"Building skill schema failed" error while deploying alexa skill using ask-cli

Question

I have a alexa skill live on amazon skill store. The interaction model of that skill was developed using the skill builder of alexa development portal and the lambda function was uploaded seperately. The skill is live and has many users.

Recently I felt the need of making some changes to that skill. Since now I am using ask-cli for all my alexa skill development, I decided to use the same for this skill too.

Here are the steps I followed:

• cloned the skill from the alexa console using command ask clone and selecting the skill I wanted to make changes in.

• made the required changes in the skill. Just changed a response, nothing special.

• tried deploying the skill again to alexa console using ask deploy command

But I got the following error:

[Error]: Building skill schema failed

I thought that the changes I made might be the cause of it, so I backtracked the changes and made the code as it was when cloned. Then I tried deployig the code again but the error remained the same.

I went through some answers on amazon portal, the most common cause of the issue for others there was some mistake in URI of skill icon URIs. To be sure, I also removed that line from skill.json and tried deploying it again but nothing changed.

Does anyone has any idea of the cause of the above mentioned issue?

Answer 1

If the cause of the issue is your image urls in Skill.json file try this hacky solution

• Goto your Alexa skill development portal and select "distribution" tab

• In that upload both small and large icon. ( Do not press save ) if its a new skill,

• right click on the image and select open image in new tab and copy the image URL. now use this URL in the skill.json

This is what i used in my skill when s3 url are all causing issue.

Answer 2

Make sure your lambda function has Alexa Skills Kit as trigger. I solved my problem by adding back the trigger (and entering the skill ID).

Answer 3

This steps might able to fix that problem.

1. Create a new skill in the developer portal.
2. Create the Lambda function for that skill then add it as an endpoint.
3. Download the skill via ask-cli
4. Copy the code in the existing skill to the new skill.
5. Re deploy using ask-cli.

It might be due to permission error or the appropriate profile used during deployment is configured incorrectly or missing.