How does it handle large, complex diffs?

For files with significant changes, the skill reads the full file context to ensure it understands the surrounding logic rather than just looking at isolated lines.

Can this skill detect leaked API keys?

Yes, it specifically scans your diffs for hardcoded secrets, tokens, and credentials using regex patterns and context analysis before they are committed.

Is this a replacement for human PR reviews?

It is intended as a pre-commit check to catch 'low-hanging fruit' issues, making the eventual human peer review more efficient by focusing on higher-level architecture.

Does it require changes to be staged?

No, it can analyze both staged (cached) and unstaged changes to give you a complete view of your current work-in-progress.

What is the difference between Quick Mode and Full Mode?

Quick Mode is a 2-5 minute sanity check focusing on obvious accidents like secrets and debug code, while Full Mode performs a deep dive into logic, edge cases, and architectural consistency.

Git Diff Reviewer

Name: Git Diff Reviewer
Author: LayerDynamics

byLayerDynamics

0•

セキュリティとテスト

Analyzes staged and unstaged Git changes to identify bugs, security risks, and incomplete code before you commit.

The Git Diff Reviewer skill acts as an automated pre-commit auditor within Claude Code, meticulously scanning your local changes for common development pitfalls. It offers two distinct modes: a lightning-fast 'Quick Mode' for sanity checks against leaked secrets and debug logs, and a 'Full Mode' for comprehensive analysis of logic, edge cases, and test coverage. This skill ensures that only production-grade, secure, and complete code enters your version control history, serving as a vital first line of defense before peer reviews.

主な機能

01Quick Mode for rapid 2-minute pre-commit sanity checks

02Identification of 'leftover' code such as debug logs, stubs, and placeholders

030 GitHub stars

04Test coverage verification for new or modified functions

05Deep logic analysis to detect edge cases and potential runtime errors

06Automated security scanning for hardcoded API keys, tokens, and credentials

ユースケース

01Performing a final sanity check to ensure no temporary debug code is accidentally committed

02Conducting a local security audit to prevent credential leakage in config files

03Validating complex logic changes and error handling paths in a large diff

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add layerdynamics/lore diff-review

For use in Claude.ai and ChatGPT