Claude Code 2.1.267
The release adds a maxEffortLevel setting to cap reasoning effort across providers, plus fixes for Cowork, mobile rendering and tmux reconnects.
The latest update to Claude Code adds a maxEffortLevel setting, configurable at the top level or per model under modelSettings, which caps the effort level used on every provider, including Bedrock, Vertex and Foundry. Users can still choose a lower level themselves. A new "--system-prompt-snapshot off" option renders the system prompt fresh on every request instead of reusing the conversation's recorded prompt, for iterating on prompt text.
The release also includes several fixes:
- Cowork scheduled tasks in the cloud no longer fail at startup for organizations whose managed settings require sandboxing.
- /context and other local command output no longer render blank on mobile clients.
- Shift+enter and option+backspace now work again after reconnecting to a tmux or ssh session inside an agent view.
- The dim last-prompt header now appears at the top of the conversation when scrolling up in fullscreen mode.
- Workflow agent() calls with large output schemas are now checked by the safety classifier instead of being refused outright in auto mode.
- A marketplace entry path containing a backslash could bypass the containment check for fetched marketplaces on macOS and Linux; this is now fixed.
- Expired AWS or Google Cloud credentials under a host app such as Claude Desktop no longer retry ten times with a generic error before failing.
Claude Code
Anthropic's agentic coding tool — Claude in your terminal, IDE, desktop app and browser, reading and editing your codebase, running commands, opening pull requests and working on tasks in the cloud.
View Claude Code →