Source: github.com/coreyhaines31/marketingskills (Corey Haines, MIT licence, 41.4k stars, 48 skills)
This lead mag covers the same 48 real skills already documented in the general directory and the founder-journey playbook, but organised a third way: by workflow chain. Instead of grouping by category or by founder stage, this shows the specific sequence of skills an AI agent runs together to execute one real campaign, from a product launch to a retention save.
The repo's skills are individually powerful but genuinely compound when chained. A single skill audits a page. A chain plans the traffic, writes the page, tests it, and measures it. This lead mag documents 6 real chains built from the 48 skills.
All skill names and trigger descriptions below are sourced verbatim from the repository README. Full step-by-step instructions inside each individual skill require installing the repo directly (raw GitHub file content is not accessible without cloning).
Why chains matter more than individual skills
Every skill in this repo reads the product-marketing context file first. That single shared dependency is what makes chaining possible: run product-marketing once, and every skill after it in a chain already knows your business, so the agent moves from one step straight into the next without you re-explaining context between them.
The 6 chains below are not officially defined in the README as named workflows; they are the natural sequences the repo's own dependency map and category structure point to when several skills clearly feed into each other for one real outcome.
Install everything before chaining
# Install the full repo (recommended for chaining, since any chain may need any skill)
npx skills add coreyhaines31/marketingskills -a claude-code
# Always run this first, before any chain
# (triggers automatically when the agent needs product/audience context, but running it explicitly once at setup is the safest start)
"Create my product-marketing context document."