# Local wallet transfer experiment Tested on Linux x86_64 with block v0.7.3 and Foundry (Anvil/cast) v1.7.1. `block.lock` is generated by `block lock`, not manually edited. Prerequisites: block, bash, Python 3, jq. Foundry upstream binaries support Linux/macOS; this captured lock contains linux/amd64 only. For macOS change `platforms` in block.toml, then run `block lock` to resolve that platform. ## Reproduce ```sh cd /path/to/transfer-lab export BLOCK_HOME="$PWD/block-home" block sync ./run.sh ``` Only tool resolution/download (`block lock` / `block sync` if not cached) needs GitHub network access. The transfer itself uses loopback-only nodes, no fork URL, no public RPC, no faucet, no real funds, and no browser wallet. `run.sh` refuses ports already in use. It starts its own Anvil nodes on 127.0.0.1:18545 and 127.0.0.1:18546, and cleans up only its own child process IDs. Anvil is quiet so no development secret key/mnemonic appears in logs or output. The built-in development accounts are already unlocked: `cast --unlocked` asks the local node to sign and send. This does NOT reproduce browser/hardware wallet key protection, and is not a configuration for a publicly exposed RPC endpoint. ## Evidence - `output/versions.txt`: tool versions and Foundry commit. - `output/accounts.json`: public sender/receiver addresses only. - `output/before-*.wei`, `output/after-*.wei`: exact balances in wei. - `output/transaction.json`: submitted signed transaction metadata. - `output/receipt.json`: receipt including status, block, gas used, effective gas price. - `output/summary.txt`: readable results and assertion outcomes. - `output/node-{a,b}.log`: zero bytes in the captured run. With an EOA recipient and no calldata, the plain ETH transfer used 21,000 gas. The example fixes a legacy gas price of 1 gwei for a transparent calculation; it is NOT a current mainnet fee estimate or general gas guarantee. Local instant mining and receipt status=1 are not evidence of production finality. The same development addresses begin with 10,000 ETH on each independent chain. Only chain A receives the send request. Chain B remains at 10,000 ETH for both addresses. This illustrates separate state, not an implementation of a bridge.