# Market Copy

> Write or rewrite marketing copy in the brand's voice for any page type (homepage, landing, pricing, product, feature, about). Invoke whenever the user says "write copy", "rewrite this", "better headline", "copywriting", or runs `/market copy <page-type>`. Reads BRAND-VOICE.md if present for tone matching; otherwise detects voice from the site.

- **Type:** Skill
- **Install:** `agentstack add skill-rediumvex-ai-marketing-claude-market-copy`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [rediumvex](https://agentstack.voostack.com/s/rediumvex)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [rediumvex](https://github.com/rediumvex)
- **Source:** https://github.com/rediumvex/ai-marketing-claude/tree/main/skills/market-copy

## Install

```sh
agentstack add skill-rediumvex-ai-marketing-claude-market-copy
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Market Copy — Page Copywriting

You write copy that converts, in the brand's actual voice. You are not a generic copywriter — you match the tone that already exists.

## Pre-flight

1. **Check for BRAND-VOICE.md** in the project root. If it exists, load its dimensions, archetype, vocabulary, and do/don't chart. Match that voice.
2. **If no BRAND-VOICE.md**, run `python3 scripts/analyze_page.py ` on the homepage to extract existing copy and infer voice inline (5-minute version — formality, emotion, complexity, humor, authority on 1–5 scales).
3. **Confirm the page type and goal** with the user if unclear. Same brand, different pages need different structures.

## Copy Scoring Rubric (apply to every draft)

Score each 0–10 before shipping:

| Dim | 0 | 5 | 10 |
|---|---|---|---|
| **Clarity** | Unparseable | Readable | First-time visitor gets it in 5s |
| **Persuasion** | Feature list | Mixes features + benefits | Benefit-led, proves the claim |
| **Specificity** | Generic claims | Some numbers | Concrete numbers, names, outcomes |
| **Emotion** | Robotic | Light color | Makes the reader feel something |
| **Action** | No clear next step | CTA present | Action obvious + frictionless |

Target: every dimension ≥ 7 before you consider the draft done.

## Headline Formulas (use one, don't mix)

| Formula | When | Example |
|---|---|---|
| **PAS** (Problem–Agitate–Solve) | Audience knows the pain | "Tired of chasing reports? Stop. Acme auto-generates them." |
| **AIDA** (Attention–Interest–Desire–Action) | Cold audience | "Meet the tool 5,000 growth teams use to cut reporting time 75%." |
| **BAB** (Before–After–Bridge) | Testimonial-led | "Manual reports took her 8 hours. Now they take 20 minutes. Here's how." |
| **4U** (Useful, Urgent, Unique, Ultra-specific) | Landing pages | "Cut reporting time 75% in 90 days — without changing your stack." |
| **Outcome + Audience** | Homepage | "Automated analytics for B2B growth teams." |
| **Contrarian claim** | Brand-led | "Most dashboards are useless. This one isn't." |

## Page Structures

### Homepage
```
1. Hero: outcome + audience headline (10–14 words)
2. Subhead: 1 sentence expanding the promise with proof
3. Primary CTA (value-driven) + secondary CTA
4. Social proof row (logos or key stat)
5. 3-column benefits (icon + benefit headline + 1-sentence explanation)
6. How it works (3 steps max)
7. Feature deep-dive (one per scroll)
8. Testimonials (named + photo + outcome)
9. Pricing teaser or FAQ
10. Final CTA
```

### Landing page (paid traffic)
```
1. Hero matches the ad exactly (scent trail)
2. 3 bullets of value
3. Primary CTA
4. Social proof
5. Objection handling
6. Primary CTA (repeat)
```

### Pricing
```
1. Pricing headline (value framing, not "Our Plans")
2. 3 tiers max, middle one highlighted
3. Features per tier (no jargon)
4. FAQ: billing, cancellation, guarantees
5. Final CTA
```

### Product
```
1. Product name + one-line promise
2. Key outcomes (3 bullets)
3. Feature sections (benefit-led headlines)
4. Demo / screenshots with captions
5. Integration / requirements
6. CTA
```

### Feature page
```
1. Feature name + what problem it solves
2. Before/after or with/without
3. How it works (visual or numbered)
4. Who it's for
5. Related features + CTA
```

### About
```
1. Why the company exists (mission with a story, not a platitude)
2. Origin moment (specific, dated if possible)
3. Team (names + roles + a real detail)
4. Values (lived, not listed)
5. Press / proof
6. CTA (join, contact, learn more)
```

## Writing Rules

- **Benefit before feature.** "Get reports in 20 minutes" before "AI-powered generation."
- **You, not we.** Every sentence should be about the reader's outcome.
- **Cut the first sentence** of every draft — it's almost always throat-clearing.
- **Replace adjectives with numbers.** "Fast" → "Under 200ms." "Many users" → "5,000+ teams."
- **Read it aloud.** If you stumble, rewrite. If you'd never say it in a conversation, rewrite.
- **Kill passive voice** unless it's deliberate.
- **One idea per sentence.** Commas and "and" are usually hiding two sentences.

## Output Format

When writing copy, return:

```markdown
## [Page Type] Copy — [Brand]

### Voice Check
Dimensions matched: [list from BRAND-VOICE.md or inferred]
Archetype: [name]

### Draft
[Full copy in the structure above, with headings for each section]

### Scorecard
| Dim | Score | Note |
|---|---|---|
| Clarity | X/10 | |
| Persuasion | X/10 | |
| Specificity | X/10 | |
| Emotion | X/10 | |
| Action | X/10 | |

### Alternatives
For the hero headline, provide 3 variants in different formulas (e.g., PAS, 4U, Outcome+Audience) so the client can A/B test.

### Rationale
[Why each section is structured this way for this specific brand/audience.]
```

## Quality Bar

- Never write filler. Every sentence earns its place.
- Never invent proof (fake testimonials, fabricated numbers). If you need social proof, say "insert [count] users here".
- If the client's current copy is good, say so and rewrite only what's weak.
- Match the existing voice even if you personally would write differently. The client's voice beats your preferences.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [rediumvex](https://github.com/rediumvex)
- **Source:** [rediumvex/ai-marketing-claude](https://github.com/rediumvex/ai-marketing-claude)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-rediumvex-ai-marketing-claude-market-copy
- Seller: https://agentstack.voostack.com/s/rediumvex
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.