12 commands. 11 workflows. 32 bundled marketing skills. Drop a full marketing department into any Claude Code project — in one command.

npx ai-cmo-agent

What This Is

Most AI marketing tools generate generic content. This agent does something different.

It researches real leads, finds real gaps, writes real emails, and builds real campaigns. Drop it into your Claude Code environment and it works like a CMO who never sleeps — one that knows your product, your ICP, and your voice.

Built for founders, solo operators, and small teams who need serious marketing output without hiring a $150k/year marketing department.

Commands

Core — Outreach & Lead Gen

| Command | What It Does | |---|---| | /cmo | Pipeline overview — active leads, outreach status, next actions | | /cmo-research [company] | Deep research on any lead — digital footprint, AI visibility gaps, fit score, personalization hook | | /cmo-email [lead] | Personalized cold email + 3 follow-ups (Day 3, Day 6, Day 9), Instantly AI-ready | | /cmo-triage | Score and prioritize a lead list by fit, response likelihood, and outreach readiness |

Extended — Full-Spectrum Marketing

| Command | What It Does | |---|---| | /cmo-ads | Ad copy at scale — Google, Meta, LinkedIn, TikTok, X. Platform-compliant, CSV-ready | | /cmo-campaign | Full paid campaign structure — audiences, bidding, retargeting, reporting | | /cmo-social | Content pillars, platform-native posts, repurposing workflows, weekly calendar | | /cmo-cro | 7-dimension conversion audit for any marketing page — quick wins + A/B hypotheses | | /cmo-launch | Product launch strategy — ORB framework, 5-phase approach | | /cmo-content | Content strategy — topic pillars, cluster map, prioritized topic table | | /cmo-pricing | Pricing strategy — value metrics, tier architecture, willingness-to-pay research | | /cmo-test | A/B test design — hypothesis formation, sample size, variant specs |

Installation

npx ai-cmo-agent

That's it. Here's what lands in your project:

your-project/
├── .agents/
│   ├── ai-cmo/
│   │   ├── IDENTITY.md                   ← Customize your agent's voice & rules
│   │   ├── product-marketing-context.md  ← Fill this out first (2 min)
│   │   ├── workflows/                    ← 11 workflow files
│   │   └── templates/                    ← Output templates
│   └── marketingskills/                  ← 32 marketing skills (bundled)
└── .claude/
    └── commands/                         ← 12 slash command files

Setup in 2 minutes

1. Run the installer → npx ai-cmo-agent
2. Fill out your product context → Edit .agents/ai-cmo/product-marketing-context.md
   • Your product name and what it does
   • Who your ICP is
   • Your pricing
   • Competitors
   • Brand voice
3. Open Claude Code → Type /cmo to see all commands

Example

/cmo-research Acme Digital Agency

## Lead: Acme Digital Agency

Type:               Full-service digital marketing agency
Niche:              Restaurants, hospitality, local retail
Digital footprint:  Active LinkedIn (2x/week), strong local SEO presence
Personalization:    Posted last week about struggling to prove ROI to clients
AI gap:             Zero of their restaurant clients appear in ChatGPT/Perplexity
                    for "best Italian restaurant in Dallas" — competitors do
Fit score:          HIGH

────────────────────────────────────────

Subject: Acme — your restaurant clients aren't showing up in AI search

Hi [First Name],

Ran a quick check — your clients are missing from ChatGPT and Perplexity
for queries like "best Italian restaurant in Dallas." Their competitors aren't.

That's the new visibility problem nobody's talking about yet. I built a tool
that finds these gaps and generates the fixes automatically.

Worth 15 minutes to see the data?

[Name]

────────────────────────────────────────

Follow-up Day 3:  [New angle — competitor comparison]
Follow-up Day 6:  [New angle — industry stat]
Follow-up Day 9:  [Final — direct ask]

How It Works

The agent reads your product-marketing-context.md before every operation. Every output — emails, ad copy, content plans — is shaped by your specific product, ICP, and voice. Not generic. Not templated.

The 32 bundled marketingskills power the extended capabilities:

• Cold email frameworks — 5 proven structures, not just one template
• Copy editing — 7-sweep quality pass before any email goes out
• RevOps scoring — point-based lead triage, not gut feel
• Ad creative — angle-organized, platform-spec compliant
• A/B test design — statistically valid, not guesswork

What It Is NOT

This agent stays in its lane.

• ✗ Not a generic content generator
• ✗ Not an ops or finance tool
• ✗ Not an engineering agent
• ✗ Not a replacement for real strategic thinking on complex brand decisions
• ✗ Does not generate actual images (coming in v2.0 — see Roadmap)

It's a marketing execution machine. Give it a target, it goes.

Roadmap

• [x] Lead research + fit scoring
• [x] Cold email personalization + follow-up sequences
• [x] Pipeline triage with RevOps scoring
• [x] Ad creative generation (copy + angles)
• [x] Paid campaign structure
• [x] Social content strategy
• [x] CRO audits
• [x] Launch strategy
• [x] Content strategy
• [x] Pricing strategy
• [ ] Image generation — actual ad creatives and social graphics via image gen API
• [ ] CRM sync — push output directly to HubSpot / Notion / Airtable
• [ ] Analytics integration — pull campaign performance back in for iteration

Built On

| | | |---|---| | Claude Code by Anthropic | The agent runtime | | marketingskills by Corey Haines | 32 battle-tested marketing skill files |

Contributing

PRs welcome. If you build a new workflow:

• Add it to the commands table above
• Document the skill it references
• Keep it in the CMO lane — marketing execution only

MIT — free forever.

npm · GitHub · marketingskills