subagent-driven-development

Dispatch fresh subagent implementers per task with two-stage spec and code-quality reviews.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/yuquanliu1-dot/Test --skill subagent-driven-development-yuquanliu1-dot
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/yuquanliu1-dot/Test/tree/main/.claude/skills/subagent-driven-development
Command: npx skills add https://github.com/yuquanliu1-dot/Test --skill subagent-driven-development-yuquanliu1-dot

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

It solves the problem of executing multi-step implementation plans reliably without context pollution, while still enforcing quality through objective checkpoints.

Core Features & Use Cases

  • Per-task fresh subagents: dispatch an isolated implementer subagent for each task so it does not inherit prior context or session history.
  • Two-stage review gates after every task: run a spec-compliance review first, then a code-quality review, iterating until both pass.
  • Structured handling of subagent states: routes results for DONE, DONE_WITH_CONCERNS, NEEDS_CONTEXT, and BLOCKED with appropriate re-dispatch or escalation.
  • Use cases: ideal when tasks are mostly independent, when you want fast iteration within the current session, and when you want consistent quality gates instead of ad-hoc checking.

Quick Start

Use subagent-driven-development to execute your implementation plan by dispatching an implementer subagent per task, then completing a spec review followed by a code-quality review before moving on.

Frequently Asked Questions about subagent-driven-development

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

FAQPage Schema
How do I execute multi-step implementation plans without context pollution in code generation?

Execute multi-step implementation plans without context pollution by dispatching isolated, fresh subagent implementers for each task, ensuring no inherited session history interferes with the current code generation workflow.

What is a two-stage code review workflow and how does it enforce spec compliance?

A two-stage code review workflow runs a spec-compliance review first, then a code-quality review, iterating through re-review loops until both pass, ensuring strict spec compliance and objective quality checkpoints.

How do I handle subagent states like BLOCKED or NEEDS_CONTEXT during workflow orchestration?

Handle subagent states like BLOCKED or NEEDS_CONTEXT during workflow orchestration by routing results through structured re-dispatch or escalation rules, preventing parallel implementer conflicts and ensuring proper review order.

When should I use subagent-driven development for test driven development and GitHub development?

Use subagent-driven development for test driven development and GitHub development when tasks are mostly independent, fast iteration is needed within the current session, and consistent quality gates are preferred over ad-hoc checking.

What is the best way to prevent parallel implementer conflicts during multi-task development sessions?

Prevent parallel implementer conflicts during multi-task development sessions by enforcing strict escalation rules and structured routing for subagent states, ensuring sequential review order and safe re-review loops until acceptance.