• CVE literacy
CVE scores are signals — proof is the decision
CVE, CVSS, EPSS, and KEV answer different questions. Builders who treat every high score as an equal fire drill burn out; builders who ignore KEV on exposed software get burned. Learn the vocabulary, then validate impact on systems you own.
Or scan your domain.
• Vocabulary
CVE
A public identifier (CVE-YYYY-NNNNN) for a known vulnerability disclosure. It names the issue so vendors and teams share vocabulary — it is not proof your deployment is exploitable.
Glossary →CVSS
A 0–10 severity score from standard metrics (attack vector, privileges, impact). High CVSS means severe in the abstract — still filtered by exposure, auth, and compensating controls in your environment.
Glossary →EPSS
Exploit Prediction Scoring System — a probability-oriented signal that a CVE will be exploited in the wild in a near-term window. Useful for prioritization; not a substitute for asset context.
KEV
CISA’s Known Exploited Vulnerabilities catalog — CVEs observed as exploited. Treat KEV hits on software you actually run as urgent patching candidates.
Why exploitability beats score theater
A CVSS 9.8 library buried behind auth on an internal tool is not the same risk as a CVSS 7.5 bug on an unauthenticated internet endpoint with public exploit code. EPSS and KEV help estimate wild exploitation; your asset inventory and proof-backed testing answer “are we actually vulnerable here?”
PandaONE’s AI VAPT and pentest phases prioritize verified impact on authorized targets. Known CVE intelligence can inform probes; it does not replace reproduction.