Secbez Docs

Severity & Confidence

How Secbez assigns severity ratings using CVSS 3.1 scoring and confidence levels to help you prioritize fixes.

Every finding has two independent ratings: severity (how bad is it?) and confidence (how sure are we?).

Severity levels

Severity is based on the potential impact if the vulnerability is exploited, calculated using the industry-standard CVSS 3.1 scoring system.

LevelCVSS ScoreMeaningExamples
Critical9.0 - 10.0Immediate, severe impact. Exploitable remotely with no authentication.Unauthenticated RCE, SQL injection exposing all data
High7.0 - 8.9Significant impact, relatively easy to exploitAuthenticated SQL injection, IDOR on sensitive data
Medium4.0 - 6.9Moderate impact or requires specific conditionsXSS requiring user interaction, missing auth on non-sensitive endpoint
Low0.1 - 3.9Limited impact or difficult to exploitInformation disclosure, missing security headers

Confidence levels

Confidence indicates how certain the analysis is that the finding is a real, exploitable vulnerability — not a false positive.

LevelMeaning
HighThe vulnerability is directly observable in the code with a clear exploit path
MediumThe vulnerability is likely real but depends on one verifiable assumption (e.g., a specific configuration)
LowThe vulnerability depends on multiple assumptions that can't be verified from the code alone

How to prioritize

Focus on findings with high severity + high confidence first — these are confirmed, impactful vulnerabilities.

Findings with high severity but low confidence may still be real but need manual verification. Don't ignore them; review them when time allows.

Low severity findings with high confidence are confirmed minor issues. Address them during routine maintenance.

On this page