andrej-karpathy – skill Lập trình cho agent AI

Karpathy Guidelines

Behavioral guidelines to reduce common LLM coding mistakes, derived from [Andrej Karpathy's observations](https://x.com/karpathy/status/2015883857489522876) on LLM coding pitfalls.

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

When to Use This Skill

• Use when writing, reviewing, or refactoring code with an LLM.
• Use when a change needs to stay surgical and avoid speculative abstractions.
• Use when assumptions, tradeoffs, and verification criteria should be made explicit.
• Use when code has become overcomplicated and needs to be simplified.

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 such as "make it work" require constant clarification.

Examples

User request: "Add validation to this form."

Better response:
1. Assumption: validation should run before submit and show inline errors.
2. Plan: add a small validator, add tests for invalid inputs, then verify existing submit behavior.
3. Scope: only touch the form component and its test file.

User request: "Refactor this service."

Better response:
1. Ask what behavior must remain unchanged.
2. Identify a concrete smell, such as duplicated parsing logic.
3. Make the smallest refactor and run the existing service tests.

Limitations

• These guidelines are behavioral guardrails, not a replacement for project-specific architecture or style rules.
• For emergency fixes, prioritize the smallest verified correction over extensive planning.
• For exploratory prototypes, some caution can be relaxed, but assumptions and verification should still be explicit.