The Claude Code Changelog

Learn how Claude Code version 2.1.317 introduces onFailure: "block" for programmatic Agent SDK hook callbacks, preventing unattended agents from failing open on unexpected errors. We also cover protections against account email leakage in tools, isolated transcripts for nested headless sessions, and full-file re-reading behavior.

Learn how to use Claude Code's blockWritesOutsideWorkingDirectories setting to lock down file write boundaries during unattended or headless runs across every permission mode.

In Claude Code 2.1.316, server-managed settings for Team and Enterprise logins now apply automatically without an approval prompt, alongside verification checks for cached settings. We examine how this impacts developer visibility into remote hooks and shell-command settings, and share tips for auditing your active configurations with claude doctor.

Show Notes

• Releases · anthropics/claude-code - GitHub: https://github.com/anthropics/claude-code/releases

Learn how to decouple worker reasoning from main sessions using the new CLAUDE_CODE_SUBAGENT_EFFORT environment variable in Claude Code 2.1.316. We also cover disabling ultrathink, blocking external file writes, and the timeline for Opus 4.8 fast mode deprecation.

Explore the updates in Claude Code version 2.1.313, focusing on subagent tool restrictions through explicit allow and disallow lists to limit blast radius and save tokens. Learn how the new environment variable allowlist secures worker environments in headless and CI workflows, along with key bug fixes for nested Git worktrees and status bar metrics.

Learn how Claude Code 2.1.311's new self-hosted runner idle-hold flags prevent background processes and dev servers from being terminated prematurely. We explore how to configure these settings, avoid common timing misconfigurations, and handle teardown caveats.

Show Notes

• Self-hosted environments reference - Claude Code Docs: https://code.claude.com/docs/en/self-hosted-environments-reference

Learn about changes in Claude Code 2.1.312 that alter default system prompts and refusal fallback behaviors for custom model deployment names on Bedrock, Vertex, Foundry, and Mantle. We also cover crucial updates to hook concurrency limits, subagent compaction settings, and a fix preventing accidental overwrites of large CLAUDE.md files.

In Claude Code 2.1.311, the ultrathink prompt keyword now actively raises the API effort parameter to xhigh for interactive terminal sessions. We also cover how compaction handles requests over 200K tokens, new telemetry behavior in VS Code, and MCP tool availability for subagents.

Discover how a prompt update in Claude Code 2.1.310 clarifies the impact of denying reads outside your working directory during auto mode, plus how to manage trusted paths with settings and the add dir flag. Also covered: fixes for headless MCP timeouts, tmux image rendering, and cloud setup script controls.

Claude Code 2.1.309 and 2.1.310 tighten sandbox and permission boundaries, including fixes for Windows hooks that could fail open. Learn how to check your version, test a blocking hook and an Edit deny rule, and diagnose Ubuntu sandbox startup issues with /sandbox.

Claude Code 2.1.310 enables the Monitor tool on Amazon Bedrock, Google Vertex, and Microsoft Foundry—even with telemetry or nonessential traffic disabled. Learn how to replace sleep-and-check loops with event-driven monitoring, configure permissions, and avoid leaving background monitors running.

The episode also covers gateway retry behavior, a compaction fix, clearer symlink permission prompts, Linux sandbox guidance, and custom Enterprise roles for cloud sessions and Remote Control.

Claude Code 2.1.309 adds a workspace-trust check to claude ultrareview. Learn how to handle the prompt, update scripts and CI with the correctly ordered --trust-workspace flag or the environment variable, and weigh the security and billing caveats. We also cover smaller updates to dangerous-command handling, sed -i, large nested CLAUDE.md files, sandboxed Bash startup files, and VS Code diagnostics.

Show Notes

• claude-code/CHANGELOG.md at main - GitHub: https://github.com/anthropics/claude-code/blob/main/CHANGELOG.md

Claude Code 2.1.307 adds the claude_code.safety_event OpenTelemetry event, letting teams track when permissions, safety checks, hooks, or the sandbox stop, prompt, or send back a tool call. Learn how to query the event for sandbox violations and recurring permission prompts, plus caveats about interpreting the counts. We also cover the release’s MCP retry and authentication fixes, checkout-hook timeouts, and pull-request review controls.

Claude Code 2.1.305 adds sandbox.allowNestedUserNamespaces for Linux and WSL2, enabling tools such as headless Chromium and rootless Podman to run inside the Bash sandbox. Learn how to enable it and why it weakens the sandbox’s file protections. The hosts also cover fixes and quality-of-life updates in 2.1.304–2.1.306, including gateway retries, /usage fixes, and retry-watchdog configuration.

Claude Code 2.1.304 expands /auto-mode-setup to more provider and gateway sessions, helping you draft trusted environment rules from project files and recent activity. Learn what the setup scans, how to review and verify its all-at-once settings changes, and why managed settings are safer for team-wide rules. Plus, a few useful fixes and quality-of-life updates from the release.

How Claude Code’s experimental /advisor pairs a lower-cost executor with a stronger model that can review the full session and offer guidance. Learn how to enable it, prompt for consultations, inspect advice, and check provider, model-pairing, and cost caveats. We also cover smaller improvements in versions 2.1.301–2.1.303.

Claude Projects is becoming a coordinator for parallel Claude Code cloud sessions: learn how to connect repositories, set project instructions, choose thread models and effort, and manage local-only work. We also cover practical limits and cost traps, including paused sandboxes, pull-request follow-up threads, and usage-limit resumption, plus smaller fixes in 2.1.301 and 2.1.302.

Claude Code 2.1.302 adds CLAUDE_CODE_AUTO_MODE_MODEL, letting users on proxy, Bedrock, or Vertex setups choose the model used by the client-side auto-mode classifier. The episode also covers a breaking default for headless sessions—explicitly set --permission-mode auto in CI—as well as server-side auto-mode compaction and smaller updates to subagents, idle compaction, hooks, image handling, and plugin startup.

Claude Code 2.1.301 adds CLAUDE_CODE_RETRY_WATCHDOG_MAX_USAGE_LIMIT_WAIT_MS, letting CI jobs fail quickly or switch to a fallback model instead of idling through a distant quota reset. Learn how to set it alongside the general retry wait cap, where to scope the watchdog variables, and what “per API request turn” means. Also covered: richer OpenTelemetry tool-decision events, safer worktree cleanup, clearer Grep and Glob errors, and several usability fixes.

Claude Code 2.1.300 adds step-aware workflow routing: mods can rewrite an agent’s prompt and choose its model, letting teams balance cost and capability across a workflow. We cover model fallbacks and allowlists, stricter hook validation, practical safeguards, and several smaller fixes—including clearer cache metrics and improved MCP tool search behavior.