codex

Independent code review is performed by invoking the skill within your project repository to analyze commits with a pass/fail gate. It operates directly on your codebase using Git context to provide intelligent feedback and suggestions.

Adversarial challenge analysis improves robustness by identifying potential issues and edge cases in your code. The AI evaluates the logic within your repository context to challenge assumptions and prevent errors before merging.

AI consultation allows you to ask any coding question regarding your existing codebase. The tool reads your project files using Bash, Grep, and Read to provide insights specific to your repository's current state.

A functioning codebase with access to a Git repository is required to run AI code analysis. The tool operates within the specific project repository context to accurately evaluate your latest commit and local file structure.

The best way to verify commit correctness is to invoke a review command within your project directory. The AI analyzes the Git repository changes, applies an adversarial challenge, and outputs a pass or fail gate.

AI consultation requires a functioning codebase and Git repository access to operate, meaning it cannot analyze code outside the current project context. It relies on repository file access rather than external runtime environments.