Function send_spawn_batch 

pub async fn send_spawn_batch(
    client: &ControlClient,
    spawn_args: SpawnArgs,
    count: usize,
    json: bool,
) -> Result<()>

Send count copies of a resolved spawn request - the same agent, task, and flags, each under its own fresh run id - and print one combined report.

This exists because spawn throughput from the CLI is otherwise bounded by process startup: each lev run invocation pays binary launch plus a socket round trip (~60 spawns/second in measurement), while the daemon itself accepts spawns as fast as they arrive. One invocation carrying the whole batch removes that bound without introducing any daemon-side cap.

count == 1 defers to send_spawn, keeping today’s single-run output shapes. A mid-batch failure stops the batch and says how many runs had already started - those runs keep running; lev ps lists them.