{"version":"2.1.300","anchor":"agent-hooks-fall-back-to-another-model-when-the-hooks-model","canonical_anchor":"agent-hooks-fall-back-to-another-model-when-the-hooks-model","heading":"Agent hooks retry on another model when their model refuses","tier":"notice","area":"Hooks","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.300\/e\/agent-hooks-fall-back-to-another-model-when-the-hooks-model","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.300","markdown":"### Agent hooks retry on another model when their model refuses\n\nWhen the model running an agent hook refuses, Claude Code now reruns the hook on a fallback model instead of cancelling it\n\n**What**\n\nA hook is a check or action you configure to run automatically at certain points in a session. An agent hook is one where a model does that work.\n\nWhen the model assigned to an agent hook refuses the request, Claude Code now runs the hook again on a fallback model and says so in a message. Before, the hook was cancelled with the error \"Agent hook did not return structured output\".\n\n**Why**\n\nA refusal from one model no longer stops an agent hook from producing a result.\n\n- Area: Hooks\n- Tier: You'll notice\n- Useful: 2\/5\n- Signal: 2\/5\n- Scope: individual\n- Heads-up: no"}