Why use subagents instead of a single session?

Using fresh subagents for each task prevents 'context pollution,' where previous code or logic interferes with new tasks, leading to cleaner and more reliable implementations.

How does the two-stage review process work?

Every task is subjected to two distinct checks: first, a spec reviewer ensures all requirements are met; second, a code quality reviewer ensures the implementation follows best practices.

Does this skill work with existing git workflows?

Yes, it is designed to work with git-worktrees and development branches to ensure all AI-driven changes are isolated and easy to merge.

What happens if a subagent finds a bug in its own work?

The workflow includes a self-review phase where the subagent identifies and fixes issues before the task is even sent to the formal review stage.

What is Subagent-Driven Development?

It is a specialized workflow where a controller AI dispatches fresh subagents to handle individual tasks from an implementation plan, ensuring task isolation and high-quality output.

Subagent-Driven Development

Name: Subagent-Driven Development
Author: AhmedElhadarey

byAhmedElhadarey

•

生産性とワークフロー

Executes complex implementation plans by dispatching specialized subagents for per-task development and multi-stage quality reviews.

This skill automates the execution of multi-step development plans within a single session by leveraging independent subagents for each task. It maintains high code standards through a rigorous two-stage review process where every task is first checked for specification compliance and then for code quality. By isolating tasks and using fresh subagent contexts, it prevents context pollution, reduces manual overhead, and ensures that features are built exactly as specified while maintaining architectural integrity and test coverage.

主な機能

01Two-stage review process: Specification compliance and code quality audits

02Seamless integration with git-worktrees for isolated development environments

03Automated task extraction and progress tracking via TodoWrite

04Fresh subagent dispatch per task to prevent context pollution

052 GitHub stars

06Integrated Test-Driven Development (TDD) and self-review cycles

ユースケース

01Executing pre-defined feature plans with multiple independent modules

02Scaling development velocity within a single Claude session without losing focus

03Refactoring codebases where tasks are decoupled but require high quality control

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ahmedelhadarey/gilfoyle subagent-driven-development

For use in Claude.ai and ChatGPT

主な機能

01Two-stage review process: Specification compliance and code quality audits

02Seamless integration with git-worktrees for isolated development environments

03Automated task extraction and progress tracking via TodoWrite

04Fresh subagent dispatch per task to prevent context pollution

052 GitHub stars

06Integrated Test-Driven Development (TDD) and self-review cycles

ユースケース

01Executing pre-defined feature plans with multiple independent modules

02Scaling development velocity within a single Claude session without losing focus

03Refactoring codebases where tasks are decoupled but require high quality control

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ahmedelhadarey/gilfoyle subagent-driven-development

For use in Claude.ai and ChatGPT