Skill

Code Review Skill

This skill should be used when the user asks for "code review", "review my changes", "review this PR", "check my code", "pre-merge review", "review diff", or mentions reviewing code quality, implementation correctness, or preparing changes for merge.

From tech-lead

Install

Run in your terminal

npx claudepluginhub ruslan-korneev/claude-plugins --plugin tech-lead

Tool Access

This skill uses the workspace's default tool permissions.

Supporting Assets

View in Repository

references/api-semantics.md

references/implementation.md

references/pyramid.md

references/security.md

references/testing.md

Skill Content

Similar Skills

skill-lookup

Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.

prompts.chat

157.5k

prompt-lookup

Searches prompts.chat for AI prompt templates by keyword or category, retrieves by ID with variable handling, and improves prompts via AI. Use for discovering or enhancing prompts.

prompts.chat

157.5k

Hook Development

10 files

Guides implementation of event-driven hooks in Claude Code plugins using prompt-based validation and bash commands for PreToolUse, Stop, and session events.

plugin-dev

83.2k

Stats

Parent Repo Stars2

Parent Repo Forks0

Last CommitMar 3, 2026

Actions

View Source View Plugin View on GitHub View README

Code Review Skill

Overview

Code review following the Review Pyramid methodology — a prioritized approach that focuses on what matters most.

The Review Pyramid

         ▲
        /|\         5. Code Style (Nit) ← Least important, automatable
       / | \
      /  |  \       4. Tests
     /   |   \
    /    |    \     3. Documentation
   /     |     \
  /      |      \   2. Implementation Semantics ← High priority
 /       |       \
/________|________\ 1. API Semantics ← Most critical, review first

Key insight: Review from bottom to top. API and implementation issues are far more important than style issues.

Two-Phase Review Strategy

Phase 1: Quick Review (Automated Checks)

Fast first pass using quick-reviewer agent (sonnet):

Linting (ruff check)
Type checking (mypy)
Security pattern scanning
Test file existence

Phase 2: Deep Review (Human-Like Analysis)

Thorough analysis using code-reviewer agent (opus):

API semantics (breaking changes, naming, contracts)
Implementation correctness (business logic, edge cases)
Security (detailed analysis)
Documentation completeness
Test quality

When to Use

Trigger	Action
"Review my staged changes"	`/review:diff staged`
"Review before I merge"	`/review:diff main --mode=full`
"Quick check my code"	`/review:diff --mode=quick`
"Deep review this feature"	`/review:diff main --task="..." --mode=deep`

Best Practices

Always provide task context with --task flag for accurate business logic review
Run quick review first to catch obvious issues
Review branch against target before merging
Fix critical issues first — API/Implementation > Style

References

pyramid.md — Detailed explanation of the Review Pyramid
api-semantics.md — API review checklist
implementation.md — Implementation review checklist
security.md — Security review checklist
testing.md — Test review checklist