What is the primary role of the Sub-Agent Orchestration Guide?

It functions as a high-level orchestrator that manages specialized sub-agents (like requirement-analyzers and task-executors) to handle complex development tasks with professional oversight.

How is code quality guaranteed?

The skill utilizes a dedicated 'quality-fixer' sub-agent that must provide an approved status by passing linting, type checks, and tests before the orchestrator performs a git commit.

Can sub-agents communicate with each other directly?

No, sub-agents cannot call each other directly. The main Claude instance acts as the orchestrator, passing structured JSON data between them to maintain state and control.

How does it decide which workflow to use?

It uses a requirement-analyzer agent to judge the scale of the task—Small (1-2 files), Medium (3-5 files), or Large (6+ files)—and assigns a corresponding multi-step workflow.

Does this skill require human intervention?

Yes. It features mandatory human checkpoints after requirement analysis, documentation creation (PRD/ADR/Design Doc), and work plan generation to ensure accuracy before execution.

Sub-Agent Orchestration Guide

Name: Sub-Agent Orchestration Guide
Author: shinpr

byshinpr

•

162

•

Productivity & Workflow

Coordinates complex coding tasks by delegating specialized roles to sub-agents based on project scale and requirement analysis.

This skill transforms Claude into a high-level project orchestrator, moving beyond simple code generation to manage a fleet of specialized sub-agents. It implements a rigorous workflow for requirement analysis, documentation (PRD/ADR/Design Docs), task decomposition, and quality assurance. By categorizing tasks into small, medium, and large scales, it ensures that every code change is backed by proper design, peer-reviewed documentation, and automated quality checks before final commitment, making it ideal for maintaining high standards in production-grade TypeScript projects.

Key Features

01162 GitHub stars

02Multi-agent coordination with specialized roles for analysis, design, execution, and fixing

03Structured JSON-based communication protocol for reliable inter-agent information transfer

04Automated quality assurance loops with mandatory quality-fixer approval cycles

05Autonomous execution mode with clear escalation triggers and human-in-the-loop checkpoints

06Scalable workflow management that adjusts complexity based on the number of impacted files

Use Cases

01Managing large-scale feature implementations across multiple files and directories

02Enforcing strict adherence to PRD and Design Doc standards in professional teams

03Automating the end-to-end development lifecycle from requirement analysis to final git commit

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add shinpr/ai-coding-project-boilerplate subagents-orchestration-guide

For use in Claude.ai and ChatGPT

Download Skill