Red-Team Lab

HiddenInk attacks itself. Each run writes a fresh payload from templates, hides it with randomised settings, checks that a person really couldn't see it, then tests whether HiddenInk catches it. Generated attacks are returned to you only and never stored.

Seed (optional)

Also test both screeners (PDF only, uses more tokens)

Guardrails: only built-in fictional resumes, every file is watermarked as a test file, payloads only target the screening decision, and this endpoint is rate-limited.