Follow Discord
Sweep 25 Sep 2026 · 19:33Z Build v2.1.283 504 read Stable v2.1.274 Latest v2.1.283 Next v2.1.283 Feeds RSS JSON llms.txt llms-full.txt Unofficial

Claude Code v2.1.281 ·

Agent tool: preserving tool results in subagents is dark-launched

The Agent tool can now keep tool results inside subagents instead of stripping them, controlled by a server-side switch

Entry
JSON All of v2.1.281
EntryKind: in v2.1.281,
ChangesSection of the release
What

The Agent tool, which starts subagents (helper copies of Claude that take on part of a task), has a new setting, preserveToolUseResultInSubagents. When it is true, the results of the tools a subagent uses (running commands, reading files and so on) are kept instead of being removed.

The setting takes its value from the tengu_velvety_sprout switch, which Anthropic's servers decide for each account. The Agent tool's own description is unchanged in this release.

Why

Where it applies, subagents hold on to more of what their tools returned.

Read from
Feature flag
tengu_velvety_sprout Not enough to say

Nothing here resolved what this flag was doing on this version, so nothing here should be read as on or off.

This account: no value returned · anonymous baseline: no value returned · compiled default in v2.1.281: off

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.

Read once, for one account on one subscription tier, against v2.1.281. It isn't a statement about your account. What a flag value here can and cannot tell you

How sure we are
Something disagreesSomething we can check disagrees with this entry, or the writer said they could not settle it.
The writer flagged doubtThe finding does not say where the kept results end up or what the reader notices from them.
Anthropic's release notes agreeImproved screen-reader output in /mcp: a disabled server is read as "off" instead of "pending"
The name it cites is new in this buildNew in this build: tengu_velvety_sprout

See this entry in the whole of v2.1.281 →

Feedback