Last reviewed
Correct answer: D. It is blocked, and nothing else happens in it until the worker is done
Explanation
The principle — the two dispatch modes trade the same thing in opposite directions: whether you wait, or whether you come back.
Why the key is correct — Anthropic's wording is specific: foreground subagents block the main conversation until complete. What stops is the conversation, not the request inside it, so nothing else happens there while the worker works.
Why the others are wrong — the block is not scoped to a single request. Staying usable is the other mode — background subagents run concurrently while you continue working — and that is a trade, not a reason to prefer it everywhere. A result arriving later belongs to the concurrent mode too; a worker that has blocked until complete is already finished.
Remember this — one mode costs you the wait, the other costs you the round trip.
Sources — Anthropic's Claude Code subagents documentation.
Sources
“Foreground subagents block the main conversation until complete.”
“Background subagents run concurrently while you continue working.”
Practise 10 questions on this topic
Take Claude Code Workflows — Timed Test 1 (10 questions) — scored instantly, explanation for every question, no login.