In 2020, ASK CLI version 2 has been published by Alexa team. We can manage our skill backend by using AWS CloudFormation.

Let’s get start!

目次

Create a new Project

We can create a new project from ask new. It’s same as v1.

Step1: Select language

% ask new
Please follow the wizard to start your Alexa skill project ->
? Choose the programming language you will use to code your skill:  (Use arrow keys)
❯ NodeJS 
  Python 
  Java

Step2: Select backend resources

This step is the most important update of v2. I think.

We can select backend resource type from Hosted / AWS CloudFormation / AWS Lambda / self-hosted. Version 1 can select only AWS Lamda.

? Choose a method to host your skill's backend resources:  (Use arrow keys)
❯ Alexa-hosted skills
  Host your skill code by Alexa (free). 
  AWS with CloudFormation
  Host your skill code with AWS services and provision with AWS CloudFormation (requires AWS account) 
  AWS Lambda
  Host your skill code on AWS Lambda (requires AWS account). 
  ──────────────
  self-hosted and manage your own hosting

Step3: Select template

Now we can select three type of example. Probably these example number will be increased.

? Choose a template to start with:  (Use arrow keys)
❯ Hello world           Alexa's hello world skill to send the greetings to the world! 
  Fact skill            Quick sample skill to manage a collection of voice data. 
  High low game         Guess a number high or low using persistent attributes. Make sure your AWS account has the permission to manage DynamoDB!

Step4: Put project information

Finally, we can put the skill name and project directory name.

? Please type in your skill name:  example-skill
? Please type in your folder name for the skill project (alphanumeric):  exampleSkill

We can get the following files in the project directory.

% tree 
.
├── LICENSE.txt
├── ask-resources.json
├── infrastructure
│   └── cfn-deployer
│       └── skill-stack.yaml
├── lambda
│   ├── index.js
│   ├── package.json
│   └── util.js
└── skill-package
    ├── assets
    │   ├── en-US_largeIcon.png
    │   └── en-US_smallIcon.png
    ├── interactionModels
    │   └── custom
    │       └── en-US.json
    └── skill.json

7 directories, 10 files

How can we customize the project?

The project seems to be a separate directory per feature.

ask-resources.json: Skill configuration (CloudFormation props / Code URI etc…)
infrastructure: backend resources definition.
( In this case, placed CloudFormation stack template)
lambda: Lambda source code
skill-package: Skill configurations. ( Interaction Model / assets / publication information / etc..)

We can go to these directory and edit these files for own skill style.

Deploy it

Deployment workflow is same as v1. Just run ask deploy command.

Then, the CLI will deploy entire resources by SMAPI and AWS CloudFormation.

How can we check the created resources from AWS CloudFormation

We can check the deployed skill information from .ask/ask-states.json.

So we can check the skill’s CloudFormation by the following command.

% aws cloudformation describe-stack-resources --stack-name $(cat .ask/ask-states.json | jq ".profiles.default.skillInfrastructure[\"@ask-cli/cfn-deployer\"].deployState.default.stackId" -r) | jq .
{
  "StackResources": [
    {
      "StackName": "example-skill-default-skillStack-1587380830650",
      "StackId": "arn:aws:cloudformation:us-east-1:99999999:stack/example-skill-default-skillStack-1587380830650/1e4f54f0-82f7-11ea-86a4-0eb40de15aba",
      "LogicalResourceId": "AlexaSkillFunction",
      "PhysicalResourceId": "example-skill-default-skillSt-AlexaSkillFunction-19SAEJO5SZMBP",
      "ResourceType": "AWS::Lambda::Function",
      "Timestamp": "2020-04-20T11:07:47.639Z",
      "ResourceStatus": "CREATE_COMPLETE",
      "DriftInformation": {
        "StackResourceDriftStatus": "NOT_CHECKED"
      }
    },
    {
      "StackName": "example-skill-default-skillStack-1587380830650",
      "StackId": "arn:aws:cloudformation:us-east-1:99999999:stack/example-skill-default-skillStack-1587380830650/1e4f54f0-82f7-11ea-86a4-0eb40de15aba",
      "LogicalResourceId": "AlexaSkillFunctionEventPermission",
      "PhysicalResourceId": "example-skill-default-skillStack-1587380830650-AlexaSkillFunctionEventPermission-14T7X7Q1K8HFM",
      "ResourceType": "AWS::Lambda::Permission",
      "Timestamp": "2020-04-20T11:08:00.883Z",
      "ResourceStatus": "CREATE_COMPLETE",
      "DriftInformation": {
        "StackResourceDriftStatus": "NOT_CHECKED"
      }
    },
    {
      "StackName": "example-skill-default-skillStack-1587380830650",
      "StackId": "arn:aws:cloudformation:us-east-1:99999999:stack/example-skill-default-skillStack-1587380830650/1e4f54f0-82f7-11ea-86a4-0eb40de15aba",
      "LogicalResourceId": "AlexaSkillIAMRole",
      "PhysicalResourceId": "example-skill-default-skillSta-AlexaSkillIAMRole-YFXEA2L5SOA4",
      "ResourceType": "AWS::IAM::Role",
      "Timestamp": "2020-04-20T11:07:44.055Z",
      "ResourceStatus": "CREATE_COMPLETE",
      "DriftInformation": {
        "StackResourceDriftStatus": "NOT_CHECKED"
      }
    }
  ]
}

Customize AWS CloudFormation

We can customize our Alexa Skill backend more easily. Because we can use AWS CloudFormation (and AWS SAM) !

This example code is add S3:GetObject action to your S3 buckets.

AWSTemplateFormatVersion: 2010-09-09
Parameters:
  SkillId:
    Type: String
  LambdaRuntime:
    Type: String
  LambdaHandler:
    Type: String
  CodeBucket:
    Type: String
  CodeKey:
    Type: String
  CodeVersion:
    Type: String
Resources:
  AlexaSkillIAMRole:
      Type: AWS::IAM::Role
      Properties:
        AssumeRolePolicyDocument:
          Version: 2012-10-17
          Statement:
            - Effect: Allow
              Principal:
                Service:
                  - lambda.amazonaws.com
              Action:
                - sts:AssumeRole
        Path: /
        Policies:
          - PolicyName: alexaExternalPolicy
            PolicyDocument:
              Version: 2012-10-17
              Statement:
                - Effect: Allow
                  Action:
                    - s3:GetObject
                  Resource: arn:aws:s3:::*
          - PolicyName: alexaSkillExecutionPolicy
            PolicyDocument:
              Version: 2012-10-17
              Statement:
                - Effect: Allow
                  Action:
                    - logs:*
                  Resource: arn:aws:logs:*:*:*
  AlexaSkillFunction:
    Type: AWS::Lambda::Function
    Properties:
      Code:
        S3Bucket: !Ref CodeBucket
        S3Key: !Ref CodeKey
        S3ObjectVersion: !Ref CodeVersion
      Handler: !Ref LambdaHandler
      Runtime: !Ref LambdaRuntime
      Role: !GetAtt AlexaSkillIAMRole.Arn
      MemorySize: 512
      Timeout: 60
  AlexaSkillFunctionEventPermission:
    Type: AWS::Lambda::Permission
    Properties:
      Action: lambda:invokeFunction
      FunctionName: !GetAtt AlexaSkillFunction.Arn
      Principal: alexa-appkit.amazon.com
      EventSourceToken: !Ref SkillId
Outputs:
  SkillEndpoint:
    Description: LambdaARN for the regional endpoint
    Value: !GetAtt AlexaSkillFunction.Arn

Conclusion

By using ASK CLI (version2), we can maintain more scalable Alexa skill using various AWS resources more easily.