skills.rest
ulpi-io
ulpi-ioOFFICIAL
Explore all skills by this author →
GitHub

claude-review

Spawn a Claude Code agent in an isolated worktree to review code changes.

13|Updated Mar 15, 2026
Search Tags:#claude#agent#worktree#quality#review#diff#code
One-click install
npx skills add https://github.com/ulpi-io/browse --skill claude-review-ulpi-io
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: claude-review
Source: https://github.com/ulpi-io/browse/tree/main/.agents/skills/claude-review
Command: npx skills add https://github.com/ulpi-io/browse --skill claude-review-ulpi-io

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Spawn a separate Claude Code agent to independently review code changes across the current session or branch, ensuring an impartial, read-only analysis.

Core Features & Use Cases

  • Independent Claude Code review in an isolated worktree to prevent unintended side effects.
  • Determine review scope: full branch, last commit, or uncommitted changes.
  • Enforce disciplined verification steps before reporting findings to avoid false positives.
  • Outputs structured findings ready for developer action and task traceability.

Quick Start

Ask Claude to review the current branch changes in a separate worktree and report findings.

Frequently Asked Questions about claude-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I perform an independent code review on a git branch using an isolated agent?▼

To perform an independent code review, you can use an agent that spawns in an isolated git worktree to analyze branch diffs without causing unintended side effects to your working directory.

What scope of code changes can a Claude agent review in a worktree?▼

A Claude agent can review varying scopes of code changes including the full branch history, the latest single commit, or current uncommitted changes to produce a structured assessment.

Can I get automated code fixes along with the review findings from the agent?▼

No, the review agent operates in a read-only mode and requires no automated changes, instead outputting structured findings ready for developer action and task traceability.

How does an isolated worktree prevent false positives during AI code analysis?▼

By spawning in an isolated worktree and enforcing disciplined verification steps before reporting findings, the agent ensures an impartial analysis that minimizes false positives.

What is the best way to review uncommitted code changes before creating a pull request?▼

The best way to review uncommitted changes is spawning a read-only Claude Code agent in an isolated worktree to assess the diff and generate structured findings for developer action.

Related Skills

Explore More Software Engineering Skills→
gufrancogufranco
11

commit

Analyze uncommitted Git changes and generate conventional commit messages.

Community
Intermediate
amitmishra-bylddamitmishra-byldd

bmad-code-review

Review code diffs adversarially and triage findings into actionable categories.

Community
Advanced
thisismynewfmail-uithisismynewfmail-ui

requesting-code-review

Verifies code changes before commit via diff analysis, tests, and linters.

Community
Advanced
AnveioAnveio
1

task-pull-request

Automate GitHub pull request creation, reviews, feedback, and merges via gh CLI.

Community
Intermediate