{"version":"2.1.287","anchor":"new-get-task-output-control-request-plus-a-flag-gating-a-ta","canonical_anchor":"new-get-task-output-control-request-plus-a-flag-gating-a-ta","heading":"Programs driving Claude Code can ask for a background task's output","tier":"use","area":"SDK","url":"https:\/\/changelogs.core-directive.com\/v\/2.1.287\/e\/new-get-task-output-control-request-plus-a-flag-gating-a-ta","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.287","markdown":"### Programs driving Claude Code can ask for a background task's output\n\nA new `get_task_output` request lets a program controlling Claude Code read a background task's output\n\n**Unclear.** It is not clear whether the request works without the feature switch, or what the parts behind the switch do.\n\n**What**\n\nThe control protocol is the set of messages a program uses to drive Claude Code, for example a program built with the Agent SDK. It has a new request, `get_task_output`, which asks for a background task's output by its task ID. The SDK client has a matching method.\n\nA related part of this, which reads the output produced so far and waits with a time limit, is behind a feature switch. In code the switch is off unless it is switched on remotely. The flag server has returned it on for this site's account in readings taken before this release.\n\n**Why**\n\nPrograms that drive Claude Code can read what a background task has produced without waiting for it to finish.\n\n- Flag `tengu_violin_fhole`: Off by default, switched on for this account (read for one account on one subscription tier against v2.1.287; this account: on, anonymous baseline: on, compiled default: off) These values were read against a different version of Claude Code, so treat them as the nearest reading available instead of one taken on this release.\n- Area: SDK\n- Names: `get_task_output`\n- Tier: Use it now\n- Useful: 5\/5\n- Signal: 3\/5"}