NotebookLM MCP Management

Tech Stack Target / Version: NotebookLM current web release, Markdown session capture, and URL-tracked research workflows.

Use this skill when research should be grounded in NotebookLM notebooks instead of a generic web search.

• Leverage native parallel subagent dispatch and 200k+ context windows where available.

Current MCP Reality

This repository already targets a real NotebookLM MCP workflow. The concrete tool surface available in this environment includes:

• get_health
• list_notebooks, search_notebooks, select_notebook
• ask_question
• add_notebook, update_notebook, remove_notebook
• list_sessions, reset_session, close_session
• setup_auth, re_auth, cleanup_data

Activation Conditions

Use symptom -> action triggers: when one matches, apply this skill and verify with the protocol below.

• Querying a specific NotebookLM notebook
• Adding a notebook from a share URL
• Managing a notebook library or switching active notebooks
• Recovering authentication or cleaning NotebookLM state
• Continuing a multi-turn research session

Recommended Workflow

1. Call get_health first to confirm authentication and server readiness.
2. Reuse an existing session when the task is the same.
3. Prefer search_notebooks or list_notebooks before asking the user to restate what is already in the library.
4. Use ask_question iteratively in the same session for deep work.
5. Use setup_auth or re_auth only when health indicates auth problems.

Library Management Rules

• Do not add or remove notebooks without explicit user confirmation.
• When adding a notebook, collect URL, description, topics, and use cases first.
• Update metadata instead of creating duplicates when the notebook already exists.

Troubleshooting

• Auth broken: get_health -> re_auth
• Stale browser state: cleanup_data(preserve_library=true) after closing browsers
• Wrong context: reset_session or switch notebooks
• Ambiguous notebook choice: search the library before creating a new one

Anti-Patterns

• Treating source content as already clean: Formatting automation will happily preserve broken or inconsistent input.
• Skipping an open-file verification pass: Documents and spreadsheets often fail in the destination app, not in the script output.
• Automating irreversible edits without checkpoints: A small mapping mistake can affect an entire workbook or document.

Verification Protocol

Before claiming "skill applied successfully":

1. Pass/fail: The Notebooklm Management workflow names the agent boundary, delegated scope, and expected return artifact.
2. Pass/fail: Context passed to helpers is minimal, task-local, and free of hidden expected answers.
3. Pass/fail: Results are integrated only after evidence, diffs, or citations are checked by the controller.
4. Pressure-test scenario: Run the workflow on two similar tasks that must not share assumptions or leaked context.
5. Success metric: Zero context leakage; every delegated output is independently reviewable.

References & Resources

Documentation

• MCP Tool Reference - Current NotebookLM MCP operations and parameters
• Troubleshooting Guide - Auth recovery, cleanup, and session issues
• Workflows - Library, query, and maintenance workflows

Scripts

• NotebookLM Helper - Local helper for library exports and reporting when MCP access is unavailable
• Scripts README - Quick commands for the helper script

Examples

• simple-query.py - Basic query pattern
• multi-turn-conversation.py - Session reuse pattern
• library-management.py - Library search and organization

<!-- MCP:START --> <!-- PORTABILITY:START -->

Cross-Client Portability

This skill is written to stay usable across GitHub Copilot, Claude Code, and Codex.

• GitHub Copilot: keep the folder in a Copilot-visible skill path or wrap the workflow in project instructions when folder discovery is unavailable.
• Claude Code: keep the folder in a local skills directory or a compatible plugin source.
• Codex: install or sync the folder into $CODEX_HOME/skills/notebooklm-management and restart Codex after major changes.

<!-- PORTABILITY:END -->

MCP Availability And Fallback

Preferred MCP Server: NotebookLM MCP

• Fallback prompt: "Use the NotebookLM MCP Management skill without MCP. Rely on the local SKILL.md, bundled references or scripts, and manual verification. Show the exact commands, evidence, and final checks you used before concluding."
• Use the NotebookLM web UI directly, capture answers in Markdown, and store session notes locally when the MCP server is unavailable.
• Preserve notebook URLs, prompt history, and manual research notes so the workflow remains reproducible.

<!-- MCP:END -->

Related Skills

• documentation-authoring: Use it when the workflow also needs drafting structured technical or product documents.
• notion-docs: Use it when the workflow also needs Notion page and database publishing workflows.
• pdf: Use it when the workflow also needs PDF extraction, generation, and layout-aware review.
• word-document: Use it when the workflow also needs Word document authoring and formatting workflows.