Definition
Plain language
The master list that gives every character a number, including many characters no screen draws.
As stated in the literature
The universal character-encoding standard assigning code points to characters plus large reserved and deprecated ranges; the gap between what's assigned a glyph and what's a valid code point enables invisible-text attacks.
Why it matters: It lets computers around the world agree on what every character means, but the gap between valid characters and drawable ones is what makes hidden-text tricks possible.
For example, the same standard that numbers the letter 'A' also numbers emoji, ancient scripts, and characters that no screen ever draws.
Heard on the show
“Out of four hundred and ninety-three content-level bypass attempts — homoglyphs, zero-width characters, the usual Unicode mischief — zero succeed.”Episode 062 — Treating Hallucinations as Exploits: A Gate-Based Architecture for Agent Safety