How is a Decision Dossier tamper-evident, and can I verify one independently? #7
|
"Tamper-proof evidence" is easy to claim and hard to substantiate. Concretely: what makes a Decision Dossier tamper-evident, what algorithms are involved, and can an auditor who doesn't trust Decionis verify one independently? |
Replies: 1 comment
|
Short answer: an Ed25519 signature over a canonical serialization of the dossier, verifiable offline against a published JWKS — with a zero-dependency verifier you can run yourself. We say tamper-evident, not tamper-proof, and the distinction is deliberate. Nothing stops bytes from being altered; what the design guarantees is that altered bytes stop verifying. The mechanism
Public keys are JWKs ( Verifying it yourself
npx @decionis/verify https://decionis.com/public/decision-dossiers/<id>Fully offline, once you hold the dossier and the key set: npx @decionis/verify --file dossier.json --jwks ./jwks.jsonExit code Honest limits — worth knowing before you rely on it
If you're evaluating this for a compliance use case and need a guarantee we haven't described here — external anchoring, a timestamp authority, customer-held keys — ask in a new discussion. We'd rather tell you what isn't there than let you assume it. Related: What is Decionis, in plain terms? · How policies are managed and versioned · Where Decionis sits in the stack Have a different architecture or integration question? Start a new discussion — we answer publicly whenever we can. |
Short answer: an Ed25519 signature over a canonical serialization of the dossier, verifiable offline against a published JWKS — with a zero-dependency verifier you can run yourself.
We say tamper-evident, not tamper-proof, and the distinction is deliberate. Nothing stops bytes from being altered; what the design guarantees is that altered bytes stop verifying.
The mechanism
canonical_document_sha256.base64url, a…