subagent-driven-development

Dispatch implementer, spec-reviewer, and code-reviewer subagents for per-task workflows.

1|Updated Jan 14, 2026
One-click install
npx skills add https://github.com/DecentralizedGeo/stac-manager --skill subagent-driven-development-decentralizedgeo
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/DecentralizedGeo/stac-manager/tree/main/.github/skills/subagent-driven-development
Command: npx skills add https://github.com/DecentralizedGeo/stac-manager --skill subagent-driven-development-decentralizedgeo

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill enables orchestrating implementation plans by dispatching fresh subagents for individual tasks within the same session, with a two-stage review after each task to ensure spec compliance and code quality.

Core Features & Use Cases

  • Per-task subagent dispatch
  • Two-stage review: spec compliance, then code quality
  • Self-review before reporting back
  • Maintains isolated context per task to avoid cross-task contamination
  • Works in a single session to accelerate iteration

Quick Start

Dispatch an implementer subagent for Task 1 and then complete the two-stage review (spec compliance and code quality) for each subsequent task.

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 task dispatch in software development?

Subagent orchestration dispatches fresh subagents for individual tasks within a single session, maintaining isolated context per task to avoid cross-task contamination while enabling rapid iteration.

What's the best way to ensure spec compliance during automated plan execution?

Spec compliance is enforced through a two-stage review process after each task, utilizing a dedicated spec-reviewer subagent alongside a code-reviewer to validate implementation against requirements.

Can I run an implementer subagent and code reviewer in the same session?

Yes, the workflow runs entirely in a single session by dispatching an implementer subagent for each task, followed by spec and code quality reviews, plus a pre- and post-task self-review process.

How do subagents handle isolated context for independent tasks?

Each subagent maintains isolated context per task to prevent cross-task contamination, ensuring that the implementation and review processes for one task do not interfere with subsequent tasks.

Does subagent-driven development work for complex plans with independent tasks?

Yes, it is specifically designed for developing complex plans where tasks are largely independent but require tight review loops for spec compliance and code quality, accelerating iteration speed.

What are the limitations of using a per-task subagent dispatch workflow?

This approach is limited to tasks that are largely independent; highly interdependent tasks may not benefit from isolated context and the sequential implementer and two-stage review process.