Marketing Copy Engine 1

Marketing Copy Engine

Your landing page converts 0.8%. Your emails get 12% open rates. Your product descriptions are generic. This skill turns any product/topic into high-converting copy using 12 proven frameworks, 8 tones, and brand-voice consistency.

What it does

• 12 copywriting frameworks — AIDA, PAS, BAB, FAB, 4Ps, QUEST, ACCA, PPP, OIC, CCP, SSS, FORREST
• 8 brand tones — professional, casual, witty, urgent, empathetic, authoritative, playful, luxurious
• 50+ templates — landing pages, hero sections, CTAs, emails, ads, product descriptions, social
• Quality gate — scores copy 0-100 (clarity, persuasion, specificity, emotion, CTA)
• A/B test suggestions — generates 3 variants per piece
• Brand voice lock — consistent tone across all output

When to use it

• You need landing page copy and don't know where to start
• Your current copy converts below 2% and you want a refresh
• You need 5 email variants for a launch sequence
• You're launching a product and need all the copy at once
• Your content sounds AI-generated and you want it to sound human
• You need product descriptions for 100 SKUs

Why it's better than ad-hoc prompting

Most "write me copy" prompts give generic output. This skill is different:

• Framework-driven — not "write something good" but AIDA / PAS / BAB
• Quality scored — every output gets 0-100 with specific improvements
• Variant generation — A/B test ready, not just one version
• Brand-locked — consistent voice across all channels
• Channel-aware — landing page ≠ email ≠ ad; different rules

Architecture

┌─────────────────────────────────────────────────────────┐
│               Agent (Claude/Cursor)                     │
│  - Asks for product, audience, channel, goal            │
│  - Selects framework + tone                             │
│  - Generates 3 variants                                 │
│  - Scores with quality gate                             │
└───────────────┬─────────────────────────────────────────┘
                │
                ▼
┌─────────────────────────────────────────────────────────┐
│          skills/marketing-copy-engine/                  │
│  scripts/                                                │
│    ├── generate.py        # Main copy generator         │
│    ├── score.py           # Quality gate (0-100)        │
│    ├── variants.py        # A/B test generator          │
│    ├── brand_voice.py     # Voice consistency checker   │
│    └── ab_calculator.py   # Sample size for A/B test    │
│  references/                                             │
│    ├── 12-frameworks.md   # Detailed with examples      │
│    ├── 8-tones.md         # Tone comparison             │
│    ├── quality-criteria.md                              │
│    └── channel-rules.md   # Landing vs email vs ad      │
│  templates/                                              │
│    ├── landing-page.md                                  │
│    ├── email-sequence.md                                │
│    ├── product-description.md                            │
│    └── social-posts.md                                   │
└─────────────────────────────────────────────────────────┘

Quick start

# 1. Install
pip install textstat language-tool-python

# 2. Generate landing page copy
python scripts/generate.py --framework PAS --tone urgent \
  --product "AI-powered code review tool" --audience "engineering managers" \
  --channel landing --out hero.md

# 3. Generate 3 A/B variants
python scripts/variants.py --framework AIDA --product "..." --audience "..." --n 3

# 4. Score your existing copy
python scripts/score.py landing-page.md

# 5. Check brand voice consistency
python scripts/brand_voice.py *.md --voice professional

The 12 frameworks (cheat sheet)

| Framework | Structure | Best for | |-----------|-----------|----------| | AIDA | Attention → Interest → Desire → Action | Landing pages, ads | | PAS | Problem → Agitate → Solution | Sales pages, emails | | BAB | Before → After → Bridge | Product launches | | FAB | Feature → Advantage → Benefit | Product pages, decks | | 4Ps | Picture → Promise → Prove → Push | Long-form sales | | QUEST | Qualify → Understand → Educate → Stimulate → Transition | B2B, high-ticket | | ACCA | Awareness → Comprehension → Conviction → Action | B2B, SaaS | | PPP | Picture → Promise → Prove | Short copy, ads | | OIC | Outcome → Insight → Change | Thought leadership | | CCP | Comprehend → Communicate → Convert | Direct response | | SSS | Star → Story → Solution | Personal brand | | FORREST | Fully → Oriented → Right → Relevant → Emotional → Story → Truth | Premium products |

The 8 tones (with examples)

| Tone | Example line | |------|--------------| | Professional | "Enterprise-grade observability for modern infrastructure." | | Casual | "We made monitoring not suck." | | Witty | "Your servers are on fire. We brought marshmallows." | | Urgent | "Last 24 hours: your competitors shipped 3 features. You shipped 0." | | Empathetic | "We've all been the on-call getting paged at 3am. We built this so you don't have to." | | Authoritative | "After 10 years building monitoring at Google, here's what I learned." | | Playful | "Lettuce help you ship faster 🥬" | | Luxurious | "Crafted for the discerning few who accept nothing less than excellence." |

The quality gate (0-100 score)

Every output is scored on:

• Clarity (25 pts) — Jargon-free, Flesch reading ease 60+
• Specificity (20 pts) — Numbers, not "many" or "fast"
• Persuasion (20 pts) — Uses framework structure correctly
• Emotion (20 pts) — At least one emotional trigger
• CTA strength (15 pts) — Clear, specific, low-friction

Sample output (PAS, urgent tone)

# Hero Section — Engineering Manager

## Headline
**Your code review queue is 47 PRs deep. Your senior engineers are burning out.**

## Sub-headline
AI-powered code review that catches the bugs your team misses — before they hit production. No more 3am pages. No more "I should have caught this in review."

## CTA Primary
**Start free — no credit card**

## CTA Secondary
See it work on your repo →

## Social proof
"Reduced our PR review time by 60%." — @alice, Staff Eng at Acme

## Pain bullets
- ⏰ 2-week average PR cycle time
- 🐛 40% of bugs caught in production, not review
- 😩 Senior engineer burnout from 6h/day of review
- 💸 Production incidents cost $14k each (your last 3: $42k)

Pricing

Single-purchase, lifetime access. $6.00.

Includes:

• 5 Python scripts (generate, score, variants, brand voice, A/B calc)
• 4 reference docs (frameworks, tones, quality, channels)
• 4 template libraries (landing, email, product, social)
• Quality gate with 5-criteria scoring
• A/B test sample size calculator
• Future updates for the same major version

Example usage

"Write a landing page for our new AI code review tool. Target is engineering managers at Series B startups. Tone: witty but professional."

The skill will:

1. Select framework (PAS for landing page)
2. Select tone (blend of witty + professional)
3. Generate 3 A/B variants
4. Score each variant 0-100
5. Pick the winner + explain why
6. Output ready-to-publish markdown

Compatibility

Works with any agent that supports the SKILL.md standard and can execute Python: Claude Code, OpenClaw, Codex CLI, Cursor, Gemini CLI, Cline, Windsurf, Aider. Tested on Linux, macOS, Windows.

Tags

marketing, copywriting, content, conversion, landing-page, email, ads, ecommerce