Message check now counts reasoning blocks, plus small internal additions
A check that scans messages now counts the model's reasoning blocks and no longer reads a stale value for tool calls; subagent side files go under a folder
Under the hoodTier: how much it should matter to you
1Useful: my rating, 1 to 5
1Signal: worth watching, 1 to 5
MessagesArea: what it touches
Internal ChangesKind: in v2.1.301,
Internal ChangesSection of the release
Unclear It is not stated what the message check decides, so the visible effect of counting reasoning blocks is unclear.
What
This release brings a few small changes inside Claude Code:
A check that scans the conversation's messages now also counts thinking blocks, which hold the model's step-by-step reasoning.
The same check now tests for tool calls (requests the model makes to run a tool, such as reading a file) using the correct value; before, it read an out-of-date one.
Side files kept alongside a subagent's transcript are now stored under a folder named "subagents". A subagent is a helper agent Claude Code starts to work on part of a task, and its transcript is the saved record of its conversation.
Several new internal counters and timestamps are tracked.
Why
These are mostly small internal bookkeeping additions. The correction to the tool-call test means that check no longer works from stale information.
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 doubtIt is not stated what the message check decides, so the visible effect of counting reasoning blocks is unclear.
