Maybe im overlooking something obvious but how do you 'simply' scan the content of Claude users their prompts?

I uh might be skewing that as I generally just use a lot of curse words with Claude by default

There’s a correlation between getting the “How’s Claude Doing This Session?” (Or whatever) and four letter words.

It’s not always then, but it often follows it.

I'm glad I'm not the only one.

I feel bad about it but sometimes it's so daft, I can't even xD

It's not my fault, they set high standards!

there are many times where I just do it myself and it thinks it did well.

Or there are global events that stress people out .. or their expectations change over time. Not that simple ;)

Good thing expectations are perfectly constant!

This might be strangely effective.