marketingskills: A Set of God-Tier Marketing Skills That Turn Your AI Into a Full Marketing Team

marketingskills GitHub repo with 40.3K stars

marketingskills: A Set of God-Tier Marketing Skills That Turn Your AI Into a Full Marketing Team

Yesterday I stumbled on a repo on GitHub: coreyhaines31/marketingskills, with 40.3K stars.

I thought it was yet another code framework, but clicking in I found it’s all Markdown files – 46 marketing Skills built for coding agents like Claude Code and Codex. Copywriting, landing-page review, SEO audits, email sequences, A/B test design – basically every marketing task you can think of is covered.

Having run a tech account for years, I’ve seen too many developers with good products die because they couldn’t sell. This project fills exactly that gap, and is worth a write-up.

marketingskills GitHub repo with 40.3K stars
The marketingskills repository on GitHub.

What Kind of Project Is This

The author, Corey Haines, comes from the marketing world – founder of Conversion Factory and writer of a marketing newsletter called Swipe Files. He is not a programmer; what he did was distill the experience he accumulated doing B2B SaaS marketing for years into a set of playbooks, formatted so an agent can use them directly.

Corey Haines author of marketingskills
Corey Haines, creator of marketingskills.

A Skill is, plainly put, a Markdown file – one file per marketing task, containing the methodology and concrete operating steps. Once installed into a project, when the agent notices we’re doing marketing-related work, it pulls out the corresponding framework and uses it. The license is MIT, so take it and modify freely.

What’s rare is it isn’t tied to any tool – Claude Code, OpenAI Codex, Cursor, and Windsurf can all install it, as long as the agent supports the Agent Skills spec.

Feature Details

All Skills First Read One Shared Product Profile

This design is the foundation of the whole repo, and I think it’s also what sets it apart from ordinary “AI marketing prompt collections.”

product-marketing skill in marketingskills
The product-marketing skill defining the shared product profile.

There’s a Skill called product-marketing with the most special status. The first time we use it, it walks us step by step through clarifying product positioning, target users, competitor situation, brand voice, and customer cases, then saves it as a context file.

After that, no matter which Skill we call – writing copy or reviewing a landing page – it reads this profile first before acting. The day the product positioning changes, edit this one file and every other Skill’s output changes with it.

Without this step, AI-written marketing content is the kind of generic platitude that works for any product if you just swap the name. With this profile, the output actually fits our own product.

Conversion-Rate Optimization Is Split Very Finely

CRO conversion optimization skills
Conversion-optimization skills in marketingskills.
  • cro: reviews any marketing page and form, then gives a prioritized fix list – is the headline clear, is the form too long, is social proof missing – item by item.
  • signup: manages the registration and trial-activation flow.
  • onboarding: manages post-signup onboarding so users taste the product’s value fast.
  • popups: conversion optimization for popups, overlays, and banners.
  • paywalls: in-app paywalls and upgrade prompts.
  • Most AI marketing tools, when asked, just say “help me optimize this page.” This project splits conversion into five Skills, each with its own review framework – a completely different level of granularity.

    Copywriting Has a Complete Pipeline

  • copywriting: writes homepage and landing-page copy using classic frameworks like PAS and AIDA, several variants at once.
  • copy-editing: revises existing copy – note it doesn’t rewrite from scratch, it polishes paragraph by paragraph, with a reason given for every change.
  • cold-email: B2B cold emails and follow-up sequences.
  • emails: automated email flows – welcome sequences, re-engagement sequences – with send rhythm and branching logic worked out.
  • social: social content for LinkedIn, Twitter/X, and similar platforms.
  • SEO Even Covers AI Search

    SEO and AI-SEO skills
    SEO and AI-search optimization skills.
  • seo-audit: technical and on-page SEO audits.
  • ai-seo: worth a special mention – it specifically optimizes content so AI search engines like ChatGPT, Perplexity, and Gemini are willing to cite us. As search traffic visibly shifts toward AI, this Skill is well positioned.
  • programmatic-seo: batch-deploys SEO pages using templates plus data.
  • schema: structured-data markup.
  • competitors: generates competitor-comparison pages to capture “XXX alternatives” search traffic.
  • Advertising, Testing, and Data Analysis

  • ads: ad campaigns on Google, Meta, and LinkedIn – helps with platform choice and budget allocation.
  • ad-creative: batch-generates ad-copy variants.
  • ab-testing: experiment design – hypothesis, variables, success metrics, all set up.
  • analytics: event tracking and attribution – tells you which events to actually track.
  • Growth and Strategy – Misc but Practical

  • marketing-ideas: 140 SaaS marketing ideas to flip through when stuck.
  • pricing: pricing and plan design.
  • launch: product-launch rhythm planning.
  • referrals: referral rebates and refer-a-friend.
  • churn-prevention: churn recovery, including cancellation flow, retention offers, and failed-charge recovery.
  • On the sales side there’s also revops and sales-enablement – the former manages lead tiering, assignment, and pipeline; the latter manages sales scripts, one-pagers, and objection-handling docs. Useful for anyone doing B2B.

    Quick Start

    Installation has no real barrier – it’s a one-line command.

    Full install:

    Install only the ones you want:

    It auto-installs into the .agents/skills/ directory and also symlinks a copy to .claude/skills/ for Claude Code compatibility.

    You can also install via Claude Code’s plugin mechanism:

    After installation, just state your needs in natural language and the agent picks the Skill itself:

    Or name one directly with a slash command, like /cro or /seo-audit.

    My suggested onboarding order: first run product-marketing once to build the product profile, then run cro to review your own homepage. After just that round, you’ll feel the gap versus asking bare AI.

    marketingskills quick start install
    Quick-start installation of marketingskills.

    One more reminder: v2.0 renamed 17 Skills, merging page-cro and form-cro into cro. If you installed an older version, after upgrading remember to clear out the old directory, otherwise the old and new sets coexist and the agent gets confused.

    My Take

    Why this repo blew up, I think, is pretty plain. It invented nothing new – it just honestly solidified the framework in a senior marketer’s head into a format an agent can read. But too few people do this: marketing methodology is everywhere online, yet organizing it into a form an agent can directly digest and that skills can call each other – this is the first.

    But don’t oversell it either. The role of these Skills is advisor, not doer. They can teach the agent how to think about a problem and produce plans and copy, but the one who clicks the button, launches the ad, and sets the price is still us. For true end-to-end automation you’d need to wire up an MCP tool chain on top.

    marketingskills framework my take
    The framework approach behind marketingskills.

    One realistic caveat must be stated clearly: the whole thing grew up in an English-marketing context. The framework itself is universal, but many specific tactics are overseas-SaaS playbooks – cold email, Product Hunt launches – and bringing them into domestic business means adapting them yourself. You can use it as a skeleton and fill in your own industry experience. Even just as a skeleton, this repo is enough – after all, it’s free.

    Open-Source Address

    https://github.com/coreyhaines31/marketingskills

    By peter_lzh

    Author of in-depth reviews of AI open-source tools; focuses on identifying high-value open-source projects and providing practical testing results as well as guidance for making choices.

    Leave a Reply

    Your email address will not be published. Required fields are marked *