subagent-driven-development

Coordinate plan execution by dispatching isolated subagents for each task.

Updated Apr 11, 2026
One-click install
npx skills add https://github.com/ilderaj/superpowering-with-files --skill subagent-driven-development-ilderaj
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/ilderaj/superpowering-with-files/tree/main/harness/upstream/superpowers/skills/subagent-driven-development
Command: npx skills add https://github.com/ilderaj/superpowering-with-files --skill subagent-driven-development-ilderaj

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Subagent-Driven Development solves the challenge of executing complex plans by dispatching isolated subagents for each task within a single session, preserving context for coordination while preventing cross-task contamination.

Core Features & Use Cases

  • Per-task subagent dispatch: allocate a dedicated agent per task to avoid context bleed and enable parallel progress where tasks allow.
  • Two-stage review workflow: spec compliance review followed by code quality review before task completion to ensure quality and adherence.
  • Plan-driven execution: coordinates tasks from a central plan with explicit prompts and governance templates.
  • Robust status handling: supports DONE, DONE_WITH_CONCERNS, NEEDS_CONTEXT, and BLOCKED with structured re-dispatch.

Quick Start

Start by loading the plan and begin dispatching a dedicated subagent for each task using the implementer, spec reviewer, and code-quality reviewer prompts.

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-by-task execution in a single session?

Subagent orchestration dispatches isolated subagents for each task within a single session, preserving context for coordination while preventing cross-task contamination. You load a central plan and allocate a dedicated agent per task using specific governance prompts.

What is the two-stage review workflow for subagent-driven development?

The two-stage review workflow enforces quality by running a spec compliance review followed by a code quality review before task completion. This ensures adherence to the plan through structured reviewer prompts before any task is marked as finished.

How do I handle blocked or incomplete tasks during subagent plan execution?

Subagent plan execution handles incomplete tasks using structured status handling with DONE, DONE_WITH_CONCERNS, NEEDS_CONTEXT, and BLOCKED states. These statuses govern execution and trigger structured re-dispatch of subagents when necessary.

When should I use isolated subagents for plan execution instead of a single agent?

You should use isolated subagents for plans with independent tasks that require strict per-task isolation. This approach prevents context bleed between tasks and enables parallel progress where the plan allows, while preserving overall coordination context.

What prompts do I need to dispatch subagents for task execution?

Dispatching subagents for task execution requires implementer-prompt.md, spec-reviewer-prompt.md, and code-quality-reviewer-prompt.md. These templates govern the execution and two-stage review workflow for each isolated subagent.