review

Reviews local code changes for defects, risks, and missing tests before pull request submission.

1|Updated May 15, 2026
One-click install
npx skills add https://github.com/mslshao/claude-harness --skill review-mslshao
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review
Source: https://github.com/mslshao/claude-harness/tree/main/project-tier/skills/review
Command: npx skills add https://github.com/mslshao/claude-harness --skill review-mslshao

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you self-review local code changes before opening or pushing a pull request. It reduces avoidable review churn by surfacing likely bugs, test weaknesses, observability gaps, and silent failure risks directly from your branch or staged diff.

Core Features & Use Cases

  • Parallel multi-agent review: Fans out one local diff to specialized reviewers for code quality, test quality, observability, and error propagation concerns.
  • Smart conditional routing: Always runs structural code review, while only invoking test, observability, and silent-failure specialists when the diff contains relevant signals.
  • Deduplicated severity reporting: Merges overlapping findings, preserves file:line citations, and returns a grouped terminal-friendly report for fast triage.
  • Use case: Run it before creating a PR to catch missing assertions in tests, risky exception handling, broken monitoring assumptions, or design issues that a careful reviewer or CI would likely flag.

Quick Start

Ask the AI to review your current branch changes with the review skill before you push.

Frequently Asked Questions about review

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

FAQPage Schema
How do I self-review local code changes before opening a pull request?

Self-review local code changes by analyzing branch-relative or staged git diffs to catch bugs, test weaknesses, and observability gaps before opening a pull request. This process grounds findings in full files and project rules to reduce review churn.

What is silent failure detection in a pre-PR code review?

Silent failure detection identifies risky exception handling and broken monitoring assumptions in changed files during a pre-PR code review. It checks error propagation paths to ensure observable failures rather than silent drops.

How to assess test quality and missing assertions in a git diff?

Assess test quality in a git diff by routing changes to specialized reviewers that check for missing assertions and test weaknesses. This conditional review only triggers when diff signals indicate test files are affected.

Can I run a local code review on staged changes without pushing to a remote repository?

You can run a local code review on staged changes without pushing by collecting git diff and git log context in a read-only workflow. The review operates entirely locally on explicit ranges or branch-relative diffs.

Does this pre-PR review work with CI pipelines or replace them?

This pre-PR review surfaces design issues, observability gaps, and silent failure risks that a careful reviewer or CI would likely flag. It works locally before pushing to reduce avoidable review churn rather than replacing CI.