After a lot of tinkering and adjusting settings, I have Vale running without errors. I had problems:

1. I had to learn about fitting Vale into the structure for my deploy.yml file for GitHub.
2. I had to update references to deprecated versions of actions used in the sources I learned from.
3. I had to solve the errors that Vale found in my documents.

I expected the first problem. That's the point of this project, to learn new skills and demonstrate them.

The second problem was an “unknown unknown.” I became aware of it when my linting jobs failed on deprecated versions (especially of Node.js version 20). Figuring out what was happening and why was a multi-step process; I first had to figure out which actions were calling a specific version of Node, then I had to find all the instances of those actions and update their references to call their latest versions.

The third problem was fully my own fault. I'd run Vale locally and seen the number of warnings it generated, mostly over reading levels. I decided to leave those issues unsolved so I could see them in the GitHub output. Unfortunately, I exceeded GutHub’s logging limits, generating an error.

With all those problems solved, I now have a CI/CD pipeline with style checking built in. My next steps are to add spell checking and checking for well-formed Markdown.

I have deep skills in documentation - researching, interviewing, experimenting, drafting, revising, polishing, and publishing in various formats - but now I am learning the ins and outs of automated docs-as-code workflows.