---
name: messaging-compliance-checker
description: Use to run any customer-facing copy through a systematic pre-publication compliance review for a legal AI assistant. Checks copy against hard rules (bible signoff, banned outcome claims, surface-specific rules), allowed/banned claim lists for both consumer and lawyer audiences, and the bridge line. Outputs a structured pass/fail verdict with flagged phrases and required edits. Apply before any marketing, in-app, social, or PR content ships.
license: MIT
metadata:
  id: messaging.compliance-checker
  category: messaging
  jurisdictions: [__multi__]
  priority: P2
  intent: [messaging, compliance, copy-review, pre-publication, marketing-review]
  related: [messaging-hard-rule-bible-signoff-required, messaging-outcome-claims-banned, messaging-outcome-claims-allowed, messaging-banned-claims-consumer, messaging-banned-claims-lawyer, messaging-bridge-line]
  source: Louis — HAQQ Legal AI (github.com/sboghossian/mini-claude-for-legal)
  version: "1.0"
---

# Messaging — Compliance Checker

## When this applies

Every piece of customer-facing copy — written by a human, drafted by an AI, or generated by a contractor or agency — must pass this compliance check before publication. "Customer-facing" includes:

- All digital advertising (search, display, social)
- Website copy (homepage, product pages, landing pages, pricing)
- Email campaigns (waitlist, onboarding, product, retention)
- In-app notifications, tooltips, and system messages
- Press releases and media statements
- Influencer briefs and sponsored content scripts
- Sales decks, product one-pagers, and partnership proposals
- Conference presentations and booth materials

This checker is **blocking**: any copy that fails a hard-rule check cannot ship until revised and re-reviewed.

---

## Inputs

| Input | Description |
|-------|-------------|
| Copy text | The full text, headline, CTA, and any disclaimers of the asset being reviewed |
| Surface / channel | Which surface the copy will appear on (homepage, Instagram ad, sales deck, press release, etc.) |
| Primary audience | Consumer (B2C), lawyer (B2B), or mixed |
| New claim flag | Whether any claim in the copy is being used for the first time (triggers bible signoff requirement) |

---

## Review Methodology — Four-Pass Check

### Pass 1 — Hard Rules

Check against [[messaging-hard-rule-bible-signoff-required]] and [[messaging-hard-rule-preapproved-press-quotes-only]].

| Rule | Check question |
|------|----------------|
| Bible signoff | Does the copy contain any claim not previously approved in the messaging bible? If yes → block until signoff obtained |
| Press quotes | Does the copy contain a quote from a third party (media, client, influencer)? If yes → verify pre-approval on file; if not on file → block |

### Pass 2 — Banned Outcome Claims

Check against [[messaging-outcome-claims-banned]].

Scan for any of the following or their semantic equivalents:

- "Will win your case" or any case-outcome guarantee
- "Guaranteed to save money"
- "Replaces your lawyer" (in any form)
- Any specific dollar amount recovered or damages claimed
- Any outcome expressed as a certainty rather than a possibility

**Action on fail:** Flag exact phrase → require replacement with [[messaging-outcome-claims-allowed]] equivalent.

### Pass 3 — Audience-Specific Banned Claims

**If consumer audience:** Check against [[messaging-banned-claims-consumer]].
**If lawyer audience:** Check against [[messaging-banned-claims-lawyer]].
**If mixed:** Apply both.

Scan for:
- UPL-triggering language (consumer surfaces)
- Displacement / commodification framing (lawyer surfaces)
- Consumer framing on professional surfaces
- Professional jargon on consumer surfaces (accessibility failure, not compliance failure)

**Action on fail:** Flag exact phrase → reference allowed alternative from the corresponding allowed-claims skill.

### Pass 4 — Bridge Line Test

Is the overall copy consistent with the bridge line ("legal understanding accessible to everyone; legal work faster for professionals")?

Ask:
1. Does it stay within the "understanding / accessibility / faster / professional primacy" frame?
2. Does it imply replacement, disruption, or commodification of the legal profession?
3. Does it make any claim the bridge line rules out?

**Action on fail:** Flag the conflicting claim → require rewrite to bring within bridge line territory.

---

## Output Format

Return a structured compliance verdict for each piece of copy reviewed:

```
COMPLIANCE REVIEW — [Asset name / surface]
Audience: [Consumer / Lawyer / Mixed]
Reviewer: [Name / AI]
Date: [YYYY-MM-DD]

VERDICT: [PASS / FAIL / CONDITIONAL PASS]

---
PASS 1 — Hard Rules
[ ] Bible signoff: [PASS / BLOCK — reason]
[ ] Press quotes: [PASS / BLOCK — reason]

PASS 2 — Outcome Claims
[ ] Banned outcomes: [PASS / FLAG — quote the flagged phrase]
  Required edit: [replacement suggestion]

PASS 3 — Audience Claims
[ ] Banned consumer claims: [PASS / FLAG — quote]
[ ] Banned lawyer claims: [PASS / FLAG — quote]
  Required edit: [replacement suggestion]

PASS 4 — Bridge Line
[ ] Bridge line consistency: [PASS / FAIL — explain]
  Required edit: [rewrite direction]

---
SUMMARY
Flagged items: [N]
Blocked items: [N]
Next step: [Revise and resubmit / Obtain signoff / Ready to publish]
```

A **CONDITIONAL PASS** means no hard blocks exist but minor revisions are recommended before publication.

---

## Escalation

| Situation | Action |
|-----------|--------|
| New claim type not in the bible | Block → refer to legal/brand for bible signoff per [[messaging-hard-rule-bible-signoff-required]] |
| Jurisdiction-specific claim (e.g., KSA bar advertising rules) | Block → refer to local bar compliance review |
| Influencer or partner copy using banned phrasing | Block → return to influencer for revision per [[messaging-surface-rule-influencer-brief]] |
| Crisis communication copy | Fast-track review — still apply all four passes but same-day turnaround |
| AI-generated copy | Same process; AI authorship does not reduce the compliance bar |

---

## Automation

This checker is designed to run as an automated pre-publish gate in the content workflow. Implementation steps:
1. Copywriter submits draft to the compliance system (CMS field, Slack workflow, or review tool)
2. System runs the four-pass check against the current versions of the linked skills
3. Output verdict is returned to copywriter with flagged items
4. Blocked items cannot be scheduled for publication until re-checked and passed
5. Conditional passes may be published after copywriter acknowledges recommendations

---

## Related skills

- [[messaging-hard-rule-bible-signoff-required]]
- [[messaging-hard-rule-preapproved-press-quotes-only]]
- [[messaging-outcome-claims-banned]]
- [[messaging-outcome-claims-allowed]]
- [[messaging-banned-claims-consumer]]
- [[messaging-banned-claims-lawyer]]
- [[messaging-bridge-line]]