I've been answering feature requests over email so I thought I'd post them here. That cuts down on the duplicates and gives people something of a roadmap of development.
I'm noticing that it's highlighting many gerunds (used as nouns, past progressive, and sometimes present progressive) as passive voice. In one or two places, it even highlighted across a long stretch because there was a being verb near the beginning and a gerund near the end.
This is definitely one of the shortcomings of the current passive voice ID script, it's not discriminating. I'm refining it to be more useful but it's one of the newest features, it literally didn't exist yesterday morning :) It also wrongly IDs other words that end in -ing, such as thing, ping, etc when they come after "to be" verbs in a sentence. There's also other instances of the passive voice that I'd like to catch, such as "Mistakes were made", and "The experiment was completed by the scientists".
Also, most of my "weak words" are in dialog, and although this is informative, it would be nice to more easily distinguish between my bland words and my characters' bland words. Perhaps there could be an option to exclude anything within quotation marks?
That is most definitely planned, I was noticing that too.
Maybe there could be highlighting for the other parts of speech besides adverbs, too. I personally don't see adverbs as bad things, so long as they add effective description. However, by highlighting all of the adverbs, I was able to see what sorts I was using and if they were repetitive. I feel like this might help with the other parts of speech too, especially if only one option was chosen at a time.
I could definitely see this being useful, especially for adjectives. The trouble is that adverbs are one of the only parts of speech that are easily machine-searchable. It's harder to nail down a rule for adjectives, etc. But I'd definitely like to develop in that direction.