Skills notebooklm

📚

notebooklm

Name: notebooklm
Author: ZhanlinCui

Low Risk ⚙️ External commands🌐 Network access📁 Filesystem access🔑 Env variables

Query NotebookLM Notebooks from Claude

Also available from: teng-lin

This skill enables Claude to query Google NotebookLM notebooks directly, providing source-grounded answers from Gemini that cite your uploaded documents, drastically reducing AI hallucinations.

Supports: Claude Codex Code(CC)

🥉 75 Bronze

Download the skill ZIP

Upload in Claude

Go to Settings → Capabilities → Skills → Upload skill

Toggle on and start using

Test it

Using "notebooklm". Ask: What is the main finding of my research notebook?

Expected outcome:

Based on the documents in your notebook, the main finding is that [specific content from your documents]. This is supported by [citation from source]. The key evidence includes [additional details from documents].

Using "notebooklm". Ask: List all the API endpoints in my project documentation

Expected outcome:

Your project documentation contains the following API endpoints: /api/users (GET, POST), /api/auth (POST), /api/documents (GET, POST, PUT, DELETE). Each endpoint is documented with request/response schemas in the OpenAPI specification section.

Security Audit

Low Risk

v1 • 2/24/2026

This skill is a legitimate Google NotebookLM integration using browser automation. The static findings are false positives triggered by: (1) shell command examples in documentation, (2) Playwright browser automation code for Google authentication, (3) network requests to Google's legitimate NotebookLM service. The combination of code execution + network + credentials is expected behavior for this type of integration skill.

Files scanned

3,425

Lines analyzed

findings

Total audits

High Risk Issues (1)

scripts/auth_manager.py:104-120 scripts/ask_question.py:66-77

Browser Automation with Credentials

Uses Playwright for browser automation with persistent authentication state. This is standard for skills that integrate with web services requiring login.

Medium Risk Issues (2)

scripts/__init__.py:53 scripts/__init__.py:65 scripts/setup_environment.py:54

Shell Command Execution in Scripts

Python subprocess calls for environment setup (venv creation, pip install, patchright install). These are required for skill initialization.

scripts/auth_manager.py:114 scripts/ask_question.py:80

Network Requests to Google Services

Hardcoded URLs for Google NotebookLM API endpoints. Required for integration functionality.

Low Risk Issues (2)

scripts/auth_manager.py:163-164 scripts/notebook_manager.py:58

Filesystem Access for Data Storage

Accesses local filesystem to store notebook library, authentication state, and browser profile. Standard for persistent skill data.

requirements.txt:10

Environment Variable Access

Uses dotenv for optional configuration. Standard skill configuration pattern.

Risk Factors

⚙️ External commands (3)

references/api_reference.md:160 scripts/__init__.py:53 scripts/setup_environment.py:54

🌐 Network access (2)

scripts/auth_manager.py:114 scripts/ask_question.py:80

📁 Filesystem access (2)

scripts/auth_manager.py:163-164 SKILL.md:193

🔑 Env variables (1)

requirements.txt:10

Audited by: claude

Quality Score

Architecture

100

Maintainability

Content

Community

Security

Spec Compliance

What You Can Build

Research Documentation Query

Ask questions about your research papers, meeting notes, or study materials stored in NotebookLM and get answers backed by citations.

Codebase Documentation Assistant

Query your codebase documentation, READMEs, and technical specs stored in notebooks to get accurate answers about your project.

Personal Knowledge Base QA

Treat NotebookLM as a personal knowledge base and ask Claude questions that retrieve specific information from your documents.

Try These Prompts

Check Authentication Status

Check if NotebookLM authentication is set up. Run: python scripts/run.py auth_manager.py status

List Available Notebooks

Show all notebooks in my library. Run: python scripts/run.py notebook_manager.py list

Query a Notebook

Ask a question to NotebookLM: python scripts/run.py ask_question.py --question "[YOUR QUESTION]" --notebook-id [NOTEBOOK_ID]. Use the follow-up mechanism to gather complete information.

Add New Notebook

Add a notebook to the library: python scripts/run.py notebook_manager.py add --url [URL] --name [NAME] --description [DESCRIPTION] --topics [TOPICS]. Use SMART ADD if content is unknown.

Best Practices

Always check authentication status before attempting queries
Use the follow-up mechanism to gather complete information before responding to user
Register notebooks with accurate descriptions and topics for better searchability
Use --show-browser flag for debugging when issues occur

Avoid

Do not call scripts directly without the run.py wrapper
Do not skip follow-up questions - always synthesize complete answers
Do not guess notebook descriptions - use SMART ADD to discover content
Do not forget to check authentication before operations

Frequently Asked Questions

How do I authenticate with Google NotebookLM?

Run python scripts/run.py auth_manager.py setup. A browser window will open for you to manually log in to your Google account. Authentication persists for 7 days.

Why do I need to use the run.py wrapper?

The run.py wrapper automatically creates and manages a virtual environment, installs dependencies, and executes scripts properly. Calling scripts directly will fail without the environment.

How many queries can I make per day?

Google NotebookLM has a rate limit of 50 queries per day on free accounts. If you hit the limit, wait 24 hours or use a different Google account.

Why does each question open a new browser?

This skill does not maintain persistent sessions. Each question creates a fresh browser context using stored authentication state. This is by design for reliability.

Where is my data stored?

All data is stored in ~/.claude/skills/notebooklm/data/ including library.json (notebook metadata), auth_info.json (authentication status), and browser_state/ (cookies).

How do follow-up questions work?

Every answer ends with 'Is that ALL you need to know?' - Claude will automatically ask follow-up questions to gather complete information before responding to you.