Implement Vale linting for abusive or lewd language

As insurance, we should implement Vale tests for our documentation to ensure that obscenities or other objectionable phrases do not accidentally get added to generated doc files. While I don't feel that our contributors (internal and external) would attempt to add words like this to our documentation, I also think that this is a low-cost, low-effort insurance effort, on the off chance of an incident.

For completeness, this rule should be set to the error level.

Mentioning @aqualls @marcel.amirault @eread for awareness. Feel free to grab this one based on your availability.