subagent-driven-development

Dispatch subagent tasks with two-stage spec and code quality reviews.

Updated Jan 19, 2026
One-click install
npx skills add https://github.com/DaFum/github-repository-vi --skill subagent-driven-development-dafum
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/DaFum/github-repository-vi/tree/main/skills/subagent-driven-development
Command: npx skills add https://github.com/DaFum/github-repository-vi --skill subagent-driven-development-dafum

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill orchestrates plan execution by dispatching fresh subagent tasks for each task within the current session, enabling rapid iteration with a structured two‑stage review (spec compliance first, then code quality).

Core Features & Use Cases

  • Per-task subagent orchestration: Assigns a dedicated subagent per task to ensure isolation and focused execution.
  • Two-stage review workflow: Enforces spec compliance followed by code quality review to guarantee adherence and quality.
  • In-session execution: Maintains context within a single session for fast feedback and reduced context-switching.
  • Self-checks & reports: Subagents perform self-review before reporting results, producing clear outputs for downstream use.

Quick Start

  1. Provide a plan containing clearly defined tasks to guide subagent execution.
  2. Invoke the skill to start per-task dispatch and the two-stage review cycle.
  3. Monitor review results and iterate as tasks are re-reviewed until acceptance.

Frequently Asked Questions about subagent-driven-development

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

FAQPage Schema
How does subagent orchestration work for software engineering workflows?

Subagent orchestration dispatches a fresh subagent for each task within a single session to ensure isolated execution. A two-stage review process then validates spec compliance and code quality before accepting the results.

How do I automate plan execution with multiple independent tasks?

Provide a plan with clearly defined tasks to the skill. It then dispatches dedicated subagents per task, maintains per-task context, and executes a two-stage review cycle for spec compliance and code quality.

What is a two-stage review workflow for code generation?

A two-stage review workflow enforces spec compliance first, then evaluates code quality. Subagents perform self-reviews before reporting results, ensuring structured validation and adherence to the original plan.

Do I need a predefined plan to use subagent task dispatch?

Yes, this skill expects a plan with clearly defined tasks as input. Without predefined task boundaries, the subagent orchestration cannot properly isolate execution or validate spec compliance.

Can I run multiple subagent tasks within a single session?

Yes, the skill applies within a single session to execute multiple independent tasks derived from a plan. It maintains per-task context for fast feedback and reduced context-switching.

What are the limitations of automated subagent reviews?

The skill relies on predefined prompt templates for implementer, spec reviewer, and code-quality reviewer roles. It cannot dynamically adapt if the initial plan lacks clearly defined tasks or boundaries for execution.