karpathy-guidelines

name	karpathy-guidelines
description	Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria. Applies to ALL code work in VSmartwatch.

Karpathy Guidelines

Behavioral guidelines to reduce common LLM coding mistakes, derived from Andrej Karpathy's observations on LLM coding pitfalls.

Source: andrej-karpathy-skills (MIT License). Lightly adapted — content is canonical.

Tradeoff: These guidelines bias toward caution over speed. For trivial tasks, use judgment.

1. Think Before Coding

Don't assume. Don't hide confusion. Surface tradeoffs.

Before implementing:

• State your assumptions explicitly. If uncertain, ask.
• If multiple interpretations exist, present them — don't pick silently.
• If a simpler approach exists, say so. Push back when warranted.
• If something is unclear, stop. Name what's confusing. Ask.

2. Simplicity First

Minimum code that solves the problem. Nothing speculative.

• No features beyond what was asked.
• No abstractions for single-use code.
• No "flexibility" or "configurability" that wasn't requested.
• No error handling for impossible scenarios.
• If you write 200 lines and it could be 50, rewrite it.

Ask yourself: "Would a senior engineer say this is overcomplicated?" If yes, simplify.

3. Surgical Changes

Touch only what you must. Clean up only your own mess.

When editing existing code:

• Don't "improve" adjacent code, comments, or formatting.
• Don't refactor things that aren't broken.
• Match existing style, even if you'd do it differently.
• If you notice unrelated dead code, mention it — don't delete it.

When your changes create orphans:

• Remove imports/variables/functions that YOUR changes made unused.
• Don't remove pre-existing dead code unless asked.

The test: Every changed line should trace directly to the user's request.

4. Goal-Driven Execution

Define success criteria. Loop until verified.

Transform tasks into verifiable goals:

• "Add validation" → "Write tests for invalid inputs, then make them pass"
• "Fix the bug" → "Write a test that reproduces it, then make it pass"
• "Refactor X" → "Ensure tests pass before and after"

For multi-step tasks, state a brief plan:

1. [Step] → verify: [check]
2. [Step] → verify: [check]
3. [Step] → verify: [check]

Strong success criteria let you loop independently. Weak criteria ("make it work") require constant clarification.

Applied to VSmartwatch

• Before writing new Dart/Python/JS code → state assumptions (state mgmt choice, DB schema shape via Prisma/SQL, error model, auth flow).
• When editing an existing Flutter screen / FastAPI router / Express controller → don't reformat, don't rename to your preferred convention. Match codebase style.
• When user says "add validation to login" → success criteria = "tests reject empty email, invalid format, password <8 chars; all pass". Don't add rate limiting / captcha unprompted (rate limit usually exists in middleware already — verify first).
• When unsure on architectural choice (vd: notify via FCM vs WebSocket, REST vs Socket.IO event) → list 2-3 options with tradeoffs, reference UC + topology, ask anh.
• For HealthGuard / health_system / Iot_Simulator_clean / healthguard-model-api — each has its own conventions overlay (rules 21-25). Read overlay before assuming.