Canonical NotFair workflow

Read ../../paid-ads/paid-ads-creative/SKILL.md completely, then follow it as the active workflow. Resolve every relative reference from that file against ../../paid-ads/paid-ads-creative/.

---

Supporting file: paid-ads/paid-ads-creative/SKILL.md

Paid Ads Creative

Read ../shared/operating-contract.md and ../shared/measurement-framework.md. Route platform execution to /notfair:google-ads-copy, /notfair:google-ads-assets, /notfair:meta-ads-creative, /notfair:paid-ads-x, or /notfair:paid-ads-linkedin as appropriate.

Work from named evidence

Start with a verified source: winning ads, search terms, customer research, approved landing-page claims, sales calls, reviews with permission, or a documented offer. Match each promise to its landing destination. Do not fabricate testimonials, ratings, prices, outcomes, availability, ownership, or product proof.

Produce a creative slate

Deliver three distinct ready_for_review concept cards. Each card contains:

• a concept ID and hypothesis: audience × motivation × angle × format × one variable;
• the evidence source and claim ledger, with unsupported claims marked needs_substantiation;
• hook, primary message, CTA, destination, and placement adaptation;
• visual/production brief, including required rights or missing assets;
• success metric, guardrail, attribution window, observation window, and failure condition.

Do not make cosmetic variants of the same idea. Test one strategic variable at a time: hook, proof, format, angle, audience, or destination. Diagnose fatigue with complete-window frequency, link CTR, CPM, and delivery context before recommending a refresh.

Publication boundary

Validate current platform specifications and policies from a verified source before publishing. A brief or generated image is not proof that a claim is permitted or that the platform accepted the asset. For platforms without a verified NotFair mutation surface, output an operator-ready brief and clear handoff rather than promising an upload.