shine-pii-safe

PII anonymization for LLM preprocessing detects sensitive fields like phone numbers, emails, and IBANs in tabular data, confirms them with you, and applies anonymization to produce a safe dataset for language model ingestion.

This skill detects and anonymizes PII including phone numbers, IBANs, emails, tax codes, and addresses in tabular datasets. It applies anonymization techniques that preserve data grouping and trends while removing the raw sensitive values.

Yes, this skill produces a local mapping file alongside the anonymized dataset for traceability. This local mapping file is not uploaded, allowing you to securely retain the ability to trace or reverse the anonymized values.

No, the data anonymization process does not upload raw data. It detects PII columns, asks for user confirmation, applies anonymization locally, and outputs an anonymized file plus a local mapping file to ensure sensitive information remains secure.

You need to anonymize PII before processing data through analytics pipelines or data-cleaning workflows when sharing datasets with AI systems. This prevents sensitive information like tax codes and addresses from leaking during language model ingestion.

The best way to clean datasets containing addresses and tax codes for safe data sharing is to use an anonymization skill that flags sensitive fields, requires user confirmation, and outputs an anonymized file that preserves data trends without exposing raw PII.