Ad Creative Generator

SkillMedia

Produce complete ad creative sets - concepts, hooks, copy, AI-generated images and videos declined per platform and format. Use for Meta, Google, LinkedIn, TikTok campaigns and product launches.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Ad Creative Generator skill

What this skill tells your AI

The instructions your AI receives, as published by navinspire-ia/navin in navin/skills/ad-creative-generator/SKILL.md and read by ahel’s review.

Turn a product or offer into a complete, ready-to-launch creative set: concepts, copy, and AI-generated visuals/videos for every placement. Handoff media buying (budgets, audiences, bidding) to paid-ads-manager.

When to use

  • Paid or organic ad creative packages
  • Product launch visual systems

When not to use

  • Long-form landing pages only (copywriting-agent)
  • Campaign calendars without creatives (campaign-manager)

Tooling

  • generate_image - key visuals, packshots, lifestyle, banners; reference images for consistency
  • generate_video - 6-10s ads (text-to-video or image-to-video)
  • visual_qa - required still-image delivery gate for dimensions, ratio, sharpness, contrast, alpha, safe zones and vision-based product/logo/text/color fidelity
  • exec + ffmpeg/Pillow - resize, overlays, concat, subtitles
  • If tools missing: deliver brief + exact prompts; tell user which Settings to enable

Workflow

1. Creative brief (never skip)

Product + key benefit, audience/ICP, platform(s), objective (awareness/traffic/conversion), brand constraints, offer/CTA, language(s). Align with campaign-manager message house.

2. Concepts before pixels

Propose 3 distinct concepts (angle + visual + hook). Angles: problem/solution, before/after, social proof, demo, lifestyle, pattern-interrupt. User picks, or ship the strongest if async.

3. Copy per placement

For the chosen concept, per platform:

  • Hook / headline (3 variants)
  • Primary text (short + long)
  • CTA matching objective
  • Hard limits: Google RSA 30/90, Meta ~125 visible, LinkedIn ~150

4. Visual generation

Build one master key visual, validate, then decline with reference_images:

PlacementRatioNotes
Feed Meta/LinkedIn1:1safe margins for text
Stories/Reels/TikTok9:16subject in middle 60%
YouTube/display16:9logo + CTA visible
Display sizes300x250, 728x90Pillow resize if needed

Leave negative space; quote any baked-in text in the prompt.

5. Video ads

6-10s: hook → benefit → CTA. Prefer image-to-video from the master for product consistency. Variants 9:16, 1:1, 16:9. Subtitles via ffmpeg (video-generation skill).

6. Delivery

  • Run visual_qa on every final still with placement requirements and all authoritative references. Rework WARN and BLOCK results. Never auto-deliver a BLOCK asset, and never claim fidelity without a reference.
  • Save under marketing/creatives/<campaign>/
  • Summary table: concept, per-placement copy, asset paths + ratios
  • Name for A/B: concept-A-hook1-9x16
  • Attach media via message when appropriate

Rules

  • One concept, many declinations - never mix concepts in one ad set.
  • Brand tokens (brand-voice-manager) override generic style.
  • Never invent claims, prices, or stats; ask or omit.
  • Legal: no competitor logos, no fake endorsements; respect platform policies.
  • Iterate masters with reference_images instead of regenerating from scratch.

Anti-patterns

  • Shipping pixels with no hooks/CTA
  • Unreadable text baked into images
  • Claiming expected CPA without account data

Signals

GitHub stars
22
Forks
4
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
ad-creative-generator
Source
github.com/navinspire-ia/navin
Ad Creative Generator (ad-creative-generator): Skill · ahel