[NOTICE] This feed is no longer maintained

2026-04-18 · Source: Claude Code Changelog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Cloud Computing & IT Infrastructure · Depth: Intermediate, long

Summary

The Claude Code changelog details recent updates across versions 2.1.105 through 2.1.117, released between April 13 and April 22, 2026. Key enhancements include the introduction of Claude Opus 4.7 with an "xhigh" effort level and auto mode for Max subscribers, alongside a new `/ultrareview` command for comprehensive cloud-based code review. Performance improvements focus on faster startup times for MCP servers and `/resume` operations, especially for large sessions. Usability features like persistent model selections, improved plugin dependency management, and enhanced terminal UI interactions (e.g., smoother scrolling, better undo behavior) are also highlighted. Security updates include stricter sandbox enforcement for dangerous path removals and improved handling of Bash deny rules. OpenTelemetry events now include more detailed command and effort attributes, and native builds on macOS and Linux utilize embedded `bfs` and `ugrep` for faster searches.

Key takeaway

For NLP Engineers and developers working with Claude Code, these updates offer significant performance gains and new capabilities. You should explore the Opus 4.7 "xhigh" effort level for enhanced model intelligence and integrate the `/ultrareview` command into your code review workflows for more comprehensive analysis. Additionally, review the improved plugin dependency management and security enhancements to ensure your development environment is optimized and secure.

Key insights

Recent Claude Code updates enhance performance, security, and user experience, introducing Opus 4.7 and advanced code review.

Principles

• Prioritize performance for large sessions.
• Enhance security through stricter sandbox rules.
• Improve user experience with persistent settings.

Method

The platform integrates new model capabilities (Opus 4.7 xhigh, auto mode) and introduces multi-agent analysis for code review via `/ultrareview`, while optimizing core operations like session resume and plugin management.

In practice

• Use `/effort xhigh` for Opus 4.7.
• Run `/ultrareview` for cloud code analysis.
• Set `CLAUDE_CODE_FORK_SUBAGENT=1` for external builds.

Topics

• Claude Code
• Opus 4.7
• Plugin Management
• Performance Optimization
• Security Enhancements

Best for: NLP Engineer, AI Engineer, Machine Learning Engineer, Software 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.
• v2.1.46 — Claude Code updates focus on performance, new models, and expanded functionality through plugins and agent capabilities.
• 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.
• 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.

Open in AIssential →

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