{"version":"2.1.314","anchor":"output-limit-recovery-reworked-new-gate-tengu-rosy-pebble","canonical_anchor":"output-limit-recovery-reworked-new-gate-tengu-rosy-pebble","heading":"Continuing a reply cut off at the length limit now uses one fixed prompt","tier":"internal","area":"Elsewhere","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.314\/e\/output-limit-recovery-reworked-new-gate-tengu-rosy-pebble","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.314","markdown":"### Continuing a reply cut off at the length limit now uses one fixed prompt\n\nWhen a reply hits the output length limit, Claude Code now continues it with one fixed prompt instead of remotely tuned ones\n\n**What**\n\nA reply can stop early when it reaches the most text the model may write in one go. Claude Code then nudges the model to continue. The choice of nudge and the point to stop retrying used to come from settings sent by a server. Those are gone, and one fixed nudge is used.\n\nA second, remotely switched prompt also exists. It applies when the cut-off reply ended in the middle of text, in the main conversation rather than a helper task. The flag server returned off for this site's account and for the anonymous baseline. The error shown when retries run out no longer has a separate case for stopping early.\n\n**Why**\n\nCut-off replies are continued in a different way, and the remote setting that tuned this has been removed.\n\n- Flag `tengu_rosy_pebble`: Off in both readings (read for one account on one subscription tier against v2.1.314; this account: off, anonymous baseline: off, 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: Elsewhere\n- Tier: Under the hood\n- Useful: 1\/5\n- Signal: 1\/5\n- Scope: individual\n- Heads-up: no"}