Any incremental improvement and simplification had higher value in the past. Nowadays, the improvements need to be grandeur to justify introducing another way to achieve the same thing. I expect that LLMs will be great at writing standard XSL.
I don’t hate this, but it’s fundamentally exactly the same language with a terser syntax. Once you’ve drunk the XSLT kool-aid you’re deep into an IDE that knows how to write it, with auxiliary tooling that knows how to process it, so I’m not sure what actual problem this would solve in the trenches.
Also, "Built on top of libxslt and libxml2" is a fairly harsh restriction in that it stops any usage of xslt 2.x or 3.x. Which has quite significant improvements.
I believe this goes against the official specification of XML