Notebooklm

Complete Google NotebookLM integration — add sources, ask questions, generate all Studio content (podcast, video, slide deck, quiz, flashcards, infographic,...

Security Scan

The skill's stated purpose (programmatic NotebookLM access) matches its instructions, but there are important mismatches and missing pieces around installation, declared credentials, and autonomous actions that could let an agent read or upload local/Drive content without clear safeguards.

Purpose & Capability

The name/description and SKILL.md consistently describe programmatic Google NotebookLM operations (add sources, ask, generate artifacts). However the skill references installing/using the third‑party library teng-lin/notebooklm-py and claims OpenClaw will install it, yet the package supplies no install spec or declared dependency — an internal inconsistency. Also the skill requires OAuth access to a Google account/Drive in practice but declares no credentials or required scopes.

Instruction Scope

The runtime instructions tell the agent to run CLI commands that add sources from URLs, local files (./file.pdf), YouTube, and Google Drive, to create notebooks, save Q&A as notes, and download artifacts. Several of these actions involve reading local files and uploading them to Google or programmatically accessing Drive and sharing permissions. The Autonomy Rules allow many of these commands to run without user confirmation (e.g., 'source add', 'create', 'ask', 'use'), which increases the risk that sensitive local or Drive data could be uploaded or exposed without clear user consent.

Install Mechanism

This is an instruction-only skill with no install spec. README/SKILL.md claim OpenClaw will install the notebooklm-py library, but there is no declared install procedure, no pinned package source, and no mention of which account or environment will perform installs. That ambiguity is a coherence issue: if the CLI isn't present, commands will fail; if the platform installs it automatically, that install step is not auditable from the skill bundle.

Credentials

The skill declares no required environment variables or primary credential, but its operations require full Google OAuth login and Drive/Docs/Sheets access in practice. Sensitive scopes (access to Drive files, read/write permission, sharing) are implied but not declared. The lack of declared credentials or scopes is disproportionate to the significant account-level privileges the skill needs to function.

Persistence & Privilege

always is false (good), but the skill allows autonomous invocation of high-impact commands (adding sources, creating notebooks, saving notes) without confirmation. Combined with implied Google account access and the ability to read/upload local and Drive files, autonomous runs could exfiltrate or publish data. The skill does not request to modify other skills or global configs, so there is no evidence of privilege escalation beyond data access.

Scan Findings in Context

[no-findings] expected: The regex scanner found no matches because this is an instruction-only skill with no code files. That absence is not evidence of safety — the SKILL.md itself contains the runtime behavior to evaluate.

What to consider before installing

This skill appears to implement programmatic access to Google NotebookLM but is missing important details. Before installing, verify the upstream library (teng-lin/notebooklm-py) yourself and confirm how OpenClaw will install it. Expect that using the skill requires logging into a Google account and granting Drive/Docs permissions; sensitive local files (./file.pdf) can be uploaded by the agent. If you have sensitive data, do not enable autonomous runs that can 'source add' or 'create' without confirmation — require manual approval for any upload/download actions. Prefer running this skill only with a dedicated Google account (least privilege) or in an isolated environment, and request the publisher to add an explicit install spec and declared OAuth scopes/requirements.

NotebookLM Skill

Full programmatic access to Google NotebookLM — including features not exposed in the web UI.

Library: teng-lin/notebooklm-py ⚠️ Unofficial — Uses undocumented Google APIs. May break without notice.

---

Prerequisites

Authentication must be active before any command:

notebooklm login          # Opens browser — log in with Google account
notebooklm list           # Confirm auth works
notebooklm auth check     # Diagnose auth issues
notebooklm auth check --test   # Full validation including network test

Verify readiness before any workflow:

notebooklm status         # Active notebook and session
notebooklm doctor         # Environment health check
notebooklm doctor --fix   # Auto-fix issues

---

When This Skill Activates

Explicit: /notebooklm, "use notebooklm", "with notebooklm"

Intent detection:

• "Create a podcast / audio overview about [topic]"
• "Summarize these URLs / documents"
• "Generate a quiz from my research"
• "Make flashcards for studying"
• "Generate a video explainer"
• "Build a presentation / slide deck"
• "Make an infographic"
• "Create a mind map"
• "Add these sources to NotebookLM"
• "Research [topic] and import sources"
• "Analyze these documents"

---

Autonomy Rules

Run without confirmation:

• notebooklm status, list, doctor — read-only
• notebooklm auth check — diagnostics
• notebooklm create — create notebook
• notebooklm use <id> — set context (⚠️ single-agent only; use -n in parallel)
• notebooklm source add — add sources
• notebooklm ask "..." — chat (without --save-as-note)
• notebooklm language list/get/set — language settings
• notebooklm profile list/create/switch — profile management

Ask user first:

• notebooklm delete — destructive
• notebooklm generate * — long-running, may hit rate limits
• notebooklm download * — writes to filesystem
• notebooklm ask "..." --save-as-note — writes a note
• notebooklm history --save — writes a note

---

Quick Reference

Task	Command
Log in	notebooklm login
List notebooks	notebooklm list
Create notebook	notebooklm create "Title"
Set active notebook	notebooklm use <notebook_id>
Show context	notebooklm status
Add URL	notebooklm source add "https://..."
Add file	notebooklm source add ./file.pdf
Add YouTube	notebooklm source add "https://youtube.com/..."
Add Drive doc	notebooklm source add-drive <file_id> "Title"
List sources	notebooklm source list
Web research (fast)	notebooklm source add-research "query"
Web research (deep)	notebooklm source add-research "query" --mode deep --no-wait
Ask a question	notebooklm ask "question"
Ask with references	notebooklm ask "question" --json
Save answer as note	notebooklm ask "question" --save-as-note
Chat history	notebooklm history
Generate podcast	notebooklm generate audio "instructions"
Generate video	notebooklm generate video
Generate cinematic video	notebooklm generate cinematic-video "description"
Generate slide deck	notebooklm generate slide-deck
Generate quiz	notebooklm generate quiz
Generate flashcards	notebooklm generate flashcards
Generate infographic	notebooklm generate infographic
Generate mind map	notebooklm generate mind-map
Generate data table	notebooklm generate data-table "description"
Generate report	notebooklm generate report --format study-guide
Check artifact status	notebooklm artifact list
Wait for artifact	notebooklm artifact wait <artifact_id>
Download podcast	notebooklm download audio ./podcast.mp3
Download video	notebooklm download video ./video.mp4
Download slides (PDF)	notebooklm download slide-deck ./slides.pdf
Download slides (PPTX)	notebooklm download slide-deck ./slides.pptx --format pptx
Download quiz	notebooklm download quiz quiz.json
Set language	notebooklm language set en

---

Sources

Supported Types

URLs, YouTube videos, PDFs, text/Markdown/Word files, Google Drive docs, audio, video, images, pasted text.

notebooklm source add "https://example.com/article"
notebooklm source add ./report.pdf
notebooklm source add "https://www.youtube.com/watch?v=VIDEO_ID"
notebooklm source add-drive <drive_file_id> "Doc Title"

notebooklm source list
notebooklm source fulltext <source_id>     # Get indexed text content
notebooklm source guide <source_id>        # Get source guide
notebooklm source wait <source_id>         # Wait for indexing
notebooklm source delete <source_id>
notebooklm source delete-by-title "Exact Title"

Web Research

# Fast (5–10 sources, seconds)
notebooklm source add-research "AI trends 2025"

# Deep — blocking (20+ sources, waits up to 5 min)
notebooklm source add-research "AI safety" --mode deep --import-all

# Deep — non-blocking (recommended for agents)
notebooklm source add-research "AI safety" --mode deep --no-wait
notebooklm research status               # Check progress
notebooklm research wait --import-all    # Wait and auto-import

# Search Google Drive
notebooklm source add-research "Project Report" --from drive

Source limits by plan: Standard 50 · Plus 100 · Pro 300 · Ultra 600

---

Chat & Q&A

notebooklm ask "What are the main themes?"

# Use specific sources only
notebooklm ask "Compare these" -s src_id1 -s src_id2

# Get answer with source citations
notebooklm ask "What is X?" --json

# Save answer as note
notebooklm ask "Summarize" --save-as-note
notebooklm ask "Summarize" --save-as-note --note-title "Research Summary"

# Continue a specific conversation
notebooklm ask "Tell me more" -c <conversation_id>

notebooklm history                         # View history
notebooklm history --show-all             # Full Q&A content
notebooklm history --save                 # Save as note
notebooklm history --save --note-title "Session Log"

---

Content Generation

🎙️ Podcast (Audio Overview)

# Formats: deep-dive | brief | critique | debate
# Lengths:  short | default | long
notebooklm generate audio "Focus on practical applications"
notebooklm generate audio "Compare two viewpoints" --format debate --length long
notebooklm generate audio -s src1 -s src2          # Specific sources
notebooklm generate audio --language en --json     # JSON output → task_id

notebooklm download audio ./podcast.mp3

🎬 Video

# Formats: explainer | brief
# Styles:  auto | classic | whiteboard | kawaii | anime | watercolor | retro-print | heritage | paper-craft
notebooklm generate video "Explain for beginners" --style whiteboard
notebooklm generate cinematic-video "Documentary-style summary"

notebooklm download video ./video.mp4
notebooklm download cinematic-video ./documentary.mp4

📊 Slide Deck

# Formats: detailed | presenter
# Lengths: default | short
notebooklm generate slide-deck --format detailed

notebooklm download slide-deck ./slides.pdf
notebooklm download slide-deck ./slides.pptx --format pptx   # ← not in web UI

# Revise individual slide (zero-indexed)
notebooklm generate revise-slide "Make title more impactful" --artifact <id> --slide 0 --wait
notebooklm generate revise-slide "Remove the table" --artifact <id> --slide 3 --wait

❓ Quiz

# Difficulty: easy | medium | hard
# Quantity:   fewer | standard | more
notebooklm generate quiz --difficulty hard --quantity more

notebooklm download quiz quiz.json                           # JSON ← not in web UI
notebooklm download quiz --format markdown quiz.md
notebooklm download quiz --format html quiz.html

🃏 Flashcards

notebooklm generate flashcards --difficulty medium --quantity more

notebooklm download flashcards cards.json
notebooklm download flashcards --format markdown cards.md

🗺️ Mind Map

notebooklm generate mind-map    # Synchronous — instant, no --wait needed

notebooklm download mind-map ./mindmap.json    # ← not in web UI

📈 Infographic

# Orientation: landscape | portrait | square
# Detail:      concise | standard | detailed
# Style:       auto | sketch-note | professional | bento-grid | editorial |
#              instructional | bricks | clay | anime | kawaii | scientific
notebooklm generate infographic --orientation portrait --style professional --detail detailed

notebooklm download infographic ./infographic.png

📋 Data Table

notebooklm generate data-table "Compare the key concepts"

notebooklm download data-table ./data.csv

📄 Report

# Formats: briefing-doc | study-guide | blog-post | custom
notebooklm generate report --format study-guide
notebooklm generate report --format briefing-doc --append "Keep under 2 pages"
notebooklm generate report "Technical white paper on AI trends"    # → auto custom
notebooklm generate report --format study-guide -s src1 -s src2   # Specific sources

notebooklm download report ./report.md

---

Generation Types Summary

Type	Command	Options	Download
Podcast	generate audio	--format, --length	.mp3
Video	generate video	--format, --style	.mp4
Cinematic Video	generate cinematic-video	same as video	.mp4
Slide Deck	generate slide-deck	--format, --length	.pdf / .pptx
Quiz	generate quiz	--difficulty, --quantity	.json/.md/.html
Flashcards	generate flashcards	--difficulty, --quantity	.json/.md/.html
Infographic	generate infographic	--orientation, --detail, --style	.png
Report	generate report	--format, --append	.md
Mind Map	generate mind-map	(sync, instant)	.json
Data Table	generate data-table	description required	.csv

All generate commands also support: -s/--source (specific sources), --language, --json, --retry N

---

Artifact Management

notebooklm artifact list
notebooklm artifact list --type audio       # Filter by type
notebooklm artifact get <artifact_id>
notebooklm artifact rename <artifact_id> "New Title"
notebooklm artifact delete <artifact_id>
notebooklm artifact wait <artifact_id> --timeout 1200
notebooklm artifact export <artifact_id> --type docs    # → Google Docs
notebooklm artifact export <artifact_id> --type sheets  # → Google Sheets
notebooklm artifact suggestions

# Batch downloads (not in web UI)
notebooklm download audio --all
notebooklm download infographic --all
notebooklm download video --latest
notebooklm download audio --all --dry-run   # Preview only

---

Notes

notebooklm note list
notebooklm note create "My research notes..."
notebooklm note get <note_id>
notebooklm note save <note_id> --title "New Title" --content "Updated content"
notebooklm note rename <note_id> "New Title"
notebooklm note delete <note_id>

---

Sharing

notebooklm share status
notebooklm share public --enable
notebooklm share public --disable
notebooklm share view-level full     # Chat + sources + notes
notebooklm share view-level chat     # Chat only
notebooklm share add [e-mail]
notebooklm share add [e-mail] --permission editor
notebooklm share update [e-mail] --permission editor
notebooklm share remove [e-mail]

---

Language Settings

notebooklm language list              # List all 80+ supported languages
notebooklm language get               # Show current setting
notebooklm language set en            # Set globally

# Per-command override
notebooklm generate audio --language de
notebooklm generate video --language fr

Important: Language is a GLOBAL setting affecting all notebooks in your account.

Common codes: en English · de German · fr French · es Spanish · it Italian · pt_BR Portuguese · ja Japanese · ko Korean · zh_Hans Simplified Chinese · ar Arabic

---

Features Not in the Web UI

Feature	Command
Batch downloads	download <type> --all
Quiz/Flashcard as JSON or Markdown	download quiz --format json
Mind map JSON export	download mind-map
Slide deck as PPTX	download slide-deck --format pptx
Individual slide revision	generate revise-slide "..." --artifact <id> --slide N
Report template append	generate report --format study-guide --append "..."
Source full-text access	source fulltext <id>
Save chat to note	ask "..." --save-as-note
Programmatic sharing	share commands

---

Common Workflows

Research → Podcast

notebooklm create "Research: AI"
notebooklm use <id>
notebooklm source add "https://en.wikipedia.org/wiki/Artificial_intelligence"
notebooklm source add-research "AI trends 2025" --mode deep --import-all
notebooklm generate audio "Focus on breakthroughs and future outlook" --format debate --wait
notebooklm download audio ./ai-podcast.mp3

Documents → Study Materials

notebooklm create "Exam Prep"
notebooklm use <id>
notebooklm source add ./lecture-notes.pdf
notebooklm source add ./textbook-chapter.pdf
notebooklm generate quiz --difficulty hard --wait
notebooklm generate flashcards --wait
notebooklm generate report --format study-guide --wait
notebooklm ask "What are the most important topics?"

YouTube → Notes + Report

notebooklm create "Video Notes"
notebooklm use <id>
notebooklm source add "https://www.youtube.com/watch?v=VIDEO_ID"
notebooklm ask "What are the main points?"
notebooklm generate report --format briefing-doc --wait
notebooklm download report ./notes.md

Full Presentation Workflow

notebooklm create "Presentation: [Topic]"
notebooklm use <id>
notebooklm source add "https://source1.com"
notebooklm source add ./data.pdf
notebooklm generate slide-deck --format detailed --wait
notebooklm download slide-deck ./slides.pptx --format pptx
notebooklm generate revise-slide "Stronger opening title" --artifact <id> --slide 0 --wait

Agent Automation (Non-blocking)

NB_ID=$(notebooklm create "Bulk Research" --json | python3 -c "import sys,json; print(json.load(sys.stdin)['id'])")
notebooklm use $NB_ID
SRC=$(notebooklm source add "https://url.com" --json | python3 -c "import sys,json; print(json.load(sys.stdin)['source_id'])")
notebooklm source wait $SRC --timeout 120
TASK=$(notebooklm generate audio "Overview" --json | python3 -c "import sys,json; print(json.load(sys.stdin)['task_id'])")
notebooklm artifact wait $TASK --timeout 1200
notebooklm download audio ./podcast.mp3

---

JSON Output Schemas

// notebooklm list --json
{"notebooks": [{"id": "...", "title": "...", "created_at": "..."}]}

// notebooklm source add "..." --json
{"source_id": "...", "title": "...", "status": "processing"}

// notebooklm generate audio "..." --json
{"task_id": "...", "status": "pending"}

// notebooklm ask "..." --json
{"answer": "...[1]", "conversation_id": "...", "turn_number": 1,
 "references": [{"source_id": "...", "citation_number": 1, "cited_text": "..."}]}

// notebooklm source list --json
{"sources": [{"id": "...", "title": "...", "status": "ready|processing|error"}]}

// notebooklm artifact list --json
{"artifacts": [{"id": "...", "title": "...", "type": "...", "status": "completed|in_progress|pending|unknown"}]}

---

Error Handling

Error	Cause	Fix
Auth/cookie error	Session expired	notebooklm auth check → notebooklm login
"No notebook context"	Context not set	notebooklm use <id> or -n <id> flag
"No result found for RPC ID"	Rate limiting	Wait 5–10 min, retry
GENERATION_FAILED	Google rate limit	Wait and retry; use web UI as fallback
Download fails	Generation incomplete	Check notebooklm artifact list
Invalid ID	Wrong ID	notebooklm list to find correct ID
RPC protocol error	Google changed API	Update notebooklm-py

On failure, offer the user: retry / skip / investigate.

Exit Codes

Code	Meaning
0	Success
1	Error (not found, failed)
2	Timeout (wait commands only)

---

Processing Times & Rate Limits

Operation	Typical Time	Timeout
Source indexing	30s – 10 min	600s
Research (fast)	30s – 2 min	180s
Research (deep)	15 – 30+ min	1800s
Mind map	instant	—
Quiz, flashcards	5 – 15 min	900s
Report, data table	5 – 15 min	900s
Audio generation	10 – 20 min	1200s
Video generation	15 – 45 min	2700s

Always reliable: Notebook/source management, chat, mind map, report, data table. May hit rate limits: Audio, video, quiz, flashcards, infographic, slide deck — retry after 5–10 min.

For agents, avoid --wait on long operations. Use artifact wait <id> in a background task.

---

Parallel Agent Safety

notebooklm use <id> writes to a shared file — unsafe in parallel workflows.

Solutions:

1. Pass -n <notebook_id> directly to each command (recommended)
2. export NOTEBOOKLM_PROFILE=agent-$ID — per-agent profile isolation
3. export NOTEBOOKLM_HOME=/tmp/agent-$ID — full isolation
4. Always use full UUIDs, not partial IDs

---

Troubleshooting

notebooklm --help
notebooklm auth check --test   # Full auth + network validation
notebooklm doctor --fix        # Auto-fix environment
notebooklm login               # Re-authenticate
notebooklm --version           # Check version

Docs: https://github.com/teng-lin/notebooklm-py/blob/main/docs/troubleshooting.md