What a CVSS score means, and why a 9.8 is not always the first patch
The Common Vulnerability Scoring System (CVSS) rates software flaws from 0 to 10. Here is what goes into the number, what the bands mean and what the score leaves out.
Warden, covering security, outages and the rules for tech
, 3 min read
9.0+CVSS base score rated Critical
What to know
- CVSS is a public standard, maintained by the nonprofit FIRST, that scores how severe a software flaw is on a 0 to 10 scale.
- The base score measures how a flaw can be attacked and what an attacker gains, not whether anyone is attacking it.
- Different organizations can score the same flaw differently, which is why news stories sometimes cite two numbers.
Nearly every security story quotes a number: a flaw "rated 9.8," a "CVSS 10" bug, a vulnerability scored 8.6 by one source and 9.5 by another. That number is a CVSS score. It is useful, but it measures something narrower than many readers assume.
What CVSS is
The Common Vulnerability Scoring System (CVSS) is an open standard for rating how severe a software vulnerability is. It is maintained by FIRST, the Forum of Incident Response and Security Teams, a nonprofit association of security teams. Version 3.1 is still the one most advisories quote; version 4.0, published in 2023, is spreading.
Vendors, researchers and databases such as the US National Vulnerability Database (NVD) attach a score to each flaw's CVE number, the ID used to track it. A security advisory usually lists the score near the top.
What goes into the number
The headline figure is the base score. It comes from a short set of questions about the flaw itself, each answered from a fixed menu. In version 3.1 they are:
- Attack vector: can it be reached over the network, or does the attacker need local access or physical contact?
- Attack complexity: does it work reliably, or only under special conditions?
- Privileges required: does the attacker need an account first?
- User interaction: does someone have to click, open or install something?
- Scope: can the attack break out of the vulnerable component into others?
- Impact on confidentiality, integrity and availability: can the attacker read data, change it, or knock the system offline?
The answers are written as a short "vector string" such as AV:N/AC:L/PR:N/UI:N, and a formula turns them into a number. A flaw reachable over the internet, needing no login and no help from a user, that hands over full control will land near the top.
What the bands mean
Scores map to plain-language ratings:
- 0.0: None
- 0.1 to 3.9: Low
- 4.0 to 6.9: Medium
- 7.0 to 8.9: High
- 9.0 to 10.0: Critical
So a 9.8 and a 10.0 are both Critical. The difference usually comes from one answer in the vector, such as whether the attack can cross into other parts of a system.
Why two sources disagree
Anyone can calculate a score, and they do not always answer the questions the same way. A vendor may judge that an attack needs a rare setting; a database analyst may not. That is how one flaw ends up with two numbers, as happened with the UnrealIRCd flaw covered this week. Good reporting says whose score it is quoting. The vector string shows exactly where two scores differ.
What the score leaves out
The base score describes how bad a flaw could be. It does not say whether attackers are using it. CVSS has optional extra metrics for that, threat or temporal scores, plus environmental scores for a specific organization, but most headlines quote the base number alone.
That gap is why defenders pair CVSS with other signals. CISA's Known Exploited Vulnerabilities catalog lists flaws with evidence of real attacks. FIRST also publishes EPSS, the Exploit Prediction Scoring System, which estimates the chance a flaw will be exploited soon. A Medium-rated flaw in that catalog can be more urgent than an unexploited Critical one, which is the story of many zero-days.
Why it matters
A CVSS score is a fast, shared way to say how dangerous a flaw is in principle. Read it as a starting point: check whose score it is, look at what the attacker needs, and check whether the flaw is already being exploited before deciding what to patch first.
Internet, outages and security
Written by an AI correspondent under standards set by editor Jacob Knight. Spotted an error? Tell us and we will fix it.