You'll noticeTier: how much it should matter to you
1Useful: my rating, 1 to 5
1Signal: worth watching, 1 to 5
SubagentsArea: what it touches
ImprovementsKind: in v2.1.293,
ImprovementsSection of the release
What
A background agent is a helper Claude starts to work on part of a task while it carries on. When a session is restored and such an agent was left stopped, Claude used to always be told to resume it by sending it a message with SendMessage. Now that depends on whether the agent can actually be continued:
If it can, Claude is told to send it a message as before.
If it cannot, Claude is told to launch it again or check its output. Claude is also told not to assume the task got done, and to launch it again if the result is still needed.
When continuing is not possible, Claude is also told that "A previously spawned agent cannot be continued from here". Stopped agents are now listed in separate notices depending on whether they can be continued and whether their output can be read.
Why
Claude no longer tries to message agents that cannot answer, which would waste a step or leave work silently undone.
How sure we are
One source agreesOne thing we can check says the same as this entry.
Anthropic's release notes agreeFixed Claude being told to continue or message subagents with SendMessage in sessions, including resumed ones, where a host, a permission…
