subagent-driven-development

Orchestrate per-task subagents with wave-based dispatch and two-stage review gates.

12|1|Updated Feb 13, 2026
One-click install
npx skills add https://github.com/ovargas/virtual-team --skill subagent-driven-development-ovargas
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/ovargas/virtual-team/tree/main/skills/subagent-driven-development
Command: npx skills add https://github.com/ovargas/virtual-team --skill subagent-driven-development-ovargas

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Distills complex task execution by outsourcing implementation to dedicated subagents, while the orchestrator coordinates reviews and progress without writing code itself.

Core Features & Use Cases

  • Fresh subagent per task: isolation and parallelism for multi-file work.
  • Two-stage review: specification-driven validation before acceptance.
  • Wave-based dispatch: controlled concurrency with conflict avoidance.
  • Progress tracking: centralized orchestration of tasks and outcomes.

Quick Start

Deploy a fresh subagent per task and enforce two-stage reviews for each work item.

Frequently Asked Questions about subagent-driven-development

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I orchestrate subagents for multi-task plans?

You orchestrate subagents by deploying a fresh, isolated subagent per task and coordinating progress through a centralized orchestrator that manages wave-based dispatch and mandatory review gates.

What is a two-stage review in autonomous agent orchestration?

A two-stage review is a specification-driven validation process enforced before acceptance, ensuring that work distributed across autonomous agents meets requirements without the orchestrator writing code itself.

Can I use wave-based dispatch to manage concurrency for complex features?

Yes, wave-based dispatch controls concurrency by distributing tasks across autonomous agents while avoiding conflicts, making it suitable for multi-file work on complex features.

When should I use an orchestrator pattern for task management?

You should use an orchestrator pattern when handling multi-task plans with complex features that require distributed planning, implementation, and reviews across isolated, autonomous subagents.

How do I ensure isolation and parallelism for multi-file work?

Ensure isolation and parallelism by deploying a fresh subagent per task, which allows autonomous agents to handle multi-file work independently while the orchestrator tracks progress.

Does the orchestrator write code during subagent task execution?

No, the orchestrator does not write code itself; it purely coordinates reviews, tracks progress, and enforces specification-driven validation gates for the work outsourced to dedicated subagents.