[Release] anthropics/claude-code v2.1.166: v2.1.166

Claude Code v2.1.166 introduces fallback model configuration for improved reliability, adds glob pattern support for tool access control, and hardens cross-session messaging security. The release includes numerous bug fixes for terminal compatibility, image processing, remote sessions, and managed settings enforcement. Key improvements enhance user experience across JetBrains IDEs, terminal protocols, and background agent sessions.

Key Points

• •Configure up to three fallback models via `fallbackModel` setting to automatically retry when primary model is overloaded or unavailable
• •Use glob patterns in deny rules (e.g., `"*"` denies all tools) for flexible tool access control; allow rules reject non-MCP globs
• •Cross-session messaging now strips user authority from relayed messages; receivers refuse relayed permission requests and auto mode blocks them
• •Disable thinking on Claude API models using `MAX_THINKING_TOKENS=0`, `--thinking disabled`, or per-model toggle; 3P providers unaffected
• •API error handling improved: retries once on fallback model for unexpected non-retryable errors; auth/rate-limit/transport errors surface immediately
• •`claude update` now announces target version before downloading for better user feedback
• •Fixed recurring 'image could not be processed' error and associated token waste from unprocessable images
• •Remote sessions no longer permanently stuck after brief backend disruptions during worker registration
• •JetBrains IDE terminal flickering (2026.1+) fixed via synchronized output; Kitty protocol Shift+non-ASCII character dropping resolved
• •PowerShell command validation hanging and orphaned daemon processes spinning at 100% CPU on macOS both fixed

Workflow Diagram

Concepts