Privara: Exploring payment-scoped selective disclosure for privacy on Stacks

Hi everyone,

I’m Samuel, building Privara, a privacy-focused payment protocol on Stacks. We’ve completed our initial grant milestones and deployed a working mainnet payment flow. I’m now exploring how to extend its privacy model, particularly around selective disclosure.

What’s implemented

Privara currently uses secp256k1-based stealth-address derivation to generate one-time recipient addresses, reducing the ability to link payments through a persistent recipient address.

The implementation also includes SIP-018-based payment authorization, sponsored transaction flows, and payment discovery mechanisms.

These provide recipient-address privacy, not full transaction confidentiality. Amounts, transaction activity, and other on-chain metadata remain observable.

What I’m exploring

The next step is payment-scoped selective disclosure: allowing a recipient to share verifiable information about a specific payment with an auditor or counterparty without exposing unrelated payments or granting access to their entire wallet history.

I’m considering a design where a recipient can generate a narrowly scoped disclosure package containing payment-specific evidence, potentially supported by encrypted payment metadata and cryptographic proofs.

This is still research, not an implemented viewing-key system.

Questions I’d appreciate feedback on:

  1. Verification: What would be the most appropriate way to bind a disclosure package to an actual Stacks transaction and recipient without exposing a reusable identity or derivation secret?

  2. Key separation: Would per-payment disclosure keys derived from a dedicated viewing secret be preferable to sharing encrypted payment metadata directly? How should this avoid leaking unrelated payment relationships?

  3. Clarity constraints: Which parts of disclosure verification should reasonably happen on-chain versus in an off-chain verifier, given Clarity’s cryptographic capabilities and execution constraints?

  4. Protocol direction: Are there ongoing Satoshi Upgrades or privacy research discussions that this should align with, particularly around viewing keys, shielded payments, or sBTC?

I’m not proposing a SIP at this stage. I’d first like to validate the threat model, understand the relevant protocol constraints, and determine whether this belongs entirely at the application layer.

I’d appreciate feedback from anyone working on Stacks Core, Clarity, wallet infrastructure, or privacy research.

Privara: GitHub - Psalmuel01/privara: Privacy-preserving SIP-010 intents payment protocol on Stacks · GitHub
Website: https://useprivara.xyz

Thanks,
Samuel

Hi Samuel, nice to see the mainnet flow live, and good to settle the threat model before any SIP.

On binding a disclosure to a real payment: the per-payment ECDH secret is the natural unit. Give the auditor that one secret plus the txid, and they can recompute the one-time address from your registered spending key and check it against the destination in that transaction. Every other payment used a different ephemeral key, so nothing else is exposed, and the viewing secret never leaves the wallet.

On what Clarity can do: as far as I know there’s no elliptic curve point math, so that derivation check has to stay off-chain. Control of the one-time address can be proven on-chain though: sign a SIP-018 message that names the auditor and the payment with the one-time key, then use secp256k1-recover? and principal-of? and compare the result with the destination.

On rotation: the M2 announcement already carries the registry epoch, which is the right hook. The registry map only keeps the current key pair, though, and the update event prints the epoch without the keys, so checking an older payment means digging the old pair out of past transaction data. Storing keys per (principal, epoch) makes that a single read.

Interesting direction. One additional approach worth exploring is zero-knowledge proofs for payment-scoped disclosure.

Instead of sharing the per-payment ECDH secret with an auditor, the recipient could generate a proof demonstrating knowledge of the payment derivation and ownership relationship, bound to a specific transaction ID and disclosure scope.

This could allow verification of recipient ownership, payment amount, or other selected attributes without exposing derivation secrets or unrelated wallet activity.

I’d keep the cryptographic verification off-chain initially, with an optional on-chain commitment or SIP-018 attestation for auditability.

The key design principle: prove the payment, not the wallet history.

one thing worth putting in the threat model early: transferability. once an auditor has the per-payment secret and the txid, they hold a check anyone can rerun, so they can pass it on to anyone, forever. the same goes for a plain zk proof, or a SIP-018 message that names the auditor: naming them doesn’t stop a third party from verifying it. scope and expiry end up as a promise, not something the crypto enforces.

if that matters for Privara’s users, designated-verifier proofs are the usual answer: the proof only convinces the named auditor, because the auditor could have produced an equally valid-looking one on their own. if it doesn’t matter, saying so plainly in the docs is enough.

a small thing from our own fair-play page: a short “what this does not prove” list (say, nothing about other payments to the same recipient, or who the sender is off-chain) got us more useful review than the parts saying what it does prove.