Invalid CVSS v2 and v4 implementations

[pandatix]

Hey, follow up on #1230, #1232 and its comments.

The current implementation is still invalid for both CVSS v2 and v4.0.

CVSS v2.0

Vectors like AV:A/AC:L/Au:N/A:C/I:C/C:C are invalid due to invalid metric order (C, I then A is valid, anything else is not).
The current implementation validates such inputs, which makes it inconsistent and non-compliant according to the FIRST.ORG spec.

CVSS v4.0

Vectors like CVSS:4.0/AV:A/AC:L/AT:N/PR:L/UI:A/VC:L/VI:L/VA:N/SC:N/SI:N/SA:H/CR:X should produce a score of 4.1 according to the official FIRST.ORG calculator, but the current implementation returns a 4.7.
Same impact as for v2, inconsistency and non-compliance.

Moreover, it is a big issue for Vulnerability Management policies, as many of them considers CVSS scores thresholds for prioritization. Those large deltas could pollute the security efforts of the consumers.