Claude Code Update: Introduces Workspace Trust Prompt and Spend-Limit Warnings

ClaudeCodeLog · x · 2026-08-08

Anthropic's CLI AI coding assistant, Claude Code, has released version 2.1.225. The update introduces a workspace trust prompt, requiring agents to verify safety when running in untrusted directories to enhance code execution security.

Additionally, it improves gateway spend-limit warnings. When the usage cap is reached, the system now clearly displays the specific limit, reset time, and operator messages. Under the hood, prompt files saw an approximate 11.6% increase in token count, and the behavior logic for the edit tool to perform exact string replacements has been further clarified.

Related event: Claude Code CLI Update Introduces Workspace Trust and Editor Fixes(2 posts)→

Original post →

More from coding & agent

coding & agent channel →