Skip to content

About

Opinionated development skills for AI coding agents

Resources

Stars

15 stars

Watchers

0 watching

Forks

Latest commit

 

History

44 Commits

Folders and files

Repository files navigation

🧩 Agentic Skills

Opinionated development skills for AI coding agents

How I build Laravel, Nuxt, Go and Python applications, refined over 20 years in the industry,
packaged as skills for any agent that reads the SKILL.md format.

License: MIT Laravel Nuxt Go Python Agents: Claude Code | Codex | Cursor

Plugins · Installation · How a Skill Loads · Contributing · Related


Each plugin is a set of skills for one framework or language: how the code is structured, which tools to reach for, and what to avoid, each with the reason behind it. An agent working in your project loads the skill that matches the task in front of it and writes the code the way the skill says.

Why Agentic Skills?

  • Opinionated on purpose. Each skill takes a position (one shape for an action, a DTO, a package, a test) and gives the reason for it. Code an agent writes from them reads as one codebase, not an average of every style it has seen.
  • Drawn from real projects. The patterns come from production applications and are checked against them. Anything that only makes sense in one project stays out.
  • Loaded when relevant. A skill's description says when to use it, so the agent reads the jobs skill when it writes a job, not on every turn.
  • Install only what you use. Each plugin installs on its own: take the Go skills without the Laravel ones.

Plugins

Plugin Description Skills
laravel Laravel development: action-oriented architecture, strict typing, DTOs, and maintainability 22 skills
nuxt Nuxt 4 + Vue 3: domain-driven, type-safe, composable-first architecture 18 skills
go Go: standard library first, small packages by concern, named types, errors handled once 8 skills
python Python 3.14: uv for everything, strict types as a gate, frozen values, Protocols as seams, nothing at import 9 skills

Each plugin's README lists its skills.

Installation

agntc (recommended)

npx agntc@latest add leeovery/agentic-skills

You're asked which plugins to install. Or install one directly:

npx agntc@latest add leeovery/agentic-skills/laravel
npx agntc@latest add leeovery/agentic-skills/nuxt
npx agntc@latest add leeovery/agentic-skills/go
npx agntc@latest add leeovery/agentic-skills/python

Update or remove them later:

npx agntc@latest update
npx agntc@latest remove leeovery/agentic-skills/laravel

See agntc for more.

Vercel Skills CLI

The plugins also install with the Vercel skills CLI:

npx skills add leeovery/agentic-skills/laravel
npx skills add leeovery/agentic-skills/nuxt
npx skills add leeovery/agentic-skills/go
npx skills add leeovery/agentic-skills/python

How a Skill Loads

Every skill is a folder with a SKILL.md: YAML frontmatter holding its name and a description of when to use it, then the guidance itself. The agent sees only the descriptions until a task matches one, then reads that skill in full. Longer material, such as a complete reference file or a setup walkthrough, sits in the skill's references/ folder, and the skill says when to read it, so the detail costs nothing until it's needed.

Contributing

Issues and pull requests are welcome: a fix, a better pattern, a question about an approach or its trade-offs, or a skill for something not yet covered. These are opinionated patterns, so open an issue to talk a significant change through before writing it.

Related

License

MIT. See LICENSE.


Built with care by Lee Overy

About

Opinionated development skills for AI coding agents

Resources

Stars

15 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages