跳到正文
cccskills
免费在 GitHub 公开

daily-ui-inspiration-capture

Collect five UI references into a dated, article-ready bundle with local images, full-page and section captures, motion studies, and detailed AI-builder prompts. Check previous collections for duplicates and document fallback reasons when video capture is unavailable.

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

Create a recurring daily UI inspiration capture. Use when the user asks to run, refresh, package, or validate dated UI inspiration bundles, especially for `articles/YYYY-MM-DD-ui-inspiration-capture/` outputs, Framer/Dribbble landing-page inspiration, motion-study screenshots/videos, AI-builder prompts, duplicate checking, or converting a project runbook into repeatable workflow.

实测评分 未实测

查看安装方法

适用场景

  • Collecting daily landing-page design references
  • Studying page animations and interactions
  • Writing detailed prompts for AI website builders
  • Packaging reference articles without duplicate examples

包含的文件(7)

  • SKILL.md7.9 KB
  • agents/openai.yaml227 B
  • demo/expected-output.md222 B
  • demo/index.html3.8 KB
  • demo/input.md193 B
  • demo/preview.jpg78.9 KB
  • demo/PROMPT.md2.1 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Daily UI Inspiration Capture

Overview

Run the current project's daily UI inspiration workflow as an article-ready bundle, not a screenshot dump or README. Treat the active project runbook as source of truth, but keep the contract below as the default unless the user updates it.

Start

Work from the content/article project named by the user. If no project is named, use the current workspace and say which directory you used.

Before collecting, read:

  • AGENTS.md, if present or supplied in the prompt
  • scripts/ui-inspiration-capture-guidelines.md, if present
  • the latest one or two articles/*-ui-inspiration-capture/content.md and manifest.json files for current shape

Check git status --short early. This repo is often dirty; keep changes scoped to the requested dated bundle or skill files.

For any browser work, use the Codex in-app browser only. Do not use Chrome. If a source blocks inspection, use the best available Codex-browser evidence or local asset fallback and say why.

Output Contract

Create a dated folder:

articles/YYYY-MM-DD-ui-inspiration-capture/

The folder should contain at minimum:

  • content.md
  • manifest.json
  • local still images
  • local video files or clear per-item video fallback reasons
  • one full-page scroll screenshot per live website reference
  • section-by-section crops from that full-page screenshot for the hero, every meaningful middle section, and the footer
  • motion-frame images when video capture succeeds

Do not create README.md.

Collect exactly 5 final references. The newer 5-reference motion-study workflow supersedes older 20-image gallery runs.

Prefer live, inspectable landing pages and product sections. Framer Marketplace is a strong source. Dribbble is best effort; if blocked by WAF or poor assets, finish with stronger Framer or direct-page candidates instead of forcing brittle workarounds.

Selection

Load previous articles/*-ui-inspiration-capture/manifest.json files before choosing finalists.

Reject candidates that repeat a normalized title, source URL, or image URL from previous captures. Rotate source searches when the feed feels stale:

  • website hero
  • saas landing page
  • product website
  • web design hero
  • recent web design
  • recent product design

Favor references with a clear first viewport, inspectable sections, strong typography, product or brand specificity, and visible motion potential. Avoid vague atmospheric thumbnails.

Capture

For each reference:

  1. Save one representative still image locally.
  2. Capture a short local video of the actual landing page or preview when possible.
  3. Include slow scroll, visible hover or click states when obvious, sticky nav behavior, reveal animation, parallax, marquee loops, carousels, cursor effects, and responsive or scroll-driven changes.
  4. Whenever capture tooling scrolls to a viewport or section before taking a screenshot or motion frame, wait 2 seconds before capturing so reveal animations, lazy media, sticky state changes, and scroll-triggered transitions have time to finish.
  5. Save one full landing-page scroll screenshot as a single tall image when the browser can render it. It should run from the first pixel of the hero to the final pixel of the footer.
  6. Cut section-by-section screenshots from that full-page image. The crops must be contiguous and lossless relative to the full image: no missing pixels between sections, no arbitrary viewport-only substitutes, and no overlap unless a sticky element makes overlap unavoidable.
  7. Save as many section images as the page has meaningful sections. At minimum include the hero, each major middle section in page order, and the footer. Label unclear middle sections as section-02, section-03, etc., while preserving exact visual order.
  8. Extract at least 4 motion progression frames when video succeeds: opening hero, hover or settled state, scroll transition, and lower or later section.
  9. If live capture is blocked, broken, login-gated, or low quality, save the best local fallback asset and document the fallback reason.

Keep all media inside the dated article folder. Use subfolders such as motion-frames/ when helpful.

Article Shape

In content.md, make the capture immediately useful to a builder:

  • open with the date and a concise pattern summary
  • embed any stitched motion-study video near the top, if created
  • for each of the 5 references, include the still image near the item heading
  • embed the local video, or state the fallback reason
  • match the established article structure from recent captures: ### Full-Page And Section Evidence, the full-page screenshot as a normal Markdown image, #### Section Crops, then each section crop as a normal Markdown image ordered from hero to footer
  • do not rename the block to Local Evidence, do not use text links, filename lists, Markdown tables, raw HTML grids, or crop-coordinate captions for the full-page/section-crop block; keep crop coordinates in manifest.json
  • show motion progression frames before the long prompt
  • describe what works, the visual pattern, and concrete motion notes
  • include a long AI-builder prompt that treats the page as reference, not a template
  • add a short note explaining why the prompt works

Write prompts with concrete instructions for layout, sections, components, typography, spacing, colors, image treatment, animation timing, scroll choreography, responsive behavior, interaction states, and things to avoid.

Manifest Shape

Use JSON. Include article-level metadata such as:

  • title
  • date
  • captureMethod
  • itemCount
  • dedupeAgainst
  • motionStudyVideo, when present
  • items

For each item, include fields such as:

  • title
  • creator
  • sourceUrl
  • pageUrl
  • imageUrl
  • imageFile
  • videoFile or videoFallbackReason
  • fullPageImage
  • sectionImages with section label, file path, source full-page image, y-start, y-end, and height when available
  • motionFrames
  • captureType
  • viewport
  • why
  • pattern
  • interactionNotes
  • animationNotes
  • prompt

Verification

Run the duplicate checker before committing:

node scripts/check-ui-inspiration-duplicates.mjs articles/YYYY-MM-DD-ui-inspiration-capture/manifest.json

Verify:

  • duplicateCount is 0
  • manifest.json has exactly 5 items
  • content.md embeds at least 5 representative still images
  • every referenced still, video, and motion-frame file exists locally
  • every live-website item has one full-page scroll screenshot and section crops for hero, all detected/meaningful middle sections, and footer
  • section crops are derived from the full-page image, are ordered top-to-bottom, and do not miss pixel ranges between adjacent sections
  • content.md follows the established full-page/section-crop structure for each live-website item: ### Full-Page And Section Evidence, rendered full-page Markdown image, #### Section Crops, and rendered section-crop Markdown images in top-to-bottom order, with no text links, filename lists, Markdown tables, raw HTML grids, Local Evidence headings, or crop-coordinate captions
  • each item has a local video or a clear fallback reason
  • successful video captures include at least 4 motion frames
  • README.md is absent

If articles/ARTICLES_INDEX.md is clean, add a single discoverability row only when that matches the current repo convention. If it is already dirty or conflicted, leave it untouched and mention that in the closeout.

Because articles/ is often ignored, use git add -f for the intended dated folder. Stage and commit only the files created or intentionally changed for this run.

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

Build woven fabric in Three.js with visible threads, realistic sheen, seams, and thick hems. The skill covers folds that preserve texture proportions, light passing through thin cloth, and variations in dye and wear.

中文简介正在准备中。当前显示英语简介。

  • Creating backlit curtains
  • Differentiating linen, silk, and wool
  • Adding hems and stitches for close-up views
MengTo/Skills69722026年10月18日 更新

Build falling leaves in Three.js and similar 3D scenes, with visible front, edge, and back views as they tumble and drift. Coordinate wind, scene depth, camera-aware recycling, and performance checks for a convincing autumn effect.

中文简介正在准备中。当前显示英语简介。

  • Adding falling autumn leaves to a 3D scene
  • Coordinating leaf tumble, sideways drift, and wind
  • Managing leaves around buildings and camera movement
MengTo/Skills69722026年10月18日 更新

Add coordinated seasonal transitions to an existing 3D scene. Blend foliage, lighting, sky, ground, snow, and particles from one shared state while preserving camera controls and interactions, so the same place remains recognizable throughout spring, summer, fall, and winter.

中文简介正在准备中。当前显示英语简介。

  • Add a four-season selector to an outdoor 3D scene
  • Show seasonal changes in an architectural walkthrough
  • Coordinate foliage, sunlight, sky, and ground appearance
MengTo/Skills69722026年10月18日 更新

Create or integrate detailed 3D models with smooth outlines and believable edges for close-up views. Prepare lower-detail versions for distant viewing, prioritize loading, and measure rendering costs in the actual scene.

中文简介正在准备中。当前显示英语简介。

  • Creating a detailed 3D model for close-up views
  • Replacing visibly faceted curves with smoother geometry
  • Preparing detailed architectural assets for rendering
MengTo/Skills69722026年10月18日 更新

Prepare detailed 3D materials that remain clear in close-up views. The skill guides texture resolution, physically based material maps, surface mapping, filtering, and progressive loading while checking download size and GPU memory use.

中文简介正在准备中。当前显示英语简介。

  • Prepare wood, stone, and fabric materials for close-up views
  • Fix texture seams, grain direction, and normal map orientation
  • Preserve floor and foliage detail at oblique or distant views
MengTo/Skills69722026年10月18日 更新

Create realistic metal surfaces in Three.js by combining structured reflections, studio lighting, scratches, fingerprints, and beveled edges. Includes polished, brushed, and aged finishes for objects such as tools, rings, and metal product parts.

中文简介正在准备中。当前显示英语简介。

  • Fixing metal that looks like gray plastic
  • Creating polished or brushed metal finishes
  • Adding patina and worn edges
MengTo/Skills69722026年10月18日 更新

查看 MengTo 的全部技能

报告此技能的问题