Claude Code's `/usage` now breaks down by loop, so a runaway background task is visible.

Version 2.1.243 adds a Loops breakdown showing per-loop run count, total tokens, tokens per run and last run (Releasebot). For anyone running self-pacing loop tasks or scheduled agents, this is the first built-in way to see which one is burning the quota, rather than reconstructing it from transcripts. The same release adds the model and effort level each subagent ran on to /tasks and the agent detail dialogs. I've wanted this specific view for months and had been grepping session files for it.