Skip to content

Skillselionwith God's help

By phase

IdeaResearch & spot demand
ValidateScope & prototype
BuildFrontend, backend, MCP
ShipTest, review, deploy
LaunchDistribution, SEO, ASO
GrowAnalytics & retention
OperateMonitoring & uptime

By type

Skills
Marketplaces
MCP servers
Workflows

Skills Marketplaces MCP Learn Brief

Browse all tools →By phase

Idea Validate Build Ship Launch Grow Operate

Explore

Skills
Marketplaces
MCP
Learn
Brief

This week for builders

Five minutes, every Monday — the tools, releases and tactics for shipping solo.

Email address

unsubscribe anytime.

Skillselion

An independent, curated directory of AI-coding tools for solo builders - skills, MCP servers, marketplaces and workflows for Claude Code, Codex and Cursor, organized by the indie build journey from idea to operate.

ResourcesBrowse Skills Browse MCP Servers Browse Marketplaces Browse Workflows Best-of shortlists Glossary State of AI Skills 2026 Weekly Brief

CommunityAbout Learn Feedback Privacy Policy Terms Cookie Policy Accessibility Cookie settings

Last updated June 10, 2026 · © 2026 Skillselion. Skillselion is an independent project and is not affiliated with Anthropic, OpenAI, Cursor, Claude, Claude Code, or Codex.

Skills/ALCHAINCYF/KARPATHY SKILL

alchaincyf/karpathy-skill

1 skill · 2.1k installs · 213 stars · GitHub

Install

npx skills add https://github.com/alchaincyf/karpathy-skill

Skills in this repo

1Andrej Karpathy PerspectiveAndrej Karpathy Perspective distills public writings and interviews into a reusable advisory persona for builders who want first-principles AI judgment instead of hype-shaped answers. Activate it when you are evaluating whether an agent feature will survive real users, debating training versus prompting, or explaining LLM limitations to stakeholders. The skill routes topics to named frameworks—march of nines for reliability, build-to-understand for learning, jagged intelligence for capability maps—and enforces disciplined role-play: a single disclaimer, hard exit phrases, and refusal to fake expertise on business or geopolitics. It is journey-wide because the same realism applies when scoping an idea, validating a prototype demo, shipping guardrails, or deciding what to automate in production. Do not invoke for routine factual AI questions; it is a perspective layer on top of your agent session. Solo builders using Claude Code or Cursor benefit when they need consistent vocabulary for Software 2.0/3.0 discussions and product reviews without rereading dozens of primary sources.2.1kinstalls