Claude Code v2.1.310 Adds Monitor Tool and Fixes Gateway Retry Logic
Claude Code v2.1.310 is out. The headline is the new Monitor tool: it now works on Bedrock, Vertex, and Foundry, in plugin monitors, and when telemetry or nonessential traffic is disabled. In those environments, long foreground sleeps are blocked as well.
Gateway and server compatibility fixes
- Requests that failed behind gateways or compatible servers rejecting the
effort-2025-11-24beta value or mid-conversation system messages now retry without those values instead of failing outright. - If an MCP tool call times out in a non-interactive session while the server's elicitation is unanswered, the error now tells Claude that no answer came back and not to repeat the call.
Cleanup, compaction, and permission fixes
- Background cleanup no longer deletes recent files such as pasted images when
cleanupPeriodDaysis set to a very large value. - Compaction after a "Prompt is too long" error no longer keeps dense text (base64, SVG) that could make the next request fail again.
- File read permission prompts now show where a symbolic link leads.
- The auto-continue timeout for a question no longer fires the moment an external editor closes after a long edit.
- Windows:
PreToolUseandPermissionRequesthooks no longer let a tool call through when Git Bash is missing and the hook command is only quoted strings, such as a quoted path.
Improved behavior
- Auto mode's prompt before the first read outside the working directories now notes when blocking those reads will make shell commands ask for permission.
- Non-interactive sessions (
-p, Agent SDK) use less memory during compaction when the transcript file is very large. - Image rendering in tmux now works with
allow-passthroughandfocus-eventson, plus kitty, Ghostty, and iTerm2 3.5+ fullscreen rendering, where only alt text showed before. - Linux
/sandboxand its startup warning now name Ubuntu's AppArmor restriction on user namespaces and how to lift it when bubblewrap can't start.
VSCode, Cloud sessions, Claude Tag, Code Review
- VSCode: Enter on "Send feedback and keep planning" now sends your comments instead of approving the plan. The panel no longer shows "Something went wrong" after "Switch branch & continue" on a cloud session from Session history's Web tab.
- Cloud sessions: Enterprise custom roles can now grant Claude Code cloud sessions and Remote Control permissions; admins get a confirmation prompt listing roles with those permissions. The environment dialog has a Choose file button for setup scripts, opens network access/setup script/env vars from rows, shows setup script output in a scrolling box, and surfaces save/archive and routine API token errors inside dialogs.
- Claude Tag:
!restartand!helpare no longer treated as ordinary messages when sent in bold, italic, or strikethrough. Admin pages survive failed bundle list refreshes, and spend limits no longer reports no channels while names load. - Code Review: The admin section retries its first load and shows a single Try again banner instead of hiding repositories and switches after a failed refresh.
If you run Claude Code behind an incompatible gateway, in a non-interactive Agent SDK pipeline, or on Linux sandboxes, this release is worth pulling. The retry fix alone removes a class of hard failures.
📖 Read the full source: GitHub Claude-Code
👀 See Also

LLM Spatial Reasoning Tested: Sokoban Benchmark Shows ChatGPT, Qwen3.7-max, Gemini 3.5-thinking Lead
A custom Sokoban benchmark tested zero-shot spatial reasoning in LLMs with strict formatting. Only ChatGPT, Qwen3.7-max, and Gemini 3.5-thinking passed. Models like Gemini 3.5-flash and Qwen3.7-plus failed due to illegal moves or deadlocks.

Bohrium AI Proteomics Competition 2026 with $13K Prize and Compute Support
Bohrium is running an AI proteomics competition in 2026 with a $13,000 prize pool, internship opportunities, and compute support. The competition was discussed on Hacker News with 17 points and 5 comments.

Understanding LLM Directive Weighting: Why Claude Sometimes Ignores Commands
A Reddit investigation reveals how Claude can ignore explicit instructions like "don't pattern match" when generating code reviews, demonstrating that LLM directives are weighted context rather than constraints.

Claude.ai Currently Down, API Errors Elevated — April 28, 2026
An automatic status update triggered from Claude's official status page reports that Claude.ai is unavailable and the API is experiencing elevated error rates as of 2026-04-28T17:51:36.000Z.