How does this skill differ from Claude's built-in code review?

This skill leverages external engines like Codex and Gemini via their CLIs to provide a 'second opinion' and uses specialized logic for parallel analysis, secret detection, and consensus scoring.

Do I need external API keys to use the AI Code Reviewer?

Yes, you must have the Codex CLI or Gemini CLI installed and authenticated on your system for the respective analysis tools to function.

Can I compare results from different AI models?

Yes, the 'analyze_code_combined' tool runs multiple models in parallel, identifies overlapping findings, and provides a confidence score based on model agreement.

Is my source code sent to external servers?

Code snippets are sent to OpenAI or Google for analysis when using the Codex or Gemini tools; however, the secret scanning tool runs locally to protect sensitive data.

AI Code Reviewer

Name: AI Code Reviewer
Author: physics91

byphysics91

安全与测试

Performs deep, multi-model code analysis and security scanning using external AI engines to identify vulnerabilities and architectural improvements.

关于

The AI Code Reviewer skill enhances Claude's native capabilities by integrating external power from OpenAI Codex and Google Gemini to provide comprehensive, multi-perspective code audits. It excels at identifying logical errors, security vulnerabilities, and performance bottlenecks through parallel execution and intelligent result deduplication. Ideal for high-stakes reviews, this skill ensures consistency across codebases while offering specialized secret detection and consensus-based scoring to provide developers with highly confident, actionable remediation steps.

主要功能

Persistent SQLite caching to reduce redundant analysis costs and latency
Structured markdown reporting with severity-based categorization and remediation
Multi-model analysis leveraging OpenAI Codex and Google Gemini for diverse perspectives
Automated secret detection for AWS, GCP, Azure, and GitHub credentials
Consensus scoring and deduplication to ensure high-accuracy findings
0 GitHub stars

使用场景

Automating high-confidence code reviews for large-scale legacy codebase refactoring
Validating complex architectural patterns using multiple AI consensus models
Performing deep security audits and credential checks before production deployments

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add physics91/claude-vibe ai-code-reviewer

For use in Claude.ai and ChatGPT

Download Skill

GitHub

关于