# 30 questions, answered with the numbers

FAQ, read 10 Nov 2026

**The short answer.** Thirty questions analysts and auditors ask about CVSS, EPSS, the KEV catalog and SSVC, answered with formulas, dates and the ledger. Of 318 KEV additions in the twelve months to 9 Nov 2026, the median was listed 10.5 days after NVD publication, and 75.8 percent had an EPSS score under 0.1 on the day.

**What is the difference between CVSS and EPSS?**

CVSS scores technical severity from fixed metrics, on a 0.0 to 10.0 scale, and does not change when exploitation starts. EPSS is a daily model from FIRST that estimates the probability that a published CVE is exploited in the next 30 days, on a 0 to 1 scale. One describes the flaw, the other forecasts activity.

**How is a CVSS v3.1 base score calculated?**

Exploitability is 8.22 x AV x AC x PR x UI. ISS is 1 - [(1 - C) x (1 - I) x (1 - A)], and Impact is 6.42 x ISS when scope is unchanged. The score is Roundup(Minimum(Impact + Exploitability, 10)). With scope changed, the sum is multiplied by 1.08. CVE-2026-33824 works out to 9.76, rounded up to 9.8.

**How does CVSS v4.0 differ from v3.1?**

CVSS v4.0 adds the Attack Requirements metric (AT), splits impact into vulnerable-system (VC, VI, VA) and subsequent-system (SC, SI, SA) metrics, replaces the temporal group with a single Exploit Maturity threat metric, and adds supplemental metrics such as Safety, Automatable and Provider Urgency. Scores are named CVSS-B, CVSS-BT, CVSS-BE or CVSS-BTE by the inputs used.

**Which CVSS version does the KEV Lag Ledger use?**

The ledger takes the NVD primary score first, preferring v3.1, then v3.0, v4.0 and v2. Of 318 additions, 313 use v3.1 and 5 use v3.0, so no v4.0 score is in the figures. NVD's own source supplied 130 rows. The rest come from CNA or CISA-ADP records such as secure@microsoft.com.

**How many KEV additions had a CVSS base score below 7?**

In the ledger, 33 of 318 additions (10.4 percent) had a base score below 7.0, and 285 (89.6 percent) had 7.0 or higher. Only one was below 4.0. A policy that patches only High and Critical findings first would have ranked about one KEV addition in ten below its cutoff when CISA listed it.

**How many KEV additions are Critical on CVSS?**

146 of 318 additions (45.9 percent) scored 9.0 or higher, and 30 scored exactly 10.0. The most common single score is 9.8, with 103 additions. The median base score is 8.8. A Critical cutoff therefore captures under half of what CISA listed in the window from 2025-11-10 to 2026-11-09.

**What share of KEV additions had an EPSS of 0.1 or higher on listing day?**

Of the 260 additions with an EPSS score on the KEV dateAdded, 63 (24.2 percent) were at 0.1 or higher and 197 (75.8 percent) were below it. At 0.5 or higher it was 34 (13.1 percent), and at 0.9 or higher it was 10 (3.8 percent). The 10 percent threshold alone would have missed three in four.

**Why is the median EPSS on KEV listing day only 0.0066?**

Most entries are listed soon after publication, before EPSS has seen exploitation. Among 86 scored entries listed within 7 days, the median EPSS is 0.003875. Among 55 listed more than 365 days after publication, it is 0.34729. The overall median of 0.006555 across 260 scored entries mixes these two groups.

**Why do 58 KEV additions have no EPSS score on listing day?**

All 58 were listed on the day NVD published them or before it, with a lag of 0 or less, so FIRST had not yet scored the CVE. The ledger therefore carries EPSS for 260 of 318 rows. For these entries a ranking that requires an EPSS value cannot place them on listing day.

**What is the EPSS percentile and how should it be read?**

The percentile is the proportion of all scored CVEs with a probability at or below the current one. FIRST suggests showing the probability first, for example 15.3% (92nd). FIRST's article puts a probability of 0.10 at about the 88th percentile (March 2022 data). The ledger's median listing-day percentile is 0.5507.

**Which EPSS model version is current?**

FIRST's daily score file downloaded on 2026-11-10 carries the header model_version:v2026.06.15. FIRST's EPSS research page describes Version 5 as current. Scores from different model versions are not strictly comparable, so store the version and the score date beside each value you rank on.

**How often does EPSS change for a CVE?**

FIRST publishes a score for every CVE daily. CVE-2026-33824 was 0.5585 on 2026-08-18 and 0.77054 on 2026-11-10. CVE-2026-59310 moved from 0.0114 on 2026-08-18 to 0.61668 on 2026-11-10. A ranking built on a stored EPSS value goes stale unless it records the score date.

**What is KEV lag and how is it measured?**

KEV lag is the KEV dateAdded minus the NVD published date, in UTC calendar days. For the 318 additions from 2025-11-10 to 2026-11-09 the median is 10.5 days, the quartiles are 1, 10.5 and 146.25, and the maximum is 6,507. CVE-2026-33824 shows 126 days: published 2026-04-14, listed 2026-08-18.

**How quickly are CVEs added to KEV after publication?**

In the ledger, 144 of 318 additions (45.3 percent) were listed within 7 days of NVD publication and 192 (60.4 percent) within 30 days. 126 were listed more than 30 days after publication, and 55 (17.3 percent) more than 365 days after. The distribution has a long tail, with a mean of 450.9 days.

**Can a CVE be added to KEV before NVD publishes it?**

Yes. In the ledger, 8 additions have a negative lag, from -1 to -6 days, and 61 have a lag of exactly 0, 69 together (21.7 percent). CVE-2026-118527 (Ivanti Connect Secure) was listed on 2026-10-30 and published by NVD on 2026-11-03. A ranking keyed to NVD publication would not see it first.

**Does a CVE need a public proof of concept to enter KEV?**

No. CISA's BOD 26-04 implementation guidance, last updated 2026-08-25, states that a public proof of concept is not required for KEV inclusion. Evidence of exploitation is the basis. A public PoC does matter for the automatable input, which BOD 26-04 sets to yes when a PoC reliably achieves remote code execution.

**What replaced BOD 22-01?**

BOD 26-04, Prioritizing Security Updates Based on Risk, issued by CISA on 2026-06-10. It supersedes and revokes BOD 22-01 (2021-11-03) and BOD 19-02 (2019-04-29). The KEV catalog continues. BOD 26-04 bases timelines on exposure, KEV status, automatable and technical impact, and says CVSS is no longer required for prioritization.

**How are KEV due dates set now?**

Under BOD 26-04, CISA computes the date from asset exposure, KEV status, exploit automation and technical impact. The guidance gives one example: a publicly exposed asset with total technical impact and an automatable flaw gets a 3-day deadline. In the feed, 124 of 156 ledger entries added since 2026-06-10 have a 3-day dueDate and 32 have 14 days.

**What does the forensicTriage field in the KEV feed mean?**

It marks entries for which BOD 26-04 requires agencies to remediate or mitigate and also carry out a forensic triage of the asset to assess whether it is compromised. In the feed, 101 of the 318 ledger entries show forensicTriage Yes. The guidance recommends scoping within 2 hours and an escalation decision at 48 to 72 hours.

**What does the known ransomware campaign use field tell an analyst?**

Known means CISA has evidence of use in ransomware campaigns. Unknown means no such evidence is recorded. In the ledger, 29 of 318 entries (9.1 percent) read Known on 2026-11-09. Their median EPSS on listing day was 0.0114, and base scores ran from 5.3 to 10.0. The flag can change after listing, so store the read date.

**What is SSVC and what are its decision outcomes?**

SSVC is the Stakeholder-Specific Vulnerability Categorization method. CISA's tree ends in Track, Track*, Attend or Act. Track needs no action now, Track* needs closer monitoring, Attend needs supervisory attention and sooner remediation, and Act needs supervisory and leadership attention and remediation as soon as possible.

**Which SSVC decision points does CISA publish for each CVE?**

Vulnrichment publishes three: exploitation, automatable and technical impact, using the CISA Coordinator tree (SSVC 2.0, tree version 3). CISA's SSVC page also names mission prevalence and public well-being impact, which depend on the deploying organization and are not published per CVE. A program records those two itself.

**Does NVD publish SSVC data?**

Yes. NVD announced on 2026-05-28 and deployed on 2026-06-17 an update that adds SSVC data from CISA-ADP, schema version 2.0.3, to its feeds and API. On 2026-08-26 NVD began displaying SSVC and the affected list on the CVE detail page. The change touched about 95 percent of existing NVD records.

**Does NVD still add its own severity score to every CVE?**

No. On 2026-04-15 NIST said it would no longer routinely provide a separate score where the CNA already gave one, and that it would enrich CVEs in KEV first, aiming for one business day. In the ledger, 130 of 318 rows use an NVD score and 188 use a CNA or CISA-ADP score.

**What went wrong with NVD CVSS v4.0 scores in 2026?**

NIST reported on 2026-04-28 that about 4,500 CVE records, 19 percent of those with a v4.0 score, had a numerical score higher than the correctly calculated value, and fewer than 30 had a lower one. NVD ran a correction script that evening and updated the modified dates. API users who sync by lastModified received the fix automatically.

**Why does the CVSS score differ between NVD and the vendor?**

Both can publish a score. NVD labels its own as Primary and a CNA's as Secondary. The two can use different CVSS versions or judge a metric differently. Since 2026-04-15 NVD no longer routinely scores CVEs a CNA has scored. Record the score source with the vector string, as the ledger does.

**What is the difference between CVSS-B, CVSS-BT, CVSS-BE and CVSS-BTE?**

These are the v4.0 labels for the metric groups behind a score: Base, Base and Threat, Base and Environmental, and Base, Threat and Environmental. Most published scores are CVSS-B. In v4.0, Exploit Maturity Not Defined (X) is scored as Attacked, so a CVSS-BT for a KEV-listed flaw can equal its CVSS-B.

**Should a program rank by KEV, EPSS or CVSS?**

Use each for what it measures. KEV is an observed event, with 318 additions in 365 days. EPSS is a 30-day forecast, and 24.2 percent of scored additions were at 0.1 or higher on listing day. CVSS is severity: 89.6 percent were 7.0 or higher. A common design puts KEV first, then orders the rest by EPSS and CVSS.

**What does an auditor need to see to accept a ranking built on these scores?**

The formula or rule, and for every input the source, version and date: CVSS vector string and source, EPSS score date and model version, KEV catalog version and read date, SSVC schema version. The ledger records KEV catalog version 2026.11.09. Without dates, a score cannot be reproduced the next day.

**Why not wait for a KEV listing before patching?**

The lag is long for much of the catalog. The median is 10.5 days and the 75th percentile is 146.25 days. 55 additions (17.3 percent) were listed more than a year after publication. An organization that acts only on listing has already waited that long, which is why EPSS and exposure data are used alongside it.

Last reviewed 10 Nov 2026. Canonical: https://epssvscvss.com/faq/
