agent-skills: Production-Grade Engineering Skills for AI Coding Agents

Introduction

The agent-skills repository, created by Addy Osmani, provides a collection of production-grade engineering skills designed for AI coding agents. This project aims to equip AI agents with the workflows, quality gates, and best practices that senior engineers employ when building software. By integrating these skills, AI agents can consistently follow robust development processes, ensuring higher quality and more reliable code across every phase of software development, from defining ideas to shipping to production.

Installation

Getting started with agent-skills is straightforward, offering flexibility for various AI agents and workflows.

The fastest path is using the open skills CLI:

npx skills add addyosmani/agent-skills            # install all 24 skills
npx skills add addyosmani/agent-skills --list     # browse before installing

You can also install individual skills:

npx skills add addyosmani/agent-skills --skill code-review-and-quality
npx skills add addyosmani/agent-skills --skill test-driven-development

For native integrations, agent-skills supports a wide range of tools. For example, with Claude Code:

/plugin marketplace add addyosmani/agent-skills
/plugin install agent-skills@addy-agent-skills

Refer to the official documentation for detailed setup instructions for Cursor, Antigravity CLI, Gemini CLI, GitHub Copilot, and other agents.

Examples

agent-skills organizes its capabilities around the software development lifecycle, offering 8 slash commands that activate specific skill sets automatically.

Here are some key commands:

What you're doing	Command	Key principle
Define what to build	/spec	Spec before code
Plan how to build it	/plan	Small, atomic tasks
Build incrementally	/build	One slice at a time
Prove it works	/test	Tests are proof
Review before merge	/review	Improve code health
Ship to production	/ship	Faster is safer

Beyond these commands, the repository includes 24 distinct skills covering various aspects of engineering. For instance:

• spec-driven-development: Guides agents to write a Product Requirements Document (PRD) before any code is written.
• test-driven-development: Enforces the Red-Green-Refactor cycle and best practices for testing.
• code-review-and-quality: Provides a five-axis review process for changes before merging.
• security-and-hardening: Focuses on OWASP Top 10 prevention and secure coding practices.

The repository also features pre-configured specialist agent personas, such as code-reviewer, test-engineer, and security-auditor, for targeted reviews.

Why Use Agent Skills?

AI coding agents often default to the shortest path, potentially skipping crucial steps like writing specifications, comprehensive testing, or security reviews. agent-skills addresses this by providing structured workflows that enforce the same discipline senior engineers bring to production code.

Each skill encodes hard-won engineering judgment, guiding agents on when to write a spec, what to test, how to review, and when to ship. These are not generic prompts, but opinionated, process-driven workflows that elevate the quality of AI-generated code from prototype-level to production-ready. The skills integrate best practices from Google's engineering culture, including concepts from "Software Engineering at Google" and Google's engineering practices guide, embedding principles like Hyrum's Law, the Beyonce Rule, and Chesterton's Fence directly into the step-by-step workflows.

Links

• GitHub Repository: https://github.com/addyosmani/agent-skills
• Skills CLI: https://github.com/vercel-labs/skills