1. A vector, and optionally a second to compare
For example the NVD's vector against the vendor's, or your environmental rescoring against the published base.
Scores
What one change would do
The base score with a single base metric set to each other value. Rows that move the severity band are marked.
| Metric | Set to | Base | Change | Severity |
|---|
2. Audit a report
In JSON, a vector is paired with a baseScore, score or cvss_score and a baseSeverity or severity in the same object. In text, with the nearest score and the severity word on the same line.
| Where | Vector | Stated | Rescored | Result |
|---|
From the command line
The same engine, no dependencies, no network: node cli.mjs 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H' scores and explains a vector; --diff '<vector>' compares two; node cli.mjs --audit report.json --strict rescores every vector in a file (or - for stdin) and exits 1 on a mismatch, for a CI step over generated advisories. Source and README.
Questions
Why do two advisories give the same vulnerability different CVSS scores?
Because they scored different vectors. A CVSS score is a function of its eight base metrics, so two different scores mean at least one metric differs: often Attack Complexity, Privileges Required or User Interaction. Paste both vectors into the compare box to see which metrics differ and how much each one moves the score on its own.
How do I check that a CVSS score matches its vector string?
Recompute it. Paste the vector to get its base score, or paste a whole report, CSV or NVD JSON into the audit box: every CVSS:3.x vector is rescored and compared with the score and severity label written next to it. When a stated score is really the temporal or environmental score, or the other CVSS version's score, the tool says so.
What is the difference between CVSS 3.0 and 3.1 scores?
Base and temporal scores are the same under both versions for every one of the 259,200 base and temporal vectors, checked exhaustively here. Environmental scores can differ by 0.1 when the modified scope is Changed, because v3.1 changed the Modified Impact formula; in a random sample of 200,000 full vectors, 6.3% differed and 0.3% moved to a different severity band.
Why is my CVSS 3.0 temporal score 0.1 higher in another calculator?
Probably floating-point rounding. CVSS 3.0 rounds up to one decimal, and a calculator that computes Math.ceil(x * 10) / 10 turns 5.0 x 0.92, which is 4.6, into 4.7 because the computer stores it as 4.6000000000000005. Of the 259,200 v3.0 base and temporal vectors, that naive rounding gets 228 temporal scores wrong and no base scores. CVSS 3.1 redefined Roundup to fix this.
Does this tool score CVSS 4.0 or CVSS 2?
No. It scores CVSS v3.0 and v3.1 vectors, and says so when you paste a v4.0 or v2 vector instead of guessing a score for it.
Is anything I paste uploaded?
No. The scoring runs in this browser tab. There are no accounts, cookies or analytics vendors, and the command-line version makes no network calls.