-run lessons to prioritize and execute tasks. Agents pull from the backlog to plan features, run automated quality checks, and integrate work into a main branch.

Can I run mutation testing and E2E validation in parallel with feature development?

Yes, the swarm architecture supports parallel execution by assigning mutation testing and E2E validation to dedicated agents. These agents operate in isolated git worktrees alongside feature agents, ensuring automated quality checks run concurrently with new code implementation.

How do I isolate AI agent coding tasks to prevent repository conflicts?

You isolate AI agent tasks by using git worktrees, which provide separate working directories linked to the same repository. This setup allows multiple agents to build features and run hardening checks independently before merging into a main integration branch.

What's the best way to automate TDD and code hardening across multiple agents?

Automating TDD and code hardening across multiple agents is best handled by a coordinated swarm that assigns specific roles for test-driven feature building and quality assurance. The system uses isolated worktrees and structured logging to track progress and apply cross-run lessons.