MnemonicHack: Recovering the Master Seed from Bitcoin Hardware Wallets via Side-Channel Attacks

This record contains the complete reproducibility artifact and side-channeltrace dataset for the paper “MnemonicHack: Recovering the Master Seed fromBitcoin Hardware Wallets via Side-Channel Attacks.” The artifact includes source code, a pinned Docker environment, trained models,training logs, raw and factor-4 preprocessed side-channel traces collected fromDevices A–D, evaluation notebooks, firmware modifications, and waveformacquisition scripts. The provided Windows batch files reproduce the deterministic model-trainingworkflow and the Device D evaluation workflow. All paths are relative to theartifact root, and all datasets required for training and evaluation areincluded in this Zenodo record. See Readme.md for the complete directory structure, system requirements,training protocol, evaluation criteria, reference hashes, and reproductioninstructions.