subagent-driven-development

Dispatch independent subagents to execute implementation plan tasks with mandatory reviews.

Updated Jan 17, 2026
One-click install
npx skills add https://github.com/brmatola/gremlins --skill subagent-driven-development-brmatola
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/brmatola/gremlins/tree/main/skills/subagent-driven-development
Command: npx skills add https://github.com/brmatola/gremlins --skill subagent-driven-development-brmatola

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill addresses the need to execute complex implementation plans by delegating work to fresh subagents per task, enabling fast iteration while preserving session context and ensuring thorough reviews after each step.

Core Features & Use Cases

  • Per-task subagents: Dispatch separate subagents for each task to isolate context and reduce cross-task contamination.
  • Two-stage reviews: After implementation, run spec compliance review followed by code quality review for each task.
  • Plan-driven execution: Read tasks from plans/active/{plan}/implementation and convert them into executable tasks with clear context.
  • End-to-end completion: Automatically mark tasks complete and trigger final reviews after all tasks.

Quick Start

Load a plan from plans/active/{plan}/implementation, dispatch an implementer subagent for each task, handle questions, perform spec and quality reviews, fix issues, and mark tasks complete. Then run the final code review for the entire implementation.

Frequently Asked Questions about subagent-driven-development

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

FAQPage Schema
How do I automate implementation plan execution with AI subagents?

You can automate implementation plan execution by dispatching independent subagents for each task. This approach isolates context to reduce cross-task contamination and enforces per-task workflows with mandatory spec and code quality reviews after each step.

What is subagent-driven development for task management?

Subagent-driven development is a task management workflow where separate subagents execute independent implementation tasks. It requires iterative spec and code reviews, applying specific implementer and reviewer prompts to ensure thorough rework before marking tasks complete.

How do I run spec compliance and code quality reviews for each task?

To run spec compliance and code quality reviews, dispatch an implementer subagent for the task, then sequentially trigger a spec-reviewer-prompt followed by a code-quality-reviewer-prompt. This two-stage review process enforces mandatory rework before the task is marked complete.

Can I use independent subagents for tasks with dependencies in a single session?

Independent subagents work best for mostly independent tasks within a single session. If your implementation plan contains tasks with heavy dependencies, the isolated subagent approach may struggle with cross-task contamination since each subagent operates with fresh context.

When should I not use subagents for implementation plan execution?

You should avoid using subagents for implementation plan execution when your tasks are tightly coupled or require shared session context. The subagent workflow is designed for mostly independent tasks and isolates context per task, which limits cross-task visibility.