ad-creative

Overview

What is ad-creative?

The ad-creative skill turns your agent into an expert performance creative strategist. It is designed to generate high-performing ad creative at scale—headlines, descriptions, and primary text that drive clicks and conversions—across major paid advertising platforms.

Use ad-creative whenever you need to:

• Create new ad copy for Google Ads, Meta (Facebook/Instagram), LinkedIn, TikTok, or Twitter/X
• Produce bulk ad variations for creative testing
• Rewrite or iterate on existing ads based on performance
• Respect platform-specific character limits and specs

For campaign strategy, budgets, and targeting use the separate paid-ads skill. For landing page messaging and on-site copy, use copywriting.

Who is this skill for?

ad-creative is built for:

• Performance marketers and growth teams running paid campaigns
• Media buyers and PPC specialists managing Google Ads, Meta, and LinkedIn
• Content and demand gen marketers who need on-brand ad copy quickly
• Agencies and freelancers who must produce ad variations at scale

If your main job is to improve ad performance through stronger creative—rather than to configure tracking, bidding, or audiences—this skill is a strong fit.

What problems does ad-creative solve?

This skill focuses on three core problems:

1. Ad ideation at scale
   Generate multiple angles, hooks, and variations for any offer, without manually writing every line.

2. Platform-compliant copy
   Use best practices and character limits aligned with the Platform Specs reference, so your Google RSAs, Meta ads, and other formats are structurally valid.

3. Performance-focused iteration
   Iterate on existing creative based on what is working—improving click-through and conversion intent instead of just rewriting text.

When is ad-creative a good fit?

Use this skill when:

• You explicitly mention “ad creative,” “ad copy variations,” “RSA headlines,” “ad iterations,” “creative testing,” or “ad performance optimization”
• You want multiple ad variations for A/B or multivariate tests
• You need platform-specific ad assets (e.g., Google RSA, Meta primary text, LinkedIn intro text)

It is not the best fit when:

• You need help choosing audiences, bids, or campaign structure (use paid-ads)
• You primarily want long-form content such as blog posts, whitepapers, or emails (use other writing skills)
• You are only editing landing page or website copy (use copywriting)

How the skill is structured

Inside the ad-creative skill directory you’ll find:

• SKILL.md – Core instructions and workflow for generating and iterating ad creative
• evals/evals.json – Example prompts and expected behavior for different platforms
• references/platform-specs.md – Character limits and structural specs for major ad platforms
• references/generative-tools.md – Guidance on using AI tools for images, video, and voiceover in ads

These references are used to keep your ad creative aligned with platform rules and to suggest compatible visual and media directions.

How to Use

Installation

To install the ad-creative skill into a compatible skills-enabled environment, run:

npx skills add https://github.com/coreyhaines31/marketingskills --skill ad-creative

This pulls the ad-creative skill from the coreyhaines31/marketingskills repository and makes it available alongside your other marketing skills.

After installation, open the skills/ad-creative/ directory and review:

• SKILL.md for the full creative workflow
• references/platform-specs.md for platform limits
• references/generative-tools.md for AI visual and media options

Recommended setup order

1. Check for product marketing context
   If your repo includes .agents/product-marketing-context.md (or .claude/product-marketing-context.md in older setups), make sure your agent reads that file first. It contains positioning, key messages, and audience details that ad-creative is designed to reuse instead of re-asking.

2. Confirm platform and format
   Clearly specify where the ads will run and which format you need:

   • Google Ads: Responsive Search Ads (RSAs), Performance Max assets, display
   • Meta: feed, stories, Reels, carousel
   • LinkedIn: single image, carousel, video ads
   • TikTok or Twitter/X: short-form video or feed formats

3. Clarify product and offer
   Provide what you are promoting: product/service, main value proposition, offer (trial, discount, demo), and key differentiators.

4. Define audience and intent
   Include your target audience (e.g., B2B SaaS founders, ecommerce marketers), their main pain points, and where they are in the funnel (cold, warm, retargeting).

5. Share existing creative, if any
   Paste in current ads or describe what has worked or failed. The skill is built to iterate on real performance data when available.

Example usage patterns

1. Generating Meta ad creative from scratch

Prompt your agent after installing ad-creative with something like:

"Use the ad-creative skill to write Meta (Facebook/Instagram) feed ads for our B2B SaaS analytics tool. Main value prop: automatic weekly reports for marketing teams. Target: marketing managers at mid-market companies. I need 4–5 different angles, each with multiple primary text, headline, and description variations that fit Meta character limits. Suggest image or video directions for each angle."

Using the evaluation patterns in evals/evals.json, the skill will aim to:

• Identify several angles (e.g., time savings, reporting accuracy, ROI, stress reduction)
• Respect recommended Meta character limits for primary text, headline, and description
• Provide multiple variations per angle
• Suggest visual directions for each concept

2. Creating Google Ads RSA copy

For Google Responsive Search Ads, use a prompt such as:

"Use ad-creative to generate Google Ads RSA copy for our CRM product. We’re targeting the keyword ‘best CRM for small business’. Include 15 headlines and 4 descriptions that follow Google RSA specs and use a mix of keyword, benefit, social proof, and CTA-focused lines."

With references/platform-specs.md, the skill can:

• Follow RSA character limits
• Suggest a good mix of headline and description types
• Prepare copy ready to paste into Google Ads

3. Iterating on underperforming LinkedIn or Meta ads

When performance is weak but you have data, prompt like this:

"Our current LinkedIn single-image ads have low CTR (0.3%). Using ad-creative, review this existing copy and propose at least 5 new angles and multiple variations per angle. Maintain our brand voice from product-marketing-context.md and respect LinkedIn intro text and headline norms."

The skill will prioritize angle-based rewrites and testing ideas rather than minor wording tweaks.

Using generative visuals and media with ad-creative

The references/generative-tools.md file outlines how to pair your copy with AI-generated visuals, video, and voice. With it, ad-creative can:

• Suggest when to use image generation vs. video generation
• Point to categories like static images, text-overlay images, short video ads, and voiceovers
• Recommend classes of tools (e.g., image generation engines or video tools) for:
  • Product mockups and variations
  • Brand-consistent templates
  • Short performance videos with or without voiceover

In practice, this means your prompts can include requests like:

"After writing my Meta ad copy, suggest image generation prompts for static ads and 15-second video prompts that match the angles."

and the skill will reference the generative-tools guidance when forming suggestions.

FAQ

When should I choose ad-creative instead of paid-ads or copywriting?

Use ad-creative when your main goal is producing or improving ad creative: headlines, descriptions, primary text, and ad variations for paid platforms. If you primarily need:

• Campaign structure, bidding, or audience strategy → use paid-ads.
• Landing pages, blog posts, or on-site messaging → use copywriting.

You can combine ad-creative with those skills in the same workspace: one to define strategy, another to write the ads, and another to optimize landing pages.

Which platforms does ad-creative support?

The skill is designed for “any paid advertising platform,” with explicit references to:

• Google Ads (including Responsive Search Ads)
• Meta (Facebook and Instagram)
• LinkedIn
• TikTok
• Twitter/X

Because it uses a platform-specs reference, it is particularly strong when you clearly name the platform and format in your prompt.

How does ad-creative handle character limits and platform rules?

ad-creative uses references/platform-specs.md as a guide for:

• Character limits for headlines, descriptions, and primary text
• Recommended combinations and mix of message types (e.g., keyword-focused, benefit-focused, social proof, CTA)

When you specify the platform (e.g., "Google RSA" or "Meta feed ad"), the skill aims to output copy that fits these constraints so you spend less time trimming or reformatting.

Can ad-creative help with creative testing and reporting?

The skill is oriented toward creative testing by:

• Generating multiple angles for each offer
• Providing several variations per angle
• Designing copy that can be compared in A/B or multivariate tests

However, it does not connect directly to ad platforms or reporting APIs. You bring performance data (e.g., CTR, CPC, conversion rate), and ad-creative helps you ideate next-round creative based on what is working or failing.

Does ad-creative generate images or videos itself?

ad-creative primarily focuses on copy. It does not directly call image or video generation APIs, but it uses references/generative-tools.md to:

• Propose visual concepts and directions
• Suggest categories of tools suitable for static images, short videos, and voiceovers
• Help you design prompts to send to your preferred image/video/voice tools

This makes it useful when you want both ad copy and clear guidance on how to support that copy with visuals.

What input should I provide for the best results?

To get strong, on-target ad creative from ad-creative, include:

• Product name and a short description
• Main value proposition and key benefits
• Offer details (trial, demo, discount, guarantee)
• Target audience and intent (cold vs. retargeting, B2B vs. B2C)
• Preferred platform(s) and format(s)
• Any performance learnings or existing ads you want to iterate from

If you maintain a .agents/product-marketing-context.md, keep it updated; the skill is built to rely on that file and only ask for missing, task-specific details.

Is ad-creative suitable for brand-new advertisers?

Yes, if you already know your basic offer and target audience, ad-creative can:

• Help you articulate angles and hooks
• Produce compliant ad copy for common formats
• Suggest visual ideas for your first campaigns

If you are still defining your product, positioning, and audience from scratch, you may want to first build out product marketing context using other strategic tools or skills, then return to ad-creative once the fundamentals are in place.