Token导航 LogoToken导航

Agent Skill · 研究检索

wtf.design-task(wtf 设计任务)

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。

wtf.design-task 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为调用外部服务;许可证信息还不明确,暂无第三方扫描结果,按安装量在研究检索类的 73,675 个 Skill 里排第 47,935。

总安装 710周安装 29GitHub Stars 3下载 227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:wtf.design-task(wtf 设计任务)
来源仓库:https://github.com/xiduzo/wtf
仓库路径:skills/wtf.design-task
安装命令:
npx skills add https://github.com/xiduzo/wtf --skill wtf.design-task
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/xiduzo/wtf --skill wtf.design-task

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

wtf.design-task 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 方案或改进组件层级。使用时需结合现有品牌、设计系统和用户任务,避免堆砌装饰元素;涉及真实页面改动时应通过截图或浏览器预览检查表现。

通过 npx skills add 命令从 GitHub 仓库安装,建议确认权限及是否会触发文件读写或构建流程。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。

来源仓库内路径:skills/wtf.design-task

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/xiduzo/wtf --skill wtf.design-task;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

    它需要调用外部服务,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

wtf.design-task 怎么安装?

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

wtf.design-task 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。

wtf.design-task 需要什么权限?

权限级别是调用外部服务。该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

wtf.design-task 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。

wtf.design-task 用的人多吗?

公开数据是累计安装 710 次,近一周安装 29 次,下载 227 次,GitHub 3 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。

wtf.design-task 安全吗?

没有查到第三方扫描结果,安全性无法从公开资料判断,安装前请自行审查源码、依赖和命令行为。

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。

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。
  • 没有第三方扫描结果,安全性只能靠自己读源码判断。
  • 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。
  • 需要处理研究检索类任务、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。
  • 要求有安全扫描结论才肯安装的团队。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
wtf.design-task

wtf 设计任务 · 本页

7103调用外部服务未明确未收录

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

Design Task

Pick up an existing Task as a designer. Core value: reads the Gherkin scenarios to identify every UI state that needs design coverage, then helps you document the design references back into the issue so developers have a single source of truth.

See references/component-spec-template.md for the expected structure when scaffolding a component spec without Figma frames.

Process

0. GitHub CLI setup

Run steps 1–2 of ../references/gh-setup.md (install check and auth check). Stop if gh is not installed or not authenticated. Extensions are not required for this skill.

Skip this step if invoked from wtf.write-task or another skill that already ran gh-setup this session.

1. Identify the Task

If the user provided an issue number in their request, use it directly. Otherwise search for recent open issues with label task to populate options and call AskUserQuestion with question: "Which Task are you designing?", header: "Task", and options pre-filled with 1–2 likely open Task issue references inferred from GitHub search (e.g. recent open issues labeled task).

Fetch the Task first, extract the Feature number from its Context section, then fetch the Feature:

gh issue view <task_number>    # Functional Description, Gherkin, Design Reference — also yields feature number
# Extract feature number, then:
gh issue view <feature_number> # User stories, ACs, visual context

2. Lifecycle check

Check whether the task already has a designed label:

gh issue view <task_number> --json labels --jq '.labels[].name'

If the designed label is present, call AskUserQuestion with:

  • question: "This task already has a designed label. Continuing will overwrite the existing Design Reference. How would you like to proceed?"
  • header: "Already designed"
  • options: [{label: "Redesign it", description: "Overwrite the existing Design Reference with a new one"}, {label: "Exit", description: "Leave the existing design as-is"}]
  • Redesign it → continue.
  • Exit → exit immediately.

If the designed label is absent, continue silently.

3. Load the design steering document

Use the Read tool to attempt reading docs/steering/DESIGN.md.

If the file exists: keep its content in context. Use its design principles, tokens, component patterns, and accessibility standards to inform every decision in this session. Do not surface it to the user — just apply it silently.

If the file does not exist, call AskUserQuestion with:

  • question: "docs/steering/DESIGN.md doesn't exist yet. This document captures your design principles, tokens, and component patterns. Would you like to create it now?"
  • header: "Design steering doc missing"
  • options: [{label: "Create it now", description: "Runwtf.steer-designbefore continuing (recommended)"}, {label: "Skip for this session", description: "Continue without it — design decisions won't reference project standards"}]
  • Create it now → follow the wtf.steer-design process, then return to this skill and continue from step 4.
  • Skip for this session → continue without it.

4. Explore the design system

Use the Agent tool with these concrete searches (run in parallel):

  • Glob('src/components/**/*', 'src/**/components/**/*', 'components/**/*') — existing UI components; note file names that match domain objects or UI states in the Task
  • Glob('**/{tokens,theme,variables,design-tokens}.{css,scss,ts,js,json}') + Grep for CSS custom property declarations (--) or Tailwind config keys — design tokens in use (colors, spacing, typography)
  • Glob('src/**/*.{stories,story}.{ts,tsx,js,jsx,mdx}') — Storybook stories as pattern references for similar screens or flows
  • Grep for figma.com URLs across all .md, .mdx, and issue body files in the repo — existing Figma references linked in related issues or docs

5. Identify UI states from Gherkin

For each Gherkin scenario in the Task:

  • Identify the UI state it represents (e.g. empty, loading, error, success, disabled, edge case)
  • Note any interaction or transition implied by the When/Then steps

List these states explicitly — this becomes the design coverage checklist.

6. Ask about design assets

Call AskUserQuestion with question: "How would you like to handle design assets for this task?", header: "Design assets", and options:

  • {label: "I have Figma frames", description: "Provide frame URLs — I'll validate coverage against Gherkin scenarios"}
  • {label: "Generate designs for me", description: "I'll use Figma MCP to generate frames from the Gherkin scenarios and design system"}
  • {label: "Scaffold a spec only", description: "No Figma — produce a text component spec from the scenarios"}
  • {label: "Partial — some states designed", description: "Provide available frames; remaining states go to generate or scaffold"}

Path A — Human provides frames: Collect frame URLs. For each Gherkin scenario from step 5, check whether a frame covers it. Flag any scenario with no matching frame as a gap. Present the coverage matrix: scenario → frame URL (or ⚠ gap). If gaps exist, call AskUserQuestion asking whether to generate the missing frames (route to Path B) or leave them as pending.

Path B — AI generates via Figma MCP: Check whether the Figma MCP tool generate_figma_design is available. If unavailable, warn the user and fall back to Path C (scaffold).

If available: for each uncovered UI state, call generate_figma_design with:

  • The Gherkin scenario as the design brief
  • Component patterns and tokens from docs/steering/DESIGN.md (loaded in step 3)
  • Any shared components identified in the parent Feature's Design Handoff (if available)

Collect the generated frame URLs and treat them as Path A frames from this point forward.

Path C — Scaffold spec only: Draft a component spec using the structure in references/component-spec-template.md, listing each state with its required UI elements and interactions. No Figma frames — this is a text-only design brief for the developer.

Partial: Collect available frame URLs, run Path A validation on covered states. For uncovered states, call AskUserQuestion asking whether to generate (Path B) or scaffold (Path C) the remainder.

7. Draft the Design Reference

Produce the content for the Design Reference section of the Task:

  • Frame URLs mapped to Gherkin scenarios (Path A/B), or scaffolded component spec (Path C)
  • Coverage matrix: scenario → frame URL or ⚠ pending
  • Component breakdown: which exist in codebase, which are new
  • Interaction notes: hover, focus, error states, transitions
  • Responsive behavior if applicable
  • Design tokens to apply

8. Review with user

Show the draft. Then call AskUserQuestion with question: "Does this cover all the states in the Gherkin?", header: "Review", and options: [{label: "Yes — looks complete", description: "Proceed to update the task"}, {label: "Missing states", description: "I want to add more coverage"}, {label: "Other changes", description: "I want to adjust something else"}].

Apply edits, then proceed.

9. Update the Task issue

Note: read the current issue body first (gh issue view <task_number>), replace only the Design Reference section with the new content, and preserve all other sections unchanged. Write the full updated body to a temp file and use --body-file.
gh issue edit <task_number> --body-file /tmp/wtf.design-task-<task_number>-body.md

Add the designed lifecycle label to mark this step complete:

gh issue edit <task_number> --add-label "designed"

Print the updated Task issue URL.

10. Offer to continue

Call AskUserQuestion with:

  • question: "What's next?"
  • header: "Next step"
  • options: [{label: "Implement this Task", description: "Runwtf.implement-taskfor this Task now (default)"}, {label: "Design another Task", description: "Design another Task for the same Feature"}, {label: "Stop here", description: "Exit — no further action"}]
  • Implement this Task → follow the wtf.implement-task process, passing the Task number in as context so the user is not asked for it again.
  • Design another Task → restart this skill from step 1, reusing the same Feature context.
  • Stop here → exit.

平台分布

Codex

38.63%
按下载量换算88

Claude

31.15%
按下载量换算71

Cursor

19.29%
按下载量换算44

Gemini CLI

8.71%
按下载量换算20

安全审计

暂无安全审计结果可展示。

权限和风险

调用外部服务

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

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