subagent-driven-development

Dispatch independent subagents per task with two-stage reviews and lifecycle signals.

Updated Mar 11, 2026
One-click install
npx skills add https://github.com/Ramachandrajoshi/Azure-Devops-Copilot --skill subagent-driven-development-ramachandrajoshi
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/Ramachandrajoshi/Azure-Devops-Copilot/tree/main/.roo/skills/subagent-driven-development
Command: npx skills add https://github.com/Ramachandrajoshi/Azure-Devops-Copilot --skill subagent-driven-development-ramachandrajoshi

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Coordinating complex plans by dispatching focused subagents per task, reducing context switches and enabling rapid iteration with built-in review loops.

Core Features & Use Cases

  • Fresh subagent per task ensures isolated context and prevents cross-task contamination.
  • Two-stage reviews after each task: spec compliance review followed by code quality review.
  • Clear task lifecycle and escalation handling (DONE, DONE_WITH_CONCERNS, NEEDS_CONTEXT, BLOCKED).

Quick Start

Run a plan with this skill to automatically dispatch a fresh subagent per task and apply 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 orchestrate plan execution with isolated subagents for each task?

Plan execution orchestration dispatches a fresh subagent per task, ensuring isolated context and preventing cross-task contamination. This approach reduces context switches and enables rapid iteration with built-in review loops.

What is two-stage review in workflow orchestration?

Two-stage review is a quality control process that evaluates task completion in sequence: first verifying spec compliance, then assessing code quality. This structured loop ensures each dispatched subagent meets both functional and quality standards.

When do I need context isolation for multi-task plan execution?

Context isolation is needed when coordinating complex plans with multiple, largely independent tasks. It prevents cross-task contamination by assigning each task to a fresh subagent, improving execution speed and quality.

How are blocked tasks or missing context handled during task dispatch?

Task dispatch handles lifecycle events through specific signals: DONE, BLOCKED, NEEDS_CONTEXT, and DONE_WITH_CONCERNS. These escalation signals manage task queue progression and flag tasks requiring intervention.

What's the best way to manage a task queue with independent subagents?

The best way to manage a task queue with independent subagents is using workflow orchestration that creates per-task subagents, applies two-stage reviews, and processes lifecycle signals to track progression and handle escalations.

Can I use subagent-driven development for tasks with interdependent steps?

Subagent-driven development applies to sessions with multiple, largely independent tasks. It is not suited for tightly interdependent steps, as context isolation prevents subagents from sharing state across tasks.