Selective disclosure and zero-knowledge: show the fact, not the data

The most powerful idea in modern digital identity is also the simplest to state: you should share only what a verifier needs, and nothing more. Two layers of cryptography make it real — selective disclosure and zero-knowledge proofs.

Selective disclosure (SD-JWT)

A credential is a set of claims. Instead of signing the raw values, the issuer signs a set of salted hashes of them. To reveal a claim, the holder discloses its salt and value; the verifier recomputes the hash and checks it against the signed set. Undisclosed claims reveal nothing — the salt makes them non-invertible.

for each claim c_i with random salt s_i:
    d_i = H( s_i || c_i )
issuer signs the set  D = { d_1, ..., d_n }

reveal claim i:  holder shows (s_i, c_i)
verify:          H(s_i||c_i) == d_i  AND  d_i in D  AND  Sig_issuer(D) ok

Zero-knowledge: a fact without the value

Selective disclosure still shows a revealed field in full. Zero-knowledge proofs go further: you prove a predicate over a hidden value. The classic example — proving “age ≥ 18” without disclosing the birth date — becomes a range proof over a signed commitment. The verifier learns only a boolean.

issuer signs commitment  C = Commit(birth_date, r)
holder proves in zero knowledge:
    exists (birth_date, r):  C = Commit(...)  AND  today - birth_date >= 18y
verifier learns ONLY the boolean  (range proof: Bulletproofs / BBS+)

With BBS+ signatures, presentations also become unlinkable — two verifiers cannot tell they saw the same person. These primitives are on our roadmap and enter as they mature and pass audit. More: Selective disclosure and SD-JWT.