Claude Code Adds ultrareview Subcommand: Automated Code Review in CI/CD

Claude Code is no longer just an interactive coding assistant. The latest CLI 2.1.120 release adds a key feature: the ultrareview subcommand, enabling it to run automated code review in CI/CD pipelines.

Core Changes

Previously, Claude Code’s code review capability was only available through interactive sessions. ultrareview changes this:

# Run in CI scripts
claude ultrareview ./src --json --output review.json

# Review results output as structured JSON, parseable by downstream tools
# Exit code: 0 means completed, 1 means serious issues found

Key features:

• Non-interactive mode: No human participation needed, suitable for CI environments
• JSON output: --json flag outputs structured results, integrable with GitHub Actions, GitLab CI
• Target specification: Supports files, directories, or entire projects as review scope
• Exit code signals: 0 for completion, 1 for issues needing fixes

CI Integration

A typical GitHub Actions integration:

name: Claude Code Review
on: [pull_request]
jobs:
  review:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Install Claude Code
        run: npm install -g @anthropic-ai/claude-code
      - name: Run Review
        run: claude ultrareview . --json --output review.json
        env:
          ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}

Cost Considerations

ultrareview is billed per token. A full review of a medium project (~500 files) is estimated to consume hundreds of thousands to millions of tokens, costing a few to tens of dollars. Recommended to enable on critical paths first, then expand.

Sources

• Claude Code CLI Changelog
• Claude Code Docs