Most LLM agents are generalists; they still make avoidable marketing mistakes because they lack compact, shareable workflows that encode domain best practices. This collection treats marketing know-how as agent-native skills — short, interlinked markdown files that give an agent the context, checklist, and heuristics needed to run real marketing tasks reliably.
What Sets It Apart
- Modular skill files that check a single product-marketing-context first, then run specialized workflows (so what: avoids inconsistent outputs by forcing a single source of product truth).
- Cross-referenced skill graph for typical marketing domains (SEO, CRO, copy, paid measurement, growth engineering) (so what: lets agents chain tasks—research → copy → experiment plan—without losing context).
- Designed for multi-agent compatibility (Claude Code, OpenAI Codex, Cursor, Windsurf and any Agent Skills–compatible agent) and toolchains (so what: you can reuse the same skills across different agent runtimes without rewriting workflows).
- Focused on pragmatic deliverables rather than abstract advice (templates for A/B tests, analytics tracking checks, copy briefs) (so what: reduces iteration time between idea and measurable experiments).
Who It's For and Trade-offs
Great fit if you are a technical marketer, growth engineer, or founder who wants an LLM agent to produce repeatable marketing outputs (landing page copy, SEO audits, A/B test setups, ad creative) and you can provide a concise product-marketing-context. Look elsewhere if you need end-to-end GUI apps, turnkey analytics integrations, or a non-technical, point-and-click marketing suite — the repo supplies agent workflows and templates, not a hosted SaaS product.
Where It Fits
Think of this as a domain-layer for autonomous marketing agents: pair the skills with an agent runtime (Claude Code, an OpenAI Codex–based agent, or an agent framework) and a small product context file to get reliable, repeatable marketing tasks. It complements conversion agencies and learning resources by codifying playbooks rather than replacing human strategy judgment.
