Public | Automated Build

Last pushed: 2 years ago
Short Description
Our Slack Hubot setup
Full Description


Hubot running on Alpine with the Slack Adapter.

Build Info


  • NodeJS: 4.12
  • NPM: 2.24.9

Default Scripts

Installed Packages

  • bash
  • supervisor
  • nodejs
  • redis
  • build-base
  • gc
  • g++
  • gcc-objc
  • libtool
  • libc6-compat
  • make
  • expat
  • expat-dev
  • python
  • wget
  • gnupg
  • tar
  • git
  • zip
  • curl
  • wget

Suggested Mounts

Mount the redis directory to avoid data reset on container replacement

  • /var/lib/redis

Mount the config directory to manage credentials/settings outside of container

  • /opt/hubot/config

Mount the scripts directory to manage any non-npm installs/simple scripts

  • /opt/hubot/scripts

Mount the external-scripts for control

  • /opt/hubot/external-scripts.json


You have a few options in how to utilize this container

Basic Start

docker run -v -e HUBOT_SLACK_TOKEN=TOKEN \
  -d hearstat/hubot-slack

Configuration File Start

docker run -v /path/to/hubot.conf:/opt/hubot/config/hubot.conf \
  -d hearstat/hubot-slack

Full Feature Start

docker run -v /path/to/hubot.conf:/opt/hubot/config/hubot.conf \
  -v /path/to/redis/save:/var/lib/redis \
  -v /path/to/external-scripts.json:/opt/hubot/external-scripts.json \
  -d hearstat/hubot-slack

Dev Mode Start

I've setup this bot to be able to switch back and forth from Slack to Shell via startup commands. This is how to get into "Dev Mode" or enable the shell adapter.

docker run -d hearstat/hubot-slack /usr/bin/devmode

then just do the following to connect to the container at the shell level

docker exec -it $container_name bash

then once in run the following


Then you can interact with hubot at the shell level



docker run -d -p 8123:8123 -p 5858:5858 \
  --name=devbot hearstat/hubot-slack devmode
docker exec -it devbot bash
coffee --nodejs --debug $(which hubot)

In Chrome navigatate to http://<docker IP>:8123/?ws=<docker IP>:8123&port=5858
For Mac users "docker IP" can be found in Kitematic.

NOTE: You should end up in a hubot REPL, if you end up in a coffee REPL you did something wrong. To exit:

coffee> process.exit()

NOTE: node-inspector currently only works in Chrome.

To set a breakpoint in coffeescript you will want to open the code in the /opt/hubot/node_modules/ directory and add at the appropriate line:


You may want to mount a directory locally so you can use your local editor. For Mac this will need to be in your /Users/<username> dir To do this consider:

mkdir -p ~/node_modules
docker run -d -p 8123:8123 -p 5858:5858 \
  -v /Users/<username>/node_modules:/opt/hubot/node_modules \
  -v /Users/ablythe/hubot/config:/opt/hubot/config \
  --name=devbot hearstat/hubot-slack devmode
docker exec -it devbot bash
# needed the first time to populate the mounted volume
# needed each time configuration changes
source ./config/hubot.conf
coffee --nodejs --debug $(which hubot)

Prod Mode Start

To switch back to slack or "Prod Mode" do the following

docker exec $container_name /usr/bin/prodmode

or just run a new container

Run Time Help

Since this container comes with a bot reload option, edit the external-scripts.json as needed and run the following

docker exec $container_name python

Note: you can just restart the container, it will re-run the same script before loading the bot.

Then in chat tell hubot to reload (my default is thebot)

@hubot reload


To build the image, do the following

docker build

A prebuilt container is available in the docker index.

docker pull hearstat/hubot-slack

Template files


Lives at /opt/hubot/config and is sourced at run time.

Add all environment variables needed to conf file. See script repos for specific settings available.

The baseline config file in container only has ADAPTER/HUBOT_NAME set.

## Bot Settings
export HUBOT_NAME='hubot' # what hubot listens to

## Slack adapter settings

# Credentials


The embedded utilizes the external-scripts.json for it's install items, I did this to simplify the process. You already have to add everything to the file regardless, so use it to install from.

Docker Pull Command
Source Repository