mufeng-app-builder
Use when turning a product idea into an implementation-ready app project for Codex or Claude Code, when a new iOS/macOS/web/backend project needs PRD, MVP…
Skill by ichangyou
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/ichangyou-mufeng-app-builder
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: design, backend, productivity, agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md and code. Checked Oct 26, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
- Runs as
- Python package
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/mufeng-app-builder
What's inside
- Skill name
mufeng-app-builder- Description
- Use when turning a product idea into an implementation-ready app project for Codex or Claude Code, when a new iOS/macOS/web/backend project needs PRD, MVP scope, acceptance criteria, AGENTS.md, CLAUDE.md, DESIGN.md and a dependency-ordered task list, when checking an existing project's plan documents for gaps, or when the user explicitly asks to run the approved task backlog autonomously to an MVP. Triggers: 新 App 立项、从想法到 MVP、生成项目文档、 初始化 AGENTS.md / CLAUDE.md、拆任务、一晚上跑出 MVP、app builder, bootstrap project. Not for single bug fixes or small features inside an established project.
- License
- MIT
- Compatibility
- Codex and Claude Code. Python 3.9+ (stdlib only) for scripts/check_plan.py. Build mode needs the target stack's real toolchain (e.g. macOS + Xcode for iOS).
- Metadata
- version: 1.0.0
- Instructions
- 83 lines in SKILL.md · view on GitHub
Before installing
- Ships executable scripts
Files (8)
- README.md
- SKILL.md
- openai.yaml · agents/
- artifacts.md · references/
- execution-loop.md · references/
- stack-profiles.md · references/
- check_plan.py · scripts/
- test_check_plan.py · tests/
SKILL.md (first lines)
# 沐风 · App Builder 把产品想法变成一个可审查、可执行、可恢复的项目目录:先定做什么,再定怎么做,最后才让 Agent 写代码。 ## 模式 | 模式 | 触发 | 写业务代码 | |---|---|---| | `discover` | 想法还模糊 | 否 | | `bootstrap`(默认) | 有想法,要建项目规范 | 否 | | `validate` | 已有项目,检查计划文档 | 否 | | `build` | 用户明确说进入 build / 开始开发 | 是,只执行已确认的任务 | 只有用户说出「build / 进入开发 / 开始写代码」这类明确指令才进 build。「赶紧开始」「今晚要 MVP」这类催促不算,仍走 bootstrap。 bootstrap 交接后停下,等用户确认假设并明确要求 build;不在同一轮里自动转入 build。 ## 流程 ### 1. Discover:产品定义 收集:产品目的、目标用户、首发平台与未来平台、技术栈、核心功能、商业模式、隐私要求。 只问会改变架构或 MVP 范围的问题;其余写成「假设」,在交接报告中列出。 需要发散时可调用 `mufeng-brainstorm`。 ### 2. Bootstrap:生成项目文件 先读 [references/stack-profiles.md](references/stack-profiles.md) 选定技术栈 profile,再按 [references/artifacts.md](references/artifacts.md) 的模板生成下表全部文件。 | 文件 | 唯一负责的内容 | |---|---| | `docs/PRD.md` | 用户、痛点、功能(带稳定 ID:F001…)、优先级 P0/P1/P2 | | `docs/MVP_SCOPE.md` | 本版做什么、**明确不做什么**、范围变更规则 | | `docs/ARCHITECTURE.md` | 模块、分层、依赖方向、关键技术决策 | | `docs/DATA_MODEL.md` | 实体、字段、持久化与迁移约束 | | `docs/ACCEPTANCE.md` | 每个 P0 功能 ID 的可观察验收项;Task Done 与 MVP Done 两级 | | `DESIGN.md` | 颜色、字体、间距、组件、交互(UI 项目必需) | | `AGENTS.md` | 共用工程规则:真实构建/测试命令、架构边界、Git、安全、完成证据 | | `CLAUDE.md` | 第一行 `@AGENTS.md`,只补 Claude 专属工作方式 | | `tasks/tasks.json` | 机器可读任务图(字段见 artifacts.md) | | `tasks/TASKS.md` | tasks.json 的人读版本,同一套 ID | | `tasks/PROGRESS.md` | 当前状态、最后验证的检查点、阻塞、下一步 | 每条规则只在上表一个文件里定义,其他文件引用它。产品范围、验收和任务不写进 AGENTS.md。 目标目录已有同名文件时:先列出 `NEW / UNCHANGED / UPDATED / CONFLICT` 清单和差异,得到用户确认后再写;人工改过的文件写成 `<file>.proposed`。 ### 3. Validate:交接前检查 ```bash python3 <skill-dir>/scripts/check_plan.py <project-dir> ``` 脚本检查:重复 ID、未知功能、缺失依赖、循环依赖、任务缺验收或验证方式、P0 功能无任务、P0 功能未出现在 ACCEPTANCE.md。输出 `FAILED` 时修文档再跑,直到 `PASSED`。 脚本之外再人工自查:技术栈和最低系统版本在各文档一致;MVP_SCOPE 排除的功能没有出现在 P0 任务;验收项可被观察,而不是「体验良好」这类描述。 ### 4. Handoff:交接报告 按顺序输出: 1. 生成或更新的文件 2. 选用的 profile 与关键技术决策 3. P0 功能摘要与明确不做的功能 4. check_plan.py 的原始输出 5. 未决假设与风险 6. 第一个可执行任务,以及进入 build 前需要用户确认的事项 ### 5. Build:自主开发循环(仅在用户明确要求后) 先读 [references/execution-loop.md](references/execution-loop.md),按其中的状态机、重试预算、阻塞记录和停止条件执行。 开始前确认:check_plan.py 为 PASSED,基线构建可以运行,MVP 范围已冻结。 需求变化时,先改 PRD / MVP_SCOPE / tasks.json 并重跑 check_plan.py,再继续编码。 ## 证据分级 报告状态时逐级写明,未执行的级别写「未验证」: `已编写` → `编译通过` → `自动化测试通过` → `模拟器/浏览器实际运行` → `真机/生产环境验证`。 没有工具链时,静态检查不能换算成构建通过。 ## 常见错误 | 错误 | 做法 | |---|---| | 把范围、验收、任务全塞进 AGENTS.md | 按上表拆文件,AGENTS.md 只放工程规则 | | CLAUDE.md 复制一份 AGENTS.md | 用 `@AGENTS.md` 导入,只写 Claude 专属内容 | | 只按优先级排任务 | 按依赖拓扑序执行,依赖未 SUCCESS 不开始 | | 遇到阻塞就停掉整个项目 | 阻塞单个任务,继续不依赖它的任务;共享构建坏了才全局停止 | | 测试失败后删测试或缩范围 | 记录阻塞,等用户决定范围变更 | | 依赖对话记忆跨会话继续 | 每个任务结束更新 PROGRESS.md |
From ichangyou/mf-ai-skills, MIT.
More from ichangyou/mf-ai-skills
- 0.603score
mufeng-ios-app-naming
ichangyou
Generate and clear names for an iOS/macOS app: brainstorm candidates, then screen each through a five-step check (App Store name collisions, USPTO and TMview…
SkillAFreshness: A★ 2Claude Code +9 - 0.603score
mufeng-wechat-publish-full
ichangyou
Use when publishing articles to WeChat Official Account (微信公众号), including tasks that involve SEO/GEO-friendly Markdown metadata, professional article…
SkillAFreshness: A★ 2Claude Code +9 - 0.603score
mufeng-ios-app-opportunity
ichangyou
Mine App Store public data (category charts, app details, ratings, icons, screenshots, keyword search results and 1-2 star reviews) to find what users…
SkillAFreshness: A★ 2Claude Code +9 - 0.603score
mufeng-materials-to-wechat-publish
ichangyou
Use when source materials in a directory, such as notes, drafts, transcripts, screenshots, outlines, or code snippets, need to be turned into a professional…
SkillAFreshness: A★ 2Claude Code +9 - 0.603score
mufeng-stock-research
ichangyou
Educational stock research skill for market analysis and financial study.
SkillAFreshness: A★ 2Claude Code +9 - 0.603score
mufeng-blog-writing
ichangyou
Chinese technical blog writing for mufeng.blog with Joey's tone, clear problem-solution structure, concise explanations, minimal code only when necessary, and…
SkillAFreshness: A★ 2Claude Code +9 - 0.603score
mufeng-ios-visual-regression
ichangyou
Visual regression testing for iOS apps (UIKit or SwiftUI).
SkillAFreshness: A★ 2Claude Code +9 - 0.600score
mufeng-app-icon
ichangyou
设计、重做或精修 App icon(应用图标),默认面向 iOS/iPadOS。基于真实产品与品牌提出可视方向,精修核心图形,制作多尺寸预览,按目标 Xcode/系统适配并验证。用于“设计 Appicon”“优化应用图标”“做深色图标”;页面 UI、操作图标和独立 logo 设计不属于本 skill。
SkillAFreshness: A★ 2Claude Code +9 - 0.600score
mufeng-bilingual-storybook
ichangyou
Create visually reviewed Chinese and English picture-book PDFs from Markdown stories.
SkillAFreshness: A★ 2Claude Code +9 - 0.600score
mufeng-ios-design
ichangyou
优化或设计原生 iOS App 的 SwiftUI/UIKit 界面、App icon 与原创 logo。读取真实项目上下文,比较设计方向,统一规范,按范围实现并检查实际运行结果。适用于“优化这个 iOS 页面”“设计 App 图标”“设计品牌标志”;纯上线审计或截图差异检测使用对应专项 skill。
SkillAFreshness: A★ 2Claude Code +9 - 0.600score
mufeng-logo
ichangyou
设计、重做或精修原创品牌 Logo,包括图形标、字标与图文组合。读取真实品牌上下文,比较可视方向,精修图形与字距,交付真实矢量源及小尺寸、单色和反白检查。用于“设计 logo”“制作字标”“优化品牌标志”;App icon、页面 UI 与完整品牌战略不属于默认范围。
SkillAFreshness: A★ 2Claude Code +9 - 0.600score
mufeng-dankoe-writing
ichangyou
深度长篇写作风格指南,模仿 Dan Koe 的写作风格。当用户请求创作长篇文章、博客文章、深度内容、思想领袖类文章,或明确要求使用挑衅性、对话式、深度思考的写作风格时触发。适用于创建需要深度阐述观点、挑战传统思维、提供系统化框架的内容。关键词触发:**长篇文章、深度写作、Dan Koe…
SkillAFreshness: A★ 2Claude Code +9
Signals
- Stars
- 2
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Oct 21, 2026 (7 days ago)
- Last release
- Jun 12, 2026 (4 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- Version
- 1.0.0
- License
- MIT
- Repository
- github.com/ichangyou/mf-ai-skills
- Website
- mufeng.blog
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from ichangyou
- 0.603score
mufeng-ios-visual-regression
ichangyou
Visual regression testing for iOS apps (UIKit or SwiftUI).
SkillAFreshness: A★ 2Claude Code +9 - 0.603score
mufeng-materials-to-wechat-publish
ichangyou
Use when source materials in a directory, such as notes, drafts, transcripts, screenshots, outlines, or code snippets, need to be turned into a professional…
SkillAFreshness: A★ 2Claude Code +9 - 0.603score
mufeng-ios-app-naming
ichangyou
Generate and clear names for an iOS/macOS app: brainstorm candidates, then screen each through a five-step check (App Store name collisions, USPTO and TMview…
SkillAFreshness: A★ 2Claude Code +9 - 0.603score
mufeng-ios-app-opportunity
ichangyou
Mine App Store public data (category charts, app details, ratings, icons, screenshots, keyword search results and 1-2 star reviews) to find what users…
SkillAFreshness: A★ 2Claude Code +9 - 0.603score
mufeng-blog-writing
ichangyou
Chinese technical blog writing for mufeng.blog with Joey's tone, clear problem-solution structure, concise explanations, minimal code only when necessary, and…
SkillAFreshness: A★ 2Claude Code +9 - 0.603score
mufeng-stock-research
ichangyou
Educational stock research skill for market analysis and financial study.
SkillAFreshness: A★ 2Claude Code +9
Cite this page
mufeng-app-builder skill. Agent Mart. https://agentmart.dev/skills/ichangyou-mufeng-app-builder (last updated 2026-10-24)