Ok I see the problem. I had Q write rules in the read me for my error trait to prevent hiding errors and now it’s telling me that was an old style of writing code that now produces clippy warnings. So what Q gave me bad code or the style changed like a few days ago?? Ugh. So many wasted cycles.

It just doesn’t read and follow directs across multiple documents very well. It is pointed to a README that has links it other library and test readmes and it’s not reading and following all instructions there or in its context. Why?

Sorry meant to delete that. I’m not sure it was really returning zero. After I reviewed how q wrote the code, though it told me git was returning zero, I think q was swallowing the error in a subshell (again & and well it’s probably Claude actually). I didn’t have time to go back and test but doubt.

I’d also like the models to stop charging me tokens for the same errors repeated over and over again but I think that’s Claude not Q. Need to dig into that more. Telling the model in context not to repeat the same mistakes doesn’t work either.

Pay as you go pricing for Q CLI like every other service as is the original premise for AWS services with a way to cap spending for those that need it.

Don’t force me to use IC. It’s very confusing and overkill. Just tried to set up new user in IC again to use Q App and was royal pain. Wasted a lot of time.