SuperAPI AI Directory

static-social-campaign-assets

Design & creation

View repository

Updated 10-15 17:20

About this skill

Static Social Campaign Assets

Use this skill when the user provides a prompt, content calendar, or asset pack and expects actual designed social posts, not strategy text. Common outputs: Instagram portrait PNGs, LinkedIn carousel slides, static campaign image sets, contact sheets, and export folders.

Core rule

Do not summarize the prompt as the deliverable. Execute it: fetch assets, build/render the posts, inspect them visually, fix issues, then return the folder and verification.

Workflow

  1. Fetch the brief and public assets.

    • Download the exact listed assets: logos, images, reference outputs, manifests.
    • If the brief has an asset manifest, inspect it before assuming the listed assets are exhaustive.
    • Preserve exact copy, spelling, capitalization, CTA text, and filenames from the prompt.
  2. Inspect the reference visually before rendering.

    • Record canvas size, background colour, safe margins, logo position/scale, headline hierarchy, body colour/size, pill treatment, image crop/card radius, CTA shape/position, and footer alignment.
    • Treat the reference as the campaign system unless the user asks for new art direction.
  1. Build a deterministic renderer.

    • Python/Pillow is a good default for exact PNG deliverables.
    • Store post copy in structured data; do not hand-edit text into multiple files.
    • Keep layout variants restrained and campaign-consistent.
    • Use only approved assets; no stock, invented images, fake illustrations, or decorative backgrounds unless the prompt explicitly allows them.
  2. Export the assets.

    • Generate every requested file at exact dimensions.
    • Generate a contact sheet for QA.
    • Print or verify filename, pixel dimensions, and count with a script.
  3. Visually QA and fix.

    • Inspect the contact sheet for repeated layout failures.
    • Spot-check full-size edge cases: split layouts, long headlines, overlaid pills, logos, footer, and tight body copy.
    • Fix text clipping, overlap, unsafe margins, unreadable tiny type, bad crops, visible logo boxes, weak spacing, or inconsistent brand feel before reporting done.

Pitfalls

  • A rasterized SVG logo can silently include a visible white square/box. Always inspect the rendered logo on the final background.
  • If the provided reference output is an allowed asset, cropping a logo lockup from that approved reference can be a faithful fallback when direct SVG rasterization is problematic. Keep the crop on the same background colour so it does not introduce a visible box.
  • Pill layout helpers must wrap relative to their current row start, not a hard-coded global margin. Otherwise split layouts can redraw pills over an image or outside the intended text column.
  • Contact-sheet QA alone can hide small issues. Full-size spot checks are required for any layout with overlaid chips, narrow columns, or long copy.
  • Do not claim the set is complete from file generation alone. Completion requires dimensions verification and visual QA.

Using design-style skills

When the user wants to use the newly installed TypeUI/awesome-design-skills library, treat those skills as style systems, not standalone asset generators. Pair a style skill such as premium, refined, editorial, bento, agentic, sleek, or storytelling with this execution workflow to produce real PNG/contact-sheet outputs. See references/typeui-design-style-skills.md for the explanation pattern, recommended pairings, and quick demonstration workflow.

Final response format

Keep handoff short:

  • folder path
  • list of generated files
  • verification summary: count, dimensions, assets used, QA result, and any fixes made