code-review-copilot

Automate code reviews on pull requests by running multi-model AI analysis that fetches diffs and evaluates correctness, security, performance, and maintainability. This Skill uses GitHub Copilot CLI to independently browse the codebase and generate a consolidated review report.

Yes, you can use multiple AI models like GPT, Claude, and Gemini for code analysis. The Skill orchestrates multi-model reviews on branches or uncommitted changes, allowing each model to browse the codebase independently and provide comprehensive feedback.

Review uncommitted changes by leveraging automated target detection to analyze your working directory. The Skill performs in-depth AI code analysis on security and performance, providing optional rebuttal and consensus rounds to resolve any detected issues.

Yes, GitHub Copilot CLI supports automated branch and commit reviews through this Skill. It automatically detects your target branches or commits, fetches the necessary code diffs, and performs deep code analysis without requiring manual oversight.

AI models reach a consensus during an automated code review through optional rebuttal and consensus rounds. After fetching diffs and analyzing code independently, the models cross-examine findings to resolve detailed feedback and confirm issue resolution.

A limitation of using AI for multi-model code analysis is that it relies on the models' ability to independently browse the codebase and fetch diffs. While it evaluates maintainability and security, complex architectural context may still require manual verification.