What problem does it solve? LLM coding assistants often overcomplicate solutions, make sweeping unrequested changes, hide confusion, and lack verifiable success criteria. This Skill provides behavioral guidelines, derived from Andrej Karpathy's observations on LLM coding pitfalls, that keep AI-generated code simple, surgical, and goal-driven. ## Core Features & Use Cases - Think Before Coding: Forces explicit statement of assumptions, surfaces multiple interpretations, and stops to ask when requirements are unclear. - Simplicity First: Enforces minimum code that solves the problem, rejecting speculative abstractions, unrequested configurability, and impossible-scenario error handling. - Surgical Changes: Restricts edits to only what the request requires, matching existing style and cleaning up only orphans created by the current change. - Goal-Driven Execution: Transforms vague tasks into verifiable success criteria with test-first loops and step-by-step verification plans. - Use Case: When asking an AI to fix a bug or add a feature, activate these guidelines so the assistant writes a reproducing test first, makes the minimal change, and avoids refactoring unrelated code. ## Quick Start Ask the assistant to apply the karpathy-guidelines skill while implementing your next code change so it states assumptions, keeps the diff minimal, and defines verifiable success criteria.
