{"version":"2.1.317","anchor":"bash-repeat-meter-sampled-by-remote-config-10-when-unset","canonical_anchor":"bash-repeat-meter-sampled-by-remote-config-10-when-unset","heading":"A measurement of repeated shell commands is now sampled by remote setting","tier":"internal","area":"Telemetry","scope":"individual","heads_up":false,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.317\/e\/bash-repeat-meter-sampled-by-remote-config-10-when-unset","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.317","markdown":"### A measurement of repeated shell commands is now sampled by remote setting\n\nA new measurement of repeated shell commands runs for a sample of sessions, set remotely and defaulting to 10% when unset\n\n**What**\n\nClaude Code has a new measurement of repeated shell commands. A remote setting decides which IDs it runs for. When the server sends nothing, the rate is 10%. A number between 0 and 1 is used as the rate, and any other value turns it off. The choice is based on the ID, so the same ID always gets the same answer. Two other checks can switch it off first.\n\n**Why**\n\nThis is a measurement or experiment around how the Bash tool, which runs shell commands, behaves.\n\n- Area: Telemetry\n- Tier: Under the hood\n- Useful: 1\/5\n- Signal: 1\/5\n- Scope: individual\n- Heads-up: no"}