spot-bot

Spot Bot is the relentless accountant of the surreptitious photos you take of each other.

• spot, spotted: Spot people by mentioning them in a message with the keyword spot or spotted and a picture of your spot.
• spotboard: Show how many times each channel member has spotted someone else. spotboard 20 will show the top 20 spotters.
• caughtboard: Show how many times each channel member has been spotted. caughtboard 20 will show the top 20 who've been caught.
• pics: View all pics of a person by tagging them in a message with the pics keyword.
• referendum: Reply referendum to a spot to start a 24-hour vote to determine if the spot will count or not.
• reset: Reset the spot record in a channel.

Tips

• Look out for ✅. If you don't see ✅, your spot didn't count!
• If you didn't spot something correctly, you can edit your message within one minute of sending it to correct the error.
• If you delete a message, the spot will no longer count!
• Spot Bot will treat each channel of the workspace separately so different teams can maintain their own spot boards.

Install

Click here to install. (I know this looks really janky, but this is actually more secure than the native slack flow which does not protect against CSRF attacks.)

By installing Spot Bot, you acknowledge that you have read and agree to our terms of use and privacy policy.

Technical details

Developed and tested on Python 3.8.10 in Ubuntu. WSL was used to develop in Linux on a Windows PC. Spot Bot was proudly union made by members of UAW 4811.

You will need to set up a virtual environment to test locally and then install the requirements in requirements.txt.

You will also need to set up the following confidential environment variables in a .env file: SPOTBOT_CLIENT_ID, SPOTBOT_CLIENT_SECRET, SPOTBOT_SECURE_LINK, SPOTBOT_SIGNING_SECRET. SPOTBOT_SECURE_LINK refers to a link for the Mongo endpoint, i.e. mongodb+srv://... Additionally, set up the WEB_CONCURRENCY variable to set the number of workers that gunicorn will use. We use WEB_CONCURRENCY=4.

Once the virtual environment is set up, the following command will start it.

source venv/bin/activate

In order to develop locally, you will need to have a Slack App configured, and also have a way for your local server to be accessed by Slack. We used ngrok during development. If you use ngrok, be sure that the port you're using matches up with the port that the server is being hosted on. More information can be found here: https://tools.slack.dev/node-slack-sdk/tutorials/local-development/. This is particularly annoying because you will have to input your bespoke ngrok public URL into the configuration settings for your Slack app each time.

You will also need to configure the MongoDB database to accept traffic from your IP address, or else you will not be able to access the database.

Once all that is configured, you can initialize the environment variables from your .env file with

. initenv.sh

And start the local server with

bash run.sh

To broadcast a message to a particular workplace, use the command with the team ID and a message.

flask broadcast T124TEAMID "Insert your message here"

To broadcast a message to every workplace, use this command. Do this sparingly.

flask systemwide_broadcast "Insert your message here"

To print a list of all installed workplaces that have since uninstalled the app, run

flask determine_inactive_teams

Prior to March 2025, the main collection in the Mongo instance did not store the Team ID and Channel ID for each instance of Spot Bot in a channel. The following command visits each workspace, identifies all the channels that Spot Bot is in, and attempts to rectify this issue by adding the Team ID and Channel ID to the given instance. This command likely can help you clean out the Mongo instance; channels that are not accessible (i.e. after you run this command they still lack a channel/team ID) likely can have their data deleted.

flask rectify_ids