← Back to Skills Marketplace
wangjiaocheng

AI Dev Workflow

by 波动几何 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
159
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install ai-dev-workflow
Description
此技能提供一个标准化、可复现的AI辅助编程工作流,通过三个有序步骤将模糊想法转化为生产就绪的代码:需求转复杂功能描述、描述加技术栈转方法调用结构、结构转完整项目代码。当用户想要开发完整软件系统、从零构建应用程序,或遵循规范的AI-人协作编程流程时,应使用此技能。
README (SKILL.md)

\r \r

AI Dev Workflow - 标准化AI辅助编程工作流\r

\r 一个标准化、可复现的三步AI辅助编程工作流,通过"样本模仿"与"提示词驱动"的有机结合,实现从模糊需求到生产代码的确定性转换。\r \r

核心原则\r

\r

  • 样本模仿:使用经过验证的文档样本引导需求扩展和技术设计,消除随机性。\r
  • 提示词驱动:提前建立元协议和架构提示词,确保所有生成物的一致性。\r
  • 渐进式展示:默认先交付架构蓝图,完整代码按需提供——让开发者掌控信息深度。\r \r

参考文档\r

\r 本技能在 references/ 目录下捆绑三份参考文档:\r \r | 文件 | 角色 |\r |------|------|\r | references/人工智能编程之复杂功能描述样本(待办任务).md | 样本1 — 展示如何编写结构化的复杂功能描述(需求点→功能点→执行点)。 |\r | references/人工智能编程之方法调用结构样本(待办任务).md | 样本2 — 展示如何编写跨平台的方法调用结构蓝图,涵盖客户端和服务端各层。 |\r | references/从灵感到实现:一个标准化、可复现的AI辅助编程工作流.md | 背景阅读 — 完整的方法论文档(原文为五步版本,本技能已简化为三步)。供人类或无程序能力的LLM阅读,用于理解背景原理和手动复制粘贴执行步骤。 |\r \r 使用方式:在执行对应工作流步骤时,将相关样本参考文档读入上下文。对于较大的参考文件,使用grep/搜索定位特定章节,而非加载整个文件。\r \r

工作流说明\r

\r 本技能已将两份提示词内置,AI可一次性获取完整上下文。实际使用流程如下:\r \r

内置提示词\r

\r 元协议(第一份提示词):\r

执行协议等待指令:复杂功能分拆成原子方法,原子方法生成完整代码,默认只展示方法调用结构但用户可以要求展示完整代码。\r \r 架构与技术栈提示词(第二份提示词):\r 做一个功能为{首次回复要求提供功能描述}的APP,使用{客户端(UI层→ViewModel→UseCase→Repository→数据源)→服务器(Controller→Service→Repository→Entity→Database)}的架构,采用{Android}客户端、{SpringBoot}服务器和{MySQL}数据库并统一使用{Kotlin}语言。\r \r 两份提示词定义了工作规则和技术路径,AI已知晓,无需用户单独发布。\r \r

实际执行步骤\r

\r

第一步:需求 → 复杂功能描述\r

\r 输入:用户提供的功能需求描述。\r \r 执行:\r

  1. references/人工智能编程之复杂功能描述样本(待办任务).md 读入上下文。\r
  2. 参考样本格式,将需求扩展为结构化的复杂功能描述文档(需求点 → 功能点 → 执行点)。\r \r 输出:复杂功能描述文档。\r \r 可修改:用户可对输出的描述进行修改、补充、调整。\r \r ---\r \r

第二步:描述 + 技术栈 → 方法调用结构\r

\r 输入:\r

  • 第一步输出的复杂功能描述(或修改后版本)\r
  • 用户填充第二份提示词中的所有变量:\r
    • {功能描述} — 第一步输出的结构化需求文档\r
    • {架构} — 前后端分层架构模式\r
    • {客户端技术} — 客户端框架\r
    • {服务端技术} — 服务端框架\r
    • {数据库} — 数据库类型\r
    • {编程语言} — 客户端/服务端编程语言(可统一或分开指定)\r \r 执行:\r
  1. references/人工智能编程之方法调用结构样本(待办任务).md 读入上下文。\r
  2. 在两份内置提示词的基础上,结合确定的技术栈,为每个功能点生成跨平台方法调用结构。\r \r 输出:方法调用结构蓝图(涵盖客户端和服务端各层)。\r \r 可修改:用户可对输出的结构进行修改、调整方法签名、增删节点。\r \r ---\r \r

第三步:结构 → 项目完整代码\r

\r 输入:第二步输出的方法调用结构(或修改后版本)。\r \r 执行:\r 根据方法调用结构,生成完整项目内容,包括:\r

  • 客户端各层代码(UI、ViewModel、UseCase、Repository等)\r
  • 服务端各层代码(Controller、Service、Repository、Entity等)\r
  • 数据库相关代码(表结构、迁移脚本、初始数据等)\r
  • 配置文件(环境配置、依赖管理、构建配置等)\r
  • 部署文件(Dockerfile、docker-compose、CI/CD配置等)\r
  • 项目结构文件(README、.gitignore等)\r \r 输出:可运行的完整项目(代码 + 配置 + 部署)。\r \r 可修改:用户可对生成的代码进行修改、优化、调整。\r \r ---\r \r

工作流总览\r

\r | 步骤 | 输入 | 内置提示词 | 样本参考 | 输出 | 可修改 |\r |------|------|------------|----------|------|--------|\r | 1 | 功能需求 | — | 复杂功能描述样本 | 复杂功能描述 | ✅ |\r | 2 | 复杂功能描述 + 技术栈 | 两份提示词 | 方法调用结构样本 | 方法调用结构 | ✅ |\r | 3 | 方法调用结构 | 两份提示词 | — | 完整项目 | ✅ |\r \r 核心理念:每个阶段的输出都可被用户修改,下一阶段基于修改后的内容继续推进。用户始终保有控制权。\r \r

核心价值\r

\r

  1. 消除模糊性:通过结构化样本,确保需求扩展和技术设计的深度与格式符合工程标准。\r
  2. 确保一致性:通过架构提示词,保证所有产出都遵循统一的最佳实践和技术栈。\r
  3. 提升可控性:通过元协议,将生成过程置于开发者的可控节奏之下,先蓝图后实现。\r
  4. 实现知识传承:将内化的架构经验沉淀为可复用的"样本"和"提示词",使最佳实践得以规模化复制。\r
Usage Guidance
This skill is instruction-only and internally consistent with its stated purpose, and it does not request credentials or install software — that reduces installation risk. Before using it: (1) be aware the skill embeds two built-in prompts that default to a specific tech stack (Android / SpringBoot / MySQL / Kotlin); if you need a different stack, explicitly override those variables when running the workflow; (2) treat generated code as a scaffold — review, test, and security-audit any production code the skill generates (it can contain bugs, insecure defaults, or license issues); (3) avoid feeding sensitive credentials or private repository secrets into the workflow or prompts (the skill itself does not request them, but downstream use could); and (4) note the skill's provenance is unknown (no homepage/author verification) — rely on code review and testing rather than implicit trust in the provider.
Capability Analysis
Type: OpenClaw Skill Name: ai-dev-workflow Version: 1.0.0 The skill bundle defines a structured AI-assisted development workflow designed to transform user requirements into full-project code through a three-step process. It utilizes reference templates in the `references/` directory to guide the AI in generating consistent architectural designs (e.g., MVVM for Android, SpringBoot for backend) and method call structures. No evidence of data exfiltration, malicious execution, or harmful prompt injection was found; the instructions are transparent and strictly aligned with the stated goal of software development productivity.
Capability Assessment
Purpose & Capability
Name/description (standardized AI-assisted dev workflow) match the contents: SKILL.md and three detailed reference docs describe samples, prompts, and a three-step pipeline that produce designs and code. There are no unrelated env vars, binaries, or install steps requested.
Instruction Scope
All runtime behavior is described in SKILL.md (read bundled references, apply two built-in prompts, produce structured outputs). This stays within the declared purpose. Note: the skill includes built-in prompts that lock a default architecture/tech-stack (Android, SpringBoot, MySQL, Kotlin); this is coherent but may be surprising if you expected multi-stack flexibility — outputs will follow those defaults unless user overrides them.
Install Mechanism
Instruction-only skill with no install spec and no code files to execute. No downloads, package installs, or archive extraction are present.
Credentials
The skill requests no environment variables, credentials, or config paths. Bundled reference files are local to the skill and used as context; no unrelated secrets are requested.
Persistence & Privilege
Flags show no 'always: true' and default autonomous invocation is allowed (normal). The skill does not request persistent system changes or access to other skills' configs.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install ai-dev-workflow
  3. After installation, invoke the skill by name or use /ai-dev-workflow
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
- Initial public release of AI Dev Workflow (v1.0.0). - Introduces a standardized, repeatable three-step workflow for AI-assisted software development: from requirements to structured description, to cross-platform method blueprint, to complete project code. - Bundles reference samples and documentation for each workflow stage to guide and standardize outputs. - Embeds two core prompt templates: protocol for task decomposition and a technical stack/architecture scaffold. - Supports user review and modification at every stage, ensuring developer control and adaptability. - Workflow designed to eliminate ambiguity, enforce consistency, and accelerate production-ready AI coding.
Metadata
Slug ai-dev-workflow
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is AI Dev Workflow?

此技能提供一个标准化、可复现的AI辅助编程工作流,通过三个有序步骤将模糊想法转化为生产就绪的代码:需求转复杂功能描述、描述加技术栈转方法调用结构、结构转完整项目代码。当用户想要开发完整软件系统、从零构建应用程序,或遵循规范的AI-人协作编程流程时,应使用此技能。 It is an AI Agent Skill for Claude Code / OpenClaw, with 159 downloads so far.

How do I install AI Dev Workflow?

Run "/install ai-dev-workflow" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is AI Dev Workflow free?

Yes, AI Dev Workflow is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does AI Dev Workflow support?

AI Dev Workflow is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created AI Dev Workflow?

It is built and maintained by 波动几何 (@wangjiaocheng); the current version is v1.0.0.

💬 Comments