addyosmani

addyosmani / agent-skills

Public

Production-grade engineering skills for AI coding agents.

agent-skills antigravity antigravity-ide claude-code cursor
50
3
100% credibility
Found Feb 17, 2026 at 22 stars 2x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Shell
AI Summary

A curated set of structured guides and workflows that enable AI coding assistants to follow professional software engineering best practices across all development phases.

How It Works

1
📰 Discover Agent Skills

You come across this helpful collection of guides that teach AI helpers to build reliable apps like pro engineers do.

2
🚀 Add to Your AI Helper

You easily add the guides to your favorite AI coding tool, like Claude or Cursor, so it's ready to use right away.

3
✨ Pick a Starting Guide

You choose a guide for your project phase, like refining an idea or planning tasks, and share it with your AI.

4
🔄 Guide Your AI Through Steps

Your AI follows the guide's clear steps, building small safe pieces, testing them, and checking quality along the way.

5
✅ Review and Polish

You have your AI review the work for quality, security, and speed, making sure everything is top-notch.

6
🚀 Launch with Confidence

Your AI handles the final checks and prepares everything for sharing with the world safely.

🎉 Enjoy Your Solid App

You now have a professional-quality app built reliably, ready for real use without worries.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 22 to 50 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is agent-skills?

Agent-skills delivers production-grade engineering workflows for AI coding agents like Claude and GitHub Copilot, packaging senior-level practices into reusable Markdown skills that guide agents through the full software lifecycle—from idea refinement and spec-driven development to testing, review, and shipping. It solves the problem of agents defaulting to quick prototypes by enforcing verification steps, anti-patterns to avoid, and quality gates, ensuring reliable web app builds. Written mostly in Shell with Markdown instructions, it integrates via plugins or slash commands like /spec, /plan, /build, /test, /review, and /ship.

Why is it gaining traction?

Unlike generic prompts, these agent skills claude and anthropic agent skills enforce structured phases with exit criteria and evidence requirements, making AI output consistent and production-ready—think context engineering for optimal agent feeds or browser testing via DevTools. Developers hook it into Claude Code, Cursor, or Copilot setups effortlessly, with pre-built personas for code review and security audits standing out for agent skills github copilot users. The opinionated, battle-tested rules from places like Vercel cut hallucination risks without bloating context.

Who should use this?

Solo devs or small teams using Claude Code or Cursor for TypeScript/JavaScript web apps, especially when scaling AI-assisted features beyond MVPs. Frontend engineers building UIs with React/Vue will like the incremental implementation and accessibility checks; backend folks get API design and CI/CD automation. Ideal for agent skills for context engineering github workflows where you need agents to handle git versioning or performance budgets reliably.

Verdict

Worth cloning for Claude or Copilot users chasing agent skills github repo quality—docs are solid, setup is dead simple—but at 17 stars and 1.0% credibility score, it's early-stage with room for more skills and community tests. Try it on a side project before team rollout.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.