Invocation Name

This is a lowercase name that is easy for Alexa to pick up as a user pronounces it. See recommendations from the Alexa invocation name guidelines.

Sheet ID

You'll be adding the ID sandwiched between "d/" and "/edit#gid=0" in your Google Spreadsheet URL to reference in the sheetId attribute in our Makeskill config.json file:

VUI Sheet

This is the name of the sheet you'll be parsing from that's designated only to VUI batches.

Multimodal Cards

This is the sheet responsible for card asset URLs for either Echo Show or Echo Spot. Pair the card batch columns accordingly with VUI sheet columns. Card assets are mapped to the VUI responses via intents. If you want to turn off multimodality, use "" as the assignment.

Custom Entities

If you're making your own slots, be sure to define the custom entities in a separate sheet and link it in the config. Makeskill will automatically render the new interaction model with your canned values.

Default Batch Letter

By default, the config file has batch A exporting for you, which is indicated by "A". Batches B, C, etc would be denoted by "B", "C", etc up to Z, respectively.

Lambda and Models Folders

In the folders of your choosing, select where the Lambda function and Interaction Model files will be generated. If using an ASK CLI template, the file paths should be ../lambda/custom and ../models respectively.

Google Secret

The googleSecret attribute references a client_secret.json in order to permit Makeskill to talk to your Google Spreadsheets.

1. Click this wizard link in step A in the Node.js Quickstart.

2. Create a project and Continue.

3. The API should be enabled. Go to credentials.

4. To add credentials to your project, select "Other UI (e.g. Windows, CLI tool)" to set where the API will be called from.

5. Select “User data” then “What credentials do I need?”

6. Use “Makeskill” as the client ID. Create client ID.

7. Now you can download credentials for your project.

   Populate your client_secret.json file with something like:

   {
      "installed":{
         "client_id":"[UNIQUE_ID].apps.googleusercontent.com",
         "project_id":"[PROJECT_ID]",
         "auth_uri":"https://accounts.google.com/o/oauth2/auth",
         "token_uri":"https://accounts.google.com/o/oauth2/token",
         "auth_provider_x509_cert_url":"https://www.googleapis.com/oauth2/v1/certs",
         "client_secret":"[SECRET]",
         "redirect_uris":[
            "urn:ietf:wg:oauth:2.0:oob",
            "http://localhost"
         ]
      }
   }
   

8. Run makeskill --config [config location] with the Spreadsheet scaffold.

9. In your terminal after running makeskill, you’ll get a prompt to visit a URL to authorize access. Copy and paste that URL into your browser to obtain a code.

10. Allow for access through the browser.

11. Copy and paste the code into your terminal where you ran makeskill.

12. Now, you should be set to make your own skill batches!

Step 1: Fill out the Google Sheets

For an individual 1-turn Skill, simply fill out your spreadsheet to mimic the following layout for sheet name "VUI". Replace with your Skill's Intents, include utterance variations, and ensure the corresponding responses line up with those intents.

Multimodality is optional in Makeskill. If enabled, you'll need a spreadsheet that looks similar to this for sheet name "Cards". See Early Support for Multimodality section for more details.

Step 2: Export the Skill

When you are finished editing your Google Spreadsheet, you'll need to run the makeskill command from the Makeskill root folder.

Export the default batch (based on config file):

makeskill --config [config location]

Step 3: Deploy

Now you're ready to deploy. Switch to the project root folder to interact with ASK CLI.

Deploy your prototype (from project root folder):

cd ..
ask deploy

You can now test locally via your device or preferred simulator.

Step 1: Prepare Your Batches

The philosophy behind Makeskill is the ability to design in bulk in order to pump out multiple Skill experiments. In this sample for sheet name "VUI", a user can export either batch A, B, or C to acquire a corresponding Skill to deploy. When designing multiple batches, the recommendation is to design 3 batch variations at a time with 1 batch being a control. If needed, Makeskill can account for batches A-Z. Currently, a developer can export one of the variations at a time.

Here, we see a sheet name "Cards" that contains public links to card assets. Each card asset column represents a batch.

Step 2: Export Your Batch

Let's say you wanted to export Batch C:

makeskill --config [config location] --batch C

Step 3: Deploy

Switch to the project root folder. Then you can deploy away.

Deploy your prototype (from project root folder):

cd ..
ask deploy

You can now test locally via your device or preferred simulator.