agentsclimarketplace

Direct response copy

Skill sanky369/vibe-building-skills/skills/marketing/direct-response-copy

Write copy engineered to drive one specific action — landing pages, sales pages, opt-in pages, ads, promotional emails — using an eight-part direct-response architecture (hook, problem, mechanism, solution, proof, objections, CTA, urgency). Use when the user says 'write my landing page / sales page / ad', 'this page isn't converting', 'rewrite my homepage copy', 'write copy for my lead magnet page', or any request for persuasive copy with a measurable action at the end. Produces the full copy draft, section-labeled against the architecture, plus hook alternatives and a conversion checklist.From its SKILL.md

Install
npx -y skills add sanky369/vibe-building-skills --skill direct-response-copy

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

SKILL.md

7.9 KB, ~1.7k tokens by cl100k_base, as published. Nobody here has run it

Direct Response Copy

Write copy whose only job is to cause one specific action — click, signup, purchase, reply. Governing principle: one page, one reader, one action. Every section either moves the reader toward that action or gets cut; persuasion comes from a named mechanism and real proof, never from volume or hype.

When to use / when not to

  • Use for any asset with a conversion goal: landing/sales/opt-in pages, ads, promotional emails, upgrade prompts.
  • Requires a positioning claim and a voice. If the positioning is missing or mushy, run skills/marketing/positioning-angles first — copy cannot rescue an undifferentiated claim. If a voice guide exists from skills/marketing/brand-voice, write within it; if not, sketch voice from intake and proceed.
  • Multi-email automated flows → skills/marketing/email-sequences (it reuses this architecture per email). Content meant to rank → skills/marketing/seo-content. Recurring relationship email → skills/marketing/newsletter.

Intake

One batch; pull answers from existing artifacts (positioning statement, voice guide, lead magnet spec) before asking:

  1. What asset, and what is the ONE action a successful reader takes?
  2. What's the offer, and what does it cost the reader (money, time, email address)?
  3. Who is the reader, and what's the #1 reason they'd hesitate? (If unknown, name the likely objection yourself and flag it.)
  4. What proof exists — testimonials, case studies, numbers, credentials, recognizable customers?
  5. Any real deadline, capacity limit, or price change?

Don't stall: with an offer and audience in hand, draft with stated assumptions rather than blocking on Q3–Q5 — but never fill Q4/Q5 gaps with invented material.

Workflow — build the eight components in order

1. Hook

Draft 3–5 hooks using different formulas, present them with one-line rationales, and have the user pick (default to your top pick if they defer):

FormulaTemplate
Curiosity"The one thing [person] gets wrong about [topic]"
Specificity"How I [specific result] in [specific timeframe]"
Benefit"[Specific benefit] without [specific drawback]"
Question"Are you [specific situation]?"
Bold statement"[Controversial claim about the industry]"

Decision rules: cold traffic → curiosity or bold; warm traffic that knows the problem → benefit or specificity; jaded market → specificity with proof baked in. A specificity hook may only use results that actually happened.

2. Problem validation

Show the reader you understand their situation before selling anything. Formula: "You're [specific situation], and it's frustrating because [specific consequence]." Rule: use the reader's own words (support tickets, reviews, sales calls) when available — felt understanding is what keeps them reading.

3. Mechanism

Explain how the result happens, using the named mechanism from the positioning statement. Formula: "Most people try [common approach]. But here's what works: [named mechanism]." Rule: the mechanism must be named ("The Micro-Launch Method"), never "our approach". No mechanism in the positioning → stop and get one from skills/marketing/positioning-angles.

4. Solution presentation

Present the offer as the mechanism made available — inevitable, not salesy. Formula: "That's why we created [offer]. It [specific benefit] by [mechanism]."

5. Social proof

Deploy the strongest proof from intake Q4: testimonials → results-bearing case studies → customer counts → awards → metrics. Hard rule: never fabricate testimonials, customer counts, or statistics. If proof is thin, substitute mechanism logic + a risk reversal (guarantee, free tier, no-card trial) and tell the user proof collection is their highest-leverage next task.

6. Objection handling

Take the #1 objection (intake Q3) plus the standard four — no time, too expensive, tried it before, won't work for me — and answer the 2–3 that fit this reader. Formula: "You might be thinking [objection]. Here's the truth: [answer]." Rule: answer honestly or reframe; never deny a true cost ("it does take 30 minutes a day — here's why that's the point").

7. Call-to-action

One action, stated identically everywhere it appears (long pages: after solution, after proof, at the end). Formulas: simple ("Start your free trial"), benefit-focused ("Get the checklist now"), curiosity ("See how it works"), low-friction ("Reply and I'll send it over"). Decision rules: high-commitment offers → low-friction or curiosity CTA; free offers → benefit-focused. Never two competing actions on one page.

8. Urgency

Use only urgency that is true: a real deadline, real capacity limit, real price change, or the honest cost of waiting ("every week without X costs you Y"). Hard rule: no fake countdowns, fake scarcity, or evergreen 'closing soon'. If nothing real exists, end on the cost of inaction and skip the section.

Then: voice pass and pressure test

Rewrite the draft through the voice guide (or stated voice assumptions), then run the persuasion checklist: specific beats vague ("increase revenue 23%" > "increase revenue" — only with a real number); proof quantified where honest; reciprocity — the page gives value before asking; authority shown, not claimed; consistency — the argument aligns with what the reader already believes about their problem.

Required output format

# [Asset type] — [Offer]
**Goal action:** [the one action] · **Reader:** [who] · **Voice:** [guide reference or stated assumption]

---
[HOOK]
...
[PROBLEM]
...
[MECHANISM]
...
[SOLUTION]
...
[PROOF]
...
[OBJECTIONS]
...
[CTA]
...
[URGENCY]
...
---

## Hook alternates
2. "..." — [formula, when to prefer it]
3. "..." — ...

## Gaps flagged
- [missing proof, unverified claims, urgency omitted because nothing real exists]

## Conversion checklist
[the quality bar below, each item marked PASS/FAIL for this draft]

Deliver the copy clean (section labels are for the user's review; note they're removable). For ads or short emails, compress: hook, problem/mechanism fused, CTA — same rules, fewer sections.

Quality bar (check before delivering)

  • One CTA action; every CTA instance asks for exactly the same thing.
  • Hook works out of context (would stop a scroll with zero surrounding info).
  • Mechanism is named; the word "approach" or "process" never stands in for it.
  • Every number, testimonial, and customer count came from the user — zero invented proof.
  • Urgency is real or absent — never manufactured.
  • Each objection answered, not dodged; true costs acknowledged.
  • Voice matches the voice guide (or stated assumption); no generic "marketing voice" drift.
  • Nothing on the page could be pasted onto a competitor's site unchanged.

Integration

  • Upstream: skills/marketing/positioning-angles supplies the claim and named mechanism (the page's spine); skills/marketing/brand-voice supplies the voice guide; skills/marketing/lead-magnet supplies the offer spec when the page sells a magnet.
  • Downstream: skills/marketing/email-sequences reuses the page's hook/mechanism/objection answers across its emails; skills/marketing/content-atomizer can atomize a strong sales page into ad variants and social posts. What crosses the boundary is the finished copy plus the flagged-gaps list.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.