A seed-generation integration flaw routed security-critical randomness through a predictable MicroPython fallback; affected seeds require migration, not only a firmware update.
Firmware seed generation reached MicroPython’s deterministic Yasmarang fallback instead of the intended hardware RNG. Hashing the output and adding a BIP39 checksum made it look uniform but could not enlarge the underlying candidate set.
Coinkite identifies Mk3 4.0.1–4.1.9 and pre-fix Mk4, Mk5 and Q seed generation as affected. Published entropy estimates differ by model and assumptions; they are security bounds and preliminary attack models, not proof that every wallet is equally easy to recover.
Updating fixes future generation only. An existing affected seed remains affected after export to another device. Independent private dice entropy or a strong unique BIP39 passphrase can add a separate barrier, but uncertain users should follow the manufacturer’s migration advisory.
For the clearest picture, read this entry together with Cryptographic entropy, Deterministic PRNG, Wallet seed generation, BIP39 checksum, BIP39 passphrase, Dice-roll entropy. The reverse links also lead from Bits of entropy, Brute-force search space, Deterministic PRNG, Reproducible firmware build.
Hashing cannot create missing entropy
Exposure follows the firmware used at seed creation
An update does not repair an old seed
Multisig helps only when a secure-key quorum remains