Token导航 LogoToken导航

Agent Skill · 前端设计

swiftui-design-principles(SwiftUI 设计原则)

用于辅助界面设计、视觉规范、排版、配色、布局和交互体验优化。它适合让 Agent 根据产品场景整理页面结构、生成 UI 方案、检查视觉一致性或改进组件层级。使用时需要结合现有品牌、设计系统和用户任务,不应只堆装饰元素;涉及真实页面改动时,应通过截图或浏览器预览检查文本溢出、对齐和响应式表现。

swiftui-design-principles 是前端设计类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为权限未标注;许可证信息还不明确,第三方扫描已通过,按安装量在前端设计类的 10,216 个 Skill 里排第 852。

总安装 10,680周安装 445GitHub Stars 10下载 3,560核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:swiftui-design-principles(SwiftUI 设计原则)
来源仓库:https://github.com/arjitj2/swiftui-design-principles
仓库路径:skills/swiftui-design-principles
安装命令:
npx skills add https://github.com/arjitj2/swiftui-design-principles --skill swiftui-design-principles
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/arjitj2/swiftui-design-principles --skill swiftui-design-principles

这个 Skill 是做什么的

前端设计类 Skill 关注页面结构、视觉规范和组件写法,输出通常是可运行的界面代码。

swiftui-design-principles 用于辅助界面设计、视觉规范、排版、配色、布局和交互体验优化。适合让 Agent 根据产品场景整理页面结构、生成 UI 方案、检查视觉一致性或改进组件层级。使用时需要结合现有品牌、设计系统和用户任务,不应只堆装饰元素。

涉及真实页面改动时,应通过截图或浏览器预览检查文本溢出、对齐和响应式表现。通过 npx skills add 命令从指定 GitHub 仓库安装并使用。

来源仓库内路径:skills/swiftui-design-principles

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/arjitj2/swiftui-design-principles --skill swif…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    权限没有明确标注,第一次使用时只给一个无关紧要的任务,观察它到底读了哪些文件和网络。

常见问题

6 条

swiftui-design-principles 怎么安装?

在终端执行 npx skills add https://github.com/arjitj2/swiftui-design-principles --skill swiftui-design…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

swiftui-design-principles 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

swiftui-design-principles 需要什么权限?

权限级别没有明确标注。第一次使用时建议只给无关紧要的任务,观察它实际读写了哪些内容。

swiftui-design-principles 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

swiftui-design-principles 用的人多吗?

公开数据是累计安装 10,680 次,近一周安装 445 次,下载 3,560 次,GitHub 10 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

swiftui-design-principles 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理前端设计类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

前端设计类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
frontend-design

前端设计 · Codex、Claude

8,823,360126,072只读MIT通过
vercel-composition-patterns

Vercel 组合模式 · Codex、Claude

3,858,81625,893需要联网MIT通过
vercel-react-native-skills

Vercel React native skills 前端 · Codex、Claude

2,468,66425,878只读MIT通过
adapt

adapt 搜索 · Codex、Claude

2,027,04023,125操作浏览器未明确通过
bolder

bolder 分析 · Codex、Claude

2,004,28823,130只读未明确通过
swiftui-design-principles

SwiftUI 设计原则 · 本页

10,68010权限未标注未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

This skill encodes design principles derived from comparing polished, production-quality SwiftUI apps against poorly-built ones. The patterns here represent what separates an app that feels "right" from one where the margins, spacing, and text sizes just look "off."

Apply these principles whenever building or modifying SwiftUI interfaces, WidgetKit widgets, or any native Apple UI.

Core Philosophy

Restraint over decoration. Every pixel must earn its place. A polished app uses fewer colors, fewer font sizes, fewer spacing values, and fewer words — but uses them consistently. Over-engineering visual elements (custom gradients, decorative borders, bespoke dividers) creates visual noise. Native components and system colors create harmony.

Attention is scarce. Keep UI copy shorter than you think it needs to be. Prefer one clear headline and one compact supporting block over repeated explanation in the title, subtitle, body, and footer. If a screen needs rationale, put it in one purposeful place instead of scattering it across the page.


1. Spacing System: Use a Consistent Grid

CRITICAL: Use spacing values from a base-4/base-8 grid. Never use arbitrary values.

Allowed spacing values

4, 8, 12, 16, 20, 24, 32, 40, 48

Bad (arbitrary values that create visual dissonance)

// WRONG - these numbers have no relationship to each other
.padding(.bottom, 26)
.padding(.bottom, 34)
.padding(.bottom, 36)
HStack(spacing: 18)
.padding(14)

Good (values from a consistent grid)

// RIGHT - predictable rhythm the eye can follow
.padding(.horizontal, 20)
.padding(.top, 8)
Spacer().frame(height: 32)
HStack(spacing: 4)  // or 8, 12, 16
.padding(.vertical, 12)
.padding(.horizontal, 16)

Standard padding assignments

  • Outer content padding: 16-20pt horizontal
  • Between major sections: 24-32pt vertical
  • Within grouped components: 4-12pt
  • Card/row internal padding: 12-16pt vertical, 16pt horizontal

2. Typography: Hierarchy Through Weight, Not Just Size

The principle

Use fewer font sizes with clear weight differentiation. Lighter weights at larger sizes; medium/regular at smaller sizes. This creates sophistication rather than visual chaos.

Recommended type scale (for a data-focused app)

RoleSizeWeightNotes
Hero number36-42pt.lightLarge but visually light -- elegant, not heavy
Secondary stat20-24pt.lightSame weight family as hero, smaller
Body / toggle label15pt.regularStandard iOS body size
Section header (uppercase)11pt.mediumWith tracking/letter-spacing
Caption / subtitle11-13pt.regularSecondary information

Bad (too many sizes, inconsistent weights)

// WRONG - 7 different sizes with no clear system
.font(.system(size: 60, weight: .ultraLight))   // hero
.font(.system(size: 44, weight: .regular))        // stat (too close to hero)
.font(.system(size: 31, weight: .ultraLight))     // percent symbol (odd ratio)
.font(.system(size: 18, weight: .regular))        // label (too big for a toggle)
.font(.system(size: 14, weight: .regular))        // header
.font(.system(size: 13, weight: .regular))        // another header
.font(.system(size: 12, weight: .regular))        // button (too small to read)

Good (clear hierarchy, fewer sizes)

// RIGHT - 5 sizes, clear purpose for each
.font(.system(size: 42, weight: .light, design: .monospaced))    // hero
.font(.system(size: 24, weight: .light, design: .monospaced))    // stat value
.font(.system(size: 15, weight: .regular, design: .monospaced))  // body
.font(.system(size: 14, weight: .regular, design: .monospaced))  // secondary
.font(.system(size: 11, weight: .medium, design: .monospaced))   // label

Font design consistency

Pick ONE font design and use it everywhere -- app AND widgets:

// If using monospaced, use it everywhere
design: .monospaced  // app views, widgets, lock screen -- all of them

// NEVER mix designs between app and widgets
// BAD: .monospaced in app, .rounded in lock screen widget

Letter spacing (tracking)

Use at most 2 values, and only on uppercase labels:

.tracking(1.5)  // section labels: "NOTIFICATIONS", "DAY", "LEFT"
.tracking(3)    // navigation/toolbar titles

Never use 3+ different tracking values like kerning(4), kerning(4.5), kerning(5) -- the differences are imperceptible but the inconsistency registers subconsciously.

Numeric formatting for identifiers

Years and other fixed identifiers should not be locale-grouped.

// RIGHT - stable, non-grouped identifier text
Text(String(year))                  // "2026"
Text(year, format: .number.grouping(.never))

// WRONG - locale grouping can render "2,026"
Text("\(year)")

3. Colors: System Semantic Colors Over Hardcoded Values

The principle

Use SwiftUI's semantic color system. It automatically handles light/dark mode, accessibility, and looks native. Hardcoded colors with manual opacity values create maintenance nightmares and look artificial.

Bad (hardcoded white with a dozen opacity values)

// WRONG - impossible to maintain, doesn't adapt to light mode
Color.black.ignoresSafeArea()           // forced dark
Color.white.opacity(0.08)               // ring background
Color.white.opacity(0.09)               // divider
Color.white.opacity(0.3)                // year text
Color.white.opacity(0.32)               // stat label
Color.white.opacity(0.42)               // percent symbol
Color.white.opacity(0.44)               // toggle tint
Color.white.opacity(0.72)               // button text
Color.white.opacity(0.88)               // toggle label
Color.white.opacity(0.9)                // stat value
Color.white.opacity(0.94)               // ring fill

Good (semantic system colors)

// RIGHT - adapts automatically, looks native, easy to maintain
Color(.systemBackground)                 // main background
Color(.secondarySystemBackground)        // card/group backgrounds
Color(.separator)                        // dividers (with optional opacity)
Color.primary                            // primary text and UI elements
.foregroundStyle(.secondary)              // secondary text
.foregroundStyle(.tertiary)               // labels, captions

When you do need opacity

Limit to 2-3 values with clear purposes:

.opacity(0.15)  // subtle background strokes
.opacity(0.3)   // separator lines
// That's it. If you need more, you're probably hardcoding what semantic colors handle.

4. Component Sizing: Proportional, Not Oversized

Progress rings / circular indicators

// App main view: 200x200 with thin stroke
.frame(width: 200, height: 200)
Circle().stroke(..., lineWidth: 3)

// Widget (systemSmall): 90x90, same stroke
.frame(width: 90, height: 90)
Circle().stroke(..., lineWidth: 3)

// WRONG: oversized ring with thick inconsistent strokes
.frame(width: 260, height: 260)    // too large, dominates screen
Circle().stroke(..., lineWidth: 9)  // background
Circle().stroke(..., lineWidth: 8)  // fill -- WHY different from background?

Stroke width consistency

Always use the same lineWidth for background and foreground strokes of the same element:

// RIGHT
Circle().stroke(background, lineWidth: 3)
Circle().trim(from: 0, to: fraction).stroke(fill, lineWidth: 3)

// WRONG - creates visual misalignment
Circle().stroke(background, lineWidth: 9)
Circle().trim(from: 0, to: fraction).stroke(fill, lineWidth: 8)

List rows and toggle rows

// RIGHT - natural sizing with proper padding
Toggle(isOn: $value) {
    Text(title)
        .font(.system(size: 15, weight: .regular, design: .monospaced))
}
.padding(.horizontal, 16)
.padding(.vertical, 12)

// WRONG - fixed oversized height
HStack {
    Text(label)
        .font(.system(size: 18))   // too big for a toggle label
    Spacer()
    Toggle("", isOn: $isOn)
        .labelsHidden()             // why hide the label? Use Toggle properly
}
.frame(height: 70)                  // way too tall

5. Grouped Content & Cards: Use System Patterns

Bad (over-engineered custom card)

// WRONG - custom gradient, overlay border, huge corner radius
VStack { ... }
    .padding(.vertical, 4)              // too tight
    .background(
        RoundedRectangle(cornerRadius: 22)   // too round
            .fill(LinearGradient(            // unnecessary gradient
                colors: [Color(white: 0.10), Color(white: 0.085)],
                startPoint: .topLeading, endPoint: .bottomTrailing
            ))
    )
    .overlay(
        RoundedRectangle(cornerRadius: 22)
            .stroke(Color.white.opacity(0.08), lineWidth: 1)  // decorative border
    )

Good (native grouped style)

// RIGHT - simple, native, works in light and dark mode
VStack(spacing: 0) {
    row1
    Divider().padding(.leading, 16)
    row2
    Divider().padding(.leading, 16)
    row3
}
.background(Color(.secondarySystemBackground))
.clipShape(.rect(cornerRadius: 10))

Key rules for grouped content

  • Corner radius: 10pt for cards/groups (matches iOS system style). Never 22pt+.
  • Dividers: Use the system Divider() with .padding(.leading, 16) for iOS-standard inset. Never build custom divider structs.
  • Card padding: 12-16pt vertical, 16pt horizontal. Never 4pt vertical.
  • Background: Color(.secondarySystemBackground) -- never custom gradients for standard cards.

6. Navigation: Use NavigationStack

// RIGHT - proper navigation with minimal toolbar
NavigationStack {
    ScrollView {
        content
    }
    .toolbar {
        ToolbarItem(placement: .principal) {
            Text("Title")
                .font(.system(size: 13, weight: .medium, design: .monospaced))
                .tracking(3)
                .foregroundStyle(.tertiary)
        }
    }
    .navigationBarTitleDisplayMode(.inline)
}

// WRONG - no navigation structure, just a ZStack
ZStack {
    Color.black.ignoresSafeArea()
    ScrollView {
        VStack {
            Text("2026").font(...) // manually placed "title"
            content
        }
    }
}

7. WidgetKit: Use Native Components

Circular lock screen widget

// RIGHT - use Gauge, it's purpose-built for this
Gauge(value: entry.fraction) {
    Text("")
} currentValueLabel: {
    Text("\(Int(entry.percentage))%")
        .font(.system(size: 12, weight: .medium, design: .monospaced))
}
.gaugeStyle(.accessoryCircular)
.containerBackground(.fill.tertiary, for: .widget)

// WRONG - manual circle drawing for lock screen
ZStack {
    Circle().stroke(Color.primary.opacity(0.18), lineWidth: 4)
    Circle().trim(from: 0, to: progress).stroke(...)
    Text(percentText)
        .font(.system(size: 14, weight: .bold, design: .rounded)) // wrong font design!
}

Rectangular lock screen widget

// RIGHT - use Gauge with linearCapacity
VStack(alignment: .leading, spacing: 4) {
    HStack {
        Text(year).font(.system(size: 13, weight: .semibold, design: .monospaced))
        Spacer()
        Text(percentage).font(.system(size: 13, weight: .medium, design: .monospaced))
            .foregroundStyle(.secondary)
    }
    Gauge(value: fraction) { Text("") }
        .gaugeStyle(.linearCapacity)
        .tint(.primary)
    HStack {
        Spacer()
        Text("\(dayOfYear)/\(totalDays)")
            .font(.system(size: 11, weight: .regular, design: .monospaced))
            .foregroundStyle(.secondary)
    }
}
.containerBackground(.fill.tertiary, for: .widget)

// WRONG - custom GeometryReader progress bar
GeometryReader { proxy in
    ZStack(alignment: .leading) {
        RoundedRectangle(cornerRadius: 2).fill(Color.primary.opacity(0.16))
        RoundedRectangle(cornerRadius: 2).fill(Color.primary)
            .frame(width: max(2, proxy.size.width * progress))
    }
}
.frame(height: 6)

Widget background

// RIGHT
.containerBackground(.fill.tertiary, for: .widget)

// WRONG - hardcoded color
.containerBackground(.black, for: .widget)

Widget family coverage

Support all relevant families -- don't skip common ones:

.supportedFamilies([
    .accessoryCircular,      // lock screen circle
    .accessoryRectangular,   // lock screen rectangle
    .accessoryInline,        // lock screen inline text
    .systemSmall,            // home screen small
    .systemMedium,           // home screen medium
    .systemLarge,            // home screen large
])

Cross-family visual consistency

Medium and large home widgets should share the same structural layout:

  • Header: year on the left, percentage on the right
  • Middle: progress bar
  • Footer: day/total right aligned

Do not re-invent hierarchy per family unless there is a hard size constraint.

Always include explicit internal padding on home widgets to avoid clipping near rounded edges:

.padding(.horizontal, 12)
.padding(.vertical, 12)

Widget memory budget (hard limit)

Widget extensions have a tight memory budget (commonly around 30 MB). Dense visualizations can be killed by EXC_RESOURCE if built from too many nested views.

// RIGHT - draw dense dot grids in one pass
Canvas { context, size in
    // draw 365/366 dots here
}

// WRONG - hundreds of nested subviews (high memory overhead)
LazyVGrid(columns: columns) {
    ForEach(1...366, id: \.self) { day in
        ZStack { Circle(); partialFillLayer }
    }
}

Timeline refresh (match data granularity)

// RIGHT - refresh at midnight for day-level data
let tomorrow = calendar.startOfDay(for: calendar.date(byAdding: .day, value: 1, to: now)!)
Timeline(entries: [entry], policy: .after(tomorrow))

// RIGHT - periodic refresh for time-of-day dependent percentages/partial fills
let refresh = Calendar.current.date(byAdding: .minute, value: 15, to: now)!
Timeline(entries: [entry], policy: .after(refresh))

// WRONG - minute-level refresh for static daily data
let tooFrequent = Calendar.current.date(byAdding: .minute, value: 1, to: now)!
Timeline(entries: [entry], policy: .after(tooFrequent))

8. Interactive Elements

Toggles

// RIGHT - use Toggle with its built-in label, tint with a single accent color
Toggle(isOn: $value) {
    Text(title)
        .font(.system(size: 15, weight: .regular, design: .monospaced))
}
.tint(.green)

// WRONG - hidden label with manual HStack layout
HStack {
    Text(label).font(.system(size: 18))
    Spacer()
    Toggle("", isOn: $isOn)
        .labelsHidden()
        .tint(Color.white.opacity(0.44))  // low-contrast tint
}

Mutually exclusive options

When options are exclusive (e.g. daily/weekly/monthly cadence), use one selected value, not three independent toggles.

// RIGHT - single source of truth
enum Cadence: String, CaseIterable { case daily, weekly, monthly }
@State private var cadence: Cadence = .daily

ForEach(Cadence.allCases, id: \.rawValue) { option in
    Button {
        cadence = option
    } label: {
        HStack {
            Image(systemName: cadence == option ? "checkmark.circle.fill" : "circle")
            Text(option.rawValue.capitalized)
        }
    }
}

// RIGHT - one preview action when content is shared
Button("Preview") { sendPreview() }

// WRONG - independent toggles allow contradictory state
Toggle("Daily", isOn: $daily)
Toggle("Weekly", isOn: $weekly)
Toggle("Monthly", isOn: $monthly)

Animated transitions for changing numbers

// Add to any Text that displays a changing numeric value
Text(String(format: "%.2f", percentage))
    .contentTransition(.numericText())

9. Interactive Editors: Centralize Geometry and State

Interactive editors (collages, crops, canvases, media framing tools, layout pickers) need stricter state and layout discipline than ordinary forms.

Presentation state

Present editor flows from payload state, not from a separate Bool plus independently-managed data.

// RIGHT - presentation only happens when payload exists
@State private var activeCropRequest: CropRequest?

.sheet(item: $activeCropRequest) { request in
    CropEditor(request: request)
}

// WRONG - the sheet can open before the backing data is ready
@State private var showCropEditor = false
@State private var selectedImage: UIImage?

.sheet(isPresented: $showCropEditor) {
    if let selectedImage { CropEditor(image: selectedImage) }
}

Shared geometry model

If the app previews pan/zoom/crop/layout live and later exports the result, use one shared geometry model for both preview and render.

// RIGHT - one source of truth for bounds and transforms
let normalized = EditorGeometry.normalizedAdjustment(adjustment, imageSize: image.size, slotSize: slotSize)
let drawRect = EditorGeometry.drawRect(for: image.size, in: slotRect, adjustment: adjustment)

// WRONG - preview and export each invent their own math
let previewOffset = ...
let exportOffset = ...

If a user can zoom out enough to reveal background, that must be an intentional part of the shared geometry model, not an editor-only exception.

Gesture coordination

Tap, long-press-drag, and pinch are not independent features. In SwiftUI they compete unless you model their relationship explicitly.

  • Use a single interaction state for the active tile/card/canvas item.
  • Decide which gesture has priority and which ones should run simultaneously.
  • Reset temporary gesture state deliberately when selection changes.
  • Prefer one coherent state machine over scattered booleans tied to individual gestures.

Fixed editor layout

If the screen must not scroll, budget vertical space top-down using a few named regions:

  • header
  • canvas stage
  • settings region
  • bottom toolbar

Keep that sizing math in one place. Don't let each subview invent its own height.

Custom headers and safe areas

If you replace the system navigation bar with a custom header:

  • Be explicit about whether the parent already respects the safe area.
  • Do not add safeAreaInsets.top reflexively; double-counting it creates obvious dead space.
  • Keep custom headers compact. They should feel like navigation chrome, not a full content section.

Settings surfaces

When an editor has several configuration modes (Layout, Border, Ratio, Background, etc.), show one active settings surface at a time instead of stacking every control on screen.

This keeps the canvas visually dominant and makes each control group easier to understand.


10. Data Model: Share Between App and Widgets

// RIGHT - one model used everywhere
struct YearProgress {
    // shared calculation logic
    static func current() -> YearProgress { ... }
}
// Used by both ContentView and widget TimelineProvider

// If percentage is shown as live progress, include time-of-day in shared math
let dayProgress = elapsedInCurrentDay / totalSecondsInDay
let elapsedDays = Double(dayOfYear - 1) + dayProgress
let fraction = elapsedDays / Double(totalDays)

// WRONG - separate snapshot structs with duplicated date math
struct YearProgressSnapshot { ... }            // in app
struct YearProgressWidgetSnapshot { ... }      // in widget extension (duplicated!)

11. Quick Checklist

Before shipping any SwiftUI view, verify:

  • All spacing values come from the grid (4, 8, 12, 16, 20, 24, 32)
  • Font sizes limited to 5 or fewer distinct values
  • One font design used consistently (including widgets)
  • Colors use semantic system colors, not hardcoded values with opacity
  • Background and foreground strokes use the same lineWidth
  • Cards use Color(.secondarySystemBackground) with 10pt corner radius
  • Dividers use system Divider() with leading padding
  • Toggle rows use Toggle's built-in label (not .labelsHidden())
  • Lock screen widgets use Gauge (not manual circle drawing)
  • Widget background uses .containerBackground(.fill.tertiary, for:.widget)
  • Year/identifier text avoids locale grouping when grouping is not desired
  • Tracking/kerning limited to 2 values max
  • NavigationStack is used (not bare ZStack)
  • Timeline refresh rate matches data granularity (midnight vs periodic)
  • Large/dense widget visuals use Canvas or similarly lightweight rendering
  • Medium and large widget families share consistent hierarchy and internal padding
  • Exclusive choices use a single selected value (not multiple toggles)
  • Percentages include time-of-day when UI implies live progress
  • No minimumScaleFactor hacks -- fix the layout instead
  • Interactive editors present from payload state, not Bool + separate data
  • Preview and export share the same geometry model for pan/zoom/crop/layout
  • Custom headers do not double-count top safe-area inset
  • No-scroll editor screens budget height through a centralized layout model
  • Multi-mode editors show one focused settings surface instead of every control at once

平台分布

Codex

35.08%
按下载量换算1,249

Claude

28.2%
按下载量换算1,004

Cursor

17.33%
按下载量换算617

Gemini CLI

9.75%
按下载量换算347

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

权限未标注

当前来源未能明确判断权限范围,默认进入异常复核队列。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills