Revert do→let/it chaining (caused 80-test regression)

The let/it wrapping changed semantics of ALL multi-command sequences,
breaking independent side-effect chains like (do (add-class) (add-class)).
Need a targeted approach — chain it only for then-separated commands.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

shared/static/wasm/sx/hs-compiler.sx

@@ -999,23 +999,7 @@
                     (hs-to-sx (nth ast 1))
                     (hs-to-sx (nth ast 2)))))
               ((= head (quote do))
-                (let
-                  ((compiled (map hs-to-sx (rest ast))))
-                  (if
-                    (= (len compiled) 1)
-                    (first compiled)
-                    (let
-                      ((last-cmd (nth compiled (- (len compiled) 1)))
-                        (init-cmds (reverse (rest (reverse compiled)))))
-                      (reduce
-                        (fn
-                          (body cmd)
-                          (list
-                            (quote let)
-                            (list (list (quote it) cmd))
-                            body))
-                        last-cmd
-                        (reverse init-cmds))))))
+                (cons (quote do) (map hs-to-sx (rest ast))))
               ((= head (quote wait)) (list (quote hs-wait) (nth ast 1)))
               ((= head (quote wait-for)) (emit-wait-for ast))
               ((= head (quote log))