What problem does it solve? LLM-generated code often suffers from over-engineering, unrequested features, hidden assumptions, and unverified changes. This Skill provides a set of behavioral guidelines, based on Andrej Karpathy's observations of LLM coding pitfalls, that keep AI-assisted coding focused, minimal, and verifiable. ## Core Features & Use Cases - Think Before Coding: Forces explicit statement of assumptions, presentation of alternative interpretations, and pausing to ask questions when requirements are unclear. - Simplicity First: Enforces minimal code for the problem, rejecting speculative abstractions, unrequested flexibility, and unnecessary error handling. - Surgical Edits & Goal-Driven Execution: Restricts changes to only what was requested, cleans up only self-created dead code, and converts tasks into verifiable success criteria with test-driven loops. - Use Case: When asking an AI to fix a bug or add a feature in an existing codebase, activate these guidelines so the AI writes a reproducing test first, touches only the necessary lines, and matches existing code style instead of refactoring unrelated code. ## Quick Start Apply the karpathy guidelines while implementing this feature so the code stays minimal, assumptions are stated, and every change is verified against a test.
