Creates a new graphics format spec (custom canvas dimensions + content rules) and publishes it to the Gooseworks library so any agent can render against it. Use when the existing community formats (carousel, story, infographic, slides, poster, chart, tweet, plus community additions) don't fit the user's canvas — e.g., LinkedIn banner 1584×396, story cover 1080×1920, event flyer 8.5×11in. Mirrors goose-graphics-create-style but for formats.
npx gooseworks install --all # then, in Claude Code, Cursor, or Codex: /gooseworks use the goose-graphics-create-format skill
Authors a new graphics format and publishes it to the central Gooseworks library
via npx gooseworks formats publish. The output is a working directory with a
gooseworks-format.json manifest plus at least one rendered example PNG paired
with an existing style.
Use when the user wants a new named format with its own content rules — a story cover, a podcast cover, a square testimonial — that maps to one of the seven built-in canvas sizes (carousel, infographic, slides, poster, story, chart, tweet). The new format defines new rules and a new slug; it inherits dimensions from one of those canvases.
Custom canvas dimensions are not supported. If the user wants a
canvas size that isn't in the seven built-ins (e.g. LinkedIn banner
1584×396, US Letter 2550×3300), tell them honestly — that requires a
code change to the goose-graphics skill pack's FORMAT_CONFIGS and is
out of scope for this skill. Suggest the closest existing canvas
instead.
Always check first: run npx gooseworks formats list (or formats search "banner" etc.) to see whether a community-published format already covers the
use case. If one fits, just use it via the regular goose-graphics flow.
goose-graphics skill must be installed in the same workspace —
this skill uses goose-graphics/screenshot/screenshot.js to render
examples and pulls style specs via npx gooseworks styles get <slug>.
Install via:
npx gooseworks install --claude --with goose-graphics--claude for --cursor or --codex as needed.) See the install
page on the hub for the canonical command:
https://skills.gooseworks.ai/skills/goose-graphicsgoose-graphics/screenshot/node_modules/ must exist). If not:
cd <path-to>/goose-graphics/screenshot && npm install && npx playwright install chromiumnpx gooseworks login) for publish
to authenticate.<working-dir>/
gooseworks-format.json # manifest (see "Manifest format" below)
example-1.png # at least one rendered example (REQUIRED)
example-2.png # optional; recommend 2–3 examples across styles
example-3.pngThe backend rejects publish with an empty examples array — at least one
rendered example is mandatory.
A new format is a new slug + new content rules over one of the seven
built-in canvas sizes — not a new canvas. Custom dimensions are not
supported: every renderer in the system (CLI screenshot tool, hub, other
agents) is locked to this allow-list, and creating a format with custom
width/height would produce something nobody can render.
Pick the canvas the new format should use:
| Canvas slug | Dimensions | Best for |
|---|---|---|
poster | 1080×1350 | Vertical hero composition, single panel |
infographic | 1080×var. | Tall vertical scroll, multi-section |
carousel | 1080×1080 | Square single panel / multi-slide |
slides | 1920×1080 | Widescreen / presentation |
story | 1080×1920 | Vertical full-screen (IG/TikTok stories) |
chart | 1080×1080 | Square data viz |
tweet | 1080×1080 | Square testimonial / social card |
Then ask the user:
The new format inherits the canvas's width/height exactly. Set the
manifest's width/height fields to the canvas's dimensions verbatim
(e.g. 1080/1350 for a poster-canvas format). Do not invent new
dimensions — the publish step will reject anything that doesn't match
one of the seven canvases.
If the user wants a canvas that isn't in the list (e.g. LinkedIn banner
1584×396), tell them honestly: that's a code change to the
goose-graphics skill pack's FORMAT_CONFIGS and is out of scope for
this skill. Suggest the closest existing canvas instead.
Suggest 2–3 descriptive names that lead with the platform / surface
(e.g., linkedin-banner, story-cover, podcast-cover, event-flyer-letter).
Slug must be lowercase-kebab-case ([a-z0-9-]+).
Collision check: run npx gooseworks formats get <slug> — if the catalog
returns a hit, the slug is taken. Suggest an alternative.
(The user may omit slug from the manifest; the backend auto-generates one.
On 409 the CLI prompts you to accept the server's suggested slug; pass
--yes at publish time to auto-accept.)
contentRulesMdThis is the load-bearing field — it teaches future agents how to fill the canvas. Cover, in this order:
Aim for at least 50 chars (the minimum) but 300–800 chars is the sweet spot.
Treat contentRulesMd as the spec a fresh agent would read to render this
format correctly without further guidance.
Pair the new format with an existing published style. Recommended: render 2–3 examples across different styles so the catalog tile demonstrates the format's flexibility.
For each example:
npx gooseworks styles list | head and choose 2–3 with
contrasting moods (e.g., matt-gray, neon-dashboard, brutalist).npx gooseworks styles get <slug>.goose-graphics rendering pipeline apply:
<style> block + Google Fonts
<link>.vw/vh/%/rem/em/clamp().html, body { width: <W>px; height: <H>px; overflow: hidden; }.goose-graphics skill pack,
passing the canvas slug (carousel/infographic/slides/poster/story/
chart/tweet — the one chosen in Step 1), not the new format's slug:
node <path-to>/goose-graphics/screenshot/screenshot.js \
--format <canvas-slug> \
--input <path-to-html> \
--output <working-dir>/example-N.png \
--font-delay 1500gooseworks-format.jsonMatch the shape documented in goose-graphics/SKILL.md §17.2. Required:
{
"name": "Story Cover",
"slug": "story-cover",
"description": "1080×1920 vertical cover slide for Instagram and TikTok stories. Single hero panel, 5-word title max, optional brand mark in the lower 10%. Use for product launches and event reminders where the viewer scrolls past in 2 seconds.",
"width": 1080,
"height": 1920,
"contentRulesMd": "## Rules\n\n- Title: 5 words max, large display, centered or top-left\n- One hero element (image, stat, or large icon) in the upper 70% of the canvas\n- Optional brand mark in the lower 10%, ≤8% of canvas height\n…",
"tags": ["story", "vertical", "social", "instagram", "tiktok"],
"examples": [
{ "file": "./example-1.png", "styleSlug": "matt-gray", "caption": "Paired with matt-gray" },
{ "file": "./example-2.png", "styleSlug": "neon-dashboard" }
]
}Constraints to respect:
name: 1–120 charsdescription: 20–1000 chars (required). See guidelines below.width / height: must match one of the seven built-in canvases
exactly (1080×1080, 1080×1350, 1920×1080, 1080×1920, or
1080×variable for infographic). Custom dimensions will render fine
locally if you stretch the rules but will fail for any other agent that
pulls the format later, since their screenshot.js only accepts the
built-in canvas slugs.contentRulesMd: minimum 50 chars (required)tags: 3–10 lowercase strings. See guidelines below.examples: minimum 1 (required); styleSlug is optional but recommended.cd <working-dir>
npx gooseworks formats publishThe CLI reads gooseworks-format.json, validates it client-side, uploads the
manifest plus the example PNGs, and registers the format in the catalog.
Slug-collision UX: if the slug is taken, the CLI prompts
Slug 'X' is taken. Use 'Y' instead? with the server's suggestion. Pass
--yes to auto-accept (useful for scripted runs):
npx gooseworks formats publish --yesSuccess output:
Published format: <slug>
https://skills.gooseworks.ai/formats/<slug>Exit codes: 0 success, 1 transient/auth (network, 401, 5xx), 2 user
error (400 validation, 403 not owner, 413 file too large, declined 409).
Amending a published format later:
npx gooseworks formats update <slug>Same manifest format; no slug-collision retry (the slug is locked).
Tell the user:
/goose-graphics --style <any-slug> --format <new-slug> --brief "..."Same shape as styles — 50–200 words, keyword-dense — but lead with canvas and content density instead of mood.
"1080×1080 square testimonial card for X feed posts."Instagram Stories, LinkedIn,
print event flyer, etc.)."Use for product launches and event reminders where the viewer scrolls past in 2 seconds.").Example:
"1080×1920 vertical canvas for Instagram and TikTok stories. Single-panel composition, 5-word title max, large hero image or stat, optional brand mark in the lower 10% of the canvas. Use for quick announcements and engagement posts where the viewer scrolls past in 2 seconds."
3–10 lowercase tags. Common format tags:
vertical, square, portrait, landscape, banner,
panoramicsingle-panel, multi-slide, multi-panelstory, feed, cover, flyer, presentation, printsocial, ad, editorial, eventlinkedin, instagram, tiktok, youtube, twitterSkip tags that just restate the slug.
examples. Render at least one — recommend 2–3 across contrasting styles.contentRulesMd. The format describes
canvas + content density; the style describes palette + typography. Keep them
separate so any style can pair with any format.examples[].file are relative to the working directory and live
on the user's machine, not in this repo.npx gooseworks formats list first — if
the format already exists, point the user at it instead of publishing a
duplicate.goose-graphics skill must be installed in the same workspace —npx gooseworks login) for publishAssemble an expert/educator motion-graphic LISTICLE video ad from a config — a spoken authoritative voiceover carries a numbered listicle while N web-animated hyperframe beats (HTML plus the Web Animations API, one branded design system of alternating tiles, big hero numerals, and glass-pill callouts) are rendered frame-by-frame via Playwright and anchored to the VO's word-level timestamps, periodic color-graded B-roll windows give visual breath, and captions burn ONLY inside those B-roll windows (2-word chunks, ASS Format header carrying a Name field so none drop) with the VO mixed under a low music bed. This is the FREE deterministic assembly stage (Playwright beat render plus ffmpeg concat plus window-masked caption burn plus VO-and-music mix plus final composite) — the VO, the music bed, and the stock B-roll come from create-vo-elevenlabs, create-music-elevenlabs, and media-proxy. Use for the vo-anchored-motion-listicle format.
Assemble a stop-motion hand-swatch-cycle product-demo ad from a config — a sequence of still PLATES (one hand swiping a single-barrel cosmetic across a cream skin-patch, the barrel + swatch changing per plate while the hand, background, crop, and lighting stay locked) is PNG→mp4 loop-encoded at each plate's own stop-motion hold (fast motion frames 150–250ms, per-shade ~380ms, hero beats 1100–1800ms), concat-demuxed with HARD cuts into a silent master, closed on a Playwright HTML-rendered branded end card (serif tagline + sans subtitle + real logo SVG over a hero BG, never AI-rendered text), and muxed with a pre-sourced music track playing under the end card with a fade tail (no VO). This is the FREE deterministic assembly stage (loop-encode + concat-demux + end-card render + music mux); the master-anchor plate, shade plates, and end-card BG come from create-image-gpt-image-fal and the track from create-music-elevenlabs. Use for the stopmotion-hand-swatch-cycle format.
Assemble a split-screen creator ad from a config — a two-zone vertical composite where a supplied AI-creator lip-sync take fills the BOTTOM ~48% while real 16:9 product/demo clips run uncropped in the TOP ~52%, each top clip contain-fit with a darkened blurred cover-scale fill of the same clip (never black bars), a 3px brand-color divider between the zones, the creator slice cover-fit per the per-scene VO timing, scenes hard-concatenated with the body audio being the concatenated creator VO slices, an end card held on the last sharp frame ~3s, then the ASSEMBLED cut transcribed with local Whisper (not the raw VO — concat drops inter-scene silence) and word-level captions burned in the chosen style. This is the FREE deterministic assembly + caption stage (two-zone composite + blurred fill + divider + hard-concat + end card + captions); the VO comes from create-vo-elevenlabs, the anchor from create-image-gpt-image-fal, and the whole-VO lip-sync from a paid VEED Fabric 1.0 take (a no-atom upstream input). Use for the split-screen-creator format.