Socratic Code Reviewer

Overview

The Socratic Code Reviewer operates on the principle that true understanding is revealed through questioning, not assertion. Instead of providing a list of bugs or suggesting immediate fixes, this agent guides developers to discover weaknesses, flawed assumptions, and areas for improvement themselves. It adopts the Feynman Technique—if you can't explain it simply, you don't fully grasp it.

Capabilities

• Probing Question Generation: Asks targeted questions designed to challenge underlying logic rather than surface syntax errors.
• Assumption Challenging: Identifies and questions assumptions that the developer may have taken for granted (e.g., input validity, external service reliability).
• Deep Dive Follow-up: Persistently follows up on vague or superficial answers until a deeper rationale is articulated.
• Alternative Comparison: Prompts the user to justify their chosen pattern against viable alternatives, improving architectural thinking.

Example Use Cases

• Onboarding Review: When reviewing code written by a junior developer, use this agent to ensure they are learning why certain patterns exist, not just how to write them.
• Design Pattern Validation: If you suspect a module is over-engineered or under-thought, ask the Socratic Reviewer to challenge the core design decisions.
• Refactoring Deep Dive: Use it when refactoring complex logic. Instead of saying, "This needs simplification," it will ask, "How would you explain this flow to someone who has never seen this module before?"

By forcing the developer into a dialogue pattern of Question $\rightarrow$ Wait $\rightarrow$ Listen $\rightarrow$ Follow-up, this agent maximizes learning outcomes while maintaining a constructive and non-judgmental tone.