Sentinel Tag
PII/PCI/PHI classification at ingestion. Automatically identifies and tags sensitive data fields before they enter the lakehouse.
What It Does
Sentinel Tag scans every field in every row and classifies sensitive data according to regulatory frameworks: PII, PCI, and PHI. It uses pattern matching (regex for SSNs, credit card numbers), ML-based entity recognition (names, addresses), and contextual analysis to achieve high-accuracy classification.
Capabilities
- PII detection (names, emails, SSNs, addresses)
- PCI detection (card numbers, CVVs, expiry dates)
- PHI detection (medical record numbers, diagnoses)
- ML-based entity recognition
- Field-level sensitivity tagging
- Regulatory framework mapping
- Confidence scoring per field
Example
const result = await mcp.call('sentinel_tag_classify', {
payload: sanitizedPayload,
stream_uuid: streamUUID,
frameworks: ['pii', 'pci', 'phi'],
});
// { "email": "PII:EMAIL", "card_number": "PCI:PAN" }
console.log(result.sensitivity_score); // 0.87 Configuration
agent: sentinel-tag
version: "1.0"
classification:
pii: enabled
pci: enabled
phi: enabled
output:
tag_format: redqueen-sensitivity-v1
include_confidence: true
min_confidence: 0.85 Related Agents
Vaccine Compiler
Payload sanitization and format normalization. Removes threats, neutralizes malicious content, and produces a clean payload for downstream processing.
The Cloak
Format-preserving encryption and tokenization. Masks sensitive fields while preserving data utility for analytics.
Ledger Guard
Cryptographic hash verification. Computes SHA-256 hashes of every payload for tamper detection and audit trail integrity.