Token导航 LogoToken导航TokenDH.com
开发操作浏览器github未标认证来源可访问许可证需确认审计提醒

bubble-io-plugins气泡 io 插件

Agent Skill

bubble-io-plugins 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。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

303

周安装

13

GitHub Stars

公开资料未说明

下载量

106
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:bubble-io-plugins(气泡 io 插件)
来源仓库:https://github.com/afaraha8403/bubble-io-plugin-boilerplate
仓库路径:skills/bubble-io-plugins
安装命令:
npx skills add https://github.com/afaraha8403/bubble-io-plugin-boilerplate --skill bubble-io-plugins
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/afaraha8403/bubble-io-plugin-boilerplate --skill bubble-io-plugins

简介

用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息。

  • 适合在需要围绕仓库状态、代码变更或协作事项进行整理时使用。
  • 可结合来源仓库和原始 README 核验具体用法,支持 Bubble.io 插件开发流程。
  • 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。
  • 适用于 Codex、Claude、Cursor、Gemini CLI 等宿主环境。

SKILL.md

Bubble.io Plugin Development — Project Rules

Project identity

This is a Bubble.io plugin development boilerplate. It provides the folder structure, coding conventions, and tooling for building plugins that run inside the Bubble.io no-code platform.

Plugins are deployed by copying code into the Bubble Plugin Editor — no build step, no npm publish.

Project structure

project-root/
  actions/
    client/                # Client-side workflow actions
      <action-name>/
        action-setup.md
        client.js
        params.json        # Optional: parameter definitions
    server/                # Server-side actions (runs on Bubble's Node.js server)
      <action-name>/
        action-setup.md
        server.js
  elements/                # Visual plugin elements
    <element-name>/
      element-setup.md
      initialize.js        # Runs once on element load
      update.js            # Runs on every property change + data load
      preview.js           # Renders placeholder in Bubble Editor
      header.html          # <head> content: CDN links, external scripts
      actions/             # Element-specific workflow actions
        <action>.js
  eslint.config.mjs        # ESLint flat config
  package.json             # ESLint scripts and dependencies
  README.md

Key architectural fact

Each local file maps 1:1 to a text field in the Bubble Plugin Editor:

Local fileBubble Editor field
initialize.jsFunction: initialize
update.jsFunction: update
preview.jsFunction: preview
header.htmlElement Header
actions/<name>.jsElement Action code
server/<name>/server.jsServer-Side Action code
styles.cssShared/Element Header (wrap in <style> tags)

Code quality expectations

When generating or editing any code in this project, follow these rules unconditionally.

Well-formatted, readable code

All code must be clean, consistently formatted, and easy to scan. This means:

  • Logical sections separated by blank lines — group related statements together (data loading, guards, rendering, event binding).
  • Descriptive variable names — avoid single-letter or cryptic abbreviations (container not c, itemCount not ic).
  • Consistent indentation — 2-space indent for all JS; match surrounding code if editing an existing file.
  • Section banners for update.js — use comment blocks (// === SECTION ===) to delimit lifecycle phases (data loading → guard → change detection → cleanup → render).
  • One concern per function — extract helpers for any logic longer than ~10 lines; define helpers *inside* the wrapper function to avoid global leaks.

Inline documentation

Every non-trivial block of code must include an inline comment explaining why it exists, not just what it does. Specifically:

  • Data loading — explain what each properties.* field contains and why it is loaded first.
  • Guards / early returns — explain the condition being checked and what would happen without the guard.
  • DOM mutations — explain the structure being built and any Bubble-specific constraints (e.g., why we use instance.canvas instead of document.body).
  • Event listeners — explain the namespace convention and why previous listeners are removed.
  • Workarounds — any Bubble quirk or browser compat hack must have a comment linking to the reason.

JSDoc comments

All functions (wrappers and helpers) must have JSDoc blocks. Follow the rules in documentation.md Section 1. Summary:

  • Wrapper functions (initialize, update, preview, actions) — include a top-level @description summarising the function's purpose, followed by @param tags for each argument (instance, properties, context).
  • Helper functions@param, @returns, and a one-line description.
  • Placement — JSDoc goes inside the wrapper, not above it (the wrapper line is stripped when pasting into Bubble).

Example (initialize wrapper):

let initialize = function(instance, context) {
  /**
   * @description One-time setup for the PLUGIN_PREFIX element.
   * Creates the root DOM container, generates a unique event namespace,
   * and initialises default exposed states.
   *
   * @param {object} instance - Bubble element instance (canvas, data, publishState, etc.)
   * @param {object} context  - Bubble context (keys, currentUser, etc.)
   */

  // ... implementation ...
};

Debug logging (verbose_logging)

When scaffolding a new element or action from scratch, ask the user once:

"Should this component include a verbose_logging toggle? This adds a boolean field in the Bubble Plugin Editor that gates all console.log output at runtime."

Do not ask on edits, reviews, refactors, or bug fixes — only on new scaffolds.

If the user accepts:

  1. Add a boolean field called verbose_logging to the element or action configuration in the Bubble Plugin Editor and document it in the relevant setup file.
  2. Gate all console.log calls behind properties.verbose_logging:
if (properties.verbose_logging) {
  console.log('[PLUGIN_PREFIX] update called', { properties });
}
  1. Log placement — add gated log statements at:

- Entry point of update.js, client actions, and server actions - After data loading completes - Before and after external API calls (server actions)

  1. console.error() in catch blocks is always unconditional — never gate error logging behind the verbose flag.
  2. initialize.js does not receive properties — verbose logging is unavailable. Use a plain console.log only for temporary init-time debugging; remove before production.
  3. preview.js and header.html run in the editor only — verbose logging does not apply.

If the user declines, omit all console.log statements. console.error() in catch blocks remains unconditionally.


Critical pitfalls — always keep in mind

These are the highest-consequence rules. Violating any of these causes hard-to-debug failures:

  1. Never catch the 'not ready' exception — Bubble uses it as control flow for data loading. If you must use try/catch, re-throw when err.message === 'not ready'.
  2. Load all data at the TOP of the function — before any DOM mutations. Bubble re-runs the entire function from the start when data arrives.
  3. Never append to document.body — use instance.canvas for all visual output.
  4. Never put API keys in client-side code — use server-side actions with context.keys.
  5. Copy only the function BODY to the Bubble Plugin Editor — not the wrapper.
  6. Prefix all CSS classes (e.g., myPlugin-root) — avoid collisions with the host app.
  7. SSA in v4 must be async — use await on .get(), .length(), and fetch().
  8. Headers only support <script>, <meta>, <link> — anything else gets auto-moved to <body>.
  9. Do NOT use $(document).ready() inside plugin functions — it breaks Bubble's dependency detection.

Which reference to load

Do not preload all files. Determine the task type, then load only the relevant reference:

  1. Determine the task:

- Writing/reviewing element runtime code (initialize.js, update.js, preview.js, header.html)? → Load bubble-platform.md - Need instance/properties/context API details, or v4 migration? → Load bubble-api.md - Working on actions (client-side or server-side)? → Load actions-guide.md - Writing, reviewing, or refactoring any JavaScript? → Load code-standards.md - Writing docs, setup files, or user-facing text? → Load documentation.md - Multiple concerns? → Load the most relevant file first, add others only if needed.

FileLoad when...
bubble-platform.mdElement lifecycle, DOM/canvas, data loading, headers, preview, events, debugging, hard limits.
bubble-api.mdinstance, properties, context API reference. BubbleThing/BubbleList types. Custom data types / API Connector App Types. Plugin API v4 migration.
actions-guide.mdClient vs server actions. When to use which. SSA Node modules, return values, option sets.
code-standards.mdESLint config, syntax rules, security, performance, error handling.
documentation.mdJSDoc, setup files, marketplace descriptions, field tooltips, changelog, publishing.

Starter templates

When scaffolding a new element or action, copy the relevant template from assets/templates/:

TemplateUse for
initialize.jsNew element — container setup, instance.data, event namespace
update.jsNew element — data-first pattern, change detection, namespaced listeners
preview.jsNew element — editor placeholder with responsive sizing
header.htmlNew element — idempotent <script> loading
client-action.jsNew client-side action
server-action.jsNew server-side action (v4 async/await)

General expectations

  1. State reasoning. When recommending a change, explain *why* — do not just state the rule.
  2. Preserve existing patterns. Before introducing a new pattern, check if the codebase already uses a convention for the same concern.
  3. No unnecessary files. Do not create files unless the task requires it. Prefer editing existing files.
  4. Linting is enforced via ESLint. Configuration lives in eslint.config.mjs (flat config format). VS Code auto-fixes on save via .vscode/settings.json (source.fixAll.eslint). Do not introduce a second formatter.

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

35.71%
按下载量换算38

Claude

27.99%
按下载量换算30

Cursor

18.28%
按下载量换算19

Gemini CLI

10.57%
按下载量换算11

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。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