When does autonomous execution stop for user input?

Execution pauses at explicit stop points such as requirement confirmation, PRD approval, Design Doc approval, and final work plan batch approval.

What happens if requirements change mid-workflow?

The guide includes a change detection checklist; if new features or constraints are detected, the autonomous flow stops and restarts from the requirement-analyzer phase.

Can subagents call each other directly?

No, to maintain control and prevent unauthorized loops, all coordination and task assignments must flow through the central orchestrator.

How does the skill determine the workflow scale?

It uses a scale determination table based on the number of files involved: Small (1-2 files), Medium (3-5 files), and Large (6+ files), each with different documentation requirements.

What is the primary role of the Orchestrator?

The Orchestrator coordinates specialized subagents like a conductor, directing investigation, analysis, and implementation tasks without performing the manual coding itself.

Subagent Orchestration Guide

Name: Subagent Orchestration Guide
Author: Tahir-yamin

byTahir-yamin

•

Productivity & Workflow

Orchestrates complex multi-agent workflows by coordinating specialized subagents through defined implementation phases and autonomous execution modes.

The Subagent Orchestration Guide transforms Claude into a sophisticated project conductor, managing a suite of specialized subagents to handle end-to-end development tasks. It provides a rigorous framework for requirement analysis, scale determination, and documentation generation including PRDs and ADRs. By implementing a 'First Action Rule' and clear 'Stop Points' for human approval, the skill ensures that autonomous execution—from task decomposition to quality assurance—remains aligned with user requirements and engineering best practices.

Key Features

01Automated task scaling (Small, Medium, Large) based on file impact and complexity

02Autonomous execution mode for seamless task decomposition and implementation loops

03Intelligent coordination of 11+ specialized subagents like quality-fixer and task-executor

04Requirement change detection that automatically triggers re-analysis and document updates

05Standardized JSON response structures for consistent communication between AI agents

063 GitHub stars

Use Cases

01Automating the creation and review of PRDs, Design Docs, and ADRs for large-scale projects

02Coordinating autonomous quality assurance and test execution across distributed codebases

03Managing complex, multi-file feature implementations with high architectural standards

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tahir-yamin/dev-engineering-playbook subagents-orchestration-guide

For use in Claude.ai and ChatGPT

Download Skill