Definition
EPSS (Exploit Prediction Scoring System) estimates how likely it is that a published vulnerability will be exploited in the wild in the next 30 days. Scores run from 0 to 1 and are updated daily, so teams can patch the flaws attackers are most likely to use first.
What EPSS is
Thousands of new vulnerabilities are published every year, and no team can fix them all at once. The Exploit Prediction Scoring System, run by FIRST, helps decide where to start. For each published CVE it gives a probability between 0 and 1 that the vulnerability will be exploited in the wild within the next 30 days.
A score of 0.02 means about a 2% chance of exploitation in that window; 0.85 means it is very likely. Scores are updated every day, because the chance of exploitation changes as exploit code appears and attackers start using a flaw.
How EPSS works
EPSS is a data model. It combines information about each vulnerability, such as its type, the products it affects, whether exploit code is public and how often it is mentioned, with data on exploitation attempts observed in real traffic. The model learns which combinations tend to come before real attacks and scores every CVE on that basis.
FIRST publishes the scores for free, together with a percentile that shows how a vulnerability compares with all others.
EPSS vs CVSS
| CVSS | EPSS | |
|---|---|---|
| Question it answers | How bad would it be if this were exploited? | How likely is it to be exploited soon? |
| Scale | 0 to 10 | 0 to 1 (probability) |
| Changes over time | Rarely | Every day |
| Based on | Properties of the vulnerability | Real-world exploitation data and signals |
The two answer different questions, so they work best together. A flaw with a high CVSS score and a high EPSS score deserves attention first. A high CVSS score with a very low EPSS score is serious on paper but less likely to be attacked soon.
Limits of EPSS
- It is about the world, not your systems. EPSS says how likely a vulnerability is to be exploited somewhere. It does not know whether the flaw is reachable or exploitable in your environment.
- It covers published CVEs only. Misconfigurations, weak passwords and logic flaws in your own applications have no EPSS score.
- It is a prediction. Low scores still get exploited sometimes, so EPSS should guide the order of work, not decide what to ignore.
For vulnerabilities that are already being exploited, the CISA Known Exploited Vulnerabilities catalog is a useful companion list.
Using EPSS to prioritise
- Start from the vulnerabilities found on systems you actually run.
- Rank them by combining severity (CVSS), likelihood (EPSS) and whether they appear on the KEV list.
- Check which of the top items are really exploitable in your environment.
- Fix those first, then work down the list as part of your remediation process.
How Plainsea relates to EPSS
EPSS estimates the chance that a vulnerability will be exploited somewhere. Plainsea answers the next question: can it be exploited in your systems? Its AI agents attempt the attack within the scope you set and report only findings confirmed as exploitable, with the evidence behind them. Read more in exploitability and AI penetration testing.