antfu/skills

Anthony Fu's curated collection of agent skills.

What it solves

It provides a shareable, on-demand knowledge base for AI agents, allowing them to access specific technical preferences, best practices, and official documentation without overloading the agent's context window.

How it works

The project uses a standardized markdown format for "skills" that agents can consume. It leverages git submodules to reference source documentation directly, ensuring that the agent's context remains reliable and up-to-date with upstream changes. Users can install these skills via a CLI tool or a skills pack.

Who it’s for

Developers who want to equip their AI agents with specific expertise in modern web stacks (such as Vue, Nuxt, Vite, and UnoCSS) or follow Anthony Fu's specific development conventions.

Highlights

  • On-demand access: Skills are pulled in as needed rather than loaded upfront, allowing for a larger total knowledge base than a single context window allows.
  • Documentation-driven: Many skills are generated directly from official source documentation.
  • Shareable: Skills are designed to be easily managed and reused across different projects.
  • Customizable: Includes a workflow for users to fork the repo and generate their own customized skill collections.

Related

  • Project
    onmax/nuxt-skills

    A collection of specialized skills for AI coding agents that provides deep, context-aware guidance for the Vue, Nuxt, and Nitro development ecosystems.

  • Project
    kangarooking/kangarooking-skills

    A collection of structured, open-source Agent Skills compatible with the Agent Skills standard, providing reusable workflows for content creation, image generation, and AI project engineering.

  • Project
    vueuse/skills

    A collection of agent skills that enables AI agents to use the VueUse library more accurately and efficiently by reducing token consumption and hallucinations.

  • Project
    vercel-labs/lat.md

    A knowledge graph for codebases written in markdown that links documentation to source code, helping AI agents and humans maintain and navigate domain context.

  • Project
    redfox-data/redfox-community

    A community‑maintained library of plug‑in “Agent Skills” for AI assistants, focused on Chinese social‑media data and generic AI tools. Each skill lives in its own folder with a `SKILL.md` description, optional scripts, and can be installed via copy‑paste, a CLI (`npx skills`), or through SkillHub/ClawHub. Skills call RedFox’s hosted APIs (requiring an API key) to fetch posts, account info, generate copy, download media, etc., enabling agents like Cursor or Claude Code to perform concrete, repeatable workflows.