Add signal/crash handling to the miner CLI

[JSONbored]

Problem: No process-level signal handling exists anywhere in the miner package — no SIGINT/SIGTERM/uncaughtException/unhandledRejection handlers. The CLI dispatches through a chain of bare process.exit() calls with no cleanup hook, so an interrupted run leaves whatever state it was mid-writing in an undefined condition.

Area: AMS / CLI

Proposal: Add top-level signal and error handlers to the CLI entrypoint that flush/close any open ledger connections and exit cleanly, rather than letting the process die mid-write.

Deliverables:

• Signal handlers registered once at CLI startup, covering every subcommand.
• A defined, tested "safe exit" behavior for each ledger.

Acceptance criteria:

• Sending SIGINT/SIGTERM mid-run leaves every ledger in a valid, non-corrupt state.
• An uncaught exception is logged and exits with a non-zero code rather than crashing silently.

Resources:

• The miner CLI's main entrypoint and dispatch chain.

Boundaries:

• Handle cleanup only — don't change any command's actual business logic.

[linear-code]

JSO-3486