Current event make me want to put out a short, incomplete list of a few things you should never do as a reviewer:
- do not take text out of context
- do not paraphrase so that the meaning changes
- do not use strawmen
- do not ask the author to write a different paper
- do not make up rules on what is “publishable” and what not
- do not try to only find fault in the paper
- do not ignore hypotheses in a paper just because you dislike them.
And here is what you should do:
- always re-check that what you wrote about the paper saying is actually what the paper says
- try to find if the paper has merit
- if you disagree with a point, give reasons and sources.
I guess that covers about 90% of reviewing etiquette.