Guide intermediate

Claude Code's New `ultrareview` Command Makes Automated Code Review Effortless

Published April 29, 2026 · Updated May 1, 2026 · by Jonathan Hildebrandt

Claude Code 2.1.120 ships `claude ultrareview`, a new CLI command that runs a deep automated review of your codebase. No IDE required. Here's what it does and how to use it.

Table of Contents
Pondero, operated by Hildebrandt AI LLC, earns a commission from some links on this page. This does not influence our editorial decisions. Read our affiliate disclosure

Claude Code’s New ultrareview Command Makes Automated Code Review Effortless

In brief: Claude Code 2.1.120 (April 28, 2026) adds claude ultrareview, a single CLI command that performs a comprehensive, automated code review across your entire project. If you ship code from the terminal, this is worth 5 minutes of your time today.

What Changed

Anthropic shipped two Claude Code releases on April 28: versions 2.1.120 and 2.1.121. The headline feature of 2.1.120 is claude ultrareview, a new top-level CLI command designed for deep, automated code review.

Before this release, reviewing code with Claude meant either:

  • Opening your IDE and using inline chat prompts, or
  • Feeding files manually to claude in the terminal and crafting your own review prompt

ultrareview replaces that with a single command that knows how to traverse your project, identify review-worthy changes, and surface actionable feedback in a structured format. Run it in any project directory and Claude handles the rest.

The 2.1.120 release also adds Windows PowerShell as a fallback when Git Bash isn’t available. This is a long-requested fix that unblocks Windows developers who previously hit errors during setup.

Meanwhile, 2.1.121 rounds out the same-day drop with MCP server alwaysLoad configuration (so you can force specific MCP servers to load on every session), PostToolUse hooks that can now replace tool output for all tools (not just a subset), and memory leak fixes for image processing.

Why It Matters

Code review is the bottleneck in most small-team workflows. You write code, it sits waiting for review, velocity drops. Automated review tools help, but the good ones (Copilot PR reviews, for example) are tied to your PR workflow. They don’t help you catch issues before you commit.

ultrareview is different: it runs locally, on demand, against whatever state your working tree is in right now. That means you can use it as a pre-commit sanity check, a way to audit a feature branch before opening a PR, or a fast second opinion when you’re working alone and don’t have a reviewer available.

For teams already deep in the Claude Code workflow, this removes an entire class of friction. Instead of context-switching to GitHub to request a Copilot review, you stay in the terminal, run claude ultrareview, get structured feedback, and iterate immediately.

The PowerShell fallback matters too. Windows adoption of Claude Code has lagged behind macOS and Linux partly because of the Git Bash dependency. Removing that barrier opens up a larger developer audience, which is relevant if you’re recommending Claude Code to teams where Windows is common.

How to Use It

Requirements: Claude Code 2.1.120 or later. Run claude --version to check; upgrade with npm install -g @anthropic-ai/claude-code@latest.

Basic usage:

# Run ultrareview against your current working directory
claude ultrareview

# Review a specific subdirectory or file set
claude ultrareview src/

# Output review to a file for sharing
claude ultrareview --output review.md

ultrareview works best when your working tree has meaningful changes. It’s designed to analyze what you’ve been building, not an empty scaffold. Run it after you’ve implemented a feature or fixed a bug, before you commit.

Combining with hooks (2.1.121): If you’ve configured PostToolUse hooks, you can now have them intercept and transform tool output from ultrareview. This is useful if you want to pipe review results into a ticket system or Slack notification automatically.

MCP alwaysLoad (2.1.121): If your review workflow depends on a custom MCP server (say, one that pulls in your team’s coding standards or security rules), you can now mark it with alwaysLoad: true in your MCP config so it’s always available when ultrareview runs.

Pricing: ultrareview is included in all Claude Code plans. Token usage will vary with project size. Expect heavier usage on large codebases.

If you’re evaluating AI coding tools for your workflow, we’ve covered the full landscape:

Claude Code is strongest for developers who spend most of their time in the terminal and want fine-grained control. ultrareview deepens that lead for teams where code quality and review speed matter.

This post is part of Pondero’s daily coverage of AI tool updates. See all guides →