Main search function with auto-domain detection

Wrap long text into multiple lines.

Save library to disk

Add random delay

Extract results list from search result dict.

Close this session and clean up resources

Format size in human-readable form

Check if valid authentication exists

Launch a persistent browser context with anti-detection features and cookie workaround.

Get size of file or directory in bytes

Save browser state to disk

Core search function using BM25

Wait for NotebookLM page to be ready

Get instructions for manual activation

Clear all authentication data Returns: True if cleared successfully

Ensure virtual environment exists and is set up

Format results for Claude consumption (token-optimized)

Main entry point for design system generation. Args: query: Search query (e.g., "SaaS dashboard", "e-commerce luxury") proje

Get the currently active notebook

Get paths that would be cleaned up Args: preserve_library: Keep library.json if True Returns: Dict

Get the correct Python executable to use

Get user by ID.

Type with human-like speed

List all notebooks in the library

Print a preview of what will be cleaned

Perform interactive authentication setup Args: headless: Run browser in headless mode (False for login) time

Lowercase, split, remove punctuation, filter short words

Basic validation of a skill

Apply reasoning rules to search results.

Detect page type from context and search results.

Find matching reasoning rule for a category.

Generate intelligent overrides based on page type using layered search. Uses the existing search infrastructure to find relevant style,

Initialize the browser session and navigate to NotebookLM

Inject cookies from state.json if available

Load CSV and return list of dicts

Load library from disk

Load reasoning rules from CSV.

Execute searches across multiple domains.

Save authentication metadata

Select best matching result based on priority keywords.

Get the current latest response text

Wait for and extract the new answer

Add a new notebook to the library Args: url: NotebookLM notebook URL name: Display name for the notebook

Ask a question to NotebookLM Args: question: Question to ask notebook_url: NotebookLM notebook URL headless: Run bro

Auto-detect the most relevant domain from query

Ensure virtual environment exists

Ensure virtual environment exists and run the requested script. This is called when any script is imported or run directly.

Build BM25 index from documents

Format design system as ASCII box with emojis (MCP-style).

Format design system as markdown.

Format design system as MASTER.md with hierarchical override logic.

Format a page-specific override file with intelligent AI-generated content.

Generate complete design system recommendation.

Get authentication information

Get a specific notebook by ID

Get library statistics

Get the virtual environment Python executable

Initialize a new skill directory with template SKILL.md. Args: skill_name: Name of the skill path: Path where the skill dire

Check if we're already running in the skill's venv

Main runner

Command-line interface for authentication management

Main entry point for environment setup

Command-line interface for notebook management

Command-line interface for cleanup management

Package a skill folder into a .skill file. Args: skill_path: Path to the skill folder output_dir: Optional output directory

Perform the actual cleanup Args: preserve_library: Keep library.json if True dry_run: Preview only, don't de

Persist design system to design-system/<project>/ folder using Master + Overrides pattern. Args: design_system: The generated de

Perform re-authentication (clear and setup) Args: headless: Run browser in headless mode timeout_minutes: Lo

Click with realistic movement

Remove a notebook from the library Args: notebook_id: ID of notebook to remove Returns: True if rem

Run a script with the virtual environment

Score all documents against query

Search notebooks by query Args: query: Search query (searches name, description, topics, tags) Returns:

Search stack-specific guidelines

Set a notebook as active Args: notebook_id: ID of notebook to activate Returns: The activated noteb

Convert hyphenated skill name to Title Case for display.

Validate that stored authentication works Uses persistent context to match actual usage pattern Returns: True if

Initialize a new browser session Args: session_id: Unique identifier for this session context: Browser conte

Initialize the authentication manager

Initialize the notebook library

Initialize the cleanup manager

Ask a question in this session Args: question: The question to ask Returns: Dict with status, quest

Create a new user.

Delete user.