Hmm, this might (or might not, hard to say what if any impact your code has on this - we know that some static site generators like jekyll have built-in “Pretty URL” features that can cause or influence this behavior) fix things for you:
enabling “Pretty URL”'s in our asset optimization, here: https://app.netlify.com/sites/flamboyant-kalam-d96ba4/settings/deploys#asset-optimization
I’d actually suggest that instead of flipping that setting on your production site - which may cause worse 301 redirects or a loop - that you try deploying a second copy of the same repo, browse without the setting set, flipping that setting, and then in the same browser you did the first browse-test of the new site with, with cache enabled, seeing if behavior is different/better. If it does, then should be safe to flip the main site as well.
If not (or if the test site gets completely broken by this), let me know and I’ll be happy to take another look!