Follow Discord
Sweep 02 Oct 2026 · 18:55Z Build v2.1.288 509 read Stable v2.1.285 Latest v2.1.287 Next v2.1.288 Feeds RSS JSON llms.txt llms-full.txt Unofficial
HomeReleasesv2.1.287entry #348

Claude Code v2.1.287 ·

Clearer wording in several messages Claude reads about tools and errors

Several messages Claude Code sends to the model about blocked tools, artifact copying and permissions have been reworded

Under the hood Internal Changes
JSON All of v2.1.287
Under the hoodTier: how much it should matter to you
1Useful: my rating, 1 to 5
1Signal: worth watching, 1 to 5
ElsewhereArea: what it touches
Internal ChangesKind: in v2.1.287,
Internal ChangesSection of the release
What

Claude Code has reworded several messages it sends to the model, the AI that does the work in a session. These are not messages you normally see. They are what Claude is told when something it tried cannot go ahead.

  • The message for when a computer cannot be reached no longer talks about "what this container can do". It now tells Claude to do the rest of the task without it and tell you what is blocked.
  • The refusal for copying from another artifact (a document Claude has produced) now names an HTML or XML document. It no longer lists SVG.
  • The refusal for an awk command that cannot be checked in advance is now written in plain language.
  • The reply sent when a permission request no longer exists, and the text for a failed launcher permission, now come from shared wording used elsewhere in Claude Code.
Why

These messages shape what Claude does next. Clearer wording makes it more likely that Claude carries on with what it can do and tells you plainly what it could not.

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 clear whether copying SVG content from another artifact is now allowed, or whether SVG is only no longer named in the message.

See this entry in the whole of v2.1.287 →

Feedback