I built 10 Claude Code skills for Meta Ads
They're free. Here's what they do: 👇
Operations:
bleed-check: finds ad sets burning cash with zero conversions, pauses them
rebalance: shifts budget from losers to winners automatically
fatigue-scan: catches creative fatigue before your CPMs spike
weekly-report: pulls KPIs, compares WoW, sends to Slack
Creative & Intelligence:
spy: scrapes competitor ads from the Ad Library, diffs weekly
bulk-creative: generates 50–500 ad variations, renders to PNG
hooks: writes 50+ copy variations using PAS, AIDA, BAB frameworks
deploy-ads: reads a manifest, creates campaigns via API in minutes
Setup & Architecture:
setup-capi: generates production-ready Conversions API code
audience-audit: finds overlap, maps funnel stages, fixes exclusions
How to use them:
Drop the .md files into .claude/commands/
Set your Meta access token
Type /bleed-check, /spy, /hooks — it just runs
~25 hours/week of manual work → automated.
No complex setup. No code to write.
Just slash commands.
Full skill files + setup guide in the article.
Comment "CLAUDE" and I'll share the link
(must be following!)
I used to drop $4K to $8K per UGC campaign on creators, studios, and reshoots.
This workflow just changed the math.
Three AI tools. One pipeline. Zero cameras.
Step 1: Claude Opus 4.8
Forget copy. We use Claude to cast our fake creator.
Give it your product, your demographic, and your setting.
Bathroom shelf.
Gym bag.
Car console.
It spits out an actor brief and scene direction that anchors every clip. The hook. The talking head. The product close-up.
Step 2: GPT-Image-2
One prompt. One face that does not exist.
No model releases. No casting calls. No "can you film this again but happier?"
The face retention is the secret sauce here. That is what makes this actually work for brand campaigns.
Step 3: Seedance 2.0
Three inputs:
Your product photo.
Your AI actor.
The motion prompt from Claude.
Out comes a 15-second clip. Natural hand motion. Subtle breathing. Lighting that looks like someone actually shot this on their phone.
The part most people miss:
You do not need five subscriptions and a dozen browser tabs.
HeyOz runs both GPT-Image-2 and Seedance 2.0 natively. Built-in image editor. Built-in video editor.
Prompt to actor to final cut to export. All in one workspace.
No freelancer invoices. No "good enough" compromises.
I wrote up the exact Claude prompt structure and the Seedance motion templates we use to keep batches consistent.
Comment WORKFLOW below and I will send it over.
Follow for Part 2 where I break down how these AI UGC variants are actually performing against human-shot content in our Meta ad accounts.
Remove your free plan.
Free users are leeches. They increase support, server costs, and make you build features your paying customers don’t want.
Less than 3% of free users ever convert.
VC startups can afford free plans because they’re optimizing for growth, not profit.
If you’re an indie founder, never give away your product for free.
The fastest way to make people trust you in 2026 is to look like you've already spoken on a stage they respect.
Nobody checks if the stage was real.
Every "expert" on this cover was generated on my laptop in an afternoon.
No conference.
No booking fee.
No photographer.
And they read as instant authority figures the second someone sees them in feed.
This is the trick high-trust niches have been quietly missing. Health, finance, coaching, B2B.
The keynote-on-a-stage shot does something a selfie never will. It signals "this person is the authority in the room" before a single word is read.
You used to need a real speaking slot and a $2,000 photoshoot to get that shot. Now you need three tools and the right order to run them:
→ Tell Claude Opus 4.8 the exact authority shot you want (a seasoned doctor mid-gesture on a wooden stage, medical slide glowing behind her)
→ Claude rewrites it into a precise prompt for GPT Image 2
→ GPT Image 2 generates a photoreal still that looks shot at a real summit
→ Seedance 2.0 turns that still into a moving clip with natural gestures and delivery
Then you pull it all into HeyOz, where GPT Image 2 and Seedance already sit together, and assemble the finished ad without an editor.
The old way: one real stage, one face, weeks of planning, thousands of dollars.
This way: any expert, any niche, any stage you can describe, done in under an hour for the cost of a coffee.
The authority is the product. The stage is just set dressing.
I wrote the whole thing up. The exact prompts, the Claude-to-GPT handoff, the Seedance settings, the assembly flow. Completely free, no email wall.
Want it? Drop "STAGE" in the comments and I'll send it your way.
(make sure you're following so it can reach you)
Claymation ads are quietly printing money on Meta right now.
2-3x the feed average CTR. Almost nobody is running them. And everyone assumes you need an animation studio and a $5,000 budget to make one.
You don't. You need the right prompts.
So I packaged the entire system into a free Claude workflow. Paste a product URL, and it spits out the full ad in minutes:
→ a 9-shot storyboard (proven problem-villain-hero arc)
→ image prompts for every single frame
→ video animation prompts ready for Kling, Veo or Runway
→ a timed voiceover script
→ music and sound direction
The clay villain reveal alone is the kind of pattern break that stops the scroll dead.
Old way: $3,000 to $5,000 and 3 weeks with a studio.
This way: $3 to $6 in credits and under an hour.
I've watched founders make their first one over a coffee break.
The skill is 100% free. No gate, no signup.
Want it? Comment "CLAY" and I'll send you the full guide
(need to be following so you can receive the DM)