Detects and scrubs common PII patterns from text using regex and Luhn checks.
// Source: https://github.com/garrytan/gbrain | License: MIT | Copyright 2026 Garry Tan // Provided via justclaw.it.com; keep this notice with the code. /** * PII scrubber for captured query text (v0.21.0). * * Runs before INSERT into `eval_candidates`. Capture is on by default, * so plaintext PII sitting in a DB column is a privacy footgun the first * time someone exports or shares a brain dump. Six regex families cover * the obvious cases: * * 1. Email addresses * 2. Phone numbers (US + international) * 3. US Social Security numbers (XXX-XX-XXXX shape, with year-like false-positive guard) * 4. JWT-shaped tokens (three base64url segments joined by '.') * 5. Bearer tokens (Authorization: Bearer <opaque>) * 6. Credit card numbers (13–19 digits with Luhn verification ... (truncated -- full source via MCP)
See the full source, get the GitHub permalink, and search for more like it.
Get a free API key