Split MLKEM schemas + vector files by test group type

[cpu]

This is a replacement for #192 with a few tweaks:

1. I went with a different naming convention (mlkem_encaps_test_schema.json and mlkem_test_schema.json for the schemas, mlkem_$SIZE_encaps_test.json and mlkem_$SIZE_test.json for the test files).
2. I reduced the flags/notes in each test file to just those that are in-use.
3. I updated CONTRIBUTING.md with guidance on splitting files by test type.
4. I updated vectorlint so CI will catch any violations of this property in the future.

Thanks to @dingxiangfei2009 for getting the ball rolling on this. 🙇

Resolves #191

[FiloSottile]

I'd argue that a mlkem_keygen_and_decaps_test also tests encaps, so they can just be called mlkem_test. It's the encaps ones that are special/limited.

[cpu]

I'd argue that a mlkem_keygen_and_decaps_test also tests encaps, so they can just be called mlkem_test. It's the encaps ones that are special/limited.

Fair point! I've reworked as suggested.

[sgmenda]

lgtm, thank you very much.