📝 CellMage Magic Command Cheatsheet

This cheatsheet provides a comprehensive reference for all CellMage magic commands and their arguments.

Core Magic Commands

1. %llm_config - Configuration Magic

%llm_config [options]

Argument	Description
-p, --persona NAME	Select and activate a persona by name
--show-persona	Show the currently active persona details
--list-personas	List available persona names
--list-mappings	List current model name mappings
--add-mapping ALIAS FULL_NAME	Add a model name mapping (e.g., --add-mapping g4 gpt-4)
--remove-mapping ALIAS	Remove a model name mapping
--set-override KEY VALUE	Set a temporary LLM param override (e.g., --set-override temperature 0.5)
--remove-override KEY	Remove a specific override key
--clear-overrides	Clear all temporary LLM param overrides
--show-overrides	Show the currently active overrides
--clear-history	Clear the current chat history (keeps system prompt)
--show-history	Display the current message history
--save [FILENAME]	Save session. If no name, uses current session ID. ‘.md’ added automatically
--load SESSION_ID	Load session from specified identifier (filename without .md)
--list-sessions	List saved session identifiers
--auto-save	Enable automatic saving of conversations to the conversations directory
--no-auto-save	Disable automatic saving of conversations
--list-snippets	List available snippet names
--snippet NAME	Add user snippet content before sending prompt (can be used multiple times)
--sys-snippet NAME	Add system snippet content before sending prompt (can be used multiple times)
--status	Show current status (persona, overrides, history length)
--model NAME	Set the default model for the LLM client
--adapter {direct,langchain}	Switch to a different LLM adapter implementation

2. %%llm - LLM Cell Magic

%%llm [options]
Your prompt to the LLM goes here...

Argument	Description
-p, --persona NAME	Use specific persona for THIS call only
-m, --model NAME	Use specific model for THIS call only
-t, --temperature VALUE	Set temperature for THIS call
--max-tokens NUMBER	Set max_tokens for THIS call
--no-history	Do not add this exchange to history
--no-stream	Do not stream output (wait for full response)
--no-rollback	Disable auto-rollback check for this cell run
--param KEY VALUE	Set any other LLM param ad-hoc (e.g., --param top_p 0.9). Can be used multiple times
--list-snippets	List available snippet names
--snippet NAME	Add user snippet content before sending prompt (can be used multiple times)
--sys-snippet NAME	Add system snippet content before sending prompt (can be used multiple times)

3. %llm_config_persistent - Ambient Mode Magic

%llm_config_persistent [options]

Accepts the same arguments as %llm_config but also enables ambient mode (all regular cells are processed as LLM prompts).

4. %disable_llm_config_persistent - Disable Ambient Mode

%disable_llm_config_persistent

Disables ambient mode, returning the notebook to normal operation.

5. %%py - Execute Python in Ambient Mode

%%py
# Python code to run
print("Hello World")

When ambient mode is active, this magic command lets you execute a specific cell as Python code.

Integration Magic Commands

1. %confluence - Confluence Wiki Integration

%confluence [identifier] [options]

Argument	Description
identifier	Page identifier (SPACE:Title format or page ID)
--cql QUERY	Confluence Query Language (CQL) search query
--max NUMBER	Maximum number of results for CQL queries (default: 5)
--system	Add content as a system message (rather than user)
--show	Display the content without adding to history
--text	Use plain text format instead of Markdown (Markdown is default)
--no-content	For CQL search, return only metadata without page content
--content	For CQL search, fetch full content of each page (makes additional API calls)

2. %jira - Jira Integration

%jira [issue_key] [options]

Argument	Description
issue_key	Jira issue key (e.g., PROJECT-123)
--jql QUERY	JQL search query instead of a specific issue
--max NUMBER	Maximum number of results for JQL search (default: 5)
--fields FIELDS	Comma-separated list of fields to include
--comments	Include issue comments
--system	Add as system message instead of user message
--show	Display content without adding to history
--text	Use plain text instead of Markdown format

3. %github - GitHub Integration

%github [repo/path] [options]

Argument	Description
repo_path	Repository owner/name or file path within a repository
--issue NUMBER	Fetch a specific issue by number
--pr NUMBER	Fetch a specific pull request by number
--search QUERY	Search for issues with the given query
--max NUMBER	Maximum number of search results (default: 5)
--comments	Include comments on issues/PRs
--system	Add as system message instead of user message
--show	Display content without adding to history
--branch NAME	Branch to use when fetching files (default: main/master)

4. %gitlab - GitLab Integration

%gitlab [repo/path] [options]

Argument	Description
repo_path	Project path or file path within a project
--issue NUMBER	Fetch a specific issue by number
--mr NUMBER	Fetch a specific merge request by number
--search QUERY	Search for issues with the given query
--max NUMBER	Maximum number of search results (default: 5)
--comments	Include comments on issues/MRs
--system	Add as system message instead of user message
--show	Display content without adding to history
--branch NAME	Branch to use when fetching files (default: main/master)

5. %webcontent - Web Content Integration

%webcontent [url] [options]

Argument	Description
url	URL of the webpage to fetch
--clean	Clean and extract main content (default behavior)
--raw	Get raw HTML content without cleaning
--method METHOD	Content extraction method: trafilatura (default), bs4, or simple
--system	Add as system message instead of user message
--show	Display content without adding to history
--include-images	Include image references in the output
--no-links	Remove hyperlinks from the output
--timeout SECONDS	Request timeout in seconds (default: 30)

6. %gdocs - Google Docs Integration

%gdocs [document_id] [options]

Argument	Description
document_id	Google Document ID
--system	Add as system message instead of user message
--show	Display content without adding to history
--search QUERY	Search for Google Docs files containing the specified term
--content	Retrieve and display content for search results
--max-results NUMBER	Maximum number of search results to return (default: 10)
--max-content NUMBER	Maximum number of documents to retrieve content for (default: 3)
--timeout SECONDS	Request timeout in seconds (default: 300)
--author EMAIL	Filter documents by author/owner email (comma-separated for multiple)
--modified-after, --updated DATE	Filter by modification date (YYYY-MM-DD or natural language)
--order-by FIELD	How to order search results (relevance, modifiedTime, createdTime, name)
--auth-type TYPE	Authentication type to use (oauth or service_account)

7. %sqlite - SQLite Storage Management

%sqlite [options]

Argument	Description
--status	Show the current state of the SQLite storage
--stats	Display statistics about stored conversations
--list	List all stored conversations
--new	Start a new conversation
--load ID	Load a specific conversation by ID
--delete ID	Delete a conversation by ID
--tag ID TAG	Add a tag to a conversation
--search QUERY	Search for conversations with content matching the query
--export PATH	Export a conversation to markdown file
--import-md PATH	Import a conversation from markdown file

8. %img - Image Processing and Integration

%img image_path [options]

Argument	Description
image_path	Path to the image file to process
-r, --resize WIDTH	Width to resize the image to (maintains aspect ratio)
-q, --quality VALUE	Quality for lossy image formats (0.0-1.0)
--show	Display the image inline after processing
-i, --info	Display information about the image
-a, --add-to-chat	Add the image to the chat session (default: always added)
-c, --convert	Force conversion to a compatible format
-f, --format FORMAT	Format to convert the image to (e.g., “jpg”, “png”, “webp”)

Example Usage

Core Commands

# Load the extension
%load_ext cellmage

# Set the default model and persona
%llm_config --model gpt-4o --persona python_expert

# Ask a question with a specific model for this call only
%%llm -m gpt-4o
Explain the visitor pattern in software design

# Enable ambient mode with specific settings
%llm_config_persistent --model gpt-4o-mini --temperature 0.7

# Execute Python code in ambient mode
%%py
import pandas as pd
print(pd.__version__)

# Disable ambient mode
%disable_llm_config_persistent

Integration Commands

# Load Confluence extension
%load_ext cellmage

# Fetch a Confluence page
%confluence TEAM:Project Overview

# Search Confluence
%confluence --cql "space = DEV AND title ~ 'Architecture'"

# Load GitHub extension
%load_ext cellmage

# Get a specific GitHub issue
%github myorg/myrepo --issue 42

# Get web content
%load_ext cellmage
%webcontent https://example.com/article

# Process and display an image
%load_ext cellmage
%img path/to/image.jpg --resize 1024 --show --info