there's a free github repo with 5,100+ stars that turns claude code into a full marketing department.

CRO audits. landing page rewrites. email sequences. ad copy. programmatic SEO. pricing strategy. 25+ marketing skills you install with one command.

i've been running it on client work for weeks. here's the setup, the output, and what it replaced.

what claude code is (30-second version)

claude code is anthropic's AI coding agent. it runs in your terminal. reads your files. executes multi-step tasks. launched february 2025. hit $1 billion in annualized revenue within six months. hit $2.5 billion ARR by february 2026.

but "coding agent" is misleading. you don't need to write code to use it. describe what you want in plain english. it does the rest.

90% of claude code's own codebase is written by the tool itself. a senior google engineer said it reproduced a year of architectural work in one hour. the stack overflow 2025 developer survey found 84% of developers use AI tools daily.

this isn't a toy. it's the fastest-growing developer tool in history. and it works for marketing.

the marketing skills repo (free)

corey haynes built a github repo that packages proven marketing frameworks into claude code skills. ogilvy. hopkins. cialdini. schwartz. the principles behind the most successful direct response copy, sales pages, and funnels of the last century.

25+ skills. install with one command. each skill loads a specific framework into claude code's context before it writes anything.

the skills include:

• copywriting (direct response principles)

• CRO audits (conversion rate optimization)

• email sequences (nurture + sales)

• ad copy (paid social and search)

• programmatic SEO (page generation at scale)

• pricing strategy (value-based frameworks)

• positioning (competitive differentiation)

• landing page rewrites (full page overhauls)

you don't prompt claude and hope for the best. the skill tells it what to think. you tell it what to analyze.

what a CRO audit looks like (90 seconds)

i fed it a real product page. 15% conversion rate from warm traffic. solid page. performing well.

the output came back in 90 seconds:

"your hero section is word soup. two headlines stacked on top of each other. the visitor's brain has to parse it in reverse. that's friction."

it caught things a $200/hr CRO consultant would catch. then it fixed them. the original headline was generic SaaS copy. the rewrite was a pattern interrupt. the AI understood why because the skill is built on principles from people who spent decades studying what makes people stop scrolling.

that's the difference between asking chatgpt to "improve my headline" and running a structured CRO skill that evaluates your page against proven conversion frameworks.

what programmatic SEO looks like (6 minutes)

i ran a programmatic SEO strategy on a real site. no spreadsheets. no prep. no data dump.

claude code pulled the sitemap. audited 440 location pages. cross-referenced google search console data.

the finding: 440 pages generating 170 total clicks over 3 months. that's 0.13 clicks per page per month. the infrastructure existed but produced nothing.

it ranked opportunities by ROI. identified which keywords had proven demand in GSC. flagged which searches google was already surfacing results for but the site had no dedicated page to capture.

then it laid out the fix: persona + service + industry pages. high intent. defensible with case studies. scales to 100+ pages.

the SEO team hadn't flagged any of it. claude code found it in 6 minutes.

how to stack skills for 10x output

the difference between AI slop and output that converts is context. most people open a chatbot, throw in a prompt, get generic output.

the skills system works differently. you stack them:

1. run the positioning skill (who you are, who you serve, why you're different)

2. feed that output into the copywriting skill (headlines, body, CTAs)

3. run the CRO skill against the draft (grades every section, suggests fixes)

4. run the SEO skill on the final version (keyword optimization, meta tags, internal linking)

each output feeds the next. by the time it writes your landing page, it already understands your positioning, your audience, and the conversion principles that should inform every line.

add MCP servers and it pulls live data. playwright MCP for competitor screenshots and copy scraping. search console MCP for keyword data. the skills tell it what to think. the MCPs tell it what to see.

the cost comparison

typical marketing setup for a mid-market company:

• marketing consultant: $5K-$15K/month

• copywriter: $5K-$15K per project

• CRO audit: $5K-$10K

• SEO strategist: $3K-$5K/month

claude code max subscription: $200/month. the skills repo: free.

it won't replace senior strategic thinking. you still need someone who understands your market and your customers. but it replaces 80% of the execution that burns time and budget every month.

how to get started in 15 minutes

step 1: install claude code npm install -g @anthropic-ai/claude-code

step 2: clone the marketing skills repo from github

step 3: load the CRO skill. feed it your highest-traffic landing page URL. run the audit.

step 4: implement the top fix. measure the result at 7 days.

step 5: repeat with email sequences, ad copy, and SEO strategy. each skill takes 30 seconds to load and 2-5 minutes to produce output that used to take a consultant a week.

we use claude code and these skills across every client engagement. it's how a small team delivers what used to take 15 people.

book a call → agentintegrator.io