Danger
Training exercise.

## What made you smile?

* Code walkthroughs seems to be powerful feature, but at the same time it requires a lot of effort. We would love to see some AI assitant that would generate chunks inside files as writing them by hand requires significant effort.


## What did you find confusing?

* It wasn't possible to upload multiple files into editor with some subfolders inside (we've tried to upload whole unpacked archive folder). As a result we've had to manually upload all files for each subfolder.
* Uploading `assets` didn't show files in file tree: ![empty-assets.png](/assets/empty-assets.5de418ffc5d7b5090f3bece3d42941bda3b41dbd9866faaef2f865d7e436bd19.43a731ee.png)
* It wasn't possible to `pin` given tab in editor
* Preview feature doesn't work with images. (Known issue)
* We've had to create a new file for prompt to have preview working
* Documentation for `Preview slot` is missing some screenshots
* We were struggling with inputs/variables:
  * Documentation is missing some more meaningful examples
  * We were struggling on how to put input between `"` character. After we've wrapped it with `"` editor started to complain about syntax in our js files
* Webview was scrolling to the area below visible page when cursor was placed on code walkthroughs
* When opening webview in new page (when scrolled to the first walkthrogh) the content is not scrolled properly
* We didn't know how to specify chunks with different languages that have different comments syntax
* When we've created guideline for edit operation variables where not working properly for one person in webview, while it was working fine for other one.
* Size of this training was overwhelming - setting variables and toggles require a lot of effort
* When we just wanted to highlight file content, we didn't have to provide chunks in file content for node, but for Python it wasn't working. We should support cases where users just link to source code file, but don't want to edit it by adding some `chunks` inside.
* It seems we have misunderstood the task - instead of creating one code walkthrough tag we used many code-walkthrough tags
* We've skipped step to highlight some code block from `Operating on image data` section becaue we didn't know how to do it. Instead we've highlighted whole file.