PII Detection Reference
RADAR detects 40 PII patterns across 6 categories with auto-discovery at runtime. Detection combines regex matching and format validation. Each detection produces a finding with a type tag and context.
Detection model
Each detection produces a finding with a confidence score (0.0–1.0) based on pattern match quality, context analysis, and cross-validation. Teams set per-category thresholds. Findings below threshold are informational; above threshold generate actionable alerts.
PII categories
RADAR's scanner auto-discovers patterns at runtime from the pii/patterns/ directory. Currently 40 patterns across 6 categories:
- Name
Base (10 patterns)- Description
SSN, email addresses, credit cards, phone numbers, IP addresses, dates of birth, passports, driver licenses, API keys, AWS access keys.
- Name
Financial (10 patterns)- Description
IBAN, BIC/SWIFT, CUSIP, ISIN, LEI, SEDOL, Bitcoin addresses, Ethereum addresses, CVV.
- Name
Healthcare (3 patterns)- Description
Medical record numbers (MRN), ICD-10 diagnosis codes, European Health Insurance Card (EHIC).
- Name
Extended EU (10 patterns)- Description
PESEL (Poland), BSN (Netherlands), CPR (Denmark), Spanish DNI/NIE, Codice Fiscale (Italy), Belgian NRN, Swedish Personnummer, Portuguese NIF, EU VAT numbers.
- Name
Legal (3 patterns)- Description
Bar numbers, case docket numbers, DUNS identifiers.
- Name
Biometric (4 patterns)- Description
IMEI, biometric template IDs, genetic RS IDs, HLA alleles.
Configuration
The PII scanner is auto-discovery based — any .py file placed in the pii/patterns/ directory is loaded at runtime. No configuration required to add new patterns.
CLI-driven threshold configuration (radar pii threshold) is planned for a future release. Today, the scanner runs with default sensitivity on all loaded patterns.