Follow Discord
Sweep 22 Oct 2026 · 21:04Z Build v2.1.309 528 read Stable v2.1.308 Latest v2.1.309 Next v2.1.309 Feeds RSS JSON llms.txt llms-full.txt Unofficial

Claude Code v2.1.246 ·

Per-task stop is first-attached-client-wins and does not survive a worker restart

The first client to attach decides per-task stop behaviour, and it is lost on worker restart.

Under the hood No documentation found New Features
JSON All of v2.1.246
Under the hoodTier: how much it should matter to you
2Useful: my rating, 1 to 5
3Signal: worth watching, 1 to 5
SDKArea: what it touches
New FeaturesKind: in v2.1.246,
New FeaturesSection of the release
What

The perTaskStopAffordance declaration is stored on a per-surface capability latch: the first client to attach sets it, and later clients cannot change it. Across worker restarts it is classified as "lost" rather than restored, so a reconnecting client must declare it again or interrupts go back to killing background tasks.

Details
  • Plumbed through the initialize config and the control-request handshake.
  • Any SDK consumer on this build can reach it by setting the field; no remote flag gates it.
Evidence

Declares that this consumer renders a per-task stop control wired to the stop_task control request, perTaskStopAffordance

Strings lifted out of the shipped bundle, so the claim above can be checked against them.

Read from
Names in the bundleperTaskStopAffordance

See this entry in the whole of v2.1.246 →

Feedback