SendMessage no longer leaks its display text into the model's…

SendMessage no longer leaks its display text into the model's view

SendMessage display text stays in the UI and no longer leaks into what the model reads.

You'll noticeTier: how much it should matter to you

3Useful: my rating, 1 to 5

3Signal: worth watching, 1 to 5

ToolsArea: what it touches

Bug FixesKind: in v2.1.232, 13 Aug 26

Bug FixesSection of the release

What

The SendMessage tool result strips its display field before serialising, so that content is shown in the UI only and no longer appears in the JSON the model reads back.

Evidence

mapToolResultToToolResultBlockParam

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

How sure we are

One source agreesOne thing we can check says the same as this entry.

Anthropic's release notes agreeType @ in the prompt to mention another Claude session by name; Claude then uses SendMessage to reach that session directly

See this entry in the whole of v2.1.232 →

The build it came inv2.1.232

13 Aug 2026 · 405 entries, this one is #263.

Use it now43
You'll notice209
Nothing to try yet41
Under the hood112

Either side of it

v2.1.231Older
v2.1.233Newer

Open the release →

Names in this entry

toolSendMessage
Messages arriving from another session now show who sent them plus a chord to expand.

Anthropic's notes

Upstream's own release notes for v2.1.232 name something this entry names, so the two are probably about the same change. Probably: the pairing is made on a shared name rather than on the words around it.

Type @ in the prompt to mention another Claude session by name; Claude then uses SendMessage to reach that session directly
SendMessage now delivers to a bare name that exactly matches one live session, instead of asking to confirm with a ref first

Is this right?