Token导航 LogoToken导航TokenDH.com
研究检索敏感数据github未标认证来源可访问许可证需确认审计通过

edge-case-identification边缘情况识别

Agent Skill

edge-case-identification 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

196

周安装

8

GitHub Stars

36

下载量

63
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

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:edge-case-identification(边缘情况识别)
来源仓库:https://github.com/owl-listener/inclusive-design-skills
仓库路径:skills/edge-case-identification
安装命令:
npx skills add https://github.com/owl-listener/inclusive-design-skills --skill edge-case-identification
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/owl-listener/inclusive-design-skills --skill edge-case-identification

简介

edge-case-identification 识别团队视为边缘但对残障用户至关重要的使用场景,推动包容性设计。

  • 涵盖输入、操作、认知、网络与输出五大类边缘情况,挑战“这只是边缘案例”的推诿逻辑。
  • 主张影响残障用户的用例不是边缘而是基本需求,提升产品普适性与合规水平。
  • 安装命令:npx skills add https://github.com/owl-listener/inclusive-design-skills --skill edge-case-identification
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Edge Case Identification

Identify the scenarios that teams dismiss as "edge cases" but that are everyday reality for users with disabilities. An edge case for the team is often a primary case for someone.

Core Principle

"That's an edge case" is the most common way accessibility gets deprioritised. If it affects people with disabilities, it's not an edge case — it's a design requirement you haven't met.

Categories of Overlooked Edge Cases

Input Edge Cases

  • User cannot use a mouse (keyboard only)
  • User cannot use a keyboard (voice or switch only)
  • User cannot use both hands (one-handed operation)
  • User pastes content instead of typing it
  • User uses autocomplete or password manager to fill fields
  • User enters content in a different script or language
  • User inputs using dictation with punctuation spoken aloud

Output Edge Cases

  • User cannot see the screen (screen reader)
  • User sees only a small portion of the screen (magnification)
  • User cannot distinguish colours
  • User cannot hear audio
  • User has screen in greyscale mode or high contrast mode
  • User has text scaled to 200%
  • User has animations disabled

Content Edge Cases

  • Extremely long names (hyphenated, multi-word, cultural naming)
  • Non-Latin characters in name or address fields
  • Empty states where no data exists yet
  • Error states when everything goes wrong at once
  • First-time use with no prior context
  • Returning after months away with forgotten context
  • Content in the wrong language for the user

Context Edge Cases

  • Session timeout during a long form
  • Connection lost mid-task
  • Battery dying during a critical flow
  • Screen rotation mid-interaction
  • System font size set to maximum
  • Browser zoom at 200% or higher
  • Ad blocker or privacy extension modifying the page

Timing Edge Cases

  • User is much slower than expected (switch user, cognitive delay)
  • User is much faster than expected (screen reader power user)
  • User needs to pause mid-task for hours or days
  • Timed session expires before user finishes
  • Content updates while user is reading it

How to Find Edge Cases

The "What If" Exercise

For each step in a flow, ask:

  • What if the user can't see this?
  • What if the user can't click this?
  • What if the user can't understand this?
  • What if the user takes 10x longer than expected?
  • What if the user leaves and comes back tomorrow?
  • What if this field contains unexpected content?

The Inversion Test

Take your happy path scenario. Invert each assumption:

  • "User is on a desktop" → mobile, or tablet, or screen reader
  • "User has fast internet" → slow, intermittent, or offline
  • "User reads English fluently" → low literacy or second language
  • "User can see the colour change" → colour blind or blind
  • "User completes this in one sitting" → interrupted, returns later

Response to "That's an Edge Case"

When someone says this, ask:

  1. How many users does this affect? (often more than assumed)
  2. How severe is the impact? (a blocked user is worse than a mildly inconvenienced one)
  3. Is this a permanent condition for some users? (then it's their daily experience, not an edge case)
  4. Would fixing this improve the experience for other users too? (almost always yes)

Assessment Questions

  1. Has the team mapped edge cases beyond the happy path?
  2. Have input, output, content, context, and timing edge cases been considered?
  3. Are there dismissed "edge cases" that are daily reality for users with disabilities?
  4. Does the test plan include scenarios for assistive technology users?

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

35.39%
按下载量换算22

Claude

29.57%
按下载量换算19

Cursor

18.62%
按下载量换算12

Gemini CLI

8.24%
按下载量换算5

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills