agentleFS
glifxyzglifxyz/glif-mcp-server/skills/app-promo-materials/SKILL.md

app-promo-materials

Skill213 starsChanged 39 days agoMITView on GitHub

Plan and make launch, store, ad and campaign graphics with Glif: real app screens turned into store listings and launch posts, or a coordinated set of product photos, ads, a short video and a jingle for any product. Use when: "App Store screenshots", "Play Store feature graphic", "launch graphics", "Product Hunt images", "social posts for the launch", "carousel", "promo banner", "static ad", "Meta ads", "ad creatives", "ad concepts", "angles to test", "UGC-style ads", "turn these reviews into ads", "product photos for my shop", "Amazon or Shopify listing images", "lifestyle shots of this product", "social media calendar", "a week of posts", "launch campaign", "poster series", "mood board", "storyboard this ad", "resize this ad for every format", "headline variations", "YouTube thumbnail", "X header", "email header". NOT for: logos and icons (use app-icons-and-logos), images inside the site or OG images (use web-visuals), demo videos (use app-demo-videos), one-off edits (use edit-images).

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use the user's screenshots and product photos if they have them.

    line 45
  2. Distinctive:

    Use demo data. Never capture real customer data, emails or keys.

    line 47
  3. Distinctive:

    If there's no settled style yet, ask Glif for 2 or 3 layout directions on one hero graphic, such as the first App Store screenshot or the …

    line 61

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Turn a product into polished promo graphics. For an app, that means real screens. For anything else, it means the user's product photos and references. The user's explicit instructions come first, within host safety rules.

1. Pick the deliverables

Pin down the subject, audience, mood, formats and any references. Ask only about gaps that would change the result. Then pick deliverables from these menus, or the user's own list.

Store and launch sizes:

DeliverableSize (px)
iPhone App Store screenshots1320×2868 or 1290×2796, portrait
iPad App Store screenshots2064×2752, portrait
Google Play feature graphic1024×500
Product Hunt gallery1270×760
YouTube thumbnail1280×720
X post1600×900
LinkedIn post1200×627
Square post1080×1080
Feed post or carousel card (4:5)1080×1350, 3 to 6 cards
Story or Reel cover (9:16)1080×1920
X header1500×500

Store rules change. Check App Store Connect and Play Console for the sizes they currently require. App Store screenshots must be PNG or JPEG with no transparency.

Campaign sets:

  • Ad creative: UGC-style creator video (9:16, script, captions), product ad video, carousel (3 to 6 cards), static banners, unboxing.
  • Product photos: studio hero on white, flat-lay, lifestyle scene, multi-angle set, detail close-up, what's in the box. Ask how many (2, 4 or 8) and keep the product identical to the user's photo.
  • Social calendar: a week or a month of posts with captions and hashtags.
  • Series: a repeatable format, such as a daily 40-second short. Save the recipe so each episode continues the same project.

Make the winner first, then resize it for each placement without cropping.

2. Gather real material

Promo graphics must show the real product. Don't invent screens, features, reviews, awards, press logos or partner logos. Turning real reviews into ads is fine when the user supplies them.

  • Use the user's screenshots and product photos if they have them.
  • For an app without screenshots, run it and capture screens with Playwright at the device size. For a 1320×2868 iPhone shot, use a 440×956 viewport with deviceScaleFactor: 3.
  • Use demo data. Never capture real customer data, emails or keys.

Upload each screen, product photo and the logo with upload_file (with shell access, pass filename and run the returned command), only when the user agrees to send them to Glif. In the brief, say: "Place the uploaded logo and screenshots exactly. Don't redraw them."

3. Write the copy and angles

Draft the angles in text first; that's free. Write one short headline per graphic, 2 to 6 words, about a benefit the screen or product really shows. For ads, write 3 to 5 headline options so the user can test them. Offer about 4 variants per piece, not dozens, unless the user asks. For video, offer a storyboard of stills before animating.

Show the copy before generating. Check names and claims against the code, README or product page. Thumbnails need one honest hook that matches the video, readable at about 120 px wide.

4. Settle the look, then make the set

A request for these graphics is approval to spend credits. If the user only asked to plan or preview, show the draft and wait. Do not ask again after a clear request.

  1. If there's no settled style yet, ask Glif for 2 or 3 layout directions on one hero graphic, such as the first App Store screenshot or the lead ad. The user picks.
  2. Continue the same project (project_id) and make the full set in that style, in one compose_project call. For each graphic, give:
    • filename, exact pixel size and the attached screen or photo it uses
    • the exact headline text, in quotes
    • layout: device frame or floating screen, headline position, background
    • shared style: brand colors (hex), fonts or mood

For a video or audio item in the set, such as a UGC clip or a jingle, write its brief in the same call the way the audio-and-video skill does: shots with camera and action, the exact script in quotes, or music mood, tempo and length. Download those results with the same curl step and check them with ffprobe.

Describe creative goals; Glif picks the models. Leave intelligence unset unless the user asks for lite, smart or genius. Ask Glif to report each output's size.

Useful follow-ups in the same project:

  • Resize: "Recompose this finished graphic for every size in the list. Don't crop the product or the text."
  • Headline variants: the same visual with each headline option.
  • Animated version: a short MP4 loop of a social card.
  • Reference layout: the user supplies an ad they like. Copy its structure only, never another brand's assets, logos or footage.

Share the projectUrl. Follow each run with get_job_status, passing the jobId as job_id. Show new media as it arrives without repeats. A pending status is not a reason to run again, and never retry a failure on your own.

5. Check and save

Download into a folder outside the shipped site, such as marketing/, with curl -fsSL "<media url>" -o <path>. Use one folder per channel, like marketing/app-store/01-home.png.

Check every file:

  • exact pixel size (file, sips -g pixelWidth -g pixelHeight or magick identify)
  • spelling of every headline, letter by letter
  • the screen or product inside matches the real thing, and the logo is unchanged
  • App Store files have no alpha channel. Flatten them locally if needed.

If a headline is misspelled or a screen got redrawn, ask Glif to fix that one file in the same project. Or set the text locally: build a small HTML page with the Glif background, the real screenshot and the headline, then screenshot it with Playwright at the exact size.

List what you made and where it is. Uploading to stores or posting is the user's call.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.