@OceanK @richland007

Just a heads up –

I was running into the same issues as you both. After hours of research I eventually discovered that the AWS IoT security policy for your certificates must be properly configured…otherwise you’ll be hit with authorization issues when clients attempt to connect to the IoT thing for subscribing, publishing, connections, etc. Once i corrected this, I had no further issues with the installation.

The change I made to my policy in ‘AWS IoT>Secure>Policies’ is below:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "iot:Publish",
        "iot:Receive"
      ],
      "Resource": [
        "arn:aws:iot:us-east-1:543211234567:topic/sdk/test/java",
        "arn:aws:iot:us-east-1:543211234567:topic/sdk/test/Python",
        "arn:aws:iot:us-east-1:543211234567:topic/topic_1",
        "arn:aws:iot:us-east-1:543211234567:topic/topic_2",
        "arn:aws:iot:us-east-1:543211234567:topic/MagicMirror:new-text",
        "arn:aws:iot:us-east-1:543211234567:topic/MagicMirror:new-images",
        "arn:aws:iot:us-east-1:543211234567:topic/MagicMirror:change-module",
        "arn:aws:iot:us-east-1:543211234567:topic/MagicMirror:new-video"
      ]
    },
    {
      "Effect": "Allow",
      "Action": [
        "iot:Subscribe"
      ],
      "Resource": [
        "arn:aws:iot:us-east-1:543211234567:topicfilter/sdk/test/java",
        "arn:aws:iot:us-east-1:543211234567:topicfilter/sdk/test/Python",
        "arn:aws:iot:us-east-1:543211234567:topicfilter/topic_1",
        "arn:aws:iot:us-east-1:543211234567:topicfilter/topic_2",
        "arn:aws:iot:us-east-1:543211234567:topicfilter/MagicMirror:new-text",
        "arn:aws:iot:us-east-1:543211234567:topicfilter/MagicMirror:new-images",
        "arn:aws:iot:us-east-1:543211234567:topicfilter/MagicMirror:change-module",
        "arn:aws:iot:us-east-1:543211234567:topicfilter/MagicMirror:new-video"
      ]
    },
    {
      "Effect": "Allow",
      "Action": [
        "iot:Connect"
      ],
      "Resource": [
        "arn:aws:iot:us-east-1:543211234567:client/sdk-java",
        "arn:aws:iot:us-east-1:543211234567:client/basicPubSub",
        "arn:aws:iot:us-east-1:543211234567:client/sdk-nodejs-*",
        "arn:aws:iot:us-east-1:543211234567:client/MagicMirror*"
      ]
    }
  ]
}

Make sure your client id defined in MirrorMirror.js matches the wildcard (“MagicMirror” in this case) that was added to iot:Connect action of the policy. This needs to be modified locally in the module and also in the uploaded lambda function!

// Setup our AWS IoT device and receive messages
app.setup = function() {
  app.device = awsIot.device({
    keyPath: __dirname + "/certs/MagicMirror.private.key",
    certPath: __dirname + "/certs/MagicMirror.cert.pem",
    caPath: __dirname + "/certs/root-CA.crt",
    clientId: "MagicMirror" + (new Date().getTime()),
    region: "us-east-1",
    host: "YOURID.iot.us-east-1.amazonaws.com",
  });

Best of luck and happy new year!

@radu_stancu @OceanK posted his json config yesterday. Just need to go back a page or two in this thread.

You’ll still need to understand what’s configured, in my opinion. Otherwise you’ll run into issue troubleshooting or extending the model. It’s pretty straight forward.

@radu_stancu AMAZON.LITERAL has been deprecated.

Check this out: https://developer.amazon.com/docs/custom-skills/literal-slot-type-reference.html

@oceank you have to understand what you’ve configured in your interaction model.

{
“name”: “TurnOffModuleIntent”,
“slots”: [
{
“name”: “moduleName”,
“type”: “LIST_OF_MODULES”
}
],
“samples”: [
“close {moduleName}”,
"close module {moduleName} ",
“hide {moduleName}”,
"hide module {moduleName} ",
“turn off {moduleName}”,
“turn off module {moduleName}”
]
},

Check and understand your interaction model and intents. Does the type “LIST_OF_MODULES” defined in your interaction model include the module that you are trying to turn off?

@OceanK @Radu_Stancu “turn off magic mirror”

@oceank magic mirror… turn off {whatever you named your mm module}. Check out ModuleNames.json in your local mm module. There’s a mapping of variables to modules.

@oceank Cool beans. I think this was referenced in the installation doc passively. But you have to use your invocation command again and then “turn off {whatever you module is named}”

@richland007 Yep. Ran into that one. The interaction model from the project is outdated. You’ll have to refactor it a bit.

AMAZON.LITERAL was deprecated and you’ll have to create a Custom literal (pretty easy to do from the UI).

Check this out: https://developer.amazon.com/docs/custom-skills/literal-slot-type-reference.html

Regarding the app id, you’re sure you copied the skill value and added it to the APP_ID variable before uploading it to Lambda? Also, did you disable skill verification in Lambda (this was mentioned in the install doc)?

@oceank does it compile? what happens when you test? That model can vary, mine has been customized a bit.

pro-tip:

A way to check and see if you’re actually subscribing to the topics that are publishing search results:

if you run the following command from a remote putty session…

cd ~/MagicMirror
Display=:0 npm start

you should eventually see that you’re subscribing to the topics!

Initially, the misconfigured security policy was preventing this from happening and authorization errors were being thrown in the background (had to dig into cloudwatch logs for this). In putty, the command would simply hang at “Attempt to connect to AWS”

Basically your youtube api search results will be published to the “new-video” queue…and if your aren’t subscribed to that queue then you wont see anything change on your monitor. Same for your google image searches and “new-images”. and so on.

@richland007 so you changed the clientId in the lambda function and uploaded it. Did you change the client id in the module as well? Can you post your security policy? You should also define your queues (modules) there.

The app id is defined in index.js of the skill. APP_ID is defined in a variable near the top of the file.

Looks like this currently:

alexa.appId = APP_ID

Also, you can verify what your skill id is by checking in the amazon developer console then selecting ‘endpoint’ (it’s an option on the left pane under interaction model and interfaces).