Regulator Eyes Skill
Marketing is written for customers but eventually read by regulators, competitors, and plaintiff's lawyers. This skill performs that hostile reading now: every claim inventoried, the substantiation each would require, and the inquiry letter that arrives when someone files a complaint. (Environmental claims have a dedicated sibling: greenwashing-self-audit.)
What This Skill Produces
- Claim inventory — every express and implied claim, including ones made by images, testimonials, and omission
- Substantiation demands — what evidence a regulator would require per claim, and whether the user has it
- The inquiry letter — the civil investigative demand / information request they could receive
- Fix-or-drop debrief — per claim: keep with evidence, reword, add disclosure, or drop
Required Inputs
Ask for these if not provided:
- The marketing material — landing page text, ad copy, emails, app store listing (paste it)
- What evidence exists — studies, data, guarantees infrastructure (or "none yet" — that's an answer)
- Jurisdiction/vertical (optional) — default to US FTC framing; flag if health, finance, or children's products (higher bar)
Framework: How an Investigator Reads
| Pass |
Looking for |
| 1. Express claims |
Direct statements: "fastest", "clinically proven", "saves 40%", "#1" |
| 2. Implied claims |
What a reasonable consumer takes away — before/afters, testimonials as typical results, comparison imagery |
| 3. Material omissions |
Conditions, fees, auto-renewals, "results not typical" realities left unsaid |
| 4. Format traps |
Fake countdown timers, dark-pattern cancellation, undisclosed endorsements/affiliates |
Risk scale: 🔴 enforcement-grade (deceptive on its face or unsubstantiated health/money claim) · 🟡 challengeable (defensible only with evidence the user must produce) · 🟢 puffery (opinion no reasonable consumer takes literally — "the best coffee in town").
Judge claims by the net impression on a reasonable consumer, not the writer's intent — that is the actual legal standard's shape.
Output Format
Regulatory Reading: [Asset] — [date]
Simulation — a plausible adversarial reading, not a prediction or legal advice.
Claim Inventory
| # |
Claim (verbatim) |
Type (express/implied/omission) |
Substantiation required |
User has it? |
Risk |
The Inquiry Letter
[A formal information request citing the specific claims, demanding the substantiation, with a response deadline — the document that starts a very bad quarter.]
Debrief — out of character
| # |
Verdict |
New wording or required disclosure |
| [keep / reword / disclose / drop for every 🔴 and 🟡] |
|
|
Confirm anything load-bearing with an advertising-law attorney — standards vary by jurisdiction and vertical.
Quality Checks
Anti-Patterns
1---2name: regulator-eyes3description: Read your marketing claims, landing page, or ad copy the way a consumer-protection investigator would (FTC/ASA framing) and draft the inquiry letter they could send. Use when asked to check my marketing claims, read this like a regulator, audit my landing page for claim risk, or is this ad compliant. Produces a claim inventory with substantiation demands, the inquiry letter, and a fix-or-drop debrief per claim.4---5
6# Regulator Eyes Skill
7
8Marketing is written for customers but eventually read by regulators, competitors, and plaintiff's lawyers. This skill performs that hostile reading now: every claim inventoried, the substantiation each would require, and the inquiry letter that arrives when someone files a complaint. (Environmental claims have a dedicated sibling: `greenwashing-self-audit`.)
9
10## What This Skill Produces
11
12- **Claim inventory** — every express and implied claim, including ones made by images, testimonials, and omission
13- **Substantiation demands** — what evidence a regulator would require per claim, and whether the user has it
14- **The inquiry letter** — the civil investigative demand / information request they could receive
15- **Fix-or-drop debrief** — per claim: keep with evidence, reword, add disclosure, or drop
16
17## Required Inputs
18
19Ask for these if not provided:
20- **The marketing material** — landing page text, ad copy, emails, app store listing (paste it)
21- **What evidence exists** — studies, data, guarantees infrastructure (or "none yet" — that's an answer)
22- **Jurisdiction/vertical** (optional) — default to US FTC framing; flag if health, finance, or children's products (higher bar)
23
24## Framework: How an Investigator Reads
25
26| Pass | Looking for |
27|---|---|
28| 1. Express claims | Direct statements: "fastest", "clinically proven", "saves 40%", "#1" |
29| 2. Implied claims | What a reasonable consumer takes away — before/afters, testimonials as typical results, comparison imagery |
30| 3. Material omissions | Conditions, fees, auto-renewals, "results not typical" realities left unsaid |
31| 4. Format traps | Fake countdown timers, dark-pattern cancellation, undisclosed endorsements/affiliates |
32
33**Risk scale:** 🔴 enforcement-grade (deceptive on its face or unsubstantiated health/money claim) · 🟡 challengeable (defensible only with evidence the user must produce) · 🟢 puffery (opinion no reasonable consumer takes literally — "the best coffee in town").
34
35Judge claims by the *net impression on a reasonable consumer*, not the writer's intent — that is the actual legal standard's shape.
36
37## Output Format
38
39---
40
41# Regulatory Reading: [Asset] — [date]
42
43> Simulation — a plausible adversarial reading, not a prediction or legal advice.
44
45## Claim Inventory
46| # | Claim (verbatim) | Type (express/implied/omission) | Substantiation required | User has it? | Risk |
47|---|---|---|---|---|---|
48
49## The Inquiry Letter
50[A formal information request citing the specific claims, demanding the substantiation, with a response deadline — the document that starts a very bad quarter.]
51
52## Debrief — out of character
53| # | Verdict | New wording or required disclosure |
54|---|---|---|
55[keep / reword / disclose / drop for every 🔴 and 🟡]
56
57*Confirm anything load-bearing with an advertising-law attorney — standards vary by jurisdiction and vertical.*
58
59---
60
61## Quality Checks
62
63- [ ] Implied claims and omissions are inventoried, not just literal sentences
64- [ ] Every 🔴 names the specific missing substantiation, not "needs evidence"
65- [ ] Puffery is honestly rated 🟢 — inflating everything to red destroys the signal
66- [ ] The letter cites the user's actual claims verbatim
67- [ ] Every red/yellow claim gets a concrete verdict with replacement wording where kept
68
69## Anti-Patterns
70
71- [ ] Do not grade intent — grade the net impression on a reasonable consumer
72- [ ] Do not invent claims the material doesn't make; the inventory quotes the source
73- [ ] Do not offer "add an asterisk" as a fix for a deceptive net impression — disclosures cure omissions, not lies
74- [ ] Do not treat testimonials as safe because they're "just customers talking" — typicality is the user's problem
75- [ ] Do not stay in character in the debrief
gabrielmoreira