h0p3/Swicky
It sounds like a great challenge, but I am sure I do not understand all the details. A few acronyms have me a little confused. Never the less I think I get the Gist.
To be clear are you searching a large collection of tiddlers containing text, tags and fields that you want to list as the result of a sophisticated compound search consisting of multiple filters?
I do feel that TiddlyWikis filters are in someways specific to the TiddlyWiki's platform requirements. Rather than let this complexity overwhelm, If I understand what you are trying to do, I would build a set of search primitives from TiddlyWiki filters that generate a set of tiddlers, then add these abstracted primitives together rather than drop down into the TiddlyWiki filters syntax, in your large compound "search criteria".
Designed carefully every search primitive could be invoked regardless of what proceeds or follows it. In which case you could specify the primitives and their order by tagging a tiddler lets call it a "Query tiddler" with the set of primitives, and the desired order (using the tag pill drag and drop ordering) and have the result "drop out at the end" or Not. This would allow you to look at the intermediate results ie Primative1 generates list A, Primative2 to uses this list to generate list B etc...
I feel a form of loop could be built that takes as its input the previous list generated by the previous primitive. If constructed correctly each step will result in a much smaller subset to test/filter and as a result the performance should be good.
In keeping with the original post I would then use the resulting list to transclude the content.
Perhaps I am way off here, but this is my brave attempt to respond to what looks like a deeply conceptual experiment, the type I love, but are best constructed in my own mind, rather than trying to understand prior work.
Regards
tony