Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →gooseworks-ai/render-mosaic-grid-reveal
Render a 'mosaic-grid-reveal' video from a config — a real-DOM FULL-BLEED N×N mosaic of real product tiles that pops in one tile at a time (scatter order, ease-out-back overshoot), the grid clears, then the brand wordmark builds line-by-line followed by a sub-label, tagline, and CTA; frame-stepped via Playwright and encoded with FFmpeg — deterministic assembly, FREE (the music bed comes from create-music-elevenlabs), so the wordmark, tile captions, and CTA stay pixel-crisp.
Passed 17 of 23 test cases with the skill, and 6 of 23 without it.
and 15 more in the full report
12 fixed6 passed either way7 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
This benchmark measured an earlier version of the skill.
One of 12 test cases the skill fixed. Same model, same prompt, with and without the skill.
What exact key names are used in the end_card object inside config.json to define the copy hierarchy for the final screen?
| Checked | Without skill | With skill |
|---|---|---|
| Uses sub, tagline_top, tagline_bottom, cta, and url key names inside end_card. | Missed | Passed |
The instructions your model reads when the skill loads.
Render the mosaic-grid-reveal format from a config. The signature of this format is a
full-bleed N×N mosaic (default 3×3) of the brand's real product stills that pops in
one tile at a time (scatter order, ease-out-back "pop"), building the whole grid; the grid
then clears (scale-up + fade) and a clean end card builds line-by-line — the brand's
real wordmark, then a small sub-label, then the tagline, then a CTA. Silent by design (a
light music bed is muxed separately via create-music-elevenlabs). The hook is RANGE — nine
maximally-distinct variants show how much choice the brand offers.
Everything is a real-DOM HTML scene frame-stepped to PNG via Playwright and encoded with FFmpeg, so the wordmark, tile captions, and CTA stay pixel-crisp (a video model would smear them). No generative video, no i2v, no AI-rendered text. This capability is FREE and deterministic — the only paid step of the format (the music bed) is a separate media cap.
config.json)width, height, fps — canvas + frame rate (default 1080×1920, 30).wordmark — brand logo SVG (full <svg> or bare <path> markup) or PNG path. Real DOM,
recolored via palette.ink. Never AI-render it. wordmark_viewbox, wordmark_width tune it.palette — bg (page colour; the demo used a warm off-white), accent + accent_deep (brand color), ink, cap.grid — cols/rows (default 3×3), inset (outer margin), gap.tiles — one per cell (length must equal cols*rows): image (real variant still),
name (caption), bg (tile colour, from choices.tile_look — the demo used a soft pastel echoing each variant), ink (deep caption color).pop_order — scatter order across cells (default corners → center → edges).timing — grid_in0, cadence, pop, hold, clear (seconds).end_card — sub, tagline_top, tagline_bottom (bold), cta, url. Brand's own approved copy only.eyebrow — optional top kicker held during the grid.scripts/config.example.json is a filled example (the Pair Eyewear "RANGE" worked example);
the tile image/wordmark paths are brand inputs bound by the orchestrator at remix time.
Its tiles, tile colours, end-card copy and music prompt are Pair's answers, not defaults.
The recipe asks the user these before any paid step; the renderer draws whatever the config says.
range — which 9 products/variants fill the grid → tiles[].image, tiles[].name. Asked of the user; the demo used 9 eyewear colourways.tile_look — tile backgrounds → tiles[].bg, tiles[].ink, palette.bg. Asked of the user; the demo used a soft pastel per variant.music — the bed's genre/mood → music.prompt (consumed by create-music-elevenlabs). Asked of the user; the demo used light upbeat marimba.test cases passed, vs 6 / 23 without
Published skills whose descriptions closely match this one, with their headline benchmark lift.
More skills from the same source and nearby in the registry, with their headline benchmark lift.