Subagent-Driven Development

Subagent-Driven Development is an AI Engineering skill for Claude Code, published by jnMetaCode in superpowers-zh.

7.7K stars729 forkson jnMetaCode/superpowers-zhAdded 2026/08/15Repository updated 2026/08/12

agent-skillsagentic-codingai-codingchineseclaude-codecode-reviewcursorgemini-clikiromcpnpm-packageprompt-engineeringskillssuperpowerstddtrae

Install in seconds

Install Subagent-Driven Development

Copy Subagent-Driven Development into your Claude Code skills folder. Run the command in your terminal, or review the source on GitHub before installing.

terminal

npx degit https://github.com/jnMetaCode/superpowers-zh/tree/main/skills/subagent-driven-development ~/.claude/skills/subagent-driven-development

Requires Node.js. Downloads this skill only — not the rest of the repository — into your Claude Code skills folder.

Without Node.js

git clone https://github.com/jnMetaCode/superpowers-zh.git

Clones the whole repository, then copy the skill’s own directory into your skills folder yourself.

In this catalog

Source file: skills/subagent-driven-development/SKILL.md in jnMetaCode/superpowers-zh
Installs to: ~/.claude/skills/subagent-driven-development
Collection: One of 20 skills cataloged from this repository
Category: AI Engineering — 2451 skills

What Subagent-Driven Development does

Subagent-Driven Development executes an implementation plan by dispatching a fresh subagent for each task, reviewing each task for spec compliance and code quality, and ending with a broad branch-wide review. Use it when executing a plan of largely independent tasks in the current session.

Subagent-Driven Development is cataloged under AI Engineering on DirSkills. Subagent-Driven Development comes from a repository tagged agent-skills, agentic-coding, ai-coding, chinese and claude-code.

Documentation

README

Excerpt

子智能体驱动开发

通过为每个任务分派一个全新的实现子智能体来执行计划：每个任务完成后做一次任务审查（规格合规性 + 代码质量），全部任务结束后再做一次覆盖整个分支的宽范围审查。

为什么用子智能体： 你把任务委派给具有隔离上下文的专用智能体。通过精心设计它们的指令和上下文，确保它们专注并成功完成任务。它们绝不应继承你会话的上下文或历史记录——你要精确构造它们所需的一切。这样也能为你自己保留用于协调工作的上下文。

核心原则： 每个任务一个全新子智能体 + 任务审查（规格 + 质量）+ 结尾宽范围审查 = 高质量、快速迭代

旁白： 工具调用之间最多说一句简短的旁白——进度账本和工具结果本身就是记录。

持续执行： 不要在任务之间停下来向你的人类伙伴确认。不间断地执行计划里的所有任务。唯一该停下的理由是：你无法解决的 BLOCKED 状态、确实妨碍推进的歧义，或所有任务已完成。"我该继续吗？"之类的询问和进度小结都在浪费他们的时间——他们让你执行计划，那就执行。

何时使用

This is the opening of the README. Read the full README on GitHub.

Frequently asked about Subagent-Driven Development

What else does jnMetaCode publish alongside Subagent-Driven Development?
Subagent-Driven Development is one of 20 skills that DirSkills catalogs from jnMetaCode/superpowers-zh, the repository it ships in. Its siblings there include Brainstorming, Chinese Code Review and Chinese Commit Conventions. Each one is a separate skill with its own page in this directory, installs the same way Subagent-Driven Development does, and is maintained by jnMetaCode in that same repository. The rest of the collection is listed on the jnMetaCode/superpowers-zh page.
How does Subagent-Driven Development compare to other AI Engineering skills?
Subagent-Driven Development ranks #420 by stars among the 2451 AI Engineering skills in this catalog. The most-starred ones next to it are Architecture Decision Records, AI-First Engineering and Agentic OS. DirSkills ranks by the star count of the repository each skill ships in, so that order reflects how popular those repositories are rather than any review of Subagent-Driven Development against them. Open each page to compare what they document and how they install.