Agent Skills
A reusable collection of Agent Skills for AI coding agents, installable via the skills CLI and published as a Fumadocs site.
gohypergiant · 已有 AI 审核记录
安装与使用
运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。
npx skills add gohypergiant/agent-skills技能说明
Agent Skills
A repository of reusable agent skills and the docs site that publishes them.
The source of truth lives in skills/. The published documentation lives in docs/content/docs/. Local harness integrations read the same skills through symlinks in .agents/skills/.
Quick Start
For your project
Skills follow the Agent Skills format and can be installed using the skills CLI.
npm:
npx skills add gohypergiant/agent-skills
pnpm:
pnpm dlx skills add gohypergiant/agent-skills
- Only select skills that are relevant for your project
- We recommend "Project" for the installation scope
- We recommend "Symlink" for the installation method
Browse the published docs locally
The documentation app lives in docs/ and uses pnpm.
cd docs
pnpm install
pnpm dev
Then open http://localhost:3000.
Rebuild local skill symlinks
bash scripts/symlink-agent-skills.sh
Run this after adding a new directory under skills/, or any time .agents/skills/ has gone stale.
What is this repository?
This repo is for creating, maintaining, and publishing agent skills.
Each skill is a directory under skills/ with its own SKILL.md and, in most cases, companion files such as CHANGELOG.md, AGENTS.md, README.md, references, scripts, or assets. The docs/ app turns that material into a browsable Fumadocs site.
If you want the user-facing introduction to skills, start with the docs site content instead of this root README:
docs/content/docs/index.mdxdocs/content/docs/getting-started.mdx
Why this repo exists
Skills are easier to maintain when the source files, local harness wiring, and published docs stay close together.
This layout gives you a few practical benefits:
- edit canonical skill content in
skills/ - expose those same directories to local agent harnesses through
.agents/skills/ - publish matching documentation from
docs/content/docs/ - k