Question about C# detection engine limitations for ML-DSA

[Chennamma-Hotkar]

Hi @n1ckl0sk0rtge,

While working on PR #385 (ML-KEM for C#), I noticed the README footnote
mentions the C# engine has "no cross-method variable tracking" as a known
limitation.

I traced this through the ANTLR grammar implementation and found that
MLDsa65.GenerateKey() follows the same single-statement static pattern
as ML-KEM, so it should work within current engine constraints.

Before I write the ML-DSA rules, I wanted to confirm: is the intended
approach to keep detection within single-method scope for now, or is
cross-method tracking on the roadmap?

Also happy to take on whatever C# work is most useful to you.