subagent-driven-development

Dispatch fresh subagents per task with isolated context and two-stage reviews.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/SeanChenR/maestro-agent --skill subagent-driven-development-seanchenr
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/SeanChenR/maestro-agent/tree/main/.agents/skills/subagent-driven-development
Command: npx skills add https://github.com/SeanChenR/maestro-agent --skill subagent-driven-development-seanchenr

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Subagent-driven development solves the complexity of coordinating implementation work by delegating each task to a fresh, isolated subagent within the same session, preventing context leakage and enabling parallel-like progress with tight coordination.

Core Features & Use Cases

  • Fresh subagents per task to ensure focused, context-isolated execution.
  • A two-stage review workflow after every task: spec compliance review followed by code quality review.
  • Plan extraction and TodoWrite-style task management to track progress and ensure all tasks are completed before moving on.
  • Standardized prompts and templates for implementer, spec-reviewer, and code-quality-reviewer to maintain consistency.
  • Robust escalation and status handling (DONE, DONE_WITH_CONCERNS, NEEDS_CONTEXT, BLOCKED) to manage blockers.
  • End-to-end workflow with finishing-a-development-branch to consolidate changes when all tasks are complete.

Quick Start

Provide a development plan and let Subagent-Driven Development dispatch a fresh implementer subagent for each task, perform per-task spec and quality reviews, and finalize all tasks by finishing the development branch.

Frequently Asked Questions about subagent-driven-development

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

FAQPage Schema
How do I dispatch subagents for task execution to prevent context leakage?▼

To prevent context leakage, dispatch fresh subagents per task to execute implementation plans with isolated context within a single session. This method ensures focused execution by assigning a new subagent to each specific task.

What is a two-stage review workflow for code and spec compliance?▼

A two-stage review workflow is a process that evaluates each completed task sequentially. It first applies a spec-compliance review to verify requirements, followed by a code-quality review to check the implementation.

How do I track task progress and manage blockers during subagent-driven development?▼

You can track task progress using auto-TodoWrite updates during task scoping and execution. The workflow also includes robust escalation and status handling with states like DONE, BLOCKED, and NEEDS_CONTEXT to manage blockers.

What's the best way to consolidate implementation changes after all development tasks are complete?▼

The best way to consolidate implementation changes is by finishing a development branch. This step finalizes all tasks and consolidates the work once the subagent-driven execution and reviews are complete.

Do I need external dependencies to orchestrate fresh subagents within a single session?▼

No, you do not need external dependencies to orchestrate fresh subagents. The workflow applies within a single session using standardized prompts for per-task implementer, spec-reviewer, and code-quality-reviewer subagents.