subagent-driven-development

Dispatch fresh subagents per task with isolated context and staged reviews.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/Silac1995/My-World-Isekai-Unity --skill subagent-driven-development-silac1995
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/Silac1995/My-World-Isekai-Unity/tree/main/.claude/skills/subagent-driven-development
Command: npx skills add https://github.com/Silac1995/My-World-Isekai-Unity --skill subagent-driven-development-silac1995

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Coordinated plan execution by dispatching fresh subagents per task with isolated context.

Core Features & Use Cases

  • Fresh subagent per task to avoid context leakage and maximize parallelism.
  • Two-stage review: spec compliance followed by code quality.
  • Supports the three prompt templates: implementer-prompt.md, spec-reviewer-prompt.md, code-quality-reviewer-prompt.md.

Quick Start

Dispatch a new subagent for the first task and start the two-stage review loop.

Frequently Asked Questions about subagent-driven-development

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

FAQPage Schema
How do I orchestrate subagents for parallel task execution?

Subagent orchestration dispatches a fresh subagent per task with isolated context, maximizing parallelism and preventing context leakage during coordinated plan execution.

How does two-stage code review work for automated task dispatch?

Two-stage code review first checks spec compliance using a spec-reviewer prompt, then evaluates code quality with a code-quality-reviewer prompt to ensure dual-review requirements are met.

What is the best way to avoid context leakage when dispatching development tasks?

The best way to avoid context leakage is dispatching fresh subagents for each independent task within a single session, ensuring isolated context and structured prompt tracking.

Do I need specific prompt templates to coordinate plan execution with subagents?

You need three prompt templates to coordinate execution: implementer-prompt, spec-reviewer-prompt, and code-quality-reviewer-prompt, satisfying per-task specification and dual-review requirements.

Can I use subagent-driven development for multiple independent tasks in one session?

You can use it for multiple independent tasks within a single session by using structured prompts, task tracking, and staged reviews to coordinate the overall plan execution.