Skip to content

Claude Code

Claude Code v2.1.169 update

Written by an AI model from the linked sources and checked against them by a second model. How we work

All about: Claude Code versions and changes →

The latest version introduces a --safe-mode flag to launch Claude Code with all custom settings disabled, including CLAUDE.md, plugins, and MCP servers. This streamlines error diagnosis.

The /cd command now allows directory changes without resetting the prompt cache. A new disableBundledSkills setting hides built-in skills and commands.

The update resolves several issues. The interface lag on macOS is fixed, as is the claude -p command hang on Windows. Problems with MCP policy application in enterprise environments are resolved. Background sessions now save launch parameters properly, and CPU usage during response streaming is reduced.

AI security

Miasma attack targets AI assistant configurations

Opening a repo with Claude Code or Cursor now carries the same risks as running an executable. The Miasma attack uses AI tool config files to auto-run malicious code when a session starts. The malware steals cloud keys, passwords, and access tokens immediately after the assistant connects to a project folder.

No comments yet. Be the first.

Join the conversation

Sign in to leave a comment. It appears after moderation.