subagent-driven-development

Dispatch isolated subagents per task with two-stage spec and code quality reviews.

115|13|Updated Mar 24, 2026
One-click install
npx skills add https://github.com/Houseofmvps/ultraship --skill subagent-driven-development-houseofmvps
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/Houseofmvps/ultraship/tree/main/skills/subagent-driven-development
Command: npx skills add https://github.com/Houseofmvps/ultraship --skill subagent-driven-development-houseofmvps

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Manually executing multi-task implementation plans leads to context bloat, inconsistent quality, and slow iteration as developers switch between tasks and miss review checkpoints. This Skill eliminates those inefficiencies by orchestrating isolated subagents for each independent task with built-in quality gates.

Core Features & Use Cases

  • Per-Task Subagent Dispatch: Assigns a fresh, focused subagent to each independent task to avoid context pollution from session history.
  • Two-Stage Review Workflow: Runs spec compliance review first, then code quality review for each task to catch gaps and issues early.
  • Task Sizing Optimization: Skips unnecessary reviews for small tasks to reduce overhead, while applying full review cycles only to large, cross-cutting tasks.
  • Use Case: A developer with a 6-task feature plan can use this Skill to dispatch dedicated subagents for each task, run automated reviews, and ensure all work meets requirements before merging, without switching contexts or missing quality checks.

Quick Start

Use the subagent-driven-development skill to execute the implementation plan in docs/feature-plan.md with per-task spec and code quality reviews.

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 a multi-task implementation plan without context bloat?

Executing a multi-task implementation plan without context bloat requires dispatching isolated subagents per independent task. This approach assigns a fresh, focused subagent to each task, preventing session history pollution and maintaining consistent quality across decoupled software development workflows.

What is the best way to automate code review for multi-task software development?

The best way to automate code review in multi-task software development is applying a two-stage review workflow per task. This method runs spec compliance checks first, followed by code quality reviews, catching gaps early before a final holistic implementation review prior to merge.

How do I set up task sizing optimization for subagent workflows?

Task sizing optimization for subagent workflows skips unnecessary reviews for small tasks to reduce overhead, while applying full two-stage review cycles only to large, cross-cutting tasks. This ensures efficient plan execution by scaling quality gates based on task complexity.

Does subagent-driven development work for parallel session handoff?

Subagent-driven development does not work for parallel session handoff. It is designed specifically for same-session software development workflows where tasks are decoupled and independent, dispatching isolated subagents per task without requiring external session handoff mechanisms.

When do I need a final holistic implementation review before merging?

A final holistic implementation review before merging is needed when executing multi-task plans with isolated subagents. It ensures all independently dispatched tasks meet overall requirements and integrate correctly, serving as a quality gate after individual per-task spec and code reviews.