Works with the AI Agents you already use
Grounded Notebook Answers
Turn Google NotebookLM into a source-bound research layer your agent queries from the terminal, with persistent login, a notebook library, and cited answers.
Free
Grounded Notebook Answers
Example session with this skill installed
I keep a NotebookLM notebook loaded with our vendor contracts and product API docs (URL: https://notebooklm.google.com/notebook/ab12cd34ef56). Register it in my local notebook library using smart discovery so the title, description, and tags come from the notebook itself, not from me. Then pull every rule about data retention, deletion requests, and audit access that appears in the contract sources, and cross-check them against the compliance doc — I need document names and passage citations for each rule, and you must tell me explicitly where the sources are silent or contradict each other. Don't answer from memory.
- Read your context and instructions
- Compiled the grounded notebook answers
Summary
Registered the notebook as "Vendor Contracts & Product API" (3 sources detected) and ran two grounded passes. Four rules confirmed with citations; retention/deletion conflict flagged; audit access is not defined in the sources.
| Finding | Document | Evidence | Confidence |
|---|---|---|---|
| 7-year contract retention | Master Services Agreement | §8.2 "records retained seven years from termination" | High |
| Deletion within 30 days of request | Data Processing Addendum | §4.1 "without undue delay, no later than 30 days" | High |
| Backup purge after 90 days | Data Processing Addendum | §4.3 "backups purged within ninety days" | Medium (no section heading in PDF) |
| Retention vs. deletion conflict | MSA §8.2 vs DPA §4.1 | 7-year retention vs 30-day deletion, no precedence clause | Gap flagged |
| Audit access rights | — | Sources silent; no auditor clause found | Not answerable |
Follow-up pass recovered the DPA §4.3 excerpt the first answer had omitted.
Next steps
- Set this notebook as default with
notebooks set-default - Resolve the MSA/DPA precedence conflict with legal before relying on either rule
- Upload the Security Addendum so audit access becomes answerable
Connects securely to your tools. The creator never sees your data.
About this skill
LLMs often hallucinate details from their training data when you need them to stick to your specific internal documents. This skill transforms Google NotebookLM into a source-bound research layer accessible directly from your terminal. It allows your agent to query your existing notebooks, retrieve answers grounded strictly in your uploaded documents, and maintain a local library of source materials.
What it does
- Grounded Q&A performs queries against NotebookLM notebooks and retrieves answers based solely on provided sources.
- Citation Extraction captures document names, section labels, and direct excerpts to back up every factual claim.
- Library Management maintains a local catalog of notebooks with IDs, titles, and tags for quick reference.
- Auth Persistence manages Google session cookies and local storage to minimize manual login interruptions.
- Automated Discovery crawls new notebook URLs to automatically generate summaries and tags for your library.
How it works
- Use the
auth statuscommand to verify your Google session andauth loginif a fresh browser profile is needed. - Add notebooks to your local registry using
notebooks addor by providing a URL for automatic discovery. - Submit questions using the
askcommand, targeting specific notebook IDs or URLs. - Review the grounded response and trigger follow-up queries if citations or details are missing.
Frameworks & tools
Built with Python 3.10+, Playwright for browser automation, and Google NotebookLM.
Why this beats prompting it yourself
Manual copy-pasting from NotebookLM to an LLM loses the active citation links and is prone to context window limits. This skill automates the browser interaction, ensuring that every answer is programmatically fetched from the grounded Gemini instance inside NotebookLM while maintaining a persistent, searchable library of your research sources.
Use cases
- Query complex legal contracts or technical specs for specific clauses with citations.
- Search across multiple documentation notebooks to find which source contains specific logic.
- Verify internal policy compliance by extracting rules directly from uploaded handbooks.
- Debug automation failures by running queries in a visible browser mode to inspect UI selectors.
Known limitations
Requires manual document uploads through the NotebookLM web UI. Browser automation introduces latency, and free Google accounts may be subject to rate limits.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse All Claude Skills.
It is browser automation, not an official API — Google UI changes, consent walls, or aggressive re-auth can break flows until fixed. Documents must already be uploaded to NotebookLM, and each query costs a few seconds plus follow-up passes for complete extraction.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
1 use
Times used to date
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 1 month ago
- Free with an Agensi account
- Accounts
Google Account
- Software
Python
Permissions
Allowed Hosts
File Scopes
Browser and network access are needed to drive your live NotebookLM session through a persistent Google login, while file writes stay limited to a local notebook catalog and browser profile state under the skill directory.