[Release] anthropics/claude-code v2.1.77: v2.1.77

Claude Code v2.1.77 introduces significant improvements to token limits for Claude Opus and Sonnet models (up to 128k), enhances sandbox filesystem controls with `allowRead` settings, and fixes numerous critical bugs affecting bash commands, memory management, API tracking, and terminal UI interactions. The release also improves performance with faster startup times and session resumption, better plugin validation, and refined agent communication workflows.

Key Points

• •Increased default max output tokens to 64k for Claude Opus 4.6 and upper bound to 128k for both Opus 4.6 and Sonnet 4.6 models
• •Added `allowRead` sandbox filesystem setting to selectively re-enable read access within `denyRead` regions
• •Enhanced `/copy` command to accept optional index parameter for copying specific assistant responses (e.g., `/copy N`)
• •Fixed compound bash command permission handling to create per-subcommand rules instead of single rules for full strings
• •Resolved auto-updater memory leak causing overlapping binary downloads and accumulating gigabytes of memory
• •Fixed `--resume` race condition that silently truncated recent conversation history
• •Fixed PreToolUse hooks bypassing `deny` permission rules and enterprise managed settings
• •Corrected Write tool line-ending conversion issues with CRLF files and directories
• •Improved performance: ~60ms faster macOS startup and up to 45% faster `--resume` with ~100-150MB less peak memory
• •Renamed `/fork` to `/branch` (with `/fork` remaining as alias) and improved agent communication with auto-resume functionality

Workflow Diagram

Concepts