2026-08-18 13:28 UTC
CVSS fails the first test of a measurement: two people scoring the same bug should get roughly the same answer.
5,564 CVEs published in 2026 carry two official CVSS scores, the CNA's and NVD's (the roughly 12% of 2026's v3-scored CVEs NVD has re-scored so far). 55% land in different severity bands; they match exactly just 10.8% of the time. 1,043 are Critical to exactly one of their two scorers.
Which feed does your queue key on, and did anyone pick it on purpose? I'd bet most of us inherited the default.
Replies (1)
-
@jbm@infosec.exchange 2026-08-19 06:54
@jgamblin@infosec.exchange "Which feed does your queue key on, and did anyone pick it on purpose? I'd bet most of us inherited the default." What do you call "the default" ? As for me, if available I always use the CNA score. They know better about their product. (Note that my use cases are 100% OSS, plus some hardware for SoC and chipsets.)