Hanselman Adversarial Code Review Skill

Run a rigorous, adversarial code review using two AI models in parallel. Cross-reference their findings to separate signal from noise, producing a consensus severity table with fix confidence ratings.

When to Use

Invoke this skill when the user asks for:

• A thorough or deep code review
• An adversarial review
• A dual-model review
• A "Hanselman review"
• A confidence-rated review
• Cross-model validation of code changes

Philosophy