{"version":"2.1.323","anchor":"truncated-output-message-now-tells-the-model-how-to-retry-a","canonical_anchor":"truncated-output-message-now-tells-the-model-how-to-retry-a","heading":"Claude is told how to retry a tool call cut off by the response limit","tier":"notice","area":"Tool Calls","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.323\/e\/truncated-output-message-now-tells-the-model-how-to-retry-a","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.323","markdown":"### Claude is told how to retry a tool call cut off by the response limit\n\nClaude Code has a new message telling Claude that a cut-off tool call did not run and to retry with shorter pieces\n\n**What**\n\nA response from Claude has a maximum length. Claude Code now has a message for when a response is cut off in the middle of a tool call (a request from Claude to use a tool, such as writing a file). The message says:\n\n- The tool call did not run and none of it was kept.\n\n- Claude should retry in a shorter response.\n\n- Long content should be split up, with the rest of a file added in later calls.\n\n**Why**\n\nThis should help Claude recover when a long response, such as writing a big file, hits the length limit.\n\n- Area: Tool Calls\n- Tier: You'll notice\n- Useful: 1\/5\n- Signal: 1\/5\n- Scope: individual\n- Heads-up: no"}