⚙️ Production-Grade Agent Skills Comparison 2026

The definitive guide to production-grade agent skills — the reusable capability layer that turns AI coding agents into reliable engineering teammates.

⭐ 404K+ Combined Category Stars
Category leader: addyosmani/agent-skills — 81,976⭐ from the Chrome DevTools team (Addy Osmani)
🏆 Category Defining 🌐 Chrome DevTools Team 🔥 Trending Aug 2026 ⭐ 404K+ Ecosystem 🤖 Agent Skills 📦 24 Skills / 8 Commands

📊 Executive Summary

🔥 The Bottom Line: Addy Osmani (Chrome DevTools) releasing agent-skills (81K⭐) is the biggest endorsement of the agent skills ecosystem yet. The skills paradigm is going mainstream. When one of the most respected voices in front-end engineering — and a member of the team that builds the browser 3.4 billion people use — publishes a production-grade skills collection for AI coding agents, the message is unmistakable: skills are the new unit of AI software engineering.

This comparison covers 10 tools spanning the full production agent skills lifecycle: skill frameworks (vercel-labs/skills, obra/superpowers), production skill packages (addyosmani/agent-skills), skill routers (reverse-skill), skill recording (microsoft/skill-recorder), skill generation (book-to-skill), and quality layers (ratchet, ponytail-improved, i-have-adhd). Combined, the category exceeds 404,000 GitHub stars.

81,976
addyosmani/agent-skills ⭐
Category Leader • +226/day
266,458
obra/superpowers ⭐
Framework Pioneer
404K+
Total Category ⭐
Across All 10 Tools
24
Skills in agent-skills
8 Slash Commands

🏆 Top 10 Production-Grade Agent Skills Tools Ranked

Filter by type:
# Tool ⭐ Stars Growth Type Score Best For
1 addyosmani/agent-skills 81,976 +226/day Production Skill Package 96 Category Defining
2 obra/superpowers 266,458 Skills Framework & Methodology 94 Framework Leader
3 zhaoxuya520/reverse-skill 19,064 +9,904/wk Skill Router Pack 85 RE & Security
4 virgiliojr94/book-to-skill 16,990 +4,596/wk Skill Generation 84 Books → Skills
5 ayghri/i-have-adhd 17,278 +3,874/wk Output Quality Layer 83 Agent UX
6 microsoft/skill-recorder 969 Skill Recording Platform 78 Enterprise Capture
7 vercel-labs/skills Skills CLI Framework 74 npx skills add
8 ponytail-improved 580 Agent Prompt Optimization 62 YAGNI Discipline
9 ratchet 412 Agent Compliance Checking 60 Policy Compliance
10 codex-as-mcp 169 MCP Bridge 55 Codex + MCP

Click column headers to sort. Use filter buttons above to narrow by tool type. Scores are composite: 40% production readiness, 25% category impact, 20% community adoption, 15% ecosystem influence.

👑 Category Leader Deep Dive

addyosmani/agent-skills — The Production Benchmark

⭐ 81,976
Markdown • +226 stars/day • 24 skills • 8 slash commands • Author: Addy Osmani (Chrome DevTools)
addyosmani/agent-skills is what happens when a world-class engineering leader applies browser-tooling discipline to the agent skills space. Built by Addy Osmani — Engineering Leader on Chrome DevTools, author of "Learning JavaScript Design Patterns" — this repository defines what production-grade means for agent skills: rigorously structured, engineering-focused, and tied to a full software delivery lifecycle via 8 slash commands.
🏆 Category Defining Score
🛠️ Production Readiness
⭐ Community Adoption
🚀 Growth Velocity (+226/day)
  • 8 slash commands: /spec, /plan, /build, /test, /review, /webperf, /code-simplify, /ship — a complete engineering workflow
  • 24 skills total: Covers spec writing, architecture, implementation, testing, code review, performance, refactoring, and shipping
  • Chrome DevTools pedigree: Credibility from one of the most respected teams in web engineering
  • Engineering-first: Skills written like production documentation — deterministic, reviewable, versionable
  • Mainstream adoption signal: 81K+ stars in weeks makes it the fastest mainstreaming agent-skills repo ever
🔥 Why This Matters: Addy Osmani is not a random open-source creator — he's an engineering leader at Google on Chrome DevTools, one of the most widely used developer tools on earth. When he ships an agent skills collection, it validates the entire category for enterprise engineering orgs that were waiting for a credible, production-grade reference. The skills paradigm just got its stamp of approval from the browser team itself.

📋 Detailed Tool Cards

#1 addyosmani/agent-skills — Production-Grade Engineering Skills ⭐ 81,976 +226/day

Markdown skills • 24 skills • 8 slash commands • Author: Addy Osmani (Chrome DevTools) • 🏆 Category Defining

The reference implementation of production-grade agent skills. Where other repos ship experimental skills, agent-skills ships a complete engineering SDLC: start with /spec to write requirements, /plan to architect, /build to implement, /test to verify, /review to critique, /webperf to optimize performance, /code-simplify to reduce complexity, and /ship to release. Each skill is written with the same care as Chrome DevTools documentation — precise, actionable, and deterministic.

  • /spec → /ship: A full delivery lifecycle encoded as eight composable slash commands
  • 24 production skills: From requirements to release — the most complete free collection
  • Chrome DevTools credibility: Authored by Addy Osmani, engineering leader on the Chrome team
  • Engineer ergonomics: Skills read like great engineering docs, not prompt hacks
  • Mainstreaming the category: 81K⭐ makes agent-skills the tipping point for enterprise adoption
Category Defining
production-grade
slash-commands
chrome-devtools

#2 obra/superpowers — Agentic Skills Framework & SDLC Methodology ⭐ 266,458

Shell • 23,825 forks • MIT License • Created Oct 2025 • The category's origin story

The project that invented the agent skills paradigm in its modern form. Superpowers treats software development as a subagent-driven skill system — every phase of the SDLC becomes a composable, reusable skill managed by specialized subagents. With 266K+ stars it remains the biggest repo in the ecosystem, and its "skills + subagents + methodology" model is the blueprint that addyosmani/agent-skills productized for production teams.

  • Subagent-driven development: Each subagent owns a domain with dedicated skills
  • Composable methodology: Chain skills into full workflows
  • Massive community: 266K+ stars, 23K+ forks, active daily development
  • SDLC integration: Brainstorming → coding → review as skill pipelines
  • MIT licensed: Free to fork, extend, and embed
Framework Pioneer
subagent-driven-development
methodology

#3 zhaoxuya520/reverse-skill — Reverse Engineering Skill Router Pack ⭐ 19,064 +9,904/week

PowerShell • 2,448 forks • MIT License • Created May 2026

The fastest-growing security skill pack in the ecosystem, growing nearly 10K stars a week. Reverse-skill is an AI-powered skill router for reverse engineering, penetration testing, and security research — it routes tasks to the right toolchain, bootstraps tools on demand, and maintains a self-evolving knowledge base. Supports Claude Code, Kiro, Cursor, Cline, and more.

  • AI-powered routing: Dispatch security tasks to the right skill automatically
  • Self-evolving knowledge base: Learns from every security research session
  • On-demand toolchains: Bootstrap RE tooling when a task needs it
  • Multi-client: Claude Code, Kiro, Cursor, Cline, and other AI coding tools
  • Explosive growth: +9,904 stars/week — the hottest skill router on GitHub
Skill Router
security
reverse-engineering
pentesting

#4 virgiliojr94/book-to-skill — Books → Claude Code Skills ⭐ 16,990 +4,596/week

Python • Created Jul 2026 • +4,596 stars/week

The most creative skill-generation tool in the ecosystem: it turns entire technical books into executable Claude Code skills. Feed it a technical book, and it extracts the methodologies, patterns, and workflows into structured skills the agent can actually apply. Growing at nearly 4,600 stars a week, it proves that skill authoring doesn't have to start from a blank page — the world's technical literature is a skill library waiting to be compiled.

  • Book → skill compilation: Convert whole technical books into agent-applicable skills
  • Claude Code native: Outputs drop into the Claude Code skills directory
  • Knowledge extraction: Distills methodology, not just facts
  • Rapid adoption: +4,596 stars/week — one of the fastest-growing skill tools of 2026
  • Ideal for teams: Turn internal runbooks and classic texts into reusable agent capabilities
Skill Generation
claude-code
books
docs-to-skills

#5 ayghri/i-have-adhd — ADHD-Friendly Agent Output ⭐ 17,278 +3,874/week

Markdown • +3,874 stars/week • Created 2026

A skill that reworks how agents present information for ADHD-friendly consumption — short, scannable, action-oriented output instead of walls of text. It's the clearest evidence yet that the skills paradigm covers human factors, not just engineering. Nearly 17.3K stars and growing ~3,900/week shows how underserved the "agent UX accessibility" layer was before skills made it codifiable.

  • Output restructuring: Forces concise, scannable, action-first agent responses
  • Cognitive load reduction: Fewer, clearer decisions per message
  • Universal applicability: Works for any agent, any task, any user
  • Massive demand signal: +3,874 stars/week proves the UX gap was real
  • Codified accessibility: Human factors as a versionable skill
Output Quality
agent-ux
accessibility
productivity

#6 microsoft/skill-recorder — Record Work → Publish as CLI Skills ⭐ 969

TypeScript • Electron desktop app • MIT License • Created Jul 2026

Microsoft's enterprise answer to skill authoring: record your on-screen workflow, and the GitHub Copilot CLI reconstructs it into a reusable Skill or Automation for Microsoft Scout, Copilot Cowork, or Copilot Studio. This is the "teach once, automate forever" pattern — no hand-writing skill files. Lower star count than the consumer hits, but the highest enterprise credibility in the category.

  • Screen recording → skill: Capture real workflows, publish as CLI skills
  • Copilot CLI reconstruction: AI turns the session into intent + ordered steps
  • Multi-target output: Microsoft Scout, Copilot Cowork, Copilot Studio
  • Electron desktop: Cross-platform recording experience
  • Enterprise backing: Microsoft docs, support, and platform integration
Skill Recording
microsoft
copilot
enterprise

#7 vercel-labs/skills — The Skills CLI Framework ⭐ npx skills add

TypeScript • Vercel Labs • The distribution layer for the skills ecosystem

Vercel's skills CLI is the package manager moment for agent skills: npx skills add installs a skill into your agent's skills directory from a registry. If addyosmani/agent-skills is the content and obra/superpowers is the methodology, vercel-labs/skills is the plumbing — a standardized install/update/remove interface that makes skills as easy to distribute as npm packages. This is the layer that turns a grassroots movement into a platform.

  • npx skills add: One-command skill installation into any agent
  • Registry-based distribution: Publish and discover skills like packages
  • Vercel ecosystem: Backed by the platform behind Next.js and v0
  • Standardization: A common contract for skill packaging
  • Framework of frameworks: Complements agent-skills, superpowers, and every skill pack
Framework
cli
vercel
package-manager

#8 ponytail-improved — Agent Prompt Optimization ⭐ 580

JavaScript • 85 forks • Community fork of the ponytail YAGNI system

An enhancement layer for the ponytail skill system: optimized skill injection, sharper reasoning chains, and stricter YAGNI enforcement. Where the 95K⭐ parent project defines the "lazy senior dev" philosophy, ponytail-improved is the tuning fork — better prompts, better tool use, lower token consumption. A small but well-loved quality layer in the stack.

  • Skill injection optimization: Smarter placement of behavioral skills in context
  • YAGNI enforcement: Harder guardrails against over-engineering
  • Token efficiency: Less context, same or better output quality
  • Reasoning improvements: Better decision chains during skill application
  • Framework-agnostic: Drops into Cursor, Claude Code, Cline, and more
enhancement
yagni
prompt-optimization

#9 ratchet — Agent Compliance Checking ⭐ 412

Python • Agent compliance and policy verification layer

The compliance guardrail for the skills era. Ratchet verifies that agents follow organizational policy — coding standards, security rules, approved toolchains — before actions are taken or code is shipped. As skills make agents more capable, ratchet makes them more trustworthy: a policy-checking layer that enterprises will bolt onto any skills framework.

  • Policy verification: Check agent output against org standards
  • Pre-action gating: Block non-compliant operations before they execute
  • Audit trail: Record what was checked and why
  • Framework-agnostic: Layers on top of any agent or skill system
  • Enterprise prerequisite: The compliance story enterprises need before wide rollout
compliance
security
governance

#10 codex-as-mcp — Codex as MCP Server Bridge ⭐ 169

TypeScript • MCP bridge • Bridges OpenAI Codex into the MCP/skills ecosystem

A small but strategically important bridge: it exposes OpenAI Codex as a Model Context Protocol server, letting MCP-based tools and skill systems drive Codex directly. As the skills ecosystem standardizes on MCP for tool connectivity, bridges like this determine which agents can consume which skills. 169 stars today; outsized ecosystem influence per star.

  • Codex ↔ MCP: Expose Codex as a standard MCP server
  • Interop: Let MCP-native clients and skill routers drive Codex
  • Standardization: Another nail in the "everything speaks MCP" coffin
  • Small but pivotal: Bridges compound the value of every skill pack
mcp
codex
bridge
interop

💡 The Skills Paradigm: Why This Ecosystem Is Exploding

💡 The Core Idea: Monolithic prompts are dead. A prompt is a one-shot conversation; a skill is a versionable, reviewable, reusable capability. Skills turn AI coding agents from "impressive demos" into "dependable engineering tools" — and that shift is why the ecosystem is growing faster than any AI category since ChatGPT plugins.

The agent skills ecosystem is exploding for five structural reasons, and Run #53's data makes each one measurable:

1. Production credibility arrived (the Addy Osmani effect)

81,976⭐ in weeks

Until now, the skills movement was led by indie hackers and framework pioneers. A Chrome DevTools engineering leader shipping addyosmani/agent-skills changed the audience: enterprise engineers who trust browser-tooling discipline now have a reference implementation they can defend in architecture reviews. The "production-grade" label went from marketing to reality.

2. Skills beat prompts on every axis

Deterministic • Reviewable • Composables

A skill is version-controlled, diffable, tested, and composable — everything a prompt is not. Teams can put skills through code review, measure their effect, and roll them back. That's the difference between "agent vibes" and "agent engineering," and it's why enterprises are standardizing on skills.

3. The distribution layer materialized

npx skills add

vercel-labs/skills gave the ecosystem its npm moment: a one-command installer backed by a registry. Combined with Claude Code's native skills directory and Cursor's rules support, installing a skill pack is now a 10-second operation. Distribution bottlenecks were the last thing holding the category back — they're gone.

4. Authoring costs collapsed

Record it • Compile it • Ship it

Three parallel authoring revolutions: microsoft/skill-recorder (record your screen, get a skill), book-to-skill (compile a book into skills), and the sheer volume of open skill packs (reverse-skill, i-have-adhd). When creating a skill is easier than writing a blog post about it, the ecosystem compounds.

5. The guardrails caught up

ratchet • ponytail-improved • i-have-adhd

Quality and compliance layers matured in parallel: ratchet for policy compliance, ponytail-improved for output discipline, i-have-adhd for human-readable output. Enterprises won't deploy capable agents without guardrails — the fact that the guardrails now exist as skills themselves is the ecosystem maturing in real time.

📈 The Trend Line

In Q1 2026 the "agent skills" category barely existed. By August 2026 it exceeds 404K stars across just the 10 tools on this page, with three repos growing at 3,800–9,900 stars per week. The comparison to npm (2015), Kubernetes (2016), and LangChain (2023) is apt — this is category-creation velocity, and the window to adopt (and build on) skills is now.

🪜 The Skills Stack: Four Layers of the Production Stack

The ecosystem isn't a grab bag of repos — it's a coherent four-layer stack. Understanding the layers explains why the pieces fit together and where the gaps (and opportunities) are.

L1

🧱 Skills Framework

The foundation: how skills are defined, packaged, installed, and executed.

  • vercel-labs/skills — the CLI/registry layer
  • obra/superpowers — the methodology layer
L1 answers: how do skills work?
L2

📦 Skill Packages

The content: production skill collections that encode real expertise.

  • addyosmani/agent-skills — 24 skills, 8 commands
  • reverse-skill — security/RE skill pack
  • i-have-adhd — output-quality skill
L2 answers: what can agents do?
L3

🧭 Skill Routers

The intelligence: deciding which skill applies to which task, and loading the right toolchain.

  • reverse-skill — AI-powered routing + toolchain bootstrapping
  • ratchet — policy gating before skill execution
L3 answers: which skill runs now?
L4

🎥 Skill Recording & Generation

The flywheel: creating new skills from demonstrations, documents, and books without hand-authoring.

  • microsoft/skill-recorder — screen recording → CLI skills
  • book-to-skill — books → Claude Code skills
L4 answers: how do skills multiply?
💡 The Stack Insight: The four layers map cleanly onto software history — L1 is the build system, L2 is the package ecosystem, L3 is the runtime/router, L4 is the compiler/recorder. Every successful platform (npm, Maven, PyPI) followed exactly this arc. The skills stack is recapitulating it at record speed, with addyosmani/agent-skills as the flagship package and vercel-labs/skills as the would-be npm. Bridges like codex-as-mcp are the interop glue that will decide how many runtimes can consume the stack.

✅ Best For Picks

Use CaseBest ToolWhy
Production engineering teamsaddyosmani/agent-skills24 battle-tested skills + full /spec→/ship lifecycle from Chrome DevTools credibility
Full SDLC methodologyobra/superpowers266K⭐ subagent-driven framework — the deepest methodology in the category
Security & reverse engineeringzhaoxuya520/reverse-skill19K⭐ AI-routed skill pack with self-evolving knowledge base, +9.9K stars/week
Training agents from existing knowledgevirgiliojr94/book-to-skillCompile books and docs into Claude Code skills at +4.6K stars/week
Enterprise skill capturemicrosoft/skill-recorderRecord workflows → publish skills for Copilot ecosystems, Microsoft-backed
Agent output readabilityayghri/i-have-adhd17.3K⭐ proof that scannable, action-first agent output is a feature, not a nicety
Distributing skills at scalevercel-labs/skillsnpx skills add — the package-manager moment for the whole ecosystem
Governance & complianceratchetPolicy checking and pre-action gating for regulated environments
YAGNI disciplineponytail-improvedOptimized prompt injection and stricter anti-over-engineering guardrails
Cross-runtime interopcodex-as-mcpExpose Codex to MCP-native skill routers and clients

🏆 The Bottom Line Recommendation

Start with addyosmani/agent-skills

For most teams in August 2026, the fastest path to production-grade agent skills is: install addyosmani/agent-skills as your baseline skill package, layer vercel-labs/skills (or your agent's native skills dir) as the installer, add ratchet or equivalent compliance gating if you're regulated, and study obra/superpowers for the methodology. Then start feeding your own knowledge in via book-to-skill-style pipelines. The stack is ready — and unlike most AI hype, it's versionable, reviewable, and shippable.

📈 Market Analysis & Timing

💡 Market Timing: The production agent skills category is in its explosive "platformization" phase. Three signals converged in the same week: a Chrome DevTools leader shipping the production reference (81K⭐), Vercel standardizing distribution, and Microsoft industrializing skill capture. When the browser team, the web platform team, and the enterprise AI team all move on the same paradigm, adoption is no longer a question of if — only how fast.

📊 Category Signals

  • Total category stars: 404K+ across the 10 tools on this page
  • Fastest growers: reverse-skill (+9,904/wk), book-to-skill (+4,596/wk), i-have-adhd (+3,874/wk)
  • Steady compounder: addyosmani/agent-skills at +226/day (≈1,600/week)
  • Enterprise validation: Microsoft (skill-recorder), Vercel (skills CLI), Google (via Addy Osmani)
  • Stage: Category Creation → Platformization — standards are forming (MCP, skills dirs, registries)
  • Monetization window: NOW — the next 6–12 months are prime for skill-marketplace, skill-authoring SaaS, and skills-focused affiliate content

🚀 Deployment & Hosting

🏠 Hosting Affiliate Links

Skill frameworks, routers, and recording pipelines need infrastructure. Self-host a skill router, run a skills registry, or deploy agent endpoints:

  • Hostinger: Best value for self-hosting skill routers and agent endpoints. VPS from $3.99/mo with 100Mbps connectivity. Deploy on Hostinger
  • DigitalOcean: Best for production skill routers and security tooling (reverse-skill, ratchet). Droplets from $6/mo. Deploy on DigitalOcean
  • Vercel: Perfect for skills registries, skill dashboards, and lightweight agent endpoints — and the home of the skills CLI itself. Generous free tier. Deploy on Vercel

💻 IDE & Development Tools

Skills are authored and consumed inside AI-native IDEs. These are the best platforms for running the skills stack:

  • Cursor: The AI-native IDE with native rules/skills support — addyosmani/agent-skills and ponytail both ship Cursor-compatible rules. Try Cursor
  • Windsurf: The fastest-growing AI IDE with multi-agent coding and deep skill-aware context. Try Windsurf
  • Claude Code: The primary runtime for the skills ecosystem — agent-skills, book-to-skill, reverse-skill, and i-have-adhd all target it first. Try Claude Code
  • GitHub Copilot CLI: Required for microsoft/skill-recorder workflow reconstruction. Try Copilot

☁️ Cloud & AI Infrastructure

Run the skills stack at scale — skill verification pipelines, compliance checking (ratchet), and enterprise recording:

  • AWS (Bedrock): Managed AI service for running skill-execution agents at enterprise scale. Explore AWS Bedrock
  • Google Cloud (Vertex AI): Enterprise agent orchestration with Gemini and open models — pairs naturally with Chrome-team skills like webperf. Explore Google Cloud Vertex AI
  • Microsoft Azure (AI Foundry): Enterprise AI infrastructure — the natural home for skill-recorder and Copilot Studio workflows. Explore Azure AI Foundry
⚡ Affiliate Disclosure: Some links on this page are affiliate links. We may earn a commission if you sign up for services through these links, at no extra cost to you.

🔗 Related Comparisons

💡 Cursor recommendation: Ai-first code editor with agentic autocomplete. Try Cursor →

Affiliate disclosure: we may earn a commission if you sign up via this link, at no extra cost to you.

💡 Cloudflare recommendation: Cdn, dns, ddos protection, and edge compute for any site. Try Cloudflare →

Affiliate disclosure: we may earn a commission if you sign up via this link, at no extra cost to you.