Compliance Checker

Created by Rob Palmer — direct-response copywriter ($523M+ tracked). Source & updates: https://github.com/robpalmer99/claude-code-copywriting-skills Free under CC-BY-4.0. Attribution requested when redistributed.

Compliance Checker Skill

This skill checks direct response copy (ads, landing pages, VSLs, emails) against platform compliance rules. Built on real-world compliance knowledge from practitioners who've been banned 300+ times and learned what actually triggers rejections.

The goal: catch compliance issues BEFORE your ad gets rejected or your account gets banned, while preserving persuasive power through compliant rewrites.

Step 1: Classify the Copy

Before checking anything, determine these four things. Ask the user if unclear.

Copy Type

• Ad (primary text, headline, description, video script)
• Landing page (bridge page, pre-sell, PDP)
• VSL script (video sales letter transcript)
• Email (broadcast, sequence, autoresponder)
• Advertorial (native content, editorial-style)

Platform

• Meta/Facebook (strictest — deep rules below)
• Google Ads (medical claims ban, bait-and-switch enforcement)
• TikTok (commercial disclosure, biometric imagery ban)
• YouTube (more lenient — before/afters allowed)
• ClickBank (affiliate network rules, income claims)

Product Category

• Supplement / ingestible (HIGHEST scrutiny on Meta — extra fine-tooth comb)
• Ecom physical (post-purchase surveys, page score penalties)
• Digital product (lower scrutiny — no post-purchase surveys)
• Coaching / service (income claim sensitivity)

Note: Digital products with physical upsells may be reclassified as ecom.

Funnel Position

• Front-end ad (strictest rules)
• Landing page (strict — product must be visible, exit nav required)
• Upsell page (more leniency — countdown timers, urgency, aggressive language OK post-purchase)

Step 2: Load Trigger Words Reference

Read references/trigger-words.md for the consolidated table of trigger words and compliant substitutions.

Step 3: Phase 1 — Checklist Audit

Run through each category. Report PASS / FAIL / WARN for each.

The 9-Point Compliance Checklist

1. Personal Attributes

• No "you" or "your" assuming the viewer has a condition (ecom/supplement ads)
• No questions implying problems ("Are you struggling with...?" = implies they have the problem)
• No singling out demographics ("women's metabolism" = personal attribute)
• No calling out conditions ("if you have diabetes" = implies they have it)
• Uses third-person framing where needed ("Many Americans..." / "Most people...")

Why this matters: Facebook considers implying you know the viewer's personal attributes (age, medical conditions, body type, financial status) a violation. Change "you" to "I" when describing pain, or use third-person framing.

Compliant patterns:

• "Many Americans are looking for ways to..." (third-person)
• "I used to struggle with..." (first-person story)
• "People working on problem areas..." (general group)

Non-compliant patterns:

• "Are you struggling with...?" (implies they have the problem)
• "If you have diabetes..." (calls out condition)
• "Women over 40 who..." (singles out demographic + age)

2. Sensational & Negative Language

• No banned negative words: ruin, destroy, kill, wreck, terrible, worse, disgusting, bad, harmful, detrimental, damaging, wipe out, butcher
• No scare tactics ("If you don't take X, you'll decrease your lifespan")
• No shocking, violent, or fear-based content
• No negative imagery descriptions (sad faces, red spots on bodies, weapons)
• Positive/neutral framing throughout

Compliant substitutions: impact, change, support, help, balance, manage, affect, influence

3. Medical & Health Claims

• All health claims use softening words (helps, supports, can, may, promotes)
• No definitive claims ("grow hair" → "help support hair growth")
• No "X results in Y days" (timelines + outcomes = definitive claim)
• No disease names used directly (diabetes, Alzheimer's, dementia, cancer)
• Scientific claims reference studies if made
• Testimonials include "results may vary" asterisk

The softening rule: NEVER use health verbs without a modifier. "Regulate blood sugar" is a claim. "Helps support blood sugar" is compliant.

Required modifiers: helps, supports, can, may, promotes, aids in, assists with

Disease name workarounds:

• "Diabetes" → "type two" (without the word "diabetes")
• "Alzheimer's/dementia" → "memory," "brain health," "brain fog"
• "Cancer" → cannot mention, must reframe entirely
• "Arthritis" → "joint health," "joint comfort"

4. Income & Financial Claims

• No guaranteed income promises
• No specific earnings claims without disclaimers
• Uses "revenue" not "earnings" where possible
• No "lifetime" access (ClickBank ban)
• Reflects actual/typical effort required

Compliant income language:

• NOT: "Let me show you how to generate XXX leads at $5 a pop"
• OK: "My unique process makes generating XXX leads for $5 possible"
• NOT: "Guarantee passive income"
• OK: "Potential income opportunity with effort and attention"

5. Before & After

• No before/after side-by-side images (Facebook ban)
• No zooming into body parts (makes users feel self-conscious)
• Compliant alternatives used if showing results

Compliant alternatives:

• First-person success stories on camera (influencer style)
• Customer testimonials describing specific benefits (text, not images)
• Product demonstration videos

Note: YouTube ALLOWS before/afters. This rule is Facebook-specific.

6. Open Loops & Clickbait

• No unresolved open loops (even if closed in body, headline can't leave them open)
• No multi-click curiosity chains / information gates
• Product name visible in post text, headline, or video watermark
• Ad accurately reflects landing page content

The open loop rule:

• NOT compliant: "Watch now to discover 5 foods that accelerate arthritis" (open loop)
• Compliant: "We reveal what foods to avoid below" (resolved)
• NOT compliant: "Find out what the 3 most effective ways to lose weight really are" (hidden behind noun)
• Compliant: "Three effective ways to manage weight" (direct statement)

7. Trigger Words

• Check all copy against references/trigger-words.md
• All flagged words replaced with compliant alternatives
• Softening modifiers present on all health/medical verbs

8. Landing Page Elements (if reviewing a landing page)

• Product info visible without watching VSL (viewer can determine what's sold by scrolling)
• VSL does not auto-play (looping GIF OK if not first 2 seconds of video)
• Exit navigation present (hamburger menu, header links, footer links)
• Homepage is legitimate (not just 2-page blog — needs substance, privacy policy, ToS)
• No fake countdown timers or fake scarcity
• Disclaimers present (privacy policy, return policy, ToS)
• "Results may vary" on testimonials
• Scientific claims have citation + full study reference at page bottom

9. Ad-to-Page Consistency

• Visual consistency (same logo/watermark across ad and page)
• Same messaging vibe (viewer feels they're in the right place)
• No bait-and-switch (ad claims match landing page content)
• Same root domain (no redirect chains to different domains — common ban reason)
• Go direct to brand URL (no affiliate-style redirect links)

Step 4: Phase 2 — Line-Level Flagging

Go through the copy line by line. For each issue found:

### Issue N [SEVERITY]
> "The exact problematic text quoted here"
**Rule:** Which compliance rule this violates and why
**Rewrite:** A compliant alternative that preserves persuasive power

Severity levels:

• HIGH — Likely rejection or account ban. Must fix before running.
• MEDIUM — Risky, could be flagged depending on reviewer. Should fix.
• LOW — Minor concern, could be flagged in stricter review cycles.

Prioritize issues by severity. HIGH first, then MEDIUM, then LOW.

Rewrite rules:

• Preserve the persuasive intent of the original
• Use compliant substitutions from references/trigger-words.md
• Maintain the copy's voice and rhythm
• The rewrite should be drop-in ready — same approximate length, same tone

Step 5: Compliance Score

After both phases, calculate:

Scoring: Start at 10. Subtract 2 per HIGH issue, 1 per MEDIUM, 0.5 per LOW. Floor at 0.

Output Format

## Copy Classification
- Type: [ad / landing page / VSL / email / advertorial]
- Platform: [Meta / Google / TikTok / YouTube / ClickBank]
- Product: [supplement / digital / ecom / coaching]
- Funnel position: [front-end ad / landing page / upsell]

## Phase 1: Checklist Audit

| # | Category | Status | Notes |
|---|----------|--------|-------|
| 1 | Personal Attributes | PASS/FAIL/WARN | [brief note] |
| 2 | Sensational/Negative Language | PASS/FAIL/WARN | [brief note] |
| 3 | Medical/Health Claims | PASS/FAIL/WARN | [brief note] |
| 4 | Income/Financial Claims | PASS/FAIL/WARN | [brief note] |
| 5 | Before/After | PASS/FAIL/WARN | [brief note] |
| 6 | Open Loops & Clickbait | PASS/FAIL/WARN | [brief note] |
| 7 | Trigger Words | PASS/FAIL/WARN | [brief note] |
| 8 | Landing Page Elements | PASS/FAIL/WARN/N/A | [brief note] |
| 9 | Ad-to-Page Consistency | PASS/FAIL/WARN/N/A | [brief note] |

## Phase 2: Line-Level Issues

### Issue 1 [HIGH]
> "quoted text"
**Rule:** explanation
**Rewrite:** compliant alternative

### Issue 2 [MEDIUM]
> ...

## Compliance Score: X/10
[Ready to run / Needs fixes / Major rewrite needed]

## Summary
- X HIGH issues, Y MEDIUM, Z LOW
- Top priority fixes: [numbered list]
- What's already compliant: [brief note]

Platform-Specific Rules

Meta / Facebook (Deep Coverage)

The 7 Deadly Sins of Facebook Ads

1. Implied negative attributes — assuming you know the viewer's problems
2. Sensational/shocking language — scare tactics, violent imagery, death language
3. Withholding information — clickbait loops that never deliver
4. Definitive medical claims — "X results in Y days," guaranteed outcomes
5. Before/after imagery — side-by-side body comparisons
6. Personal attributes — singling out gender, age, condition
7. Negative language framing — ruin, destroy, kill, wreck, terrible

Meta's 3-Tier Review System

Digital vs Ecom Scrutiny on Meta

Compliant Headline Rewrites (Meta Examples)

Facebook vs YouTube Comparison

Google Ads (Light Coverage)

• Prohibited: medical claims, "free trial," "miracle cure," "unproven remedy"
• All claims must be supported by scientific evidence
• Bait-and-switch: Landing page must align perfectly with ad
• Consequences: Immediate account suspension without warning; potential permanent ban
• Google is less nuanced than Meta — they tend to ban first, ask questions never

TikTok (Light Coverage)

• Mandatory commercial content disclosure for all promotional posts
• Claims must be proven with in-frame disclosures
• Medical props or "biometric imagery" banned
• Hormone-related keyword ads removed
• Supplement advertisers must show regional certification before approval
• More focused on disclosure requirements than language policing

YouTube (Light Coverage)

• More lenient than Facebook — before/afters allowed, more direct claims
• Same FTC rules apply — no deceptive claims, income disclaimers required
• Many things banned on Facebook are fine on YouTube
• Still: no outright false claims, no medical diagnosis language

ClickBank (Light Coverage)

• Never use "lifetime" for lifetime access
• No personalized coaching/services claims that cannot be delivered at scale
• Cannot claim specific income from techniques taught
• Reasonable assumption: Average person can achieve supplemental income with effort
• Sales pages must clearly explain what customers receive
• Salesmanship encouraged, but do not over-embellish

FDA Language Rules (Health/Supplement Specific)

For health and supplement copy, FDA rules apply on top of platform rules:

• Cannot claim to treat, cure, diagnose, or prevent any disease
• Structure talk is OK: Describe anatomy and body functions ("supporting healthy liver function")
• Function talk is OK: "Helping brain function in a healthy way"
• Conditional language required: "may," "can," "helps support"
• Abstract nouns safe: Happiness, wellness, vitality, freedom (cannot be measured = cannot be disputed)

Specific FDA language flags:

• "High blood sugar" → use "erratic blood sugar"
• "Blood sugar spike" → use "blood sugar fluctuations"
• "Managing your blood sugar" → use "sustaining your blood sugar"

Pro tip: Read FDA warning letters — they show the exact language the FDA does not like. These are essentially a cheat sheet for what to avoid.

The Compliance Philosophy

Compliance is a creative constraint that produces better copy. When you cannot rely on hype, you need:

• Better stories
• Stronger mechanisms
• More credible proof
• More nuanced emotional appeals

The best modern DR copy is MORE sophisticated because of compliance. Smart copywriters treat compliance as a competitive advantage — those who understand it can scale while others get banned.

The golden rule: Write for conversion, then edit for compliance. Do not let compliance thinking kill the creative process — get the persuasive draft right first, then run it through this checklist.

Being compliant and being persuasive are not mutually exclusive. The best copy does both.