{"version":"2.1.300","anchor":"forked-sub-agent-model-selection-now-uses-a-model-permission","canonical_anchor":"forked-sub-agent-model-selection-now-uses-a-model-permission","heading":"Forked helper agents get their model passed in a different way","tier":"internal","area":"Subagents","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.300\/e\/forked-sub-agent-model-selection-now-uses-a-model-permission","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.300","markdown":"### Forked helper agents get their model passed in a different way\n\nForked helper agents now receive the main session's model as a separate settings layer, and the effort setting is dropped unless a condition holds\n\n**Unclear.** It is not clear whether forked helpers now actually run on a different model or effort level than before.\n\n**What**\n\nA subagent is a helper that Claude Code starts to work on part of a task. A forked subagent starts from a copy of the current session. Three things change in how one is set up:\n\n- The main session's model now reaches the helper as its own separate layer of settings. Before, it was set on the helper directly.\n\n- The effort setting, which controls how much reasoning the model does, is now left out unless a particular condition holds.\n\n- The chosen model is cleaned up into a standard form before it is used.\n\n**Why**\n\nThis may change which model and which effort level a forked helper ends up using.\n\n- Area: Subagents\n- Tier: Under the hood\n- Useful: 1\/5\n- Signal: 1\/5\n- Scope: individual\n- Heads-up: no"}