How do I limit the code sent for review?

The skill provides file selection techniques using glob patterns (e.g., --file 'src/**' --file '!**/*.test.*') to ensure you only send relevant source files and stay within token limits.

What kind of tasks benefit most from a Second Opinion?

Critical technical decisions benefit most, specifically security-sensitive code, major refactoring involving multiple modules, and complex architectural design choices.

Why should I use a second model for code review?

Different LLMs have unique training biases and reasoning strengths. Cross-validation helps identify edge cases, security vulnerabilities, or architectural flaws that a single model might overlook.

Does this skill require external API keys?

If you use the Oracle CLI in API mode, you will need a key for the secondary provider (e.g., OpenAI). However, you can also use browser-based modes or manual templates that don't require additional keys.

Second Opinion

Name: Second Opinion
Author: xiaobei930

byxiaobei930

•

보안 및 테스팅

Cross-validates architectural decisions, security logic, and complex refactoring by leveraging secondary AI models for high-stakes code reviews.

The Second Opinion skill is designed to mitigate risks in critical development phases by facilitating cross-model validation. Whether you are finalizing a major system architecture, conducting a security audit, or troubleshooting a persistent bug, this skill provides the framework to package context and solicit structured feedback from alternative LLMs like GPT-4 or Gemini. It supports automated workflows via the Oracle CLI as well as manual review templates, ensuring that complex technical decisions are scrutinized from multiple perspectives to improve reliability and code quality.

주요 기능

015 GitHub stars

02Specialized prompt templates for architecture, security, and debugging

03Automated context packaging for external model review

04Structured risk assessment and alternative solution reporting

05Integration with Oracle CLI for multi-engine AI validation

06Selective file globbing to exclude tests and snapshots from audits

사용 사례

01Collaborative debugging for complex issues that require a fresh perspective

02Validating high-level architectural designs and technical stack migrations

03Performing deep security audits on authentication and data processing logic

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add xiaobei930/claude-code-best-practices second-opinion

For use in Claude.ai and ChatGPT

Download Skill