What happens if the code reviewer subagent finds an error?

If critical or important issues are found, the skill dispatches a follow-up subagent to fix those specific issues before the workflow is allowed to proceed to the next task.

What is Subagent-Driven Development?

It is a specialized workflow for Claude Code where a main session coordinates multiple short-lived subagents to perform specific coding tasks, followed by automated review cycles.

How does this differ from standard plan execution?

Unlike standard execution which might happen in a separate session, this stays in your current session while providing the quality benefits of fresh context and automated peer review.

How does this skill prevent context pollution?

By dispatching a completely fresh subagent for each task in the plan, the skill ensures that the AI doesn't get confused by previous implementation details or irrelevant history.

Is a plan required to use this skill?

Yes, this skill is designed to execute a structured plan. It typically requires the 'writing-plans' skill to create the roadmap before execution begins.

Subagent-Driven Development

Name: Subagent-Driven Development
Author: franroa

byfranroa

生産性とワークフロー

Executes complex implementation plans by dispatching dedicated subagents for individual tasks with automated quality gates and code reviews.

概要

This skill optimizes the development workflow within Claude Code by automating the execution of multi-step plans through a modular subagent architecture. Instead of a single long-running session that might suffer from context drift, it dispatches a fresh subagent for every individual task, ensuring isolated and focused implementation. By integrating mandatory code review steps between tasks and a final architectural validation, the skill maintains high code quality and Test-Driven Development (TDD) standards while enabling rapid, continuous progress without requiring human-in-the-loop intervention between subtasks.

主な機能

Automated Quality Gates: Integrates mandatory code review subagents after every task completion to catch issues early.
0 GitHub stars
TDD Enforcement: Encourages subagents to follow Test-Driven Development patterns for verified, bug-free implementations.
Feedback Loop Management: Automatically dispatches follow-up agents to resolve critical issues identified during reviews.
Isolated Subagent Dispatch: Assigns a fresh subagent to every task to prevent context pollution and logic drift.
Session Continuity: Executes complex plans within the same active session without requiring manual handoffs.

ユースケース

Managing large-scale refactorings that require high quality-control and sequential verification.
Executing multi-file feature implementations derived from a pre-defined architectural plan.
Automating the end-to-end TDD cycle for complex components while maintaining a clean session history.

概要

主な機能

Automated Quality Gates: Integrates mandatory code review subagents after every task completion to catch issues early.
0 GitHub stars
TDD Enforcement: Encourages subagents to follow Test-Driven Development patterns for verified, bug-free implementations.
Feedback Loop Management: Automatically dispatches follow-up agents to resolve critical issues identified during reviews.
Isolated Subagent Dispatch: Assigns a fresh subagent to every task to prevent context pollution and logic drift.
Session Continuity: Executes complex plans within the same active session without requiring manual handoffs.

ユースケース

Managing large-scale refactorings that require high quality-control and sequential verification.
Executing multi-file feature implementations derived from a pre-defined architectural plan.
Automating the end-to-end TDD cycle for complex components while maintaining a clean session history.