What happens if a subagent fails a task or a review?

If issues are found during the review gate, the skill dispatches a follow-up fix subagent to address critical and important issues before the workflow proceeds to the next task.

Does it support Test-Driven Development (TDD)?

Yes, subagents are specifically instructed to implement tasks following TDD patterns, including writing and verifying tests before reporting back.

How does this differ from standard plan execution?

Unlike standard execution, this skill uses a fresh subagent for every single task to prevent context pollution and enforces mandatory code review checkpoints between every step.

Are there required dependencies for this skill?

Yes, it integrates with and requires the writing-plans, requesting-code-review, and finishing-a-development-branch skills to complete the full development lifecycle.

Subagent-Driven Development

Name: Subagent-Driven Development
Author: Kingly-Agency

byKingly-Agency

0•

Productivity & Workflow

Orchestrates sequential task execution using fresh subagents and automated code reviews to ensure high-quality, iterative code delivery.

Subagent-Driven Development is an advanced workflow for Claude Code designed to execute complex implementation plans with high precision. By dispatching a dedicated, fresh subagent for every individual task, the skill eliminates context pollution and prevents AI drift during long sessions. Each task completion is guarded by a mandatory code review gate, ensuring that bugs or architectural missteps are caught and corrected immediately. This approach provides a rigorous, automated pipeline that balances rapid iteration with production-grade quality controls, making it the ideal choice for developers building modular features or refactoring critical systems.

Key Features

01Integrated fix cycles for immediate application of reviewer feedback

020 GitHub stars

03Real-time Todo management and progress tracking within the session

04Automated code review gates between tasks to catch issues early

05Final architectural validation and branch completion workflows

06Isolated subagent dispatch per task to maintain clean context

Use Cases

01Executing multi-step implementation plans for complex feature development

02Automating TDD workflows where each unit must pass testing before proceeding

03Refactoring legacy codebases where incremental quality gates are required

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add kingly-agency/kingly-claude-adapter subagent-driven-development

For use in Claude.ai and ChatGPT

Download Skill