[Release] anthropics/claude-code v2.1.88: v2.1.88

Claude Code v2.1.88 is a maintenance release focused on stability, performance, and bug fixes across rendering, tool execution, and session management. Key improvements include flicker-free rendering, prompt cache fixes for long sessions, structured output schema reliability, and numerous UI/UX enhancements. The release addresses critical issues like memory leaks, file handling on Windows, voice mode permissions, and terminal compatibility across platforms.

Key Points

• •Enable flicker-free rendering with `CLAUDE_CODE_NO_FLICKER=1` environment variable for improved alt-screen experience
• •New `PermissionDenied` hook allows models to retry denied auto-mode commands by returning `{retry: true}`
• •Fixed prompt cache misses in long sessions caused by tool schema bytes changing mid-session
• •Resolved nested CLAUDE.md re-injection bug that occurred dozens of times in long file-reading sessions
• •Fixed Edit/Write tools on Windows: eliminated CRLF doubling and restored Markdown hard line break handling (two trailing spaces)
• •Corrected `StructuredOutput` schema cache bug causing ~50% failure rate in multi-schema workflows
• •Eliminated memory leak where large JSON inputs were retained as LRU cache keys in long-running sessions
• •Prevented out-of-memory crashes when Edit tool processes very large files (>1 GiB)
• •Fixed LSP server zombie state — server now auto-restarts on next request instead of requiring session restart
• •Improved error messaging: 'Rate limit reached' now shows actual API entitlement errors with actionable hints

Workflow Diagram

Concepts

Artifacts (3)

CLAUDE_CODE_NO_FLICKER=1

PermissionDenied hook:
  - Fires after auto mode classifier denials
  - Return {retry: true} to allow model retry
  - Enables conditional command re-execution

{
  "showThinkingSummaries": true
}