A-05 ISOLATED MULTISIG EXPERIMENT

Requirements: Python 3.11+, cryptography (measured with 46.0.4),
and verified Bitcoin Core 31.1 bitcoind / bitcoin-cli binaries.
The published measurement used Windows, Core v31.1.0, and public dummy inputs.

Run the non-node safety checks:
  python test_lab_safety.py

Run the experiment only if you authorize creation and internal use of a temporary
local RPC authentication cookie by Core and bitcoin-cli:
  python multisig-recovery-lab.py --bin-dir /path/to/bitcoin/bin --output recovery-results-new.json --approved-ephemeral-rpc

The script creates a fresh temporary regtest directory, disables peer networking,
and does not read your normal Bitcoin directory. It accepts no real wallet input.
Never use its public dummy keys for real funds or enter real seed words into it.
The flag acknowledges only this experiment, not access to any existing wallet.

The successful run stops the node and removes its temporary directory before
writing results. On failure it attempts to stop the node and retains temporary
files; it does not print cookie values. Do not publish retained runtime files.

Read the limitations in recovery-results.json and the article. Signatures use
known reference-script/previous-output data and derived dummy private keys.
This is not a hardware-wallet, mnemonic-entry, or product interoperability test.
Elapsed times vary; 13.3 seconds is one observation, not a performance target.