jupyterlab_galaxahub_motd_extension

Shows the GalaxaHub message of the day in a JupyterLab tab. When the lab starts, the extension pulls the welcome content that the user's groups carry and the broadcasts sent to the user. It opens a Message of the day tab when the hub has something to show, and stays silent when it has not.

Features

• Welcome tab on start - opens the Message of the day tab and makes it current when the hub carries at least one welcome entry or one notification for the user
• Silent without a hub - opens nothing and logs one console line when a hub URL setting is empty, or the hub has no motd extension, cannot be reached, or has nothing to show
• Two columns - the entry cards on the left, the Notifications column 380 px wide on the right; a tab narrower than 800 px stacks them
• One card per group - markdown rendered by the lab's own markdown renderer, an HTML package shown in a sandboxed iframe at its hub address, sized to fit its page
• Notifications catch-up - the broadcasts sent to all users and those naming the user, newest first, each marked by an icon in its type's colour, with the relative time and an all or direct marker
• Palette command - Message of the day: Open reopens the one tab and pulls again
• Token stays on the server - a server extension calls the hub with the lab's API token and passes Etags through, so the browser never holds the token and a repeated pull answers 304

Requirements

• JupyterLab >= 4.6
• A GalaxaHub hub carrying galaxahub-motd-extension; without it the extension shows nothing
• The two hub URLs in the lab's Jupyter config, as Server configuration states; without them the extension shows nothing

Install

pip install jupyterlab_galaxahub_motd_extension

Settings

Server configuration

The lab's Jupyter config names the two hub URLs the extension reads. Both settings are empty by default, and while either one is empty the extension shows nothing.

• Files - jupyter_server_config.py or jupyter_lab_config.py (or their .json form) in a Jupyter config directory; jupyter --paths lists the directories
• Not read - the jupyter_server_config.d/ directories, which carry only extension enabling
• CLI - jupyterlab-galaxahub-motd asks the running lab server for both values, so a --config file on the lab's command line applies to it too
• Token - the extension sends the lab's JUPYTERHUB_API_TOKEN to both URLs, so both must belong to the hub that spawned the lab

For a GalaxaHub lab, where JUPYTERHUB_API_URL is the hub API of the lab:

import os

hub = os.environ.get("JUPYTERHUB_API_URL", "")
if hub:
    c.GalaxaHubMotd.motd_api_url = f"{hub}/extensions/motd"
    c.GalaxaHubMotd.notifications_api_url = f"{hub}/user-notifications"

Server routes

The server extension proxies three hub reads under its own prefix and adds nothing to them. It reads the two URLs from Server configuration and JUPYTERHUB_API_TOKEN from the lab's environment.

• The hub's status, body, Content-Type, Etag and Cache-Control pass through, and the browser's If-None-Match goes to the hub; a matching Etag answers 304 with no body
• One answer for an absent hub - a hub 404 (no motd extension), a hub that cannot be reached, and an empty URL setting all answer 204 No Content with Cache-Control: no-cache; the server log states which of the three it was
• Any other hub status, 403 included, passes through unchanged
• Settings - GET /jupyterlab-galaxahub-motd-extension/settings answers motd_api_url and notifications_api_url as the running server holds them; the CLI reads it
• Hub side - docs/design-api.md states what a hub must answer on the three routes, for a developer who writes the hub side without GalaxaHub

Command line

The package installs jupyterlab-galaxahub-motd, which prints the same message of the day in a lab terminal. It asks the running lab server for the two URLs, so whatever configured the lab applies, and makes the three hub reads itself with JUPYTERHUB_API_TOKEN. It finds the lab server through JUPYTER_SERVER_URL, which JupyterLab sets in every terminal.

jupyterlab-galaxahub-motd show                   # terminal text, rich entries, notifications
jupyterlab-galaxahub-motd notifications --json   # one JSON document
jupyterlab-galaxahub-motd terminal 2>/dev/null   # in a lab startup script

A lab startup script prints the motd with the last line and holds no motd URL and no token. With no motd, or a hub that fails, stdout stays empty.

jupyterlab-galaxahub-motd --help lists the commands, the configuration, the environment variables and the exit codes, and each command's --help carries examples.

The repository carries an agent skill, .agents/skills/jupyterlab-galaxahub-motd-extension/SKILL.md. pip install puts a copy in share/jupyter/agents/skills/ under the Python environment. No agent reads that directory, so a lab image links it into the agent skills directory, with the Python that runs the lab:

mkdir -p ~/.agents/skills
ln -s "$(python -c 'import sys; print(sys.prefix)')/share/jupyter/agents/skills/jupyterlab-galaxahub-motd-extension" ~/.agents/skills/jupyterlab-galaxahub-motd-extension

Agents that read .agents/skills also find it in a clone of this repository; to make it available to Claude Code everywhere, link it into the skills directory from the clone:

ln -s "$PWD/.agents/skills/jupyterlab-galaxahub-motd-extension" ~/.claude/skills/jupyterlab-galaxahub-motd-extension

Uninstall

pip uninstall jupyterlab_galaxahub_motd_extension