Skip to content
Agent Orchestration
Skill

/karpathy-guidelines

Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code. Derived from Andrej Karpathy's LLM coding pitfalls.

BOOST
From plugin
claude-lane-stack
12464 skills24 agents5 commands12 hooks
+1
Install
$ npx -y skills add VKirill/claude-lane-stack --skill karpathy-guidelines --agent claude-code

How it fires

How this skill gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
  • You can call itInvoke it directly when you want it.
  • Slash command/karpathy-guidelines

Context preview

The summary Claude sees to decide when to auto-load this skill.

Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code. Derived from Andrej Karpathy's LLM coding pitfalls.

SKILL.md

karpathy-guidelines.SKILL.md
name: karpathy-guidelines
description: Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code. Derived from Andrej Karpathy's LLM coding pitfalls.
license: MIT

Karpathy Guidelines

Derived from [Karpathy on LLM coding pitfalls](https://x.com/karpathy/status/2015883857489522876). Bias: caution over reckless speed. Trivial tasks — use judgment.

1. Think Before Coding

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

  • State assumptions explicitly. If uncertain, ask (Lane Pilot writer: see below).
  • Multiple interpretations → present them; don't pick silently.
  • Prefer simpler approaches; push back when warranted.
  • Unclear → stop, name the confusion, ask.

Inside a Lane Pilot writer thread (`LANE_PILOT_AGENT_TYPE` set) nobody answers mid-task, and a question in prose is not recognised: when the contract is unclear or contradictory, change no files and make the first line of your answer `NEEDS_HUMAN: <one question for the project owner>`. Everything you can look up or decide yourself (see Autonomy zone) you decide.

2. Simplicity First

Minimum code that solves the problem. Nothing speculative.

  • No features beyond the ask.
  • No single-use abstractions or fake "flexibility".
  • No error handling for impossible cases.
  • 200 lines that could be 50 → rewrite.

3. Surgical Changes

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

  • Don't "improve" adjacent code, comments, or formatting.
  • Don't refactor what isn't broken.
  • Match existing style.
  • Unrelated dead code → mention, don't delete.
  • Remove only what **your** change made unused.

**Test:** every changed line traces to the user request.

4. Goal-Driven Execution

Define success criteria. Loop until verified.

  • "Add validation" → tests for invalid inputs, then make them pass.
  • "Fix the bug" → failing test first, then green.
  • Multi-step: `step → verify: command` list.

Strong criteria → autonomous loop. Weak ("make it work") → clarify first (in a Lane Pilot writer thread: the `NEEDS_HUMAN:` first line above).

Autonomy zone (MAY)

Inside the task scope, decide tech details, edit order, and local design without asking. Do not ask the user what you can look up in the repo or docs.

Read more
Ships withclaude-lane-stack

Claude Lane Stack is a self-hosted toolkit for coordinating coding agents through CLI processes, task contracts, Git workspaces, and verification receipts (bin/run-controller:117-159, schemas/task-v2.schema.json:1-60).

Get the whole plugin

Other skills on claude-lane-stack.