Token导航 LogoToken导航

Agent Skill · 研究检索

canvas-page-definition(画布页面定义)

canvas-page-definition 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。

canvas-page-definition 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,权限为需要联网;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 47,337。

总安装 737周安装 31GitHub Stars 1下载 258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:canvas-page-definition(画布页面定义)
来源仓库:https://github.com/drupal-canvas/skills
仓库路径:skills/canvas-page-definition
安装命令:
npx skills add https://github.com/drupal-canvas/skills --skill canvas-page-definition
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/drupal-canvas/skills --skill canvas-page-definition

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。

canvas-page-definition 定义 Canvas 页面的标准格式,确保本地 JSON 规范与 Drupal Canvas 平台一致。用于页面预览与推送/拉取同步流程,要求页面文件位于指定目录且满足标题、结构等基本约束。

适用于 Workbench 工作流中的页面开发,需配合 canvas.config.json 配置 pagesDir 路径。修改页面定义前应备份原文件,避免因格式错误导致同步失败或内容丢失。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/canvas-page-definition

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/drupal-canvas/skills --skill canvas-page-defin…;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

    它需要需要联网,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

canvas-page-definition 怎么安装?

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

canvas-page-definition 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。

canvas-page-definition 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

canvas-page-definition 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。

canvas-page-definition 用的人多吗?

公开数据是累计安装 737 次,近一周安装 31 次,下载 258 次,GitHub 1 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。

canvas-page-definition 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。
  • 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
canvas-page-definition

画布页面定义 · 本页

7371需要联网未明确通过

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

Canonical definition

A Canvas page is a JSON page spec stored in the repository's configured pagesDir.

Workbench uses these files for local page previews, and the Canvas CLI push/pull workflow uses them to sync pages between the local project and Drupal Canvas.

Page specs should fully mirror what authors can build in Drupal Canvas.

Minimum contract (MUST)

Every Canvas page MUST satisfy all checks below:

  • The page file exists in the configured pages directory (pagesDir in canvas.config.json; default ./pages)
  • The page file is a JSON object
  • The page includes a non-empty title
  • The page includes a path that starts with / and is unique within the project
  • The page includes an elements object with at least one element
  • Every element in elements includes a discovered component type
  • Any slots entries reference element IDs defined in the same file

If any item is missing, the page is incomplete for Canvas usage.

Location and naming

Author the canonical local page specs in the configured pages directory (pagesDir in canvas.config.json). If pagesDir is not set, Workbench defaults to the top-level pages/ directory.

  • Place each page in <pages-dir>/<page-name>.json
  • Keep page files flat in the configured pages directory; nested files are ignored
  • Use a stable, descriptive filename that mirrors the page title in a filename-safe way
  • Use a human-readable title inside the JSON file for the actual page title

Examples:

  • <pages-dir>/homepage.json
  • <pages-dir>/about-us.json

Treat files in the configured pagesDir as the source of truth for local page definitions that can be previewed in Workbench and synced with Canvas via the CLI.

Authored page work is incomplete unless the page is previewable from the configured pagesDir and ready for validation.

Page file format

Each page file must be a JSON object with:

  • title: the actual page title
  • path: the page's URL path (e.g., /about). Must start with / and be unique within the project. Other pages link to this page using this value.
  • elements: an object of authored page elements

The page file may also include:

  • description: a short plain-text summary of the page. Surfaced as page metadata (e.g., for listings or SEO). Leave as an empty string if not set.

Each entry in elements must include:

  • type: a discovered component type

Each entry may also include:

  • props: root props for that element
  • slots: a map of slot names to arrays of element IDs from the same file

Workbench preserves elements insertion order when building the page. Put top-level sections in the order you want them rendered.

Use the discovered machine-readable component type for each element, such as card, heading, or js.hero, depending on what Workbench finds in the local project.

Linking between pages

To link to another page in the project, reference the target page's path.

Example

{
  "title": "About",
  "path": "/about",
  "description": "",
  "elements": {
    "hero": {
      "type": "js.hero",
      "props": {
        "heading": "About us",
        "text": "Learn how our team plans, builds, and ships Canvas experiences.",
        "layout": "left_aligned"
      }
    },
    "spacer-lg": {
      "type": "js.spacer",
      "props": {
        "height": "large"
      }
    },
    "body-section": {
      "type": "js.section",
      "props": {
        "width": "normal"
      },
      "slots": {
        "content": ["body-copy"]
      }
    },
    "body-copy": {
      "type": "js.text",
      "props": {
        "content": "<p>Canvas pages should use real component composition rather than custom page wrappers.</p>"
      }
    }
  }
}

Use stable, descriptive element IDs unless you are intentionally preserving IDs from another source such as an exported Canvas page.

Image props in pages

When a component prop is defined as an image in component.yml, page JSON must pass that prop as a single image object. Match the component prop name exactly and use the Canvas image shape inside props.

{
  "type": "js.card",
  "props": {
    "image": {
      "src": "https://placehold.co/400x300/png",
      "alt": "Card placeholder image",
      "width": "400",
      "height": "300"
    },
    "heading": "Featured card"
  }
}

Supported authored page image sources:

  • External http:// or https:// URLs
  • data:image/... URLs

Do not use local file paths in page JSON. CLI media upload from local files is not supported.

If page media reconciliation or page sync adds resolved image data such as width, height, or _provenance, preserve that data unless you are intentionally changing the image source.

Format constraints

The page-spec format only supports discovered component elements plus their props and slots.

That means page specs cannot directly represent:

  • Hand-written React wrappers such as PageLayout
  • Raw HTML layout elements such as <div> or <section>
  • className-driven spacing or custom wrapper styling
  • Inline code duplication of an existing component's markup

If you need a layout primitive, look for an existing component first. If none exists, create it as a component before using it in a page spec.

Validation

Validate every authored page before finishing.

npx canvas validate

平台分布

Codex

35.16%
按下载量换算91

Claude

29.28%
按下载量换算76

Cursor

17.34%
按下载量换算45

Gemini CLI

8.65%
按下载量换算22

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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