网站截图预览

SKILL内容创作
内容创作 未标注权限 许可证不明确
安装这个技能

将当前工作区或指定 URL 中的成品网站整理成可分享的本地作品集。自动发现页面、生成桌面端与移动端截图,并提供离线展示画廊及多种可下载的社交分享卡片。适用于网站展示、作品打包、截图和分享,不负责网站部署或发布。

一句话选型结论

按资料字段整理

网站截图预览是内容创作类的 Agent Skill,安装方式未标注,权限级别未标注,许可证不明确,有可直接复制的安装命令。技能本质是给模型读的说明书,效果取决于模型与客户端;装完先用一个小任务验证,再放进正式流程。

分类内容创作 安装方式未标注 权限级别未标注权限 许可证许可证不明确
累计安装 未标注 本周 未标注
GitHub 星标 未标注 仓库 未标注
下载量 未标注 技能市场下载数
权限级别 未标注权限 安装前请确认风险范围
许可证 许可证不明确 未说明

安装

未标注
在支持 Skills 的客户端里执行,装完问一句「你有哪些技能」验证
skillhub install user_71358182/website-showcase
  1. 1
    确认客户端支持

    确认你的客户端支持 Skills(Claude、Codex、Cursor 等),并升级到较新的版本。

  2. 2
    拿到技能文件

    执行上面的命令安装;也可以把仓库里的 SKILL.md 手动放进客户端的 skills 目录。

  3. 3
    刷新技能列表

    重新打开客户端,或刷新技能列表,让新放进来的技能被加载。

  4. 4
    验证是否生效

    问一句「你有哪些技能」,确认列表里出现它,再拿一个小任务试跑。

技能档案

内容创作
字段来自技能仓库与公开索引,中文说明由本站标注
作者 / 维护方user_71358182仓库 未标注 · 版本 未标注
宿主客户端未标注技能可被这些客户端加载
安装方式未标注安装包 community
权限级别未标注权限安装前请确认风险范围
许可证未说明许可证不明确,商用前请自行确认条款
仓库可用性available技能仓库当前是否可访问
最近核验未标注上游最后核验时间 · 数据更新 1791671827
正文体量6.7 KBSKILL.md 原文大小,越大说明说明书写得越细

典型使用场景

3 类
按分类整理的常见用法,具体以 SKILL.md 为准
写文章与脚本

按主题产出初稿与分镜脚本。

改稿与润色

把草稿改顺、统一语气。

多平台文案改写

同一内容按平台改成长短不同的版本。

选择建议

装之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 有可直接复制的安装命令,不用自己翻文档找入口。

先想清楚的地方

  • 没有标注权限级别,实际能做什么要装好后确认。
  • 许可证不明确,商用或二次分发前先到仓库确认条款。
  • 没有做站内核验,内容以作者仓库为准。
  • 技能只是说明书,实际效果取决于模型与客户端,装完先用小任务试一遍。

更适合

  • 需要在内容创作场景里给模型补一套做法的人。
  • 愿意先读一遍 SKILL.md、确认它会让模型做什么的人。

不太适合

  • 对许可证要求很严、必须明确可商用的人。
  • 只想装完就自动变强、不打算看说明书的人。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
技能权限级别安装方式累计安装宿主客户端安装命令
网站截图预览当前 未标注权限 未标注 未标注 未标注 有
AI 视频生成 调用外部服务 GitHub 仓库 431.3 万 Codex、Claude、Cursor 有
视频帧 联网访问 ClawHub 市场 262.1 万 OpenClaw 有
Markdown 转换器 写入文件 ClawHub 市场 135.3 万 OpenClaw 有
哔哩哔哩全合一 接触敏感数据 ClawHub 市场 64.1 万 OpenClaw 有
Remotion video toolkit 命令行 浏览器控制 ClawHub 市场 63.9 万 OpenClaw 有

SKILL.md 全文

6.7 KB
模型实际读到的说明书原文,来自技能仓库

Website Showcase

Convert a working website into presentation assets. Reuse the versioned standard showcase and card-preview template; each run supplies only project data and screenshots. Anonymous PostHog analytics are enabled by default from config/analytics.json; use --no-analytics when the user asks to disable tracking for a run.

Browser Tool Detection

Before resolving the source, inspect the tools that are actually callable in the current session and select one browser backend for navigation, session access, and visual checks:

  1. Use Chrome MCP tools matching mcp__claude-in-chrome__* when the target requires an authenticated browser session or other state already present in the user's Chrome profile.
  2. Otherwise use Playwright MCP tools matching mcp__plugin_playwright_playwright__* for public or local websites and interactive visual checks.
  3. Use playwright-cli when it is callable and the task needs a repeatable scripted or batch flow, or when neither MCP backend is available.

The prefixes above are capability checks, not commands to invent. If the preferred backend is unavailable, continue with the next available backend and report the fallback only when it affects access or coverage. If none of the three is callable, use the bundled website-showcase.mjs browser path for public or local sources; stop and report the missing authenticated-session capability for a protected source. Use one backend consistently for a flow unless authenticated state or a backend failure requires switching.

The selected browser backend complements the bundled CLI rather than replacing the output pipeline. Use it to reach pages, reuse session state, inspect navigation, or verify captures. Keep website-showcase.mjs generate as the canonical step for copying the validated template and writing showcase-data.json. When authenticated state cannot be passed to the bundled CLI, capture the confirmed routes with Chrome MCP and record the image paths in the external capture manifest defined by references/showcase-contract.md.

Workflow

  1. Run Browser Tool Detection, then resolve the source. If the user has not already specified one, ask them to choose:
  • Current environment: find the generated website in the current workspace, detect its start command, and use an existing server when available.
  • Service URL: use the supplied HTTP(S) address and skip project detection and server startup.

For current-environment mode, run:

powershell node <skill-dir>/scripts/website-showcase.mjs detect --project <project-dir>

Prefer a server that is already running. Otherwise start the detected command in a persistent terminal and use the exact URL printed by the server. Do not change application code merely to make capture convenient. The source step is complete when one reachable entry URL and its source mode are known.

  1. Discover candidate pages. Use the bundled discovery command for public or local sources that it can access:

powershell node <skill-dir>/scripts/website-showcase.mjs discover --source <environment|service> --url <url> --out <work-dir>/discovery.json

Read the manifest and use the selected browser backend to verify ambiguous routes or discover pages that depend on the active browser session. For authenticated sources inaccessible to the bundled CLI, inspect navigation with Chrome MCP and create an equivalent manifest that satisfies the showcase contract. Present a compact L1/L2 page tree. Explicitly identify pages excluded by default, navigation failures, the discovery limit, and any routes the selected backend could not access.

  1. Stop for page confirmation. The user decides which pages represent the work. This gate is complete only when every requested page is either selected or deliberately excluded. If the user already gave an exact route list, treat that as confirmation.

  2. Capture screenshots and bind them to the standard template in a new output directory. For sources accessible to the bundled CLI, run:

powershell node <skill-dir>/scripts/website-showcase.mjs generate --manifest <work-dir>/discovery.json --out <output-dir> --route / --route /pricing

For an authenticated source captured with Chrome MCP, import its capture manifest instead of launching a new browser session:

powershell node <skill-dir>/scripts/website-showcase.mjs generate --manifest <work-dir>/discovery.json --capture-manifest <work-dir>/captures.json --out <output-dir> --route / --route /account

Omit --route only when the user confirmed every page marked defaultSelected. Never overwrite an existing output directory; choose a new name unless the user explicitly requests replacement. Generation copies the validated assets/showcase.html unchanged and writes project-specific fields to showcase-data.json. Analytics failure must not fail screenshot generation; report its status separately.

  1. Check the run-specific artifacts only: every confirmed route has the requested device captures, files are non-empty, and failures is empty. Use the selected browser backend to open the served showcase and verify that the generated assets load. Serve the output directory over localhost for the user.

The showcase UI and the Single image, Device pair, and Multi-page card layouts are prevalidated template behavior. Repeat browser/visual validation only when the template version changed, a capture is blank or malformed, the generated data violates the template contract, or the user explicitly asks for validation.

  1. Report the output path, served preview URL, source mode, selected browser backend, selected page count, screenshot count, capture failures, and analytics status. Keep the preview server running so the user can inspect it.

Runtime

The CLI requires Node.js and the skill-local playwright dependency. If the import is missing, run npm install in the skill directory. The CLI falls back to an installed Chrome or Edge when Playwright's bundled Chromium is absent. If none is available, run npx playwright install chromium in the skill directory, then retry the failed command.

Read references/showcase-contract.md when deciding capture scope, output behavior, privacy, or analytics. Do not add hosting, accounts, cloud storage, community features, or automatic social publishing to this workflow.

常见问题

5 问
答案由本站依据条目字段整理
网站截图预览怎么安装?

在支持 Skills 的客户端里执行:skillhub install user_71358182/website-showcase。安装方式为未标注。装完刷新技能列表或重开客户端。

网站截图预览支持哪些客户端?

资料没有标注宿主客户端,但只要客户端支持 Skills(Claude、Codex、Cursor 等),一般都能加载。

网站截图预览权限高吗,安全吗?

权限级别为未标注权限。安装前先看它会读取或写入哪些内容。

网站截图预览收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于技能本身和所用的模型或客户端,正式使用前按作者说明确认。

网站截图预览用什么许可证?

许可证标注为 未说明(许可证不明确)。商用或二次分发前请到仓库确认条款原文。

热门专题

按场景挑技能
全部专题 →

技能内容以作者仓库为准;分类、权限级别与许可证状态的中文说明为中文整理,最近核验 未标注。