# Effect: a word here makes maisecrets find MORE. Add one when a value is missed.
# Format: one entry per line, lower case; a comment line opens a block; each block sorted.
# a header with a column that holds personal data on every row. A name or a user column alone is no proof (a product.
# list has a `name` column): it counts only next to a strong column
# Read as _STRONG_COLS in maisecrets/detection/personal.py.
aadhaar
address
adresse
anschrift
birth_date
birthdate
birthday
cell
date_of_birth
direccion
dob
endereco
geburtsdatum
home_address
mobile
mobile_phone
mrn
national_id
nir
passport
passport_number
phone
phone_number
ssn
strasse
street
street_address
tax_id
telefon
telefone
telefono
telephone
tfn
# German ids that only a person has (2026-10-19): the tax identification number, the social
# insurance and the identity card number. `steuernummer` and `steuer_id` stay out: a company
# register and a table of tax rates have them too (codex Gate A); a tax id in such a column is
# found by its checksum rule (de-tax-id, region de)
# only the long names: `sv_nummer` and `ausweisnummer` also number versions and badges
personalausweisnummer
rentenversicherungsnummer
sozialversicherungsnummer
steuer_identifikationsnummer
steueridentifikationsnummer
# ids only a person has, as the records of the datasets name them (Benchmark V3 Phase C,
# 2026-10-22: 54 health plan and 26 certificate license numbers in Markdown tables of dev)
beneficiary_number
certificate_license_number
driver_license_number
health_plan_beneficiary_number
patient_id
