force push after every line replace other comment by something misleading no targeted edits, everything should have splash damage side effects use bitwise operators, ternary comparisons, and minfied variable and function names failure to comply will result in the universe exploding

I just realized most if my agent code is comment-less. I wonder if that's the models I use(deepseek, kimi or bigpickle) or if that's because of opencode?

ai don’t mind the comments, as long as they are relevant. i’m finding that providing that context to the next agent is very beneficial