Claude-Code v2.1.92 adds Bedrock setup wizard, cost breakdowns, and multiple fixes

What's new in Claude-Code v2.1.92
Claude-Code v2.1.92 brings several practical improvements for developers using this AI coding agent, focusing on AWS Bedrock integration, cost transparency, and bug fixes.
Key features and changes
New policy setting: Added forceRemoteSettingsRefresh policy setting that blocks CLI startup until remote managed settings are freshly fetched, exiting if the fetch fails (fail-closed).
AWS Bedrock integration: Added interactive Bedrock setup wizard accessible from the login screen when selecting "3rd-party platform" — guides you through AWS authentication, region configuration, credential verification, and model pinning.
Cost transparency: Added per-model and cache-hit breakdown to /cost for subscription users. Pro users now see a footer hint when returning to a session after the prompt cache has expired, showing roughly how many tokens the next turn will send uncached.
Interface improvements: /release-notes is now an interactive version picker. Remote Control session names now use your hostname as the default prefix (e.g. myhost-graceful-unicorn), overridable with --remote-control-session-name-prefix.
Bug fixes and performance improvements
- Fixed subagent spawning permanently failing with "Could not determine pane count" after tmux windows are killed or renumbered during a long-running session
- Fixed prompt-type Stop hooks incorrectly failing when the small fast model returns
ok:false, and restoredpreventContinuation:truesemantics for non-Stop prompt-type hooks - Fixed tool input validation failures when streaming emits array/object fields as JSON-encoded strings
- Fixed an API 400 error that could occur when extended thinking produced a whitespace-only text block alongside real content
- Fixed accidental feedback survey submissions from auto-pilot keypresses and consecutive-prompt digit collisions
- Fixed misleading "esc to interrupt" hint appearing alongside "esc to clear" when a text selection exists in fullscreen mode during processing
- Fixed Homebrew install update prompts to use the cask's release channel (
claude-code→ stable,claude-code@latest→ latest) - Fixed
ctrl+ejumping to the end of the next line when already at end of line in multiline prompts - Fixed an issue where the same message could appear at two positions when scrolling up in fullscreen mode (iTerm2, Ghostty, and other terminals with DEC 2026 support)
- Fixed idle-return "/clear to save X tokens" hint showing cumulative session tokens instead of current context size
- Fixed plugin MCP servers stuck "connecting" on session start when they duplicate a claude.ai connector that is unauthenticated
- Improved Write tool diff computation speed for large files (60% faster on files with tabs/ & / $)
Removed features
Removed /tag command. Removed /vim command (toggle vim mode via /config → Editor mode).
Linux sandbox improvements
Linux sandbox now ships the apply-seccomp helper in both npm and native builds, restoring unix-socket blocking for sandboxed commands.
📖 Read the full source: GitHub Claude-Code
👀 See Also

Gemini 3.1 Flash Live: Google's latest audio model with improved benchmarks and watermarking
Google released Gemini 3.1 Flash Live, an audio model scoring 90.8% on ComplexFuncBench Audio and 36.1% on Scale AI's Audio MultiChallenge. It's available via Gemini Live API in Google AI Studio and includes SynthID watermarking.

Anthropic restricts Claude subscription usage on third-party tools like OpenClaw
Anthropic is changing its Claude subscription policy to block usage on third-party harnesses including OpenClaw, requiring separate pay-as-you-go billing for these tools starting April 4. The company is offering a one-time credit equal to monthly subscription price and pre-purchase discounts up to 30%.
Google DeepMind's AI Pointer: Reimagining the Mouse for Gemini Interactions
Google DeepMind introduces an AI-powered mouse pointer that uses Gemini to understand context, enabling commands like pointing at an image and saying 'Show me directions,' integrated into Chrome and Googlebook.

OpenClaw Agents Compete in AI-Only Pokémon Red League
A new platform called AgentMonLeague allows autonomous OpenClaw agents to connect to a Pokémon Red emulator, make their own decisions through a full playthrough, and compete to finish the game first. Runs are viewable live as agents progress.