可爱的可视化 AI 构建工具
要求 AI 将项目源代码导出或记录,识别生成的假设,列出缺失的移动资产和应用身份。
问题
源代码要求
AI 只有了解当前应用的实际源代码、构建输出、资产和 native 假设时,才能准备出有用的移动交付。
要求 AI 将项目源代码导出或记录,识别生成的假设,列出缺失的移动资产和应用身份。
repo-first 助手可以直接修改文件,但仍应返回可审查的移动交付,而不是假设 native 发布工作已完成。
现有的应用需要相同的检查:构建输出、路由、认证重定向、资产路径、原生插件需求和发布通道边界
解决方案
解决方案
AI 指令
将此内容粘贴到 AI 构建器中,等待 web 应用程序工作。它告诉代理程序检查、文档化和准备移动交付,不假设本地工具访问。
AI 移动发布说明
You are preparing this existing web app to become an installable iOS and Android app.
Do not assume this builder can run local tooling. If this environment cannot create native projects or run local tools, create a clean handoff document instead.
Inspect the app and return:
1. The framework, package manager, production build process, and build output folder.
2. Whether the app can be exported or committed to a real Git repository.
3. The app name, bundle ID or package ID suggestion, app icon and splash-screen source, and any missing brand assets.
4. Auth, redirect, deep-link, storage, camera, push, payment, geolocation, file, or notification features that need native plugin planning.
5. Routing, environment variables, API URLs, and asset assumptions that could break inside a native WebView.
6. A proposed Capacitor configuration, including appId, appName, webDir, and required plugins, without inventing secrets.
7. A release-channel plan: production, staging, preview, and rollback expectations for web-layer updates.
8. A list of tasks that must happen outside the AI builder, such as Apple signing, Android keystore setup, native sync, store records, and signed builds.
9. A docs/mobile-release-handoff.md file with the exact checklist, open questions, and files that a developer or build service should handle next.
10. A short summary of what you changed, what you could not do in this environment, and the next human action.
交付清单
目标不是一键自动化。目标是清晰可审查的交付,开发人员或发布服务可以继续。
要求 AI 识别框架、构建输出、路由模型、资产、环境变量和本机敏感特性。
让应用程序身份、Capacitor 配置、插件需求、权限、图标和启动屏幕输入,以及缺失的本机假设。
要求一个名为 docs/mobile-release-handoff.md 的文件,包含更改的文件、开放的问题、未发明的秘密以及超出 AI 构建器的任务。
审查可以通过 web 层更新的内容、需要本机构建的内容以及等待签名或存储访问的工作。
用户信号
最强大的结果来自于 AI 不被要求以魔法方式发布移动应用。它被要求检查 web 应用、准备配置并为发布步骤编写交接文档。
常见的 AI web-to-mobile feedback