craft-squad

Coordinate a design-first pipeline across Claude, Codex, and a local LLM.

17|Updated Mar 30, 2026
One-click install
npx skills add https://github.com/alexiolan/craft-skills --skill craft-squad
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: craft-squad
Source: https://github.com/alexiolan/craft-skills/tree/main/skills/craft-squad
Command: npx skills add https://github.com/alexiolan/craft-skills --skill craft-squad

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Coordinates a full design-first AI pipeline (brainstorm, plan, develop, test, and report) using Claude, Codex, and a local LLM to coordinate multi-agent workflows and reduce hand-off friction.

Core Features & Use Cases

  • End-to-end design-first pipeline with triple-agent orchestration (Claude + Codex + local LLM) for high-stakes features.
  • Optional adversarial review via codex-plugin-cc to provide skeptical second opinions when available.
  • Flexible prerequisites and deployment paths to support LM Studio, Codex CLI, and profile-based gating for experiments.

Quick Start

Describe your feature goal with /craft-squad to launch the full multi-agent design pipeline.

Frequently Asked Questions about craft-squad

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

FAQPage Schema
How do I coordinate multi-agent workflows for complex feature development across Claude and Codex?

Multi-agent orchestration for complex feature development is handled by a design-first pipeline coordinating Claude, Codex, and a local LLM. It automates brainstorming, planning, development, testing, and reporting to reduce hand-off friction across high-stakes web apps, APIs, or UI components.

What is an AI-driven design pipeline and when do I need multi-agent orchestration for software engineering?

An AI-driven design pipeline is a structured process automating brainstorm through reporting using multiple agents. Multi-agent orchestration is needed for high-stakes features with multiple integration points where rigorous cross-agent review and coordination are valuable for complex web apps or APIs.

Does multi-agent code review work with LM Studio and Codex CLI deployments?

Multi-agent code review works with LM Studio and Codex CLI deployments through flexible prerequisite paths. The pipeline ensures deployment parity across these environments, enabling local LLM integration alongside Claude and Codex for coordinated cross-agent feature development.

How do I start a multi-agent design pipeline for brainstorming and testing a feature?

To start the multi-agent design pipeline, describe your feature goal using the /craft-squad command. This launches the full orchestration workflow across Claude, Codex, and a local LLM to automate brainstorming, planning, development, testing, and reporting for your complex feature.

Can I get an adversarial review of my code during the AI pipeline if codex-plugin-cc is present?

Adversarial review is available during the AI pipeline when codex-plugin-cc is present. This optional feature provides skeptical second opinions to enhance rigorous code review and cross-agent coordination for high-stakes features with multiple integration points.

Do I need a specific profile configuration to gate multi-agent orchestration experiments?

Profile gating uses a .craft-profile configuration to manage multi-agent orchestration experiments. This profile-based gating supports flexible prerequisites and deployment paths for coordinating Claude, Codex, and local LLM workflows across different experimental environments.