Follow Discord
Sweep 12 Nov 2026 · 23:00Z Build v2.1.323 540 read Stable v2.1.311 Latest v2.1.323 Next v2.1.323 Feeds RSS JSON llms.txt llms-full.txt Unofficial

Claude Code v2.1.323 ·

Clearer wording for tool calls that may not have run

When a tool call has no result, the model is now told plainly it may not have run, with advice to retry or check first

You'll notice Improvements
JSON All of v2.1.323
You'll noticeTier: how much it should matter to you
1Useful: my rating, 1 to 5
1Signal: worth watching, 1 to 5
Tool CallsArea: what it touches
ImprovementsKind: in v2.1.323,
ImprovementsSection of the release
What

A tool call is a single action the model asks Claude Code to take, such as running a command. When one ends with no result, for example after an interruption, the model now gets one of two messages:

  • "As far as this session knows, this call did not run."
  • "This call has no result. It may not have run, or not completely."

Claude Code then adds advice either to call it again or to check first, unless that advice is switched off for the case at hand.

Why

What the model is told after an interrupted action shapes whether it repeats it. The second wording warns that the action may have partly happened, so the model is steered to check before running it again.

See this entry in the whole of v2.1.323 →

Feedback