跳到正文
cccskills
免费在 GitHub 公开

magicpath

Find and inspect UI components in MagicPath, add them to an app, and adapt their behavior and styling. Apply themes or create and edit canvas components from code, including work based on an existing repository interface.

中文简介正在准备中。当前显示英语简介。

原文查看原文说明

Use when the user mentions MagicPath, designs, UI components, themes, canvas selections, or repo-to-canvas UI work; run magicpath-ai to search, inspect, install, or author components.

查看安装方法

适用场景

  • Find UI components and add them to an app
  • Apply a MagicPath theme to an interface
  • Create or edit a component on the canvas
  • Bring an existing repository interface into MagicPath

包含的文件(4)

  • SKILL.md6.1 KB
  • references/cli-reference.md3.5 KB
  • references/working-with-embedded-browsers.md1.9 KB
  • references/working-with-repositories.md3.1 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

MagicPath

MagicPath is a canvas and component platform. Use this skill when the user mentions MagicPath, designs, UI components, themes/design systems, team projects, selected canvas items, or bringing local/repository UI into a MagicPath canvas.

Always run MagicPath CLI commands as:

npx -y magicpath-ai <command> -o json

Use JSON output for data-returning commands and -y for non-interactive installs.

First Step

Run:

npx -y magicpath-ai info -o json

If the user is not authenticated, run:

npx -y magicpath-ai login
npx -y magicpath-ai whoami -o json

Pick the Workflow

  • Find or install a MagicPath component: search/list, confirm the right component, inspect it, then add/adapt it.
  • Work with the current canvas: use selection -o json for selected components/images, or active-project -o json for the open project.
  • Use team work: run list-teams -o json, then pass --team "<nameOrId>" to project, search, theme, or member commands.
  • Use a theme/design system: list-themes -o json, then get-theme <id-or-name> -o json; apply CSS variables, fonts, and prompt guidance in the target app.
  • Create or edit canvas components from code: use code start, edit only allowed files, then code submit --wait.
  • Bring an existing repo UI into MagicPath: follow Working with repositories.
  • Keep a MagicPath project open inside Codex's Browser when doing canvas work: follow Working with embedded browsers.

Find and Confirm Components

  1. If the user refers to a selected design/component/image, run selection -o json.
  2. If the user refers to the project they have open, run active-project -o json.
  3. Otherwise search or browse:
npx -y magicpath-ai search "button" -o json
npx -y magicpath-ai list-projects -o json
npx -y magicpath-ai list-components <projectId> -o json

Search/list results include generatedName, project context, owner fields, and often previewImageUrl. Use previews when visual context matters.

Stop and ask for confirmation before installing or editing unless the user gave an exact generatedName, selected canvas item, or component/project id.

Install Into an App

Use this when MagicPath is the source and the user's app is the destination.

  1. Inspect first:
npx -y magicpath-ai inspect <generatedName> -o json
  1. Read the target code before installing. Understand current props, callbacks, validation, layout, data flow, styling system, and accessibility behavior.
  2. For React/TypeScript apps, install:
npx -y magicpath-ai add <generatedName> -y -o json
  1. Import and render the installed component using the returned importStatement and usage.
  2. Adapt the installed source in src/components/magicpath/<name>/:
    • Replace static text and mock data with props or real project data.
    • Wire events, loading, error, empty, disabled, focus, and keyboard states.
    • Make fixed dimensions responsive.
    • Preserve existing behavior when replacing an existing component.
    • Match the app's styling and state-management patterns.

Do not run add just to read code. Use inspect for read-only source. For non-JS projects, inspect and translate the design into the target framework instead of running add.

Create or Edit Canvas Components

Use this when the MagicPath canvas is the destination.

npx -y magicpath-ai code start --project <projectId> --dir <workdir> --name "Component Name" --width <px> --height <px> -o json
npx -y magicpath-ai code start --component <componentId> --dir <workdir> -o json
npx -y magicpath-ai code submit --dir <workdir> --wait -o json

Rules for code work:

  • Run code start before writing files so the canvas shows the pending work.
  • Edit only src/App.tsx, src/index.css, src/components/generated/**, and temporary image assets under assets/**.
  • Usually leave src/App.tsx alone except for the theme value.
  • Put real implementation in src/components/generated/<Name>.tsx; split larger pieces into sibling files there.
  • Use Tailwind v4 through src/index.css; do not add tailwind.config.js.
  • Keep output responsive, centered, and free of device/browser mockups unless explicitly requested.
  • Build one screen per component. For related multi-view flows, use local React state inside one component; for independent screens, create separate components in separate workdirs.
  • Make interactive surfaces actually interactive: controlled inputs, real handlers, toggles, tabs, dialogs, form validation, hover/focus/disabled states, and useful transitions.
  • If selected canvas images are returned by code start, use the downloaded assetPath, not the short-lived accessUrl.
  • If code submit fails, fix only allowed files and resubmit.

code context is read-only. Do not use it as the submit path.

Teams, People, and Ownership

  • list-teams -o json: discover teams/workspaces.
  • list-members --team "<team>" -o json: resolve people to user ids.
  • list-projects --team "<team>" -o json: see team projects only.
  • list-components <projectId> --created-by <userId> -o json: find work by a person in a team project.

Personal projects are private to their owner unless shared. Do not search another person's personal work; search team projects instead.

Project and Share Links

Use share when you need a URL without opening a browser:

npx -y magicpath-ai share <generatedName> -o json
npx -y magicpath-ai share <projectId> -o json

Use view only when intentionally opening the OS browser:

npx -y magicpath-ai view <generatedName>
npx -y magicpath-ai view <projectId>

Never run view commands in parallel.

References

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Review charts, diagrams, and dashboards for accessibility, including color contrast, keyboard access, and text alternatives. Get specific guidance for preserving key findings in screen readers, static exports, and reduced-motion views.

中文简介正在准备中。当前显示英语简介。

  • Reviewing dashboard colors and keyboard access
  • Writing chart alt text and long descriptions
  • Preparing accessible figures for PDF reports
openai/plugins75032026年10月8日 更新

Give a photo collection a consistent look using Adobe tools to adjust color, tone, presets, and optional crops. Review a before-and-after sample before processing the full set, then receive links to the edited images.

中文简介正在准备中。当前显示英语简介。

  • Giving travel photos a consistent look
  • Unifying photos taken in mixed lighting
  • Applying a warm or cinematic style across a photo set
openai/plugins75032026年10月8日 更新

Visualize a logo, sketch, or design on products such as mugs, shirts, and business cards. Review a pilot image, generate a visually consistent set, and collect the results on a shareable Firefly Board.

中文简介正在准备中。当前显示英语简介。

  • Preview a logo on mugs and shirts
  • Turn a hand-drawn sketch into product mockups
  • Create consistent brand presentation images
openai/plugins75032026年10月8日 更新

Prepare images and videos for social platforms using Adobe tools. Create image crops that keep the subject in view, expand backgrounds when available, preview a full-set sample, and resize videos while preserving their aspect ratio.

中文简介正在准备中。当前显示英语简介。

  • Creating image variants for multiple social platforms
  • Cropping photos for Instagram Stories
  • Preparing LinkedIn images and YouTube thumbnails
openai/plugins75032026年10月8日 更新

Create flyers, social posts, business cards, and other visuals from Adobe Express templates. Choose a template, customize its text and background color, and receive a design link, with additional editing and export options depending on the available tools.

中文简介正在准备中。当前显示英语简介。

  • Creating an event flyer
  • Designing an Instagram post
  • Customizing business card text and colors
openai/plugins75032026年10月8日 更新

Turn an uploaded video into three highlight edits with Adobe Quick Cut. Choose a target length and style, then compare versions that feature different moments and download the ones you prefer.

中文简介正在准备中。当前显示英语简介。

  • Creating a short highlight video for social media
  • Condensing interview footage into a highlight reel
  • Comparing three edits with different moment selections
openai/plugins75032026年10月8日 更新

查看 openai 的全部技能

报告此技能的问题