skills.rest
xiaobin
xiaobinCOMMUNITY
Explore all skills by this author →
GitHub

subagent-driven-development

Distribute implementation tasks to fresh subagents with two-stage reviews.

1|Updated Apr 22, 2026
Search Tags:#automation#workflow#planning#development#subagent#code-review
One-click install
npx skills add https://github.com/xiaobin/superpowers-kimi --skill subagent-driven-development-xiaobin
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: subagent-driven-development
Source: https://github.com/xiaobin/superpowers-kimi/tree/main/.agents/skills/subagent-driven-development
Command: npx skills add https://github.com/xiaobin/superpowers-kimi --skill subagent-driven-development-xiaobin

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Coordinating implementation tasks across independent subagents without cross-contamination of context, enabling fast, high-quality execution within a single session.

Core Features & Use Cases

  • Fresh subagent per task to prevent context leakage and promote focused execution.
  • Two-stage reviews after each task: spec compliance first, then code quality review.
  • Supports both sequential and parallel task flows with clear status updates and rework loops.

Quick Start

Kick off the process by dispatching the first task to a fresh subagent and start the two-stage review loop.

Frequently Asked Questions about subagent-driven-development

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

FAQPage Schema
How do I isolate context for each implementation task when automating development workflows?▼

You isolate context by distributing implementation tasks to fresh subagents, ensuring each task operates independently without cross-contaminating context from other tasks. This mechanism promotes focused execution and prevents context leakage.

What is the best way to run automated code reviews for isolated development tasks?▼

Automated code reviews for isolated development tasks are best handled through a two-stage review process: spec compliance review first, then code quality review, applied to each task executed by a fresh subagent.

Can I execute subagent tasks in parallel or do they have to run sequentially?▼

Subagent task execution supports both parallel and sequential flows, allowing you to choose the mode that fits your development workflow while maintaining clear status updates and rework loops.

How do I start a subagent-driven development workflow for my project?▼

You start the workflow by dispatching the first implementation task to a fresh subagent, which then enters a two-stage review loop consisting of spec compliance and code quality checks before proceeding.

Why does context leakage happen between tasks and how does a fresh subagent prevent it?▼

Context leakage happens when tasks share the same execution environment, causing interference. A fresh subagent prevents this by providing an isolated context for every single task, ensuring high-quality execution.

Does this development workflow support self-review and status reporting?▼

Yes, the workflow supports structured prompts for implementer, spec reviewer, and code quality reviewer roles, plus built-in self-review and status reporting to track task progress and rework loops.

Related Skills

Explore More Software Engineering Skills→
bitovibitovi
9

subagent-driven-development

Automate task execution and iterative code reviews with subagents.

Official
Advanced
SentioLabsSentioLabs
1

implement

Delegate arc implementation tasks to isolated subagents with TDD workflows.

Official
Advanced
bobmatnycbobmatnyc
146

Requesting Code Review

Dispatch a code-reviewer subagent to review changes against plans and requirements.

Community
Basic
wibus-weewibus-wee
54

dispatching-parallel-agents

Dispatch specialized agents with isolated context to execute independent tasks in parallel.

Community
Intermediate