{"version":"2.1.290","anchor":"remote-subagent-permission-requests-backgrounded-cancel-tre","canonical_anchor":"remote-subagent-permission-requests-backgrounded-cancel-tre","heading":"Dismissing a background subagent's permission prompt now counts as a denial","tier":"notice","area":"Subagents","scope":"individual","heads_up":true,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.290\/e\/remote-subagent-permission-requests-backgrounded-cancel-tre","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.290","markdown":"### Dismissing a background subagent's permission prompt now counts as a denial\n\nCancelling a permission prompt from a remote subagent running in the background now denies the request instead of stopping the turn\n\n**Unclear.** It is not clear what decides whether the answer is reported to Anthropic.\n\n**What**\n\nA subagent is a helper copy of Claude that takes on part of a task. When a remote subagent asks for permission and you cancel the prompt, Claude Code now responds based on where the subagent is running:\n\n- If it is running in the background, the request is answered with \"User denied permission\" and the current turn carries on.\n\n- Otherwise the answer is \"User aborted\", as before.\n\nBefore, cancelling always gave \"User aborted\" and could interrupt the turn. The permission dialog also now has a short grace period on input. Claude Code also reports to Anthropic how each request was answered: whether it was allowed, whether it was dismissed, and whether the subagent was in the background.\n\n**Why**\n\nDismissing a prompt from a background subagent no longer stops the work you are doing in the foreground.\n\n- Area: Subagents\n- Tier: You'll notice\n- Useful: 2\/5\n- Signal: 2\/5\n- Scope: individual\n- Heads-up: yes"}