By surfing the forums I found that a lot of users have problems validating their ssml. Constant issues are misplaced tags, unclosed tags, missing brackets or extra spaces inside of the tags.

Because error message says only, that input was invalid users don't know what was invalid and they keep asking the same thing - to explain them what is wrong with their ssml, which most of the time is one of the problems I mentioned before.

My suggestion would be to add a better validation for ssml, including 'Voice & Tone' tab. It should give better messages to at least help users to pinpoint what could be wrong, because current message is not really helpful.

Not sure if this is a bug or is by design. If I have the output text of <speak> 2 is > 3 </speak> this works and Alexa will read 2 is greater than 3. However if I have the output text of <speak> 2 is < 3 </speak>, where I would expect 2 is less than 3, I get an error "The input was incompatible" and Alexa will instead respond There was a problem with the requested skills response.

This appears to be an issue because it sees the < as the beginning of a closing tag for </speak>. Is there anyway around this to support this "<" character? My use case is I get back numbers from a DB that could be #, ># or <#. All cases of <# fail due to this.

I would like to propose a change to the speek API. It should be possible to specify the output language of the text to testing is easier. If you specify short answers like "no" (or "nein" in german) or output english domain names, the automatic test declares them as "english" or any other language which does not meet the skills requirements as everything has to be in the correct language. This is very annoying, because you have to change phrases or example invocations to include german sounding words in a domain name for example to pass the automatic test before submition.

When a user invokes a Skill on a Fire TV Cube, provide a way for our fulfillment webhook to determine that the device is a Fire TV Cube. We know we can determine that the skill was invoked on a device that has a screen, but we need a way to determine that it is not an Echo Show or Spot, but instead a TV connected to a FireTV Cube.

Then, provide a way to determine whether or not the TV power is on or off. If it is off, provide a way for us to tell the Cube to turn on the power to the TV so that video will display.

I've created a smart home skill for a device that needs some time between each actuation.

I know in advance if a device is ready or not to be controlled again.

I would like to provide a nicer feedback to the user by replying to control request with an error indicating that the device need some time to execute a new command in the cases where my device cannot be controlled when the request arrives.

This command should be translated in something like:
The device X is temporarily busy please try again later.

The ASK SDK already provide few errors that could potentially cover this usecase: https://developer.amazon.com/docs/device-apis/alexa-errorresponse.html#error-type-values

• ENDPOINT_BUSY


• RATELIMITEXCEEDED

Both sounds like potential good candidates.

However all the error I've tried are translated in something like: "The device X is not responding"

This gives the user the idea that there is something wrong whit his device instead of attempting to explain what is the actual reason why he cannot control the device.