Claude Marketing Skills: 32 Picks for Claude Code

One repo changed how I think about AI for marketing. coreyhaines31/marketingskills has 45K stars and 50 skills organized into a full marketing stack. CRO, copywriting, SEO, paid ads, analytics, retention, growth engineering, and sales operations. But the clever part is not the skills themselves. It is the architecture: every skill reads from a shared .agents/product-marketing.md file. Define your product, audience, and positioning once. Every skill output respects it.

That is the difference between "AI for marketing" (paste context, get generic output) and an actual marketing ops system (define strategy once, execute everywhere).

Here is what that looks like in practice. Without the shared context, you prompt Claude "write me a landing page headline" and get:

Transform Your Business with Our Innovative Solution

With product-marketing context defined and the copywriting skill installed:

Ship 3x faster. The 847 engineering teams on [Product] merge 40% more PRs per week - without adding headcount.

Same model. The skill pulled your positioning (developer tools), your proof point (847 teams, 40% metric), and your direct voice from the shared context file. That is what "define once, use everywhere" means in practice.

I cataloged the full marketing skills ecosystem. The coreyhaines31 repo is the centerpiece, but there are dedicated skills for ad copy, competitor analysis, email sequences, LinkedIn optimization, and AI-native SEO (GEO) scattered across a dozen other repos. Here is everything worth installing.

What Is the Shared Context Pattern and Why Does It Matter?

Before installing any marketing skill, understand the pattern that makes them 10x more useful.

The best marketing skill repos use a shared context file. In coreyhaines31/marketingskills, this is the product-marketing skill. It creates a context document that captures:

• Product positioning: What you sell, who you sell to, the problem you solve
• ICP definition: Firmographics, psychographics, buying triggers
• Competitive landscape: Key competitors, your differentiation
• Brand voice: Tone, vocabulary, personality
• Proof points: Customer quotes, case studies, metrics

Every other skill in the collection reads this context automatically. The copywriting skill pulls your positioning into landing page copy, while cro checks pages against your ICP. For topic-cluster planning, content-strategy uses your business context and competitive landscape.

Set this up first. Run the product-marketing skill and answer the questions. Every subsequent skill becomes 10x more relevant.

What Marketing Skills Should You Install?

CRO & Conversion Optimization

SEO, GEO & AI Visibility

The ai-seo skill stands out because most SEO skills stop at Google. It optimizes for AI citation, including the format and structure that gets your content pulled into ChatGPT, Perplexity, and Gemini answers. That is where organic traffic is shifting.

Content Marketing

What makes these useful: The content-strategy skill builds a full content pillar architecture with hub pages, spoke articles, and an internal linking map grounded in your positioning. The content-creator skill's repurposing workflow is the time saver: one long-form piece becomes 8-12 distribution assets (LinkedIn post, email teaser, Twitter thread, social cards) without manual rewriting.

Paid Advertising

What makes these useful: The ads skill selects platforms based on your ICP (LinkedIn for B2B enterprise, Meta for B2C, Google for high-intent), recommends budget allocation, and structures campaign testing. The ad-copy-writer from cognyai generates platform-specific variants that respect character limits and format constraints, saving the manual work of adapting a LinkedIn ad to Google's headline character cap.

Email Marketing & Lifecycle

What makes these useful: The emails skill generates complete flows, including sequence length, timing, subject lines, and the job of each email. The churn-prevention skill focuses on voluntary and involuntary churn with cancel flows, targeted save offers, and failed-payment recovery.

Analytics & Attribution

What makes these useful: The analytics skill builds a measurement plan that covers event tracking, funnel analysis, and multi-touch attribution. The gtm-metrics skill from chadboyda calculates the metrics a board actually cares about: CAC payback period, NRR, and growth accounting (new vs expansion vs contraction vs churn revenue).

Growth Engineering & PLG

Social Media & Brand

Competitive Intelligence

Where Does the Data Come From?

One thing the tables above gloss over: a lot of these skills assume the agent can already see keyword volume, SERP results, backlinks, ad libraries, or company data. The seo skill writes a good brief, but only if it has real volume numbers. competitor-profiling is guesswork without backlink and ad data. Normally that means a Semrush, Moz, or Apollo subscription per data source, plus an API key the agent has to hold.

The workaround we use is treg, a source-available tool registry (Apache 2.0 with extra terms) that sits in front of about 2,800 endpoints across those providers and bills per call on one token. The agent searches the catalog for the job ("keyword volume for these 300 terms"), reads the price, and calls it. Our team built it for our own SEO loops, and it is self-hostable if you would rather run it on your own keys.

npx skills add superdesigndev/treg -s treg
treg mcp install   # adds the MCP server to Claude Code

Two things worth knowing before you rely on it. Calls are metered from a prepaid balance, so batch keyword lists into one call instead of one per term. And your own connected accounts always take priority over the shared keys, so if you already pay for a provider, connect it and those calls are free.

Which Repos Have the Best Marketing Skills?

How Do You Stack Skills for a Growth Workflow?

The framing that has taken over on YouTube this summer is "build an AI marketing team," and it is the right one. The method most of those walkthroughs converge on is four steps: list the marketing tasks you actually repeat every week, turn each one into a skill, group related skills into agent roles that do not overlap (a researcher, a writer, an analyst), then tell Claude in your project context who is on the team and when to reach for an agent versus a single skill. The stack below is that team, written as a pipeline.

Strategy & positioning:
  → product-marketing (define product, ICP, positioning ONCE)
  → competitor-profiling (competitive landscape)
  → marketing-loops (recurring workflows with checks, saved state, and stopping conditions)

Acquisition:
  → seo-audit + ai-seo (traditional SEO diagnosis + AI visibility)
  → ads + ad-copy-writer (paid channels)
  → content-strategy + content-creator (content engine)
  → social + linkedin-automation (social distribution)

Conversion:
  → cro (audit existing pages and forms)
  → copywriting (new page copy with variant generation)
  → landing-page (full page generation)

Retention & expansion:
  → emails (lifecycle flows)
  → churn-prevention (cancel flows, save offers, dunning)
  → analytics (funnel analysis, cohort analysis, attribution)

Sales handoff:
  → revops (scoring, routing, handoff)
  → sales-enablement (decks, one-pagers, battle cards)

Every skill in this stack reads from the product-marketing context file. Change your positioning once, and every output across every channel updates.

Install them where every tool can see them. v2.0 moved the context file to .agents/product-marketing.md for a reason: .agents/skills/ and AGENTS.md are the locations Cursor, Codex, and the rest read too, while .claude/ is Claude Code only. This matters more than it sounds. Tobi Lutke, Shopify CEO, said on August 25, 2026 he is considering banning Claude Code at Shopify until it reads AGENTS.md and .agents/skills, because a team where each person's tool reads a different config ends up with a split-brain marketing setup. Put the skills in .agents/, point CLAUDE.md at AGENTS.md, and the marketer on Cursor gets the same cro audit you get.

How Do You Fork and Customize Skills for Your Brand?

Firecrawl's team wrote about this: they rarely use the coreyhaines31 skills as-is. Instead they fork individual skills and customize them with their specific brand voice, audience, and positioning. The copywriting and CRO skills "especially require tuning before they produce output that actually fits your brand." (They reviewed it back when CRO was still split into page-cro and form-cro.)

The workflow:

1. Install the full coreyhaines31 collection
2. Run product-marketing to define your context
3. Test each skill with a real task
4. Fork the skills that need customization into your project's .claude/skills/ directory
5. Add your gotchas list (brand-specific things the AI gets wrong)

The gotchas list is where the value compounds. "Never use the word 'revolutionary' - our brand is understated." "Always lead with the specific number, not the percentage." "Our competitor X was acquired - position against their new parent company, not the old brand."

Related Content

• Anthropic Built 300+ Claude Code Skills Internally. Here's What They Learned. - How to build your own skills + the complete skills directory by category
• Claude Code for Sales - 60+ sales skills mapped to every pipeline stage
• Claude Code for Data Scientists - 35+ data science skills for EDA, visualization, and modeling
• Claude Code for DevOps - 150+ skills for Docker, K8s, Terraform, and CI/CD
• How to Become an AI-Native Company - What it looks like when the marketing skills above stop being a session and become a loop that runs without you
• Claude Code for Solopreneurs - The 15-skill stack that replaces 5 hires

Start Here

Install the coreyhaines31 collection: npx skills add coreyhaines31/marketingskills -a claude-code. Run the product-marketing skill first. Then test cro on your homepage. That single audit will show you the difference between generic AI advice and context-aware marketing ops.

Want to build your own marketing skills from scratch? The shared context pattern in this guide is just the beginning. Encoding YOUR brand voice, YOUR conversion data, YOUR competitive positioning into custom skills is where the real compounding happens. Agent Skills 101 walks you through the SKILL.md spec, building a skill from scratch, testing and iterating, and making it work across Claude Code, Cursor, and 50+ agents. 25 lessons.

Start Agent Skills 101 →