{"version":"2.1.274","anchor":"new-bash-first-prompt-steer-variant-experiment-slot","canonical_anchor":"new-bash-first-prompt-steer-variant-experiment-slot","heading":"New bash-first prompt steer variant experiment slot","tier":"soon","area":"Prompt Experiments","scope":null,"heads_up":null,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.274\/e\/new-bash-first-prompt-steer-variant-experiment-slot","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.274","markdown":"### New bash-first prompt steer variant experiment slot\n\nA new bashFirstSteerVariant experiment slot was added, gated by tengu_cozy_teapot and overridable via CLAUDE_CODE_COZY_TEAPOT\n\n**What**\n\nThe per-session experiment-memoization system gained a new field, `bashFirstSteerVariant`, joining existing experiment slots like `preReadLineDropped`, `leanPrompt`, `modelForPrompt`, and `delegationCostsFirst`. It defaults to \"strict\" and can be controlled by the `tengu_cozy_teapot` gate or overridden with the `CLAUDE_CODE_COZY_TEAPOT` environment variable.\n\n**Why**\n\nThis is infrastructure for testing variants of the 'bash-first' steering wording that pushes Claude toward using Bash for edits, allowing the behavior to be adjusted per session without a code change.\n\nGate state\n\n- Flag `tengu_cozy_teapot`: Not enough to say (read for one account on one subscription tier against v2.1.274; this account: no value returned, anonymous baseline: no value returned, compiled default: not a boolean we can read) 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: Prompt Experiments\n- Names: `tengu_cozy_teapot`, `CLAUDE_CODE_COZY_TEAPOT`\n- Tier: Nothing to try yet\n- Useful: 2\/5\n- Signal: 3\/5\n- Present in the build but not switched on"}