copy-chief
Skill2 starsChanged 3 months agoMITView on GitHub
Copywriting for sales pages - headlines, product copy, CTAs, microcopy, objection handling. Use for any customer-facing words.
In this repository18 other files it gives its agents.
- art-director-reviewSkill
- b2b-landingSkill
- brand-voiceSkill
- checkout-trustSkill
- composition-gridsSkill
- demo-assetsSkill
- design-tokensSkill
- experimentation-measurementSkill
- layout-compositionSkill
- marketing-strategySkill
- motion-performanceSkill
- navigation-searchSkill
- pixel-qaSkill
- post-purchase-lifecycleSkill
- product-page-croSkill
- promotions-eventsSkill
- typography-craftSkill
- visual-identitySkill
The file
We split this skill into the rules it gives an agent
11 rules. About 0 are common, 10 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈10
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 37
[Mobile headline check]: Test hero copy at a 375px viewport: headline, subhead, and primary CTA must all render above the fold with zero …
- Distinctive: line 41
[FAB ladder]: Write every feature block as Feature (factual spec) → Advantage (the comparative edge it creates) → Benefit (the reader's …
- Distinctive: line 48
[Benefit-forward verb]: Prefer verbs that frame what the user receives over verbs that frame what they must do: "Get" beat "Order" by …
- Distinctive: line 57
[FAQ = objection tool]: Every FAQ entry must map to one specific unresolved objection from Rule 15.
- Distinctive: line 61
[Above-fold proof]: Place at least one proof element (logo bar, single named testimonial, or aggregate rating) above the fold.
- Distinctive: line 62
[Attribution required]: Use named, attributed testimonials — real name + title/company, photo where possible — never anonymous quotes.
- Distinctive: line 63
[Specific numbers]: Use precise, non-rounded numbers in proof copy ("11,247 businesses," not "thousands of customers") — specificity reads …
- Distinctive: line 66
[Live-proof authenticity]: If using live-activity notifications ("Sarah in Austin just signed up"), show only real, recent events tied to …
- Very specific: line 70
[Charm pricing default]: Use charm pricing (.99/.97 endings) as the default for price-sensitive, transactional/low-consideration offers.
- Distinctive: line 74
[No price-gating self-serve]: Never hide price behind "Contact us" for a self-serve or low-consideration offer — it reopens the …
- Distinctive: line 81
[Reading level]: Write body copy and microcopy at a 5th-7th grade reading level.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Copy Chief — Sales Copywriting Rules
When to use
- Writing or reviewing a hero headline, subheadline, or CTA button copy for a sales/landing page.
- Writing product page copy: feature blocks, benefit statements, pricing tables.
- Writing or auditing objection-handling copy: FAQ sections, guarantee/risk-reversal copy, comparison content.
- Writing or reviewing testimonials, review displays, ratings, and other social-proof copy.
- Writing microcopy: form labels, helper text, error messages, empty states, confirmation copy.
- Any pass where the deliverable is customer-facing words meant to move a visitor toward a lead, signup, or sale.
Voice interface (contract with brand-voice)
The brand-voice skill defines HOW the brand sounds and hands this skill a voice card as a
fixed input. The precedence contract, declared identically in both skills: on the exact line of a
primary CTA, a form label, a displayed price, or an error message, THIS skill's clarity rules win,
full stop — no voice guardrail overrides them there. Everywhere else (headlines, body copy, empty
states, hero taglines) brand-voice governs register and style, and headlines must still clear this
skill's one-line clarity test. When drafting copy for a brand that has a voice card, load
brand-voice alongside this skill; when there is no voice card yet, flag that gap instead of
inventing a voice ad hoc. — brand-voice (Precedence section)
Rules
Headlines & awareness stage
- [Awareness match] Identify the primary traffic's awareness stage before writing the headline. Unaware/Problem-Aware → lead with the problem, no product name in line 1 ("Still tracking invoices in a spreadsheet?"). Solution-Aware → name the solution category + your edge. Product-Aware/Most-Aware → product name + differentiator + offer, all in line 1. [FRAMEWORK] — Schwartz
- [One-line test] Every headline must pass: "if a visitor reads only this line, do they know exactly what's sold and to whom?" Reject any headline that could paste onto a competitor's page unchanged ("Improve your workflow" fails; "The invoicing tool freelancers use to get paid in 3 days, not 30" passes). [FRAMEWORK] — Shapiro
- [Headline is load-bearing] Assume 80-90% of visitors read only the headline + CTA and never reach body copy (direction corroborated by 2 independent sources; exact % varies 80-90%). Write the headline to carry the complete value proposition, not curiosity bait. [TREND-2026] — Copyblogger/Neil Patel, SEO Sherpa
- [Subhead discipline] Cap the subheadline at 1-2 sentences with exactly two jobs: state what the product is, and give one reason the headline's claim is credible. If it runs to a paragraph, cut it. [FRAMEWORK] — Shapiro
- [Claim specificity] Replace superlatives ("best," "revolutionary," "seamless") with a checkable number, named comparison, or named mechanism. Test: could a competitor paste this claim onto their own page unchanged? If yes, rewrite it. [FRAMEWORK, moderate confidence] — marketingexamples.com
- [Mobile headline check] Test hero copy at a 375px viewport: headline, subhead, and primary CTA must all render above the fold with zero scrolling (~83% of landing-page traffic is mobile). [TREND-2026] — SEO Sherpa
FAB / BAB frameworks
- [FAB ladder] Write every feature block as Feature (factual spec) → Advantage (the comparative edge it creates) → Benefit (the reader's emotional/practical payoff). End on the Benefit line; never stop at Feature or Advantage. [FRAMEWORK] — Copy Brothers
- [BAB structure] For transformation-style offers, structure the opening as Before (1 sentence, specific current pain) → After (1 sentence, achieved state) → Bridge (product positioned as the mechanism connecting them). [FRAMEWORK] — Copy Brothers
- [Feature block cap] Cap the primary page at 3-6 feature blocks. Each block = one benefit-stated header (not a feature label) + one paragraph carrying feature → benefit → objection handled + one real product image or GIF (no stock art). [FRAMEWORK] — Shapiro
CTA verb + deliverable rules
- [Verb + deliverable] Lead CTA copy with a specific action verb plus a stated deliverable ("Get my template," "Start my free trial") rather than a generic verb ("Submit," "Learn More," "Click Here"). One cited PPC test added this kind of relevance for a 68% lift. [TREND-2026] — Unbounce
- [Benefit-forward verb] Prefer verbs that frame what the user receives over verbs that frame what they must do: "Get" beat "Order" by 38.26% in one B2B test. [TREND-2026] — Unbounce
- [Pronoun: A/B, don't assume] Test first-person ("Start my trial") against second-person ("Start your trial") phrasing per audience rather than shipping a fixed rule — the widely cited "90% lift" figure is an unverified outlier, not an expected result. Micro-copy tests of this class are only valid above
experimentation-measurement's traffic floor (~1,000 sessions / 30 conversions per variant per month); below it, ship the second-person default and move on. [TREND-2026, low confidence] — experimentation-measurement R1-2 - [One offer, one CTA] Ship exactly one primary offer and one primary CTA per page. Multi-offer/multi-link pages measurably underperform single-offer pages (magnitude is source-dependent; direction is reliable). [TREND-2026] — SEO Sherpa
- [CTA touch target] Size every CTA button to a 44×44 CSS px floor (Apple HIG; Material 48dp) with ≥8px clearance from any adjacent tappable element. WCAG 2.2 AA's legal minimum is 24×24 CSS px — that's the compliance floor, not the target; ship 44. [LAW] — WCAG 2.2 SC 2.5.8, Apple HIG, Material Design
Objection sequence
- [Objection order] Map and answer objections in this cognitive order before drafting page copy: (1) legitimacy — "is this real?", (2) relevance — "does this work for my situation?", (3) mechanism — "how does it work?", (4) results — "are the numbers real?", (5) peer validation — "what do others say?", (6) pricing/terms transparency, (7) risk — "what if I get burned?". A page answering only 2-3 of these converts only the warmest visitors. [FRAMEWORK] — InstantPress
- [Objection placement] Place legitimacy + relevance signals (logos, named testimonials, "built for [persona]" language) within the first screen. Place mechanism/results detail mid-page. Place risk-reversal copy (guarantee, refund terms, trial terms) directly beside or below the final CTA — never buried only in an FAQ. [FRAMEWORK] — InstantPress, Shapiro
- [FAQ = objection tool] Every FAQ entry must map to one specific unresolved objection from Rule 15. Delete generic filler entries ("What is your company?") that don't resolve a numbered objection. [FRAMEWORK] — InstantPress
Social proof formats
- [Above-fold proof] Place at least one proof element (logo bar, single named testimonial, or aggregate rating) above the fold; place deeper proof (case studies, longer testimonials) mid-page. Never stack two redundant proof types back-to-back in the same scroll position. [FRAMEWORK] — Shapiro, WiserNotify
- [Attribution required] Use named, attributed testimonials — real name + title/company, photo where possible — never anonymous quotes. One cited test: adding 3 attributed testimonials lifted purchases 34%. [TREND-2026] — WiserNotify
- [Specific numbers] Use precise, non-rounded numbers in proof copy ("11,247 businesses," not "thousands of customers") — specificity reads as more credible. [TREND-2026, directional] — WiserNotify
- [Reviews are load-bearing] Display third-party or on-site reviews wherever the platform allows it; two independent sources report a large conversion lift from visible reviews (roughly 270% for everyday goods, up to 380% for higher-consideration items). [TREND-2026] — Spiegel Research Center (via WiserNotify), Crazy Egg
- [Review metadata] Surface review count and average rating near the price/CTA cluster; label verified purchases where the platform supports it. [TREND-2026] — Crazy Egg
- [Live-proof authenticity] If using live-activity notifications ("Sarah in Austin just signed up"), show only real, recent events tied to an actual timestamped log. Fabricated or looped activity feeds are a legitimacy risk (Rule 15, objection #1), not a proof asset. [TREND-2026] — WiserNotify
Pricing psychology (with caveats)
- [Charm pricing default] Use charm pricing (.99/.97 endings) as the default for price-sensitive, transactional/low-consideration offers; cited lift is roughly 24-35% vs. round pricing. [TREND-2026] — MIT/U Chicago 2003, Cornell (via SHNO)
- [Premium override] For premium/luxury positioning, use round numbers instead of charm pricing — .99 endings signal value-hunting and can undercut premium positioning. [TREND-2026] — SHNO
- [Anchoring] When showing tiers, always place a visible reference point next to the target tier: a higher list price with strikethrough, or a higher-priced tier alongside it. Cited anchoring effect on perceived value is roughly +32%. LEGAL FLOOR: any strikethrough/was price must clear
promotions-eventsR24-R30 (FTC 16 CFR 233.1, Cal. B&P 17501's 90-day rule, EU Omnibus 30-day lookback — $50M/$7M settlement precedent behind fictitious reference prices). A reference price you never actually charged is not an anchor, it's a lawsuit. [TREND-2026] — SHNO; promotions-events R24-R30 - [Decoy tier, replication caveat] A decoy tier must be genuinely inferior, priced at or near the target tier's price. Ariely's classic test: a 68%/32% cheap-vs-bundle split with no decoy flipped to 84% choosing the bundle once a decoy was added — but the effect is reported to shrink with realistic (non-lab) stimuli. Validate with your own A/B test before relying on the lab magnitude — subject to
experimentation-measurement's traffic floor and sample-size procedure (its R1-5); below the floor, don't ship a decoy on faith in the lab number. [TREND-2026, replication caveat] — GrowMeOrganic, SHNO; experimentation-measurement R1-5 - [No price-gating self-serve] Never hide price behind "Contact us" for a self-serve or low-consideration offer — it reopens the pricing-transparency objection (Rule 15, #6) instead of resolving it. Reserve "Contact us" pricing for genuinely custom/enterprise deals. [FRAMEWORK]
Microcopy & error messages
- [Validation timing] Validate fields inline on blur (when the user leaves the field) — never on first keystroke, never submit-only. Once an error shows, clear it live as the user corrects it. [FRAMEWORK] — Baymard usability testing
- [Required-field labeling] Mark required fields explicitly; don't rely on marking only optional ones. Leaving required fields unmarked produced a 32% skip rate in testing. [FRAMEWORK] — Baymard
- [Error copy content] Error copy must state what's wrong and how to fix it in plain words, and must never cause layout reflow that hides other fields or the CTA. [FRAMEWORK] — Baymard, UX consensus
- [Reading level] Write body copy and microcopy at a 5th-7th grade reading level; cited benchmark shows 11.1% conversion at that level vs. 5.3% for college-level copy on a comparable offer. Run copy through a readability checker before shipping. [TREND-2026] — SEO Sherpa
- [Field minimization] Cut form fields to the minimum required for the immediate transaction — dropping from 11 fields to 4 is cited at a 120% conversion lift. Every extra field is a labor tax on the copy around it. [TREND-2026] — SEO Sherpa
Checklist
- Headline matches the traffic's awareness stage (Rule 1) and passes the one-line test (Rule 2).
- Headline alone carries the full value proposition — not dependent on body copy being read (Rule 3).
- Subheadline is 1-2 sentences, states the product + why the headline is credible (Rule 4).
- Every superlative claim replaced with a number, named comparison, or mechanism (Rule 5).
- Hero (headline + subhead + primary CTA) verified visible with zero scroll at 375px (Rule 6).
- Every feature block follows Feature → Advantage → Benefit and ends on Benefit (Rule 7).
- Transformation offers use Before → After → Bridge in the opening (Rule 8).
- Page has 3-6 feature blocks, each with header + paragraph + real image/GIF (Rule 9).
- Every CTA uses verb + deliverable, not a generic verb (Rule 10).
- CTA verbs are benefit-forward ("Get") rather than cost-forward ("Order") by default (Rule 11).
- Pronoun choice (first/second person) is flagged for A/B test, not assumed (Rule 12).
- Page has exactly one primary offer and one primary CTA (Rule 13).
- CTA buttons measure ≥44×44 CSS px with ≥8px clearance (Rule 14).
- All 7 objections (legitimacy → risk) are addressed somewhere on the page (Rule 15).
- Legitimacy/relevance signals are in the first screen; risk-reversal sits beside the final CTA (Rule 16).
- Every FAQ entry maps to a named objection; filler Qs removed (Rule 17).
- At least one proof element is above the fold; proof types aren't stacked redundantly (Rule 18).
- Testimonials carry real name + title/company (+ photo), never anonymous (Rule 19).
- Proof numbers are precise/non-rounded, not vague ("thousands") (Rule 20).
- Reviews are displayed on-page wherever the platform allows it (Rule 21).
- Review count/rating sit near price/CTA; verified-purchase labels used if available (Rule 22).
- Any live-activity notification is tied to a real, timestamped event (Rule 23).
- Pricing model (charm vs. round) matches the offer's positioning (Rules 24-25).
- Tiered pricing shows a visible anchor next to the target tier (Rule 26).
- Any decoy tier is genuinely inferior and priced near the target tier, flagged for A/B validation (Rule 27).
- Price is visible for self-serve offers — no "Contact us" gate on low-consideration products (Rule 28).
- Form fields validate on blur with live error-clearing, not on keystroke or submit-only (Rule 29).
- Required fields are explicitly marked (Rule 30).
- Error copy states the problem + fix and never reflows the layout (Rule 31).
- Body copy and microcopy score at a 5th-7th grade reading level (Rule 32).
- Form field count is trimmed to the immediate transaction's minimum (Rule 33).
Anti-patterns
- Vague, category-less headlines ("Reimagine the future of X") — fails Rule 2's one-line test; reads confident but tells the visitor nothing checkable. Do this instead: name the product category, the customer, and the value in line 1.
- Generic, low-commitment CTA verbs ("Submit," "Learn More," "Click Here") — fails Rule 10. Do this instead: pair a specific verb with a stated deliverable.
- CTA or hero copy hidden behind scroll-triggered reveals or scroll-hijacked motion — native scroll-driven reveals are fine (content reachable without JS, honors
prefers-reduced-motion); scroll-hijacking that overrides scroll position or speed is forbidden on sales pages. Do this instead: keep the primary CTA reachable without hunting or waiting on an animation. - Fake urgency or fake scarcity (countdown timers that reset on refresh, "only 3 left" that never changes) — contradicts Rule 23's authenticity requirement and reopens the legitimacy objection (Rule 15, #1) instead of resolving it. Do this instead: show only real, timestamped events, or drop the tactic entirely.
- Testimonial carousels that auto-rotate before a visitor can finish reading — undermines Rule 19's readable-proof standard. Do this instead: pause on hover/tap, or use a static grid.
- Anonymous or unattributed testimonials — fails Rule 19. Do this instead: always attach a real name + title/company, photo where possible.
- Price hidden behind "Contact us" on a self-serve offer — fails Rule 28 and reopens the pricing objection. Do this instead: show a starting price or a transparent range.
- Walls of unstructured text with no subheads, bullets, or bolding — fails Rules 3 and 32; skimmers (the majority of visitors) get nothing to latch onto. Do this instead: break every block into a benefit header + paragraph (Rule 9).
- Validating on every keystroke, or only at final submit — fails Rule 29; produces false-positive errors mid-typing or delays feedback until the user has already moved on. Do this instead: validate on blur, clear errors live.
- Marking only optional fields and leaving required fields to be inferred — fails Rule 30; measured at a 32% skip rate. Do this instead: mark required fields explicitly.
- Error messages that reflow the layout or hide the field/CTA they refer to — fails Rule 31. Do this instead: reserve inline space for error text so nothing shifts.
- Stacking multiple redundant proof types in the same scroll position (logo bar directly above 3 testimonials directly above a rating badge) — fails Rule 18's placement discipline. Do this instead: spread proof types across the page — above the fold, mid-page, beside the CTA.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

