You'll noticeTier: how much it should matter to you
2Useful: my rating, 1 to 5
3Signal: worth watching, 1 to 5
SubagentsArea: what it touches
ImprovementsKind: in v2.1.293,
ImprovementsSection of the release
What
A subagent is a helper agent Claude starts to work on part of a task, sometimes in the background. Claude can sometimes send it a follow-up message, but not every caller has the tool to do that.
Agent results for completed, background-launched and ordinary runs now include canContinueAgent, based on whether the caller's tools allow continuing agents. It stays true only if both the agent's side and the caller's side allow it. The agent tool's output schema gains this field, and results also carry callerCanContinueAgents.
Background launches receive ownerCanContinueAgents, the peers listing is given callerCanSend, and the prompt for the agent tool takes a new continueAvailable option.
The completion notice adds "Send it a message ... to retrieve its result" only when canContinueAgent is not false. Before, it was always added when there was no output file.
The "still running" notice mentions messaging the agent only in the same case. Otherwise it says only "You will be notified when it completes."
Why
In setups where tools are restricted, Claude is no longer told to use a messaging tool it does not have, which avoids failed attempts and confusion.
How sure we are
One source agreesOne thing we can check says the same as this entry.
Anthropic's release notes agreeFixed a custom agent named worker being shown as "Agent" when it starts, and the agent detail dialog's title losing the agent type once the…
