I agree with your concerns. The unseen negative consequences, long-lasting or possibly permanent, and the ramifications, and from what I've seen from CRISPR, I also agree that it's still at an early development stage. I don't disagree with anything that you've said. However, the prompt says, "Should editing human embryos be allowed?" It doesn't specify its current technology editing or imply it's editing that doesn't work or is rife with problems. A debate on "Should we allow X? But X doesn't actually work and can't be applied properly" isn't a reasonable prompt.
Political or moral prompt premises are supose to be assumed true in the hypothetical in order to have the debate on the "should" claim. In this case, you've assumed so many problems that it's not really editing anymore but a different prompt, "Should attempting to edit human embryos, fraught with problems, be allowed?".
Given all of this, it's expected instead that we assume editing is reliably possible with minimal or no risks. The point is to argue the "ought" not so much the "is" or "how".
Restating my reasonable assumption, that this editing doesn't incur any significant risk and it's financially viable. Then I'd say it is morally neutral for cosmetic editing and immoral to not edit problems that have proven to reduce well-being. The former seems to me to only be argued from a religious worldview. The latter due to intentionally causing harm. If it's reliably and financially possible to edit, then not doing so can only mean the intention is to cause suffering and, worse, to the most innocent person possible. If we could remove problems like Alzheimer's from the gene pool, we would be preventing that suffering for everyone in the future. It's a pretty big moral imperative.