What bothers me is the last few lines where it notes something "important" that you NEED to pay attention to. Just to find out that it is something you already mentioned somewhere in the conversation or that is not related to the current work.

I agree -- the name of this plugin is confusing and confused.

The skill simply demands concise and well-formatted responses from an agent. It is something demanded by anybody daily-driving agents for their actual job since >75% of the text output from agents is fluff. THis would better be named `/i-wont-read-that-heap-of-garbage`

I often use Google Deep Research when I want it to actually go out and find an answer for me, but the prompt must have a high school essay requirement in there. Nothing I can do prevents it from giving me a page of background that tries to make the question sound important and interesting, even when it isn't.

I usually follow up with an "I'm not reading all that" and make it summarize.

I'm not paying for it, so when they get wordy and stupid, I just respond with something akin to "turn that response into like half a paragraph."

Or worse, buries an important detail somewhere in the text. Tangentially, over the weekend I had claude ask me if I wanted to continue with UIM-001. I couldn't find UIM-001 in any document it referenced creating or any text it had previously generated. Claude seems to be the worst about all of this stuff.

Also, somehow over the weekend it responded with these sections all clearly laid out - What landed, Decisions I made and recorded, Two findings, and What you need to do. Not sure why it can't do that all the time.

[flagged]