Token导航 LogoToken导航TokenDH.com
研究检索需要联网clawhub未标认证来源可访问clear审计通过

notebooklm-youtube-skillNotebookLM youtube 技能

Agent Skill

notebooklm-youtube-skill 用于查找、检索和筛选相关信息,适合在 OpenClaw 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。

总安装

2,203

周安装

90

GitHub Stars

公开资料未说明

下载量

713
OpenClaw

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

MIT-0

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:notebooklm-youtube-skill(NotebookLM youtube 技能)
来源仓库:https://github.com/x-rayluan/notebooklm-youtube-skill
安装命令:
openclaw skills install notebooklm-youtube-skill
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 OpenClaw 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。

ClawHubOpenClaw
openclaw skills install notebooklm-youtube-skill

简介

自动从 YouTube 视频创建 NotebookLM 笔记本。

  • 提取人物信息并研究相关背景资料。notebooklm-youtube-skill 属于研究检索类 Skill,可作为该场景下的辅助能力补充。
  • 适用于视频内容深度分析与知识整理。
  • 需提供有效的 YouTube URL 作为输入源。
  • 注意遵守 YouTube 内容使用政策。

SKILL.md

name
notebooklm-video-research
description
Automate NotebookLM notebook creation from YouTube videos. Given a YouTube video URL, extract information about people featured in the video, research them online, create a new NotebookLM notebook with the video and research as sources, and generate an audio overview. Uses screenshot-first automation to reliably target UI elements.

NotebookLM Video Research Skill

Automate NotebookLM notebook creation from YouTube videos with automatic research and Audio Overview generation.

Core Principle

Screenshot before every action — verify UI state, identify correct elements, confirm success.

📸 Screenshot → 👀 Analyze → 🎯 Target → ✅ Verify → 🖱️ Execute → 📸 Screenshot → ✓ Confirm

Quick Start Workflow

Phase 1: Research

# 1. Get video info
Control Chrome:open_url("https://www.youtube.com/watch?v=VIDEO_ID")
Control Chrome:get_page_content()

# 2. Research people mentioned
web_search("[Person Name] background career")

# 3. Create research document (keep under 500k chars)

Phase 2: NotebookLM Automation

# 1. Navigate
Control Chrome:open_url("https://notebooklm.google.com/")

# 2. Create notebook
# Click "Create new" button

# 3. Add YouTube source
# Click "Websites" → Enter URL → Click "Insert"

# 4. Add research document  
# Click "Add sources" → "Copied text" → Paste content → Click "Insert"

# 5. Generate audio
# Click "Audio Overview" in Studio panel

Critical: Element Targeting

⚠️ The Main Trap

NotebookLM has multiple textareas. Generic selectors grab the wrong one!

SIDEBAR (wrong)          CENTER MODAL (correct)
┌──────────────┐         ┌──────────────────┐
│ Search box   │         │ URL/Text input   │
│ query-input  │         │ "Paste any links"|
└──────────────┘         └──────────────────┘

Correct Selectors

ElementSelector
URL inputtextarea[placeholder="Paste any links"]
Text inputtextarea[placeholder="Paste text here"]
Sidebar (AVOID)textarea[placeholder*="Search"]

Angular Event Pattern

// Simple .value = doesn't work! Use this:
var setter = Object.getOwnPropertyDescriptor(
    HTMLTextAreaElement.prototype, 'value'
).set;
setter.call(textarea, 'content');
textarea.dispatchEvent(new Event('input', {bubbles: true}));
textarea.dispatchEvent(new Event('change', {bubbles: true}));

Button State Check

var btn = document.querySelector('button');
if (btn && !btn.disabled && !btn.className.includes('disabled')) {
    btn.click();
}

Complete Code Snippets

Enter YouTube URL

var urlTextarea = document.querySelector('textarea[placeholder="Paste any links"]');
if (urlTextarea) {
    var setter = Object.getOwnPropertyDescriptor(
        HTMLTextAreaElement.prototype, 'value'
    ).set;
    setter.call(urlTextarea, 'https://www.youtube.com/watch?v=VIDEO_ID');
    urlTextarea.dispatchEvent(new Event('input', {bubbles: true}));
    urlTextarea.dispatchEvent(new Event('change', {bubbles: true}));
}

Enter Research Text

var pasteTextarea = document.querySelector('textarea[placeholder="Paste text here"]');
if (pasteTextarea) {
    var setter = Object.getOwnPropertyDescriptor(
        HTMLTextAreaElement.prototype, 'value'
    ).set;
    setter.call(pasteTextarea, RESEARCH_CONTENT);
    pasteTextarea.dispatchEvent(new Event('input', {bubbles: true}));
    pasteTextarea.dispatchEvent(new Event('change', {bubbles: true}));
}

Click Button by Text

var buttons = Array.from(document.querySelectorAll('button'));
var btn = buttons.find(b => b.textContent.includes('Insert'));
if (btn && !btn.disabled) {
    btn.click();
}

Verify State

// Check what textareas exist
var textareas = Array.from(document.querySelectorAll('textarea'));
textareas.map(t => ({
    placeholder: t.placeholder,
    value: t.value.substring(0, 50)
}));

Checkpoint Summary

StepVerify BeforeVerify After
Navigate-Home page loaded
Create notebookCreate button visibleNotebook opened
Add URLCorrect textarea identifiedURL in center, not sidebar
Click InsertButton ENABLEDSource added
Add textCorrect textarea identifiedText in center, not sidebar
Click InsertButton ENABLED2 sources shown
Audio OverviewStudio panel visibleGeneration started

Anti-Patterns

❌ Don't✅ Do
document.querySelector('textarea')textarea[placeholder="Paste any links"]
Click without checking stateVerify button enabled first
Assume action succeededScreenshot to confirm
Rush through stepsWait for async operations

Wait Times

ActionWait
Page navigation3s
Dialog open2s
YouTube source processing8-10s
Text source processing5s
Audio generation5-10 minutes

Error Recovery

Wrong input targeted:

  1. Clear sidebar: document.querySelectorAll('.query-input').forEach(el => el.value = '')
  2. Re-target with specific selector

Button disabled:

  1. Re-dispatch events on textarea
  2. Verify content actually entered

Modal didn't open:

  1. Wait longer
  2. Click trigger button again

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

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

补充不同宿主或平台的使用分布数据

能力 5

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

OpenClaw

94.49%
按下载量换算674

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

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。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