Hyper

RAG & Memory 💻 Python ⚖️ MIT 🟢 Actively maintained
83 stars

Marketing Skills for AI Agents is a collection of markdown skill files that teach AI agents how to execute marketing tasks end-to-end, such as paid ads, social media, SEO, and competitor research. It runs on the Hyper MCP, which provides 100+ integrations and built-in tools, and is designed for developers using AI agents in coding environments like Claude Code, Cursor, and others.

Hyper MCP: 200+ Marketing Tools and Integrations for Every Agent

🎬 Hyper MCP: 200+ Marketing Tools and Integrations for Every Agent · Hyper AI

✨ Key features

  • 24 skills covering ads, social, SEO, research, creative, email, analytics
  • Runs on Hyper MCP with 100+ direct integrations
  • Built-in tools for scraping, image/video generation, and browser automation
  • Guardrails and permissions for action-by-action approval
  • Auto-detects agent and installs skills in the right directory
  • Skills compose for multi-step marketing campaigns

🎯 Use cases

  • Plan and launch Google Ads, Meta, TikTok, and other ad campaigns
  • Research competitors and customer sentiment from social and review sites
  • Generate ad creatives and video content in brand style
  • Build email lifecycle programs and cold outreach campaigns
  • Pull analytics insights from GA4, Search Console, and BigQuery

📦 Installation

🧰 Requirements: Requires Node.js and an agent that supports the Agent Skills spec; Hyper MCP account and API keys for integrations as needed.

npx skills add hyperfx-ai/marketing-skills

Auto-detects your agent and drops the skills in the right directory. Then ask your agent something a skill knows about:

"Plan a Google Ads search campaign for example.com."

install one
npx skills add hyperfx-ai/marketing-skills --skill google-ads
install all 24
npx skills add hyperfx-ai/marketing-skills --all

Other install options:

install globally (available across every project)
npx skills add hyperfx-ai/marketing-skills -s seo-research -g
Claude Code plugin marketplace
/plugin marketplace add hyperfx-ai/marketing-skills
/plugin install hyper-marketing@hyperfx-marketing-skills
manual copy
cp -r skills/google-ads ~/.claude/skills/         # Claude Code
cp -r skills/google-ads ~/.agents/skills/         # Cursor / Codex / generic
git submodule
git submodule add https://github.com/hyperfx-ai/marketing-skills.git .agents/marketing-skills

🚀 Usage

Skill discovery is automatic. Invoke explicitly with /google-ads, /cold-email-outreach, /seo-research, etc. Example:

"Pull all the active Facebook ads for 'meal kit delivery' into a database."meta-ads-library

"Run a 50-prospect cold-email campaign to Heads of Growth at US Series A SaaS companies."cold-email-outreach

"Pull last month's GA4 traffic by channel and flag any conversion event that dropped >20%."analytics-insights

"Build a weekly competitor digest for [comp-a, comp-b, comp-c] — site changes, social, search rank, mentions."competitor-intel

⚠️ Good to know

Some skills are in beta (e.g., hyper-cli) and the roadmap lists upcoming skills not yet available.

❓ FAQ

What is a skill?

A skill is a markdown file (SKILL.md) with workflows, decision rules, and example tool calls for your agent to run specific marketing functions end-to-end. The tools come from the Hyper MCP.

How do I install a specific skill?

Use npx skills add hyperfx-ai/marketing-skills --skill <skill-name> to install one skill, or --all to install all 24.

What integrations are available?

The Hyper MCP provides 100+ direct integrations including Google Ads, Meta, TikTok, LinkedIn, Amazon, Pinterest, Gmail, Klaviyo, GA4, BigQuery, HubSpot, Shopify, and more.

Can skills be used together?

Yes, skills can compose because they all run against the same MCP. For example, a campaign can use meta-ads, google-ads, tiktok-ads, and ad-creative-generation together.

📊 Repository

Stars★ 83
Forks🍴 14
Open issues🐛 3
Last commit🕒 Aug 25, 2026
Created📅 Apr 2026
Language💻 Python
License⚖️ MIT

🤖 Overview, features, install steps and FAQ were generated from the project's README on Sep 4, 2026. Always check the original source before running commands.