Why is the two-stage review important?

It ensures the code first meets the functional requirements (spec compliance) before auditing the architectural standards (code quality). This prevents wasting time polishing code that doesn't actually solve the requested problem.

How does this differ from standard plan execution?

Standard execution often happens in a single linear context. This skill remains in your current session but dispatches fresh subagents for each task to prevent 'context pollution' and includes a mandatory two-stage review for every task.

When should I avoid using this skill?

You should avoid this skill if your tasks are tightly coupled or interdependent, or if you do not yet have a structured implementation plan ready to follow.

Does it support Test-Driven Development (TDD)?

Yes, subagents are explicitly instructed to follow TDD principles for each task, implementing and verifying tests before proceeding to the review phase.

Subagent-Driven Development

Name: Subagent-Driven Development
Author: tianyuanztm

bytianyuanztm

0•

생산성 및 워크플로

Executes complex implementation plans by orchestrating specialized subagents for task fulfillment and multi-stage quality reviews.

This skill optimizes the execution of multi-step implementation plans within a single Claude Code session by delegating independent tasks to fresh subagents. By utilizing a fresh subagent approach for each task, it prevents context pollution and maintains high development velocity. The core of the workflow is a rigorous two-stage review process: every task is first validated for specification compliance and then for code quality, ensuring that final deliverables are both functionally accurate and architecturally sound without requiring constant human intervention.

주요 기능

01Fresh Subagent Dispatching to prevent context bloat and confusion

02Two-Stage Review Cycle covering spec compliance and code quality

03Automated Progress Tracking via integrated TodoWrite functionality

040 GitHub stars

05Integrated TDD workflow for reliable and tested task completion

06Pre-implementation Questioning to resolve ambiguities before coding begins

사용 사례

01Implementing complex API endpoints that need both functional and quality validation

02Executing a feature roadmap with multiple independent components in a single session

03Refactoring modular codebases requiring strict adherence to technical specifications

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tianyuanztm/myplan subagent-driven-development

For use in Claude.ai and ChatGPT

주요 기능

01Fresh Subagent Dispatching to prevent context bloat and confusion

02Two-Stage Review Cycle covering spec compliance and code quality

03Automated Progress Tracking via integrated TodoWrite functionality

040 GitHub stars

05Integrated TDD workflow for reliable and tested task completion

06Pre-implementation Questioning to resolve ambiguities before coding begins

사용 사례

01Implementing complex API endpoints that need both functional and quality validation

02Executing a feature roadmap with multiple independent components in a single session

03Refactoring modular codebases requiring strict adherence to technical specifications

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tianyuanztm/myplan subagent-driven-development

For use in Claude.ai and ChatGPT