@bender2k14 so right now the process is as follows:

There is an internal export of plugin functionality that allows using the plugin code within its documentation (with some considerations, since the plugin and the documentation have two separate sets of dependencies). This is used on the docs to verify and document new functionality.

So for each new functionality (a component, post-processor, etc):

• It is first coded, the code is tested via the docs. This means testing the feature and documenting it almost happen simultaneously
• Then I typically do a commit titled "bump version" (example), in which merely the version of the package is increased.
• Then I run npm publish.

To be honest features of the package haven't had many surface consistency issues or similar things to raise the alarm for automated testing (unlike the CLI, for example). It is mostly seeing if something "looks good" since they mostly don't have much complex functionality (this might not be the case in the future). Still, some automated testing ensuring some basic functionality perhaps wouldn't hurt. Besides that, I guess the main thing that can be automated is publishing when there is a bump version commit, so I am unsure how much could this particular library benefit from automated publishing process (since it is already pretty simplistic).