Token导航 LogoToken导航

Agent Skill · 前端设计

ux-friction-analyzer(ux 摩擦分析仪)

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。

ux-friction-analyzer 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 7,486。

总安装 371周安装 15GitHub Stars 98下载 116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:ux-friction-analyzer(ux 摩擦分析仪)
来源仓库:https://github.com/curiositech/some_claude_skills
仓库路径:skills/ux-friction-analyzer
安装命令:
npx skills add https://github.com/curiositech/some_claude_skills --skill ux-friction-analyzer
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/curiositech/some_claude_skills --skill ux-friction-analyzer

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。

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。适合根据产品场景整理页面结构、生成 UI 方案或改进组件层级。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。注意该技能侧重用户体验分析,不适合后端逻辑或数据库设计。

来源仓库内路径:skills/ux-friction-analyzer

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/curiositech/some_claude_skills --skill ux-fric…;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

ux-friction-analyzer 怎么安装?

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

ux-friction-analyzer 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。

ux-friction-analyzer 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。

ux-friction-analyzer 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。

ux-friction-analyzer 用的人多吗?

公开数据是累计安装 371 次,近一周安装 15 次,下载 116 次,GitHub 98 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。

ux-friction-analyzer 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
ux-friction-analyzer

ux 摩擦分析仪 · 本页

37198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

UX Friction Analyzer

A comprehensive skill for analyzing and optimizing user experience through cognitive psychology, ADHD-friendly design, and flow state engineering.

Activation

Use this skill when:

  • Designing new interfaces or user flows
  • Auditing existing UX for friction points
  • Optimizing for neurodivergent users (ADHD, autism)
  • Simulating user journeys before building
  • Reducing cognitive load in complex applications

Trigger phrases: "analyze UX", "friction audit", "user journey", "ADHD-friendly", "optimize flow", "reduce cognitive load"


Core Frameworks

1. ADHD-Friendly Design Principles

Apply these patterns to ALL interfaces:

PrincipleImplementationWhy It Matters
Progressive DisclosureShow one task at a time; hide future stepsPrevents overwhelm, maintains focus
Context PreservationAuto-save every keystroke; never lose workReduces anxiety about losing progress
Gentle RemindersStatus updates, not alarms; no red urgencyAvoids panic, maintains calm
Pause & ResumeSession state persists across days/weeksRespects inconsistent schedules
Minimal DistractionsSingle focus area; dim non-active panelsReduces competing stimuli
Chunked ProgressVisual cards/steps, not endless scrollingCreates completion dopamine hits
Predictable NavigationSame layout always; no surprisesReduces reorientation cost
Calm Mode OptionReduced animations, muted colors on demandAccommodates sensory sensitivity

2. Gestalt Psychology

Apply these perception principles:

PROXIMITY
─────────
Elements close together = perceived as related
White space creates natural boundaries

┌─────────┐  ┌─────────┐     ┌─────────┐  ┌─────────┐
│ Related │  │ Related │     │ Other   │  │ Other   │
│ Item A  │  │ Item B  │     │ Group A │  │ Group B │
└─────────┘  └─────────┘     └─────────┘  └─────────┘
     ↑ CLOSE = GROUPED            ↑ SEPARATE = DISTINCT

SIMILARITY
──────────
Same color/shape/size = perceived as related function

┌──────┐  ┌──────┐  ┌──────┐     ┌──────┐  ┌──────┐
│ BLUE │  │ BLUE │  │ BLUE │     │ CORAL│  │ CORAL│
│ Save │  │ Copy │  │ Edit │     │ Del  │  │ Clear│
└──────┘  └──────┘  └──────┘     └──────┘  └──────┘
     ↑ SAME = Related actions         ↑ DIFFERENT = Destructive

CONTINUITY
──────────
Eye follows lines/paths naturally

Step 1 ──→ Step 2 ──→ Step 3 ──→ Complete
   ●──────────●──────────●──────────●

CLOSURE
───────
Brain completes incomplete shapes
Use for progress indicators, loading states

[ ████████░░░░░░░░ ] 50% - brain "sees" the end

3. Cognitive Load Theory

Three types of mental load to manage:

TypeDefinitionStrategy
IntrinsicTask complexity itselfCan't eliminate; acknowledge it
ExtraneousPoor design adding effortELIMINATE THIS - your job
GermaneLearning/understandingMinimize for repeat users

Working Memory Limits:

  • 7±2 items maximum (Miller's Law)
  • 4 chunks optimal for complex tasks
  • Micro-breaks every 25 minutes

Reduce Extraneous Load By:

  • Removing unnecessary choices
  • Using recognition over recall
  • Providing smart defaults
  • Eliminating decorative elements that don't inform

4. Fitts' Law

Time to acquire target = f(Distance / Size)

IMPLICATIONS FOR BUTTONS:
─────────────────────────

  ┌───────────────────┐          vs          ┌──┐
  │     GENERATE      │                      │Go│
  │                   │                      └──┘
  └───────────────────┘
         ↑                                     ↑
  44px+ touch target                   Hard to hit
  Easy to acquire                      Frustrating

MINIMUM SIZES:
- iOS: 44x44 CSS pixels
- Android: 48x48 CSS pixels
- Desktop: 32x32 minimum, 44x44 preferred

EDGE TARGETS ARE INFINITE:
  ┌─────────────────────────────────────────────────────┐
  │ ■ LOGO                                    MENU ■   │
  │                                                     │
  │   Screen edges = can't overshoot                   │
  │   Place critical actions at corners/edges          │
  │                                                     │
  │ ■ HELP                                  EXPORT ■   │
  └─────────────────────────────────────────────────────┘

ICON + LABEL > ICON ALONE:
- Larger target area
- Reduced ambiguity
- Faster acquisition

5. Flow State Engineering

Key Metrics:

  • 15-25 minutes to enter flow state
  • 23 minutes to recover from interruption
  • 40% productivity loss with frequent interruptions
  • Only 41% of work time spent in flow (McKinsey)

Flow Conditions:

  1. Clear goals for the current task
  2. Immediate feedback on actions
  3. Balance between challenge and skill
  4. No anxiety about failure

Preserve Flow By:

  • Background processing (don't block UI)
  • Push notifications when ready (bring user back faster)
  • Quick re-orientation panels after breaks
  • Auto-save eliminating "save anxiety"
  • Undo everything (confidence to experiment)

Analysis Methodology

Step 1: Create Decision Tree

Map every user path with probabilities:

                    ┌─────────────┐
                    │ USER LANDS  │
                    └──────┬──────┘
                           │
           ┌───────────────┼───────────────┐
           ▼               ▼               ▼
     ┌──────────┐   ┌──────────┐   ┌──────────┐
     │ Action A │   │ Action B │   │ Action C │
     │  (40%)   │   │  (45%)   │   │  (15%)   │
     └────┬─────┘   └────┬─────┘   └────┬─────┘
          │              │              │
          ▼              ▼              ▼
        [Next]         [Next]         [Next]

For each edge, record:

  • Probability (%)
  • Friction score (1-10)
  • Time to complete (seconds/minutes)
  • Cognitive load (low/medium/high)

Step 2: Simulate User Journeys

Create detailed simulations for each persona:

Template:

TIME    ACTION                           COGNITIVE STATE           FRICTION
─────────────────────────────────────────────────────────────────────────────
0:00    [User action]                    [Mental state]            Low/Med/High
        └─ [System response or UI shown]

0:15    [Next action]                    [How they feel]           Low/Med/High
        └─ [What happens]
        └─ PROBLEM: [Friction point if any]

...continue...
─────────────────────────────────────────────────────────────────────────────
TOTAL TIME: X minutes
FRICTION POINTS: N (list them)
ABANDONMENT RISKS: N (critical moments)
DELIGHT MOMENTS: N (positive surprises)

Personas to simulate:

  1. Expert User - Knows the system, moving fast
  2. New User - First time, needs guidance
  3. Distracted User - Context switching, interruptions
  4. Explorer - No goal, seeing what's possible
  5. Completer - Trying to finish, hitting obstacles

Step 3: Friction Analysis Matrix

Quantify and prioritize:

Friction PointUsers AffectedSeverity (1-10)Fix DifficultyPriority Score
[Issue 1]X%NEasy/Med/HardHIGH/MED/LOW
[Issue 2]X%NEasy/Med/HardHIGH/MED/LOW

Priority Formula:

Priority = (Users Affected × Severity) / Fix Difficulty

Step 4: Impedance Mapping

Compare current vs ideal:

TASK                          CURRENT IMPEDANCE     IDEAL IMPEDANCE
────────────────────────────────────────────────────────────────────
[Task 1]                      Low (X sec)           ✓ Optimal
[Task 2]                      Medium (X sec)        Could be Y sec
[Task 3]                      HIGH (X min)          Should be Y sec

Step 5: Time-Loss Analysis

Calculate context switch costs:

Action                        Frequency    Time Lost Each    Total Impact
─────────────────────────────────────────────────────────────────────────
[Interruption type 1]         X/session    Y min             Z min
[Interruption type 2]         X/session    Y min             Z min
─────────────────────────────────────────────────────────────────────────
TOTAL CONTEXT SWITCH LOSS                                    Z min/session

Optimization Patterns

Immediate Fixes (Low Effort, High Impact)

  1. Giant CTA on Landing <button class="cta" style="min-height: 60px; min-width: 200px;"> Primary Action <span class="subtext">Supporting text</span> </button>
  2. Visible Edit Affordances

- Show pencil/edit icons by default, not just on hover - Add tooltips: "Click to edit"

  1. Auto-Fill Prompts

- After user completes 1 item manually, offer to auto-complete rest - "Want me to fill in the remaining X items?"

  1. Floating Action Buttons

- Critical actions always visible (not buried in menus) - Bottom-right for mobile thumb zone

  1. Progress Indicators

- Show "Step X of Y" always - Visual progress bar at top

Medium-Term Improvements

  1. Re-Orientation Panels ┌─────────────────────────────────────────────┐ │ Welcome back! Here's where you left off: │ │ │ │ ✓ Step 1: Complete │ │ → Step 2: In progress (60%) │ │ ○ Step 3: Not started │ │ │ │ [Continue where I left off] │ └─────────────────────────────────────────────┘
  2. Keyboard Shortcuts

- Number keys for mode switching (1, 2, 3...) - Cmd+Enter for primary action - Escape for cancel/close

  1. Background Processing

- Never block UI for long operations - Show progress, allow user to continue - Push notification when complete

  1. Smart Defaults

- Pre-fill based on user history - Remember last-used settings - Suggest most common option first

Long-Term Vision

  1. Predictive UI

- Anticipate next action based on patterns - Pre-load likely next screens - Suggest before user asks

  1. Personalized Complexity

- Simple mode for new users - Power user mode unlocks over time - User controls their complexity level

  1. Accessibility Suite

- High contrast mode - Reduced motion option - Screen reader optimization - Keyboard-only navigation


Checklist for New Features

Before shipping any feature, verify:

Cognitive Load

  • Can user complete with ≤4 things in working memory?
  • Are there unnecessary choices that could be defaults?
  • Is recognition used instead of recall?

ADHD-Friendly

  • Can user pause and resume without losing context?
  • Are there gentle progress indicators (not anxiety-inducing)?
  • Is the interface calm (not visually noisy)?

Fitts' Law

  • Are primary buttons ≥44px tall?
  • Are destructive actions away from common paths?
  • Do buttons have labels, not just icons?

Flow Preservation

  • Does any action block the UI for >2 seconds?
  • Can long operations run in background?
  • Is there a clear "done" state?

Error Recovery

  • Can every action be undone?
  • Are error messages actionable (not just "Error")?
  • Is auto-save enabled?

Example Analysis Output

When running this skill, produce a document with:

  1. Executive Summary - Key findings in 3 bullets
  2. Decision Tree - All user paths with probabilities
  3. User Journey Simulations - 3-5 personas, full timeline
  4. Friction Matrix - Prioritized issues table
  5. Optimization Recommendations - Immediate/Medium/Long-term
  6. Implementation Checklist - Specific changes to make

Integration Points

  • web-design-expert: Implement UX recommendations visually
  • adhd-design-expert: Deep neurodivergent design patterns
  • frontend-developer: Technical implementation of fixes
  • diagramming-expert: Create user flow diagrams

Sources


Core Philosophy: Every click, every second of confusion, every moment of "where am I?" is friction stealing from your users. Design for the distracted, optimize for the overwhelmed, and everyone benefits.

平台分布

Codex

31.9%
按下载量换算37

Claude

29.27%
按下载量换算34

Cursor

19.99%
按下载量换算23

Gemini CLI

10.03%
按下载量换算12

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