138% more reported CVEs per day in 2026 than last year.New: always-on Fixboxes. Vulnerable dependencies, fixed and verified.

Vulnerabilities

CVSS calculator: score and understand CVSS 4.0 and 3.1

Pick the metrics or paste a vector string, and the calculator gives you the score, the severity and a link you can share. Below it: what each metric means, what changed in CVSS 4.0, why the same CVE gets different scores, and how to use CVSS without letting it set your priorities on its own.

Updated

The Common Vulnerability Scoring System (CVSS) describes how severe a vulnerability is, as a score from 0.0 to 10.0 and a vector string that records every judgment behind it. It is maintained by FIRST. CVSS 4.0 is the current version, published in November 2023; CVSS 3.1, from 2019, is still widely used in CVE records and tools.

The calculator follows FIRST's own scoring for both versions. CVSS 4.0 scores come from a port of FIRST's reference calculator, so they match it to the decimal. Paste a vector from an advisory to see how it was scored, change a metric to see what moves the number, or copy the link to share a score: the vector is kept in the address.

Version
Score CVSS-B0.0Severity: None
CVSS-B
Base metrics only
MacroVector
002201
Vector stringCVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:N/SC:N/SI:N/SA:N
A CVSS 4.0 or 3.1 vector, starting with CVSS:4.0/ or CVSS:3.1/.
Base metrics
The vulnerability itself, the same in every environment. Set by the supplier or the CNA.

Exploitability

Attack Vector (AV)
Exploitable over the network, up to and including the internet.
Attack Complexity (AC)
No target-specific evasion is needed; the attack works repeatably.
Attack Requirements (AT)
The attack works against all or most instances of the vulnerable system.
Privileges Required (PR)
No account or access to settings or files is needed.
User Interaction (UI)
The attacker can exploit it without anyone else doing anything.

Vulnerable system impact

Confidentiality (vulnerable system) (VC)
No loss of confidentiality in the vulnerable system.
Integrity (vulnerable system) (VI)
No loss of integrity in the vulnerable system.
Availability (vulnerable system) (VA)
No impact on availability of the vulnerable system.

Subsequent system impact

Confidentiality (subsequent systems) (SC)
No impact beyond the vulnerable system.
Integrity (subsequent systems) (SI)
No impact beyond the vulnerable system.
Availability (subsequent systems) (SA)
No impact beyond the vulnerable system.
CVSS 4.0 scores use a port of FIRST's reference calculator (github.com/FIRSTdotorg/cvss-v4-calculator), Copyright (c) 2023 FIRST.ORG, Inc., Red Hat, and contributors, under the BSD 2-Clause license. CVSS 3.1 scores follow FIRST's CVSS 3.1 specification.

What CVSS measures, and what it does not

A CVSS score answers one question: if this vulnerability is exploited, how bad is it, given how it can be reached and what it lets an attacker do? It does not say how likely exploitation is, whether you ship the affected code, or how much the affected product matters to your business.

FIRST says so directly. The CVSS 4.0 user guide states that Base scores measure the severity of a vulnerability and should not be used alone to assess risk. A 9.8 in a library you do not ship is not a risk to you; a 6.5 in your internet-facing login service may be the most urgent thing on your list.

CVSS 4.0 vs 3.1: what changed

CVSS 3.1 (2019) CVSS 4.0 (2023)
Metric groups Base, Temporal, Environmental Base, Threat, Environmental, Supplemental
Attack conditions Attack Complexity (AC) covers both defeating mitigations and conditions such as race conditions Split in two: Attack Complexity (AC) for defeating security measures, the new Attack Requirements (AT) for conditions of the deployment, such as winning a race or being on the network path
User interaction None or Required None, Passive (the user does something ordinary, like visiting a page) or Active (the user must take a deliberate step, like opening a file or dismissing a warning)
Impact beyond the vulnerable system Scope (S): Unchanged or Changed, which inflates the score Scope removed. Separate impacts on the vulnerable system (VC, VI, VA) and on subsequent systems (SC, SI, SA)
Threat Exploit Code Maturity, Remediation Level, Report Confidence Exploit Maturity (E) only: Attacked, POC or Unreported. Not Defined counts as Attacked
Safety Not covered Modified subsequent integrity and availability can be set to Safety (S), and a Supplemental Safety metric
Score names Base, Temporal and Environmental score CVSS-B, CVSS-BT, CVSS-BE or CVSS-BTE, naming the groups behind the number
How the score is computed A formula of weighted metric values Expert-ranked groups of vectors (MacroVectors) and interpolation within them

The two versions are not interchangeable. The same vulnerability usually scores differently in 4.0 and 3.1, so compare a 4.0 score only with other 4.0 scores, and keep the version next to the number.

The four metric groups

  • Base: the intrinsic characteristics of the vulnerability, the same everywhere and over time. Set by whoever publishes the vulnerability: the vendor, a CVE Numbering Authority (CNA) or a database such as NVD. Base metrics are the only mandatory part of a vector.
  • Threat: in 4.0, one metric, Exploit Maturity. It records whether attacks are reported, a proof of concept is public, or neither. Left undefined it assumes the worst case, Attacked, so a Base score is a reasonable worst case and Threat information can only lower it.
  • Environmental: how much confidentiality, integrity and availability matter for the affected system (CR, IR, AR), and Modified versions of the Base metrics for your deployment: a service that is not reachable from the internet, a mitigation that removes an impact.
  • Supplemental (4.0 only): Safety, Automatable, Recovery, Value Density, Vulnerability Response Effort and Provider Urgency. They describe the vulnerability further but never change the score.

FIRST's guidance is that Threat and Environmental metrics belong to the people who run the software, not to the supplier who publishes the advisory. A supplier's score is CVSS-B; your own assessment, with what you know about exploitation and your deployment, is CVSS-BTE.

How a CVSS 4.0 score is calculated

CVSS 3.1 multiplies weights: each metric value has a number, and fixed equations turn them into a score. CVSS 4.0 works differently. FIRST's experts sorted the possible vectors into 270 MacroVectors, groups of vectors of comparable severity, defined by six equivalence sets (EQ1 to EQ6) over the metrics, and ranked them. Each MacroVector has a score in FIRST's lookup table. A vector gets its MacroVector's score, adjusted down by how far it is from the most severe vector in that group.

That is why there is no short formula for CVSS 4.0, and why calculators that approximate it disagree. The calculator on this page shows the MacroVector next to the score; its six digits are the levels of EQ1 to EQ6, 0 being the most severe.

Worked example: regreSSHion (CVE-2024-6387)

CVE-2024-6387 is a race condition in OpenSSH's server (sshd) that an unauthenticated remote attacker may be able to trigger by failing to authenticate within a set time; FIRST's example scores it as code execution that could compromise the whole system. Red Hat, the CNA that published it, scored it CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H, 8.1 High, and NVD's own assessment gives the same vector. FIRST uses it as an example in its CVSS 4.0 examples document, scored three ways:

Assessment What changes Score
CVSS-B Network attack (AV:N); defeating ASLR (AC:H) and winning a race (AT:P); no privileges or user; full confidentiality, integrity and availability impact on the server 9.2 Critical
CVSS-BT A public proof of concept, no reported attacks at the time: E:P 8.2 High
CVSS-BTE The mitigation is applied (LoginGraceTime 0): code execution is gone, but anyone can now exhaust the connections. MAC:L, MAT:N, MVC:N, MVI:N, MVA:L 5.5 Medium

The CVSS-BT and CVSS-BTE scores are FIRST's; the CVSS-B score is what the same Base metrics give without Threat metrics, from FIRST's algorithm. Select a row to load it into the calculator. Two lessons: Attack Requirements lets 4.0 separate "hard because of mitigations" from "hard because of a race", which 3.1 folded into one AC:H, and a mitigation you have applied is an Environmental change that can move a critical finding to medium. Record it as such rather than arguing with the Base score.

Why the same CVE has different CVSS scores

Look up a CVE in two tools and you may see two scores. The usual reasons:

  • Different scorers. The CNA that publishes a CVE can include a CVSS score. NVD may add its own assessment; its records list both, marking NVD's as Primary and the CNA's as Secondary. CISA adds scores to some records as an Authorized Data Publisher (ADP). They do not always agree.
  • Different versions. A CNA may publish a 4.0 score while a database or scanner shows 3.1.
  • Different products. FIRST explicitly allows more than one Base score for a vulnerability when the affected platforms differ, for example where one operating system has a mitigation another lacks. Linux distributions often score a library CVE for their own build of it.
  • Threat and Environmental metrics. A score that includes them describes one situation, not the vulnerability.

A well-known case is curl's SOCKS5 heap overflow, CVE-2023-38545. The curl project rated it High in its own advisory, without a CVSS score. NVD scored it CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H, 9.8 Critical, while CISA's ADP entry scored it 8.8 High with user interaction required (UI:R). The difference is one judgment about how an attacker gets curl to connect, and it moves the finding from one band to another.

The practical rule: read the vector, not just the number. It tells you which judgment differs, and whether the assumption holds in your product. Record which source and version your process uses, and stick to it.

Using CVSS with EPSS, KEV and VEX

CVSS is one input to prioritization, and the one most often overused. Sorting a backlog by CVSS puts a large share of it at High or Critical, and most of those vulnerabilities are never exploited. A sounder order:

  1. Confirm you are affected. Check that the vulnerable package is in what you ship. Where it is not, or the vulnerable code cannot be reached, say so in a VEX statement instead of scoring it.
  2. Exploited first. A CVE in CISA's Known Exploited Vulnerabilities catalog is being used in attacks, whatever its score.
  3. Then likelihood. EPSS estimates the probability of exploitation in the next 30 days. Do not multiply it by CVSS; use the two side by side.
  4. Then severity in context. Use CVSS, ideally with your own Environmental metrics, to rank what is left and to break ties.

The full method is in vulnerability prioritization, and how to cut the list down before ranking it is in reducing vulnerability noise.

Where CRACI fits

A CVSS score is only useful once you know the vulnerable package is in your product. CRACI records the packages each GitHub Actions job actually fetched, including transitive dependencies, and each package shows its advisories (CVE, GHSA and other IDs) with their severity. Build history shows every past build that contained a vulnerable dependency, direct or transitive, so you know which releases carry it and for how long. Security teams triage the findings and send them to the teams that own the fix. See vulnerability tracking.

When CRACI is not the answer

  • Scoring your own vulnerabilities. Writing a CVSS vector for a flaw in your own code is analyst work: this calculator and FIRST's specification are the tools. Publishing it is your CNA's or advisory process's job; see coordinated vulnerability disclosure and CSAF.
  • Builds outside GitHub Actions. CRACI works with GitHub Actions today. GitLab CI and Jenkins are on the roadmap.

CVSS: frequently asked questions

What is a CVSS score?

A number from 0.0 to 10.0 that describes how severe a vulnerability is, from a set of metrics such as how it can be reached, what an attacker needs and what they can do. The vector string records the metric values behind the score. CVSS is maintained by FIRST.

How do I calculate a CVSS 4.0 score?

Choose a value for each of the 11 Base metrics, or paste a vector string, in a calculator that follows FIRST's algorithm, like the one on this page. CVSS 4.0 has no short formula: the score comes from FIRST's table of MacroVector scores and an interpolation within the MacroVector, so approximations can differ.

What are the CVSS severity ratings?

The same in CVSS 4.0 and 3.1: None 0.0, Low 0.1 to 3.9, Medium 4.0 to 6.9, High 7.0 to 8.9 and Critical 9.0 to 10.0.

What is the difference between CVSS 3.1 and 4.0?

CVSS 4.0 adds Attack Requirements (AT), splits user interaction into Passive and Active, replaces Scope with separate impacts on the vulnerable system and on subsequent systems, reduces the Temporal group to one Threat metric (Exploit Maturity), adds Supplemental metrics that do not change the score, and names scores by the groups used: CVSS-B, CVSS-BT, CVSS-BE and CVSS-BTE. Scores from the two versions are not comparable.

What do CVSS-B, CVSS-BT, CVSS-BE and CVSS-BTE mean?

They name the metric groups a CVSS 4.0 score was calculated from: Base only (CVSS-B), Base and Threat (CVSS-BT), Base and Environmental (CVSS-BE), or all three (CVSS-BTE). FIRST asks that every published 4.0 score carries one of these names.

Why does NVD show a different score from the vendor?

NVD makes its own assessment, which can differ from the CNA's: a different judgment on one metric, such as whether user interaction is required, can move a score from High to Critical. NVD records list both, marking NVD's own as Primary and the CNA's as Secondary. Compare the vectors to see which judgment differs.

Is a high CVSS score the same as high risk?

No. CVSS measures severity if the vulnerability is exploited. Risk also depends on whether it is being exploited (CISA KEV), how likely exploitation is (EPSS), and whether the vulnerable code is in your product and reachable. FIRST says Base scores should not be used alone to assess risk.

Know which builds contain a vulnerability

A CVSS score says how bad a vulnerability is. Book a demo to see which of your GitHub Actions builds contain it, from a record of what each build fetched.

Book a demo