subagent-driven-development

Execute development plans by dispatching sub-agents with two-stage reviews.

8|Updated Jan 13, 2026

Search Tags:#automation #tdd #development #code review #task execution #subagent

Ready to run

npx skills add https://github.com/hxk622/TokenDance --skill subagent-driven-development-hxk622

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/hxk622/TokenDance/tree/main/backend/app/skills/builtin/superpowers/subagent-driven-development
Command: npx skills add https://github.com/hxk622/TokenDance --skill subagent-driven-development-hxk622

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill streamlines complex development tasks by breaking them down into independent sub-tasks, each handled by a dedicated AI sub-agent, ensuring high-quality code through a rigorous two-stage review process.

Core Features & Use Cases

Task Decomposition: Automatically breaks down an implementation plan into manageable, independent tasks.
Sub-agent Dispatch: Assigns a fresh AI sub-agent to each task for focused execution.
Two-Stage Review: Implements a spec compliance review followed by a code quality review for each task.
Iterative Refinement: Manages feedback loops where sub-agents fix issues identified during reviews.
Use Case: Developing a new feature by providing a detailed plan. The skill will create sub-agents to implement each part, review them against the spec, check code quality, and iterate until all tasks are complete and approved.

Quick Start

Use the subagent-driven-development skill to execute the implementation plan located at docs/plans/feature-plan.md.

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 development tasks using subagents?▼

Automating development tasks with subagents involves dispatching a dedicated AI sub-agent for each task in an implementation plan. This skill handles task decomposition and assigns independent sub-agents to execute each part.

How does the two-stage code review process work for subagent development?▼

The two-stage code review process runs after each sub-agent completes a task. It first checks specification compliance to ensure requirements are met, then evaluates code quality, triggering iterative fixes if issues are found.

Can I use TDD principles with AI-powered subagent execution?▼

Yes, you can use TDD principles with AI-powered subagent execution. The skill ensures adherence to Test-Driven Development practices within the dispatched sub-agents during the implementation process.

What is the best way to break down a complex implementation plan into independent tasks?▼

Breaking down a complex implementation plan is handled automatically by the skill through task decomposition. It divides the plan into manageable, independent tasks for focused execution by individual sub-agents.

Does subagent-driven development manage iterative fixes during code review?▼

Yes, subagent-driven development manages iterative fixes during code review. It creates feedback loops where sub-agents automatically address and fix issues identified during the specification and quality reviews.

How do I start executing an implementation plan with a subagent workflow?▼

To start executing an implementation plan, provide a detailed plan document like a feature-plan markdown file. The skill will read it, create sub-agents for each task, and iterate until all tasks are complete and approved.