subagent-driven-development

Dispatch fresh subagents per task with mandatory code reviews.

Updated Nov 3, 2025

Search Tags:#workflow #development #quality assurance #code review #task management #AI agents #subagents

Ready to run

npx skills add https://github.com/ayourtch/kimichat --skill subagent-driven-development-ayourtch

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/ayourtch/kimichat/tree/main/skills/subagent-driven-development
Command: npx skills add https://github.com/ayourtch/kimichat --skill subagent-driven-development-ayourtch

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill enables high-quality, fast-iteration development by dispatching fresh subagents for each task in an implementation plan, with mandatory code review between tasks.

Core Features & Use Cases

Task-Specific Subagents: Dispatches a new subagent for each independent task, preventing context pollution and ensuring focus.
Continuous Code Review: Integrates requesting-code-review after each task, catching issues early before they compound.
Quality Gates: Ensures all critical and important issues are fixed before proceeding to the next task.
Use Case: When executing a multi-task implementation plan, use this skill to dispatch a subagent for Task 1: Implement API endpoint, review its work, then proceed to Task 2: Add database integration.

Quick Start

I'm using the subagent-driven-development skill to execute the user-onboarding implementation plan.

Frequently Asked Questions about subagent-driven-development

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

FAQPage Schema

How do I prevent context pollution when running multiple development tasks with AI agents?▼

Subagent-driven development dispatches a fresh subagent for each independent task, isolating context and focus. This prevents earlier tasks from contaminating the reasoning state of later tasks, ensuring each receives dedicated attention and clean execution.

Can I automate code review between tasks in a multi-task implementation plan?▼

Yes. This approach integrates mandatory code review after each task completion, catching issues before they compound. Quality gates ensure all critical and important issues are fixed before proceeding to the next task.

What's the best way to execute a multi-task implementation plan while maintaining quality?▼

Dispatch a subagent for each task, enforce post-task code review, and apply quality gates. This workflow balances fast iteration with quality assurance by reviewing work after each discrete task before moving forward.

How does per-task subagent instantiation improve development speed?▼

Fresh subagents per task eliminate context carryover, reduce reasoning overhead, and enable parallel quality checks. Each subagent focuses narrowly on its assigned task, shortening iteration cycles without sacrificing review discipline.

Do I need test-driven outputs for this workflow?▼

Yes. This skill satisfies requirements for test-driven outputs alongside per-task subagent instantiation and mandatory code review, ensuring committed changes meet quality standards before integration.