Agile style standup bot for hubot ala tender.

Create a room (or channel on IRC) for standup (existing one is okay) and invite hubot to the room if necessary.

First, you tell hubot who is a member for a particular team for the standup, using the roles commands with "(who) is a (team) member".

Let's take "engineering" team for example.

miyagawa: hubot miyagawa is an engineering member
hubot: Ok, miyagawa is an engineering member
miyagawa: hubot john is an engineering member
hubot: Ok, john is an engineering member
miyagawa: hubot davidlee is an engineering member
hubot: Ok, davidlee is an engineering member

You can create as many teams as you want.

Note: This setup process hopefully won't be necessary in the future, when Hubot properly implements an API to get the list of online users in a room. So that everyone in a standup room == standup participants.

Hubot won't schedule the standup for you (yet), you have to start it by yourself when it's time.

miyagawa: hubot standup for engineering
hubot: Ok, let's start the standup: miyagawa, john, davidlee
hubot: john: your turn

Hubot remembers who should participate the standup, and will tell whose turn is the next. Tell what you did yesterday, will do today, anything blocked. and say "next" (or "done" or "that's it") when you're done.

john: Done some pretty nice hack yesterday.
john: I will work on another cool stuff today.
john: I'm not blocked
john: hubot next
hubot: davidlee: your turn

When the user is offline or away for a second, tell hubot to skip the user.

miyagawa: hubot skip davidlee
hubot: Will skip davidlee
hubot: miyagawa: your turn

Once the last user is done, hubot will tell you how long the standup was.

miyagawa: I'm working on some nice stuff, and will continue doing so today.
miyagawa: hubot next
hubot: All done! Standup was 5 minutes and 24 seconds.

This hubot script depends on roles.coffee script. You're recommended to use redis-brain.coffee to persist the team information.

In your package.json, add the following line to the dependencies:

   "hubot-standup": "git://github.com/miyagawa/hubot-standup.git"

Then run npm install and create a symbolic link from scripts directory:

ln -s ../node_modules/hubot-standup/src/scripts/standup.coffee scripts/

Add the symlink file to the git repository if necessary (for Heroku deployment).

By symlinking to standup-yammer.coffee in addition, the bot will post the standup archive to Yammer. You need to set a valid Yammer OAuth2 token to HUBOT_STANDUP_YAMMER_TOKEN environment variable.

Here's how to get a valid Yammer OAuth2 token with the standard OAuth2 authorization flow.

See Yammer documentation for more details.

• Register a new application on Yammer at https://www.yammer.com/<DOMAIN>/client_applications/new. Leave the callback URLs empty
• Take notes of your consumer_key and consumer_secret
• Make a new bot user on Yammer (optional). This is the user who will post archives as.
• Sign in as the new bot user on Yammer if necessary
• Go to https://www.yammer.com/dialog/oauth?client_id=<consumer_key>
• There's an authorization dialog. Authorize the app
• Look at the URL bar and there's a code=<CODE> query parameter in there, copy that.
• curl https://www.yammer.com/oauth2/access_token?code=<CODE>&client_id=<consumer_key>&client_secret=<consumer_secret>
• you'll get a big JSON that contains access_token -> token

Now set the token to HUBOT_STANDUP_YAMMER_TOKEN and Hubot will ask which group ID the log should be posted to. Use the group ID 0 to turn off the feature for a group.

Tatsuhiko Miyagawa

MIT License