Agent DailyAgent Daily
releaseintermediate

[Release] anthropics/claude-code v2.1.121: v2.1.121

By ashwin-antgithub
View original on github

Claude Code v2.1.121 introduces significant usability and performance improvements, including a new `alwaysLoad` MCP option for tool availability, enhanced plugin management with pruning, and improved terminal UI with scrollable dialogs and better URL handling. The release addresses critical memory leaks affecting image processing and session history, fixes OAuth and authentication issues across multiple platforms, and adds support for Workload Identity Federation and custom redirect URIs. Multiple terminal rendering bugs are resolved, and performance optimizations reduce file descriptor usage and startup time.

Key Points

  • Added `alwaysLoad` option to MCP server config to make all tools from a server always available, skipping tool-search deferral
  • Introduced `claude plugin prune` command to remove orphaned auto-installed plugin dependencies with cascading uninstall support
  • PostToolUse hooks can now replace tool output for all tools via `hookSpecificOutput.updatedToolOutput`, extending beyond MCP-only functionality
  • Fixed critical unbounded memory growth (multi-GB RSS) when processing many images and leaks in `/usage` command affecting large transcript histories
  • Enhanced terminal UI: dialogs are now scrollable with arrow keys/PgUp/PgDn, fullscreen mode prevents scroll jumping, and wrapped URLs are clickable
  • Added type-to-filter search box to `/skills` for easier skill discovery in long lists without scrolling
  • Improved MCP server resilience with auto-retry up to 3 times on transient startup errors instead of staying disconnected
  • Fixed Bash tool becoming permanently unusable when the startup directory is deleted or moved mid-session
  • Added support for X.509 certificate-based Workload Identity Federation (mTLS ADC) for Vertex AI authentication
  • Resolved OAuth and authentication issues including Microsoft 365 MCP duplicate prompt parameters and claude.ai connector silently disappearing on transient errors

Found this useful? Add it to a playbook for a step-by-step implementation guide.

Workflow Diagram

Start Process
Step A
Step B
Step C
Complete
Quality

Concepts