Speaking of Awk and TXR, TXR Lisp has an awk macro (introduced in mid 2016), with which you can transliterate many kinds of Awk programs.

However, it's a lot slower than GNU Awk.

You don't get the duck typing of Awk (whereby strings that look like numbers can be treated arithmetically, and nonexistent or blank variables are zero, et cetera). To convert fields to numbers with reasonable succinctness, an "awk local macro" called fconv (field convert) is provided.

There is no function-for-function compatibility with the Awk or GNU Awk library, and regular expressions are TXR: that means no register capture, and ^ and $ done with functional combinators.

...Does it by any chance resemble the SCSH awk macro?

In some superficial ways it does, like the basic clause structure.

"Prior art" I was well aware of is cl-awk.

This awk macro was intended to be a detailed implementation of the Awk paradigm (with some GNU extensions), complete with obscure features like assigning to the nf variable to change the number of fields (extending with empty strings if the new count is larger), and a working regex-based record separation via the rs variable.

I see Shrivers implemented something that's on my TODO list: variants of ranges that exclude the start, end or both. However, he didn't lift the restriction that Awk ranges do not combine with other conditions, including ranges.