subagent-driven-development

Dispatch a subagent per task and run two-stage reviews for spec compliance and code quality.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/Sjw2005/https---gitee.com-song-jiawei-2005-yolopro --skill subagent-driven-development-sjw2005
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/Sjw2005/https---gitee.com-song-jiawei-2005-yolopro/tree/main/.agents/skills/subagent-driven-development
Command: npx skills add https://github.com/Sjw2005/https---gitee.com-song-jiawei-2005-yolopro --skill subagent-driven-development-sjw2005

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Subagent-Driven Development addresses the challenge of executing complex plans by isolating work into independent subagents per task, enabling two-stage reviews after each task to ensure adherence to specifications and code quality.

Core Features & Use Cases

  • Dispatch a dedicated subagent for each task to maintain isolated context and reduce cross-task interference.
  • Enforce two-stage reviews after every task: spec compliance first, then code quality, reducing rework.
  • Coordinate whole plans with parallelizable, structured workflows for faster delivery.
  • Ideal for multi-step implementation plans, experiments, and iterative development in knowledge work.

Quick Start

Dispatch a per-task subagent for the current plan and trigger the two-stage review workflow.

Frequently Asked Questions about subagent-driven-development

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

FAQPage Schema
How do I automate plan execution with isolated subagents for each task?

Automating plan execution with isolated subagents involves dispatching a fresh subagent for each task, which prevents context leakage and reduces cross-task interference during structured workflow orchestration.

What is a two-stage review pipeline for task dispatch and why use it?

A two-stage review pipeline enforces spec compliance first, then evaluates code quality after every task is executed by a subagent, reducing rework and ensuring adherence to project specifications.

How do I execute parallelizable multi-step development plans without context leakage?

You can execute parallelizable plans by decomposing them into independent tasks and dispatching dedicated subagents for each task, enforcing structured review workflows with provided templates to maintain isolated contexts.

Does subagent-driven orchestration work for highly interdependent tasks in a development plan?

Subagent-driven orchestration targets plans with largely independent tasks that can be executed in parallel; highly interdependent tasks are not ideal due to the isolated context designed to prevent cross-task interference.

When should I not use subagents for orchestrating code generation workflows?

You should avoid using subagents for orchestration when your development plan lacks largely independent tasks, as the workflow relies on isolated contexts and parallel execution to prevent cross-task interference and context leakage.