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/MULTICA AI/ANDREJ KARPATHY SKILLS

multica-ai/andrej-karpathy-skills

1 skill · 3.4k installs · 171k stars · GitHub

Install

npx skills add https://github.com/multica-ai/andrej-karpathy-skills

Skills in this repo

1Karpathy GuidelinesKarpathy Guidelines is an agent skill that encodes Andrej Karpathy’s observations about common LLM coding failures into repeatable behavioral rules for writing, reviewing, and refactoring code. Solo and indie builders install it when they want Claude Code, Cursor, or Codex to stop guessing requirements, shipping 200-line solutions for 50-line problems, or “cleaning up” files they did not need to touch. The skill walks the agent through stating assumptions, choosing the minimum viable change set, presenting interpretations instead of picking one silently, and defining how success will be verified before code lands. It is not a linter or a test runner; it is procedural knowledge that shapes how the model plans and edits. Because triggers cover implementation and review alike, Prism tags it journey-wide with Ship/review as the primary shelf. Use it whenever agent output feels over-engineered, vague, or unnecessarily broad—especially before you accept a large diff you did not fully specify.3.4kinstalls