I don't consider it wise to skip a stratum of metaredundancy and relate clause meanings directly to context, more or less ignoring co-text since it's the co-textualised meanings that construe field, tenor and mode.
Strata are not levels of metaredundancy. Metaredundancy is a redundancy on a redundancy. For example, semantics is metaredundant on the redundant relation between lexicogrammar and phonology. The relation between two adjacent strata is one of redundancy not metaredundancy.
(2) "clause meanings and skipping a stratum"
The meanings of clauses are meanings realised by clauses, and so the meanings realised by clauses are of the semantic stratum. There is no "skipping a stratum" when relating lexicogrammar to semantics to context.
(3) "it is co-textualised meanings that construe field, tenor and mode"
It is meaning that construes/realises context (field, tenor and mode). "Co-textualised" meanings are only at the instance pole of the cline of instantiation — in texts — where they realise specific contexts of situation. There are no "co-textualised" meanings at the system pole of the cline of instantiation, and so: no construal of the field, tenor and mode of the context of culture by them. A theory of semantics must be accountable at all points along the cline of instantiation, not just at the instance pole.