The PR's YAML linting steps are defined in taxonomy/.github/workflows/lint.yml. Both yamllint and yq are used to enforce several requirements. It is difficult for the user to run all these checks locally without some serious hacking. Currently, it is necessary to push updates to qna.yaml files upstream and see if they pass. This is slow, tedious, and wasteful of upstream compute resources.

Also, I personally had a PR commit fail the lint without sufficient detailed output for me to diagnose the problem. I had to pip install yamllint and run the same command in lint.yml to see the error. (The problem was lines that were too long.)

Implementing a lab lint command that uses the same lint.yml definitions would make the developer experience much faster and less tedious.