[Release] anthropics/claude-code v2.1.169: v2.1.169

Claude Code v2.1.169 introduces safe mode for troubleshooting, a `/cd` command for directory switching without breaking prompt cache, and a `disableBundledSkills` setting to control skill visibility. The release includes numerous bug fixes for MCP policy enforcement, UI responsiveness, remote session handling, and background agent reliability. Performance improvements reduce CPU usage during streaming and fix stalls on macOS and Windows.

Key Points

• •Added `--safe-mode` flag to disable all customizations (CLAUDE.md, plugins, skills, hooks, MCP servers) for troubleshooting
• •New `/cd` command allows changing working directory mid-session without breaking prompt cache
• •Added `disableBundledSkills` setting and environment variable to hide bundled skills and slash commands from the model
• •Fixed MCP policy enforcement (`allowedMcpServers`/`deniedMcpServers`) on reconnect, IDE configs, and first-session scenarios
• •Resolved ~30-50ms UI stall on macOS for users with claude.ai credentials and slow `claude -p` on Windows
• •Fixed Remote Control reconnection issues after OAuth token refresh and stale permission prompts on remote session reattach
• •Improved background agent reliability: now preserve flags across retire→wake cycles and respect project-level settings
• •Enhanced error messages for API key authentication and improved TaskCreate validation with automatic malformed input repair
• •Reduced CPU usage during response streaming and spinner animations; restored 5-minute idle timeout for Vertex/Foundry
• •Fixed Git Credential Manager popup on Windows startup and improved slash-command menu color contrast for accessibility

Workflow Diagram

Concepts

Artifacts (5)

claude --safe-mode

export CLAUDE_CODE_SAFE_MODE=true

export CLAUDE_CODE_DISABLE_BUNDLED_SKILLS=true

export API_FORCE_IDLE_TIMEOUT=0

/cd <new-directory-path>