subagent-driven-development

Orchestrate independent subagents to execute implementation plans with two-stage reviews.

6|2|Updated Nov 16, 2025

Search Tags:#automation #orchestration #subagent #two-stage-review #task-dispatch #plan-management #ai-planning

Ready to run

npx skills add https://github.com/EricOo0/stock-trading-platform --skill subagent-driven-development-ericoo0

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/EricOo0/stock-trading-platform/tree/main/.codex/superpowers/skills/subagent-driven-development
Command: npx skills add https://github.com/EricOo0/stock-trading-platform --skill subagent-driven-development-ericoo0

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill automates the orchestration of independent subagents to execute multi-step implementation plans within a single session, enabling fast iteration while preserving quality through structured reviews.

Core Features & Use Cases

Two-stage reviews: spec compliance review followed by code quality review after every subtask.
Fresh subagent per task: avoids context pollution and ensures clear boundaries.
Self-review loop: each subagent performs a self-review before reporting back.
Task planning and execution: dispatches tasks from a centralized plan and tracks progress until completion.

Quick Start

Activate a plan file and start per-task subagents with automatic two-stage reviews.

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 task dispatch for multi-step implementation plans?▼

Automating task dispatch involves orchestrating independent subagents to execute modular tasks from a centralized plan, tracking progress until completion while avoiding context pollution through fresh subagent allocation per task.

What is a two-stage review process for AI coding subagents?▼

A two-stage review process evaluates subagent output by first checking spec compliance against the plan, followed by a code quality review, ensuring structured quality control before a subagent reports back after its self-review loop.

Can I use independent subagents for rapid iteration across complex plans?▼

Yes, you can use independent subagents for rapid iteration across complex plans by dispatching modular tasks per subagent, which preserves quality through strict spec-review loops while enabling fast execution within a single session.

How do I prevent context pollution when running multiple AI coding tasks?▼

Preventing context pollution requires dispatching a fresh subagent per task, ensuring clear boundaries and isolated execution environments for each modular task before the results are integrated during the final step.

What is the best way to manage spec compliance and code quality reviews in AI orchestration?▼

The best way to manage spec compliance and code quality reviews is using a structured two-stage review loop where each subagent performs a self-review before reporting back, ensuring strict adherence to implementation plans.

When should I not use subagent orchestration for implementation plans?▼

You should not use subagent orchestration for implementation plans lacking clear modular task boundaries, as this approach relies on dispatching independent subagents and performing structured spec reviews per isolated task.