Can I run Codex and Gemini reviews simultaneously?

Yes, the skill is designed to run both tools in parallel and present a combined report, highlighting where the two models agree or differ in their assessment.

How does it handle security reviews?

When the security focus is selected, the skill uses specialized prompts and can automatically trigger dependency scans if manifest files like package.json or requirements.txt are modified.

Do I need external tools installed to use Second Opinion?

Yes, this skill requires either the OpenAI Codex CLI or the Google Gemini CLI to be installed and configured on your system to perform the external reviews.

Can I review specific branches instead of the whole repo?

Absolutely. You can choose to review uncommitted changes, a specific commit via its SHA, or a diff between your current branch and the main branch.

Second Opinion Code Review

Name: Second Opinion Code Review
Author: trailofbits

bytrailofbits

•

2,992

•

Security & Testing

Conducts independent, multi-model code reviews using OpenAI Codex and Google Gemini CLI to identify bugs, security flaws, and performance issues.

The Second Opinion skill provides a robust validation layer for your development workflow by piping code changes to external LLMs for an unbiased review. It can analyze uncommitted work, branch diffs, or specific commits, leveraging the unique strengths of models like OpenAI Codex and Google Gemini. By integrating project-specific context from CLAUDE.md and offering specialized focus areas like security and error handling, it ensures your code meets high standards before it ever reaches a human reviewer.

Key Features

01Dual-model reviews using OpenAI Codex and Google Gemini CLI

02Support for uncommitted changes, branch diffs, and specific commit SHAs

03Automated integration of project conventions from CLAUDE.md or AGENTS.md

04Parallel execution with a consolidated summary of findings

052,992 GitHub stars

06Specialized focus modes for security, performance, and error handling

Use Cases

01Performing a final automated audit on a feature branch before opening a pull request

02Verifying uncommitted changes for logic errors or security vulnerabilities before committing

03Cross-referencing code review findings across different AI models to catch edge cases

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add trailofbits/skills second-opinion

For use in Claude.ai and ChatGPT

Download Skill