---

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]]