RedHub AIRedHub AI
Tool Gate · How it works

How the gate works

PII Detection for Structured Data

This tool grades each column of a structured dataset for exposed PII. Each signal is scored 0–2; the verdict is the weakest one — then a gate that can only make it worse.

The 8 signals

01customer_id (not PII)
02full_name (low)
03email (moderate)
04phone (moderate)
05ssn (regulated)top trigger
06dob (moderate)
07plan (not PII)
08signup_ip (low)
SAFE TO USE
every PII column meets its class minimum
REMEDIATE
a column is under-protected, or a low/moderate column is exposed
HOLD
a regulated column is left exposed
Regulated-exposed gate · worsen-only

One regulated column left exposed holds the whole dataset - HOLD regardless of how many columns are clean. 83% handled with an open SSN column is still a HOLD.

HOLD

Why: Every column is handled except the SSN, which is regulated and stored unmasked - one regulated column exposed - so the dataset is HOLD even at 83% of PII handled. weakest signal 0/2 · 7 of 8 columns handled, 83% of PII.

Fix firstssn - mask, tokenize, or encrypt the SSN column before the dataset moves anywhere; a regulated field in the clear holds everything.

A clean sheet didn’t save it. The verdict is the weakest signal — and a top trigger is dispositive on its own — so it names the one thing to fix first. Grades handling you describe against each class's minimum, not your regulatory obligations. Not legal advice.