Claude Code v2.1.31
Released Feb 4, 2026 (6 months ago).
221 issues reference this build
4 still open
217 resolved (98%)
"References this build" means the issue body mentions running 2.1.31 (from the
bug-report template). It reflects report volume, not a proven defect count — an issue may mention a
version without being caused by it. First-party data from the anthropics/claude-code tracker.
Top issues reported on v2.1.31
Ranked by community engagement (reactions and comments).
- [BUG] macOS sandbox blocks Security.framework TLS verification, breaking Go binaries (gh, terraform, etc.)
- [BUG] /resume is broken — session history inaccessible since v2.1.31 (3 root causes identified)
- /compact fails with 'Conversation too long' when context limit is reached
- [Bug] TypeError in insights command: Object.keys() called on undefined value
- [Bug] Claude Code TUI completely freezes on Windows startup without input
- [Bug] File autocomplete suggestions disappear when navigating subdirectories with @ syntax
- [BUG] CLAUDE.md/memory file edits show "Writing memory" without diff - cannot see what changed
- [BUG] Repeated API 500 "Internal server error" during single Claude Code session (3 occurrences)
- [BUG] SessionStart hooks break stdin on Windows Terminal
- [BUG] sessions-index.json stops being updated after v2.1.31 — new sessions invisible to /resume
- User interrupt during streaming can freeze whitespace-only text block into history, causing persistent 400 errors
- [BUG] Memory saved to undocumented MEMORY.md instead of CLAUDE.md
- Claude in Chrome Extension Not Connecting on Windows 11
- [BUG] Named sessions not appearing in /resume picker
- Bug: HTTP MCP client ignores no_proxy environment variable
- [Bug] Anthropic API Error: Internal Server Error (500)
- [BUG] Ctrl+V in /resume conflicts with Windows paste - cannot be rebound
- [BUG] /feedback cancellation upon "N" symbol
- Claude Code excessive limits usage in recent updates. Limits have either been reduced, or Claude Code is incorrectly counting tokens
- claude resume shows empty list + Feature request: session naming
- [BUG] Getting incorrect "exceeded the 32000 output token maximum"
- [Bug] Anthropic API Error: Internal server error (500)
- [BUG] CLAUDE.md edits show collapsed "memory" UI instead of inline diffs like other files
- [Bug] classifyHandoffIfNeeded error affects BUILT-IN agents (Explore, Plan) - not just custom subagents
- api Errors: 500
- Claude-in-Chrome MCP: extension executes but CLI receives 'not connected' error (Windows)
- [Bug] Session resume with -c/-r flags causes input to freeze
- Bug: /resume command shows incomplete session list
- [Bug] Subagent Skills Leak into Main Context After Compaction - extremely costly bug that breaks everything
- [BUG] /feedback command exits prematurely saying "feedback canceled"
Changelog
What's changed
- Added session resume hint on exit, showing how to continue your conversation later
- Added support for full-width (zenkaku) space input from Japanese IME in checkbox selection
- Fixed PDF too large errors permanently locking up sessions, requiring users to start a new conversation
- Fixed bash commands incorrectly reporting failure with "Read-only file system" errors when sandbox mode was enabled
- Fixed a crash that made sessions unusable after entering plan mode when project config in
~/.claude.jsonwas missing default fields - Fixed
temperatureOverridebeing silently ignored in the streaming API path, causing all streaming requests to use the default temperature (1) regardless of the configured override - Fixed LSP shutdown/exit compatibility with strict language servers that reject null params
- Improved system prompts to more clearly guide the model toward using dedicated tools (Read, Edit, Glob, Grep) instead of bash equivalents (
cat,sed,grep,find), reducing unnecessary bash command usage - Improved PDF and request size error messages to show actual limits (100 pages, 20MB)
- Reduced layout jitter in the terminal when the spinner appears and disappears during streaming
- Removed misleading Anthropic API pricing from model selector for third-party provider (Bedrock, Vertex, Foundry) users