v2.1.105

2025-02-22 · Source: Claude Code Changelog · Field: Technology & Digital — Software Development & Engineering, Artificial Intelligence & Machine Learning, Cybersecurity & Data Privacy · Depth: Intermediate, long

Summary

Recent updates to Claude Code, spanning versions 2.1.105 through 2.1.114, introduce significant enhancements across user experience, security, and model capabilities. Version 2.1.114 primarily addressed a crash in the permission dialog. Version 2.1.113 brought a native Claude Code binary for the CLI, added `sandbox.network.deniedDomains` for granular network control, and improved UI elements like fullscreen scrolling and multiline input. Security was a major focus, with stricter Bash deny rules, safer handling of `find -exec`, and enhanced macOS path protections. Version 2.1.111 notably introduced Claude Opus 4.7 xhigh, offering new effort levels and auto mode for Max subscribers. It also added `/ultrareview` for parallelized code review and `/less-permission-prompts` for generating allowlists. Earlier updates in 2.1.110 and 2.1.108 included a `/tui` command for flicker-free rendering, push notification capabilities, and a recap feature for session context, alongside numerous bug fixes and performance improvements.

Key takeaway

For CTOs or VPs of Engineering evaluating AI-assisted development tools, the introduction of Claude Opus 4.7 xhigh and features like `/ultrareview` in Claude Code 2.1.111 offers significant advancements in performance and code quality assurance. Your teams can now leverage more intelligent models and automated, parallelized code review, potentially streamlining development workflows and enhancing security posture through improved sandbox controls. Consider integrating these updates to optimize resource allocation and accelerate project delivery.

Key insights

Recent Claude Code updates enhance security, user experience, and model capabilities, notably introducing Opus 4.7 xhigh and advanced review tools.

Principles

• Granular control over network access enhances security.
• Parallelized multi-agent analysis improves code review efficiency.

Method

The `/ultrareview` skill performs comprehensive cloud-based code review using parallel multi-agent analysis and critique, invokable for current branches or specific GitHub PRs.

In practice

• Use `sandbox.network.deniedDomains` to block specific network access.
• Employ `/effort xhigh` for Opus 4.7 to balance speed and intelligence.
• Run `/ultrareview` to automate comprehensive code analysis.

Topics

• Claude Opus 4.7
• Code Review
• CLI Enhancements
• Sandbox Security
• Permission Management

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Engineer, Software Engineer, Machine Learning Engineer

Related on AIssential

• v2.1.110 — Recent Claude Code updates enhance security, user experience, and model capabilities, notably with Opus 4.7 xhigh and advanced code review.
• v2.1.109 — Claude Code updates enhance security, user experience, and model capabilities, notably with Opus 4.7 xhigh and improved CLI.
• Introducing Claude Opus 4.7 — Claude Opus 4.7 offers significant advancements in software engineering, multimodal understanding, and long-running task consistency with enhanced safety.
• Best practices for using Claude Opus 4.7 with Claude Code — Opus 4.7 optimizes agentic coding through adaptive thinking, calibrated effort levels, and refined interaction patterns.
• Claude Opus 4.7: Most Powerful Coding Model Ever! Beats EVERYTHING! (Fully Tested) — Claude Opus 4.7 significantly advances coding, web development, and vision, but at a higher token cost.
• OPUS 4.8!!! (also maybe GPT5.6??) — Opus 4.8 enhances AI agentic capabilities and cost-efficiency through parallel processing and optimized inference.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Claude Code Changelog.