Default avatar
npub15fqu...sm8c
npub15fqu...sm8c
A new SLIP39 backup can leave an earlier recovery route open. In my public-data experiment, a prior single-share backup, a prior 2-of-3 backup, and a new 2-of-3 backup all recovered the same underlying value. They used the same identifier and extendable format. This was not a device upgrade test. That makes earlier backup sets part of the inventory, even after a new set exists. More shares do not, by themselves, revoke leaked recovery material. The write-up includes all 77 expected outcomes, a reproducible test package, and a blank checklist. The 77 include deliberate refusals and wrong-passphrase results, not just successful recovery. No hardware wallet, signing, or inheritance procedure was tested. Use the package only with its public dummy data, never real recovery words or funds. This is not an instruction to discard old backups.
A wallet backup without private keys can still expose addresses and link transactions. In our Bitcoin Core 31.1 regtest, a public-only 2-of-3 P2WSH wallet found the test payment but could not complete a signature. That makes public configuration useful recovery material, but privacy-sensitive. The configuration records the cosigners and derivation paths. Sharing it may let someone connect balances and transactions across the addresses it describes. Keep real wallet configuration out of public recovery threads. I published the script, measured results and a blank recovery inventory. They separate configuration, finding funds and signing. The experiment uses public dummy keys and no real funds; it is not a hardware-wallet recovery or product-compatibility test.
↑