Claude Code Update: Terminal Layout Controls, Telemetry Diagnostics, and Session Stability Fixes

What changed#

The latest release introduces several configuration controls, diagnostic utilities, and targeted patches for conversation state management.

Terminal and interface adjustments include a new maxProseWidth setting that constrains Claude’s prose output in wide displays while preserving full-width rendering for tables and code blocks. The /feedback drafts view now displays a hover-activated scrollbar when operating in fullscreen mode.

Operational visibility improves through startup notifications and updated /status and claude doctor outputs, which now enumerate telemetry variables present in project settings that were either ignored or actively disabled. For Chrome integration, the allowClaudeInChromeWithManagedMcp managed setting permits claude --chrome to coexist with an exclusive managed-mcp.json, and error messaging now explicitly identifies Chrome-blocking conditions.

Infrastructure reliability receives a bump via store.readiness_grace_seconds, added to the Claude apps gateway to keep /readyz endpoints healthy during brief Postgres failovers.

Session continuity sees extensive refinement. Conversations containing undecryptable web search results — typically injected through third-party gateways — no longer trigger 400 request failures. The --continue and --resume flags have been patched to stop resending historical messages in altered formats, a behavior that previously caused the API to discard prior reasoning chains. Extended thinking now persists correctly when developers invoke immediate slash commands like /model, /rename, or /artifacts mid-session, and survives relaunches even when a new --tools argument excludes built-in capabilities used earlier.

Error handling and recovery logic has been tightened. When the API returns an “Invalid data in redacted_thinking block” response, Claude Code now strips the corrupted thinking payload and automatically retries the request once. Compaction failures triggered by refused summarization calls gracefully fall back to an alternate model. The “Effort xhigh isn’t available with thinking turned off” crash following a safety-driven model switch is addressed as a standalone patch without additional retry layers. In SDK-hosted environments like Claude Desktop, unanswered Fable usage-credits prompts no longer force unwanted model switches; the turn now terminates cleanly, and Remote Control clients receive the corresponding notice.

Authentication, policy enforcement, and permission parsing received multiple corrections. Stale login refresh locks no longer block operations for up to a minute after a background process terminates. Sessions spawned during active sign-in refreshes now properly retry organization policy fetches. macOS symlink resolution for CLAUDE.md and rule files across /Network, /.vol, and /home paths is corrected. Bash permission rules containing a mid-pattern :* now apply uniformly across all configuration sources, paired with a startup warning explaining the match behavior. Managed settings now validate boolean lock keys like disableClaudeAiConnectors individually rather than failing silently, and reject partial invalid blocks without discarding valid nested values. Repository, user, and --add-dir skill manifests can no longer self-preapprove tools via allowed-tools under allowManagedPermissionRulesOnly. Amazon Bedrock and Bedrock Mantle safeguard blocks now emit message IDs alongside request IDs. Finally, Vertex AI web search support extends to newly released models previously unrecognized by the client.

Why it matters#

These updates shift focus from feature expansion to workflow stability. The session continuity patches directly mitigate context loss and silent API drops that frequently break long-running coding agent loops. By allowing extended thinking to survive slash command invocations and tool-list mismatches, developers can chain complex reasoning steps without manual context reconstruction. The diagnostic improvements (/status, claude doctor, startup notices) give automation engineers explicit visibility into configuration drift and telemetry gaps that historically caused opaque behavior. Permission validation hardening prevents accidental privilege escalation or silent config overrides in managed environments, while the refined retry logic for thinking blocks and compaction reduces the operational overhead required to recover from transient API state corruption.

What to do next#

1. Audit your project configurations using claude doctor and /status to identify flagged telemetry variables or misconfigured managed settings.
2. Test maxProseWidth in your preferred terminal emulator to balance readability against screen real estate.
3. Validate managed policy enforcement by intentionally introducing mistyped boolean lock keys (e.g., disableClaudeAiConnectors) to confirm granular rejection behavior.
4. Verify Fable integration in SDK-hosted setups to ensure usage-credit prompts terminate cleanly without triggering unexpected model switches.
5. If operating on macOS, review CLAUDE.md and rule file paths for legacy symlink references pointing to /Network, /.vol, or /home.
6. Update to this release to capture the gateway readiness grace period, improved Vertex AI model recognition, and the consolidated permission parsing logic.