How is the feedback structured?

Issues are categorized into 'Critical' (must fix before committing), 'Warnings' (should be addressed), and 'Suggestions' (minor architectural or performance improvements).

Can it analyze code that hasn't been staged yet?

Yes, it evaluates both staged changes (ready to commit) and unstaged changes (current edits) to give you a full picture of your work in progress.

What kind of security vulnerabilities can it detect?

It audits for common risks including SQL injection, cross-site scripting (XSS), command injection, and the accidental inclusion of hardcoded credentials.

Will it catch linting and style errors?

The skill is specifically designed to focus on logic, security, and architecture; it avoids nitpicking style issues that are better handled by automated linters.

How does the Review skill access my code changes?

The skill utilizes built-in git commands like 'git diff' and 'git status' to identify exactly what has changed in your working directory and provide context-aware feedback.

Automated Git Code Review

Name: Automated Git Code Review
Author: igrybkov

byigrybkov

0•

Collaboration Tools

Analyzes staged and unstaged git changes to identify bugs, security vulnerabilities, and code quality improvements before committing code.

The Review skill serves as an AI-powered peer reviewer that integrates directly into your git workflow. By examining current changes through git diffs and status updates, it provides a comprehensive audit of logic errors, security flaws like hardcoded secrets, and architectural best practices. This skill is particularly useful for developers who want to catch critical issues early, ensure high-quality commits, and receive actionable feedback categorized by severity—from must-fix critical errors to optional performance suggestions.

Key Features

01Detection of logic errors, race conditions, and edge case handling

02Comprehensive analysis of both staged and unstaged git modifications

030 GitHub stars

04Prioritized feedback loops with Critical, Warning, and Suggestion tiers

05Security auditing for SQL injection, XSS, and hardcoded credentials

06Context-aware reviews using git status and repository file reading

Use Cases

01Refinement of complex logic changes where off-by-one or async errors might hide

02Security auditing of new features to identify potential injection or data handling risks

03Pre-commit sanity checks to ensure no accidental debug code or secrets are pushed

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add igrybkov/dotfiles review

For use in Claude.ai and ChatGPT