Platform

Models & Algorithms

Every score, formula and algorithm the module uses, with live calculators

GovernanceSecurity posture scoreMaturity index (NIST CSF 2.0)Secure-coding skill rating (Elo)

Edge & identityIP risk scoreImpossible travelZero-trust access decisionBot probabilityToken-bucket rate limit

Endpoint & dataDevice trust scoreVulnerability priorityData risk score

Operations & riskAlert triage (Bayesian)Behaviour anomaly (EWMA z-score)Isolation forest anomaly scoreAnnualised loss expectancy (FAIR, quick)Alert correlation similarityDetection and response times

Cryptography & privacyMosca’s inequality (post-quantum urgency)Key life usedCustodian quorum (k of n)Differential privacy for published statisticsPrivacy-preserving IP handling (Whistleblower)

IP risk score

Edge & identity

z = −4.2 + 3.1·rep + 2.4·tor + 1.6·hosting + 0.9·ln(1 + v/10) + 2.0·g
p = σ(z) = 1 / (1 + e^(−z))
rep = min(1, feed hits ÷ 3)   tor, hosting ∈ {0, 1}
v = requests per minute   g = 0 allow · 0.5 challenge · 1 block (geo rule)
decision: block if p ≥ 0.85 · challenge if p ≥ 0.60 · else allow

Calculator · IP risk score

Feed reputation0.67

Tor exit (0/1)0

Hosting or VPN (0/1)1

Requests per minute120

Geo rule (0, 0.5, 1)0.50

Risk

94 / 100

Block

z2.79

p0.942

Algorithm

1Step 1

Look up the IP in 9 feeds

2Step 2

Classify the network (Tor, hosting, VPN, residential, mobile)

3Step 3

Measure velocity over a sliding minute

4Step 4

Apply the platform geo rule

5Step 5

Score with the logistic model and pick the action

Validation and sources

AUC on labelled edge traffic

0.96

Coefficients refitted

weekly (logistic regression)

Open-source basis

scikit-learn logistic regression · feeds from Spamhaus, FireHOL, CrowdSec, abuse.ch

Where it is used

IP intelligence

Geo-blocking & IP Intelligence › IP intelligence

open →

Geo policy

Geo-blocking & IP Intelligence › Geo policy

open →