subagent-driven-development

Dispatch fresh subagents per task with two-stage spec and code quality reviews.

Updated Mar 3, 2026
One-click install
npx skills add https://github.com/jacexh/skills --skill subagent-driven-development-jacexh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/jacexh/skills/tree/main/skills/subagent-driven-development
Command: npx skills add https://github.com/jacexh/skills --skill subagent-driven-development-jacexh

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill eliminates context pollution and inconsistent implementations when executing multi-step implementation plans by dispatching a fresh implementer subagent per task and enforcing a two-stage review sequence to catch spec gaps and quality issues early.

Core Features & Use Cases

  • Per-task fresh subagents: Launch an implementer subagent for each independent task so work stays isolated and reproducible.
  • Two-stage review gates: Require a spec compliance reviewer first, then a code quality reviewer to prevent over/under-building and maintain code standards.
  • TDD and git-friendly flow: Mandates tests, commits, self-review, and integrates workflows like git worktrees and final branch finishing for safe merges.
  • Use case: Execute a feature plan split into independent tasks, have each task implemented, verified against spec, and quality-reviewed before marking complete.

Quick Start

Execute the implementation plan docs/plans/feature-plan.md using subagent-driven-development so each independent task runs an implementer, a spec reviewer, and a code-quality reviewer before marking it complete.

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 execution for an implementation plan without context pollution?

Automate task execution by dispatching a fresh implementer subagent per task to keep work isolated, then enforcing a two-stage review process for spec compliance and code quality before marking tasks complete.

How does subagent task orchestration work with test-driven development?

Subagent task orchestration integrates with TDD by mandating tests, commits, and self-review within each isolated implementer subagent, ensuring iterative fixes and correctness before final branch finishing.

What is the best way to ensure spec compliance when executing multi-step feature plans?

Ensure spec compliance by running a dedicated spec reviewer subagent immediately after the implementer subagent finishes each task, catching gaps before a separate code quality reviewer evaluates the implementation.

Can I use git worktrees with automated subagent code review workflows?

Yes, automated subagent code review workflows integrate with git worktrees to provide safe merges, executing independent tasks in isolated branches while the two-stage review process validates spec compliance and code quality.

Why does multi-task implementation suffer from inconsistent code quality across independent tasks?

Multi-task implementation suffers from inconsistent quality due to context pollution; dispatching a fresh implementer subagent per task and enforcing a sequential spec and code-quality review gate prevents over and under-building.

When should I not use a fresh subagent per task for implementation?

Avoid dispatching a fresh subagent per task when implementation tasks are highly interdependent or lack specific spec requirements, as the two-stage review process requires independent tasks needing tests, commits, and iterative fixes.