Nephthys is the bot powering #summer-of-making-help and #identity-help in the Hack Club Slack! Below is a guide to set her up for developing and here's a list of some of her features :)
You can tag tickets in the private tickets channel or with the macro ?tag <tag_name>. This will DM the people who are specialised in responding to those issues and have it show up in their assigned tickets.
You can assign yourself to get notified for specific tags on the app home
Sometimes it’s nice to be able to do things quickly... Here’s where macros come in! Send one of the following messages in an open thread and something will happen
?resolve- the ticket gets closed. Equivalent of hitting i get it now?identity- redirect to #identity-help?faq- redirect to the FAQ?hii- silly message :3?shipcertqueue- tell them to wait and vote because there's a backlog of ships?fraud- redirect to fraud dept?thread- remove the reaction and all Nephthys replies to unclutter duplicates- more to come?? feel free to PR your own into hackclub/nephthys or tell me what you want
Tickets that have been not had a response for more than 3 days will automatically be closed as stale. The last helper to respond in the thread gets credit for closing them
Stale ticket handling is not working at the moment, but more features for dealing with stale tickets are planned.
At midnight UK time each day, you get to see the stats for the day in the team channel! Helpers can also see more detailed stats at any time on the app home for the bot!
When you send a message in a help thread, that thread is assigned to you and it is up to you to resolve it. You can see a list of threads waiting for you on the app home - just select the Assigned Tickets tab at the top
- Python (3.13 or later)
- uv
- A Slack workspace where you have permissions to install apps
- Tunneling tool (for local development)
- A PostgreSQL database (you can run one in Docker using the command below)
docker run --name hh-postgres -e POSTGRES_PASSWORD=postgres -p 5432:5432 -d postgres- Go to https://api.slack.com/apps and click "Create New App".
- Choose "From an app manifest" and select your workspace.
- Copy and paste the manifest in
manifest.yml. - Review and create the app.
- In the "Basic Information" section, note down the
App Id,Client Id,Client Secret,Signing Secret. - Go to "OAuth & Permissions" and install the app to your workspace. Note down the "Bot User OAuth Token".
-
Clone the repository:
git clone https://github.com/hackclub/nephthys cd nephthys -
Install dependencies:
uv sync source .venv/bin/activate # for bash/zsh source .venv/bin/activate.fish # for fish source .venv/bin/activate.csh # for csh source .venv/bin/activate.ps1 # for powershell -
Copy the
.env.samplefile to.env:cp .env.sample .env -
Edit the
.envfile and fill in the values:
-
Start your tunneling tool and expose the local server. (Not needed in socket mode with
SLACK_APP_TOKENset)Note the HTTPS URL you get.
-
Update your Slack app's request URLs:
- Go to your Slack app's settings.
- In "Event Subscriptions" and "Interactivity & Shortcuts", update the request URL to your HTTPS URL followed by
/slack/events. - In "OAuth & Permissions", update
Redirect URLsto your HTTPS URL followed by/slack/oauth_redirect.
-
MAKE SURE YOU CHANGE THE COMMAND - DO NOT USE THE SAME COMMAND
-
Install pre-commit hooks:
uv run pre-commit install -
Start your database and update the database schema:
uv run prisma db push uv run prisma generate -
Start the application:
nephthys
Your Slack app should now be running and connected to your Slack workspace!
Contributions are welcome! Please feel free to submit a Pull Request.
The codebase contains some scripts in the nephthys/scripts/ directory to help with development and testing. They are documented below.
add_dummy_data.py is a script that adds a bunch of dummy (i.e. fake) support ticket records to the database, for stress-testing/performance testing.
Usage: uv run nephthys/scripts/add_dummy_data.py <num_records>
- Ensure you run it after the
nephthyshas been run at least once (and once the DB has been initialized) - It takes a while to run (adding 20k records takes ~50 seconds on my machine)
- Don't run this in production, obviously
This project is licensed under the MIT License.