SuperCMO Skills
With SuperCMO skills, an agent creates ad photos and videos from a product image and brief, seeking approval before generation.
- 306 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- The agent runs paid generation and submits assets to selected providers; local MCP reads keys from configured storage.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/supercmo-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The agent runs paid generation and submits assets to selected providers; local MCP reads keys from configured storage.
This is third-party code. Review the repository files before installing.
What it does
SuperCMO Skills guide agents through ad production from a product photo or page. The agent analyzes the product, proposes a concept, writes a script and plans shots. After approval, connected image, video and voice models generate material for assembly. The bundle installs skills with a local MCP server. Users choose their own vendor keys or managed SuperCMO access and review both the concept and the final product depiction.
Who it is for. Marketing teams producing product creatives with Claude Code, Cursor or Codex who want a plan before paid generation.
Good fit when
- A product photo and brief need a new ad video.
- The script and shot sequence need approval before generation.
- A campaign needs several product-image concepts.
- Generated clips need assembly with narration into one video.
Not a fit when
- You expect generation without keys or paid model access.
- Product depiction must be perfect without visual review.
- The team lacks rights to photos and references submitted to providers.
Example request
Propose an ad script from this product photo and show generation dry runs without making paid requests.Limitations
Installation needs uv on PATH because the local MCP runs through uvx. Generation uses paid requests; agree the concept and inspect a dry run before spending. Own-key requests go directly to vendors, while managed access routes through SuperCMO; choose the appropriate path for your data. Anonymous usage counts are enabled by default and can be disabled with `SUPERCMO_TELEMETRY=false` or `DO_NOT_TRACK=1`.
How to disable. Disable the skills and local MCP in the client. Revoke unused vendor keys or managed access separately; generated media remains in the project.
FAQ
What follows installation?
Run supercmo-setup or ask the agent to configure keys and connectivity.
Why are tools missing?
Restart the client to reload its MCP connection.
Can I preview a request without spending?
Yes. Set `dry_run` to have MCP show the exact request without calling the generation API.
Where do my own keys go?
Use local `~/.supercmo/.env` or the MCP process environment, outside chat.
Does copying skills also configure MCP?
No. skills.sh copies instructions; the native installer wires MCP for the client.
Discussion
No questions or reviews yet.
Related
SkillsA skill library takes an agent from audience and offer definition to ad ideas, funnel structure and a final copy clarity check.
E-commerce Visual Copywriting
ecommerce-visual-copywriting
A skill that turns product data into a plan for main images, the product page and A+ content, with on-image copy, generation prompts and a claims check
Nine skills for taking a product overseas: Meta, Google and TikTok ads, cold email, UGC, social media and a 90-day strategy.
A product marketing skill set from customer research and positioning to launch, with shared context and approval steps.
Ten Claude Code skills for small business social media: brand, calendar, captions, visuals and a monthly performance review
Eight skills for paid social creative strategy: review mining, messaging angles, hooks and ad formats from Motion's strategy lead
README in short
The collection covers campaigns, storyboards, product photography and video. The hosted SuperCMO app has additional features distinct from the skill bundle.
README badge
Original SKILL.md
--- name: generating-ad-videos description: Makes a product ad video — a product showcase with no one on screen, or a story or lifestyle commercial where a presenter or actor plays a role in the brand's spot, with or without a voiceover. Use when the user wants a product ad, commercial, TV ad, brand film, or product-hero video. Not for a creator or customer sharing their own take on a product, like a review, unboxing, try-on or testimonial — that is generating-ugc-videos; not for a video with no product being sold — that is generating-videos. license: Apache-2.0 metadata: version: "0.3.0" category: creative summary: "Produces a polished, cinematic product commercial in your brand's voice — a no-actor product showcase or an actor-led story ad, at any length. It storyboards every clip to hold your real product identical across cuts, writes the voiceover and casts the actor when the ad calls for one, then stitches it into one ready-to-run spot you approve step by step before anything expensive renders." --- # Product ads Turn a product into a finished commercial video. One flow, with two optional steps: a voiceover (on by default) and an actor (only where a person is on camera). ## Workflow **The ad runs on `seedance-2.0-fast` unless the user named a model, and is delivered `9:16` unless the user explicitly asked for another ratio.** ### Step 1: Read what you have - **A product image or URL** → hand it to `analyzing-products` for what the product is, how a person physically uses it, which parts open or move, and what must stay identical wherever it appears. - **Run `image_analysis` on every image supplied**, not only the product — what each one shows, and whatever the product facts leave out. - **What each image is for comes from the brief, not from its contents.** A person in a frame does not make it the presenter, and a second object does not make it a prop. - **No product** → don't guess at one. It becomes the first thing Step 2 asks for. ### Step 2: Interview **Skip this when the brief already settles the ad** — a clear product, a clear intent, a length. Otherwise ask once, bundled into a single message, always with a free-text way out. | Ask | When | | --- | --- | | **The product** — a link or a photo | Neither was supplied. Offer to wait for an upload; a photographed product beats a described one. | | **How long** | The brief doesn't say. Offer 15s, 30s, 45s or 60s, and let them type their own. | | **Whether a presenter or actor is on camera** | The brief doesn't say whether a person appears in the ad. | | **Whether a voiceover carries it** | The brief doesn't say. The default is yes — a voiceover over the picture; they can pick music and effects only. Any words are always a voiceover, never spoken on camera. | When the user waives questions, take 15s, no actor, a voiceover on, say in one sentence what you took, and keep going.