Skip to content
Back to skills

living-canvas

ASecurity

Turn anything (GitHub repo, URL, doc, concept, topic, one sentence) into a beautiful animated canvas: README-ready animated SVG and/or interactive canvas page.

  • 4 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 7, 2026
ai-agentsgonodegit

Works with

  • cursor
  • terminal
  • cli

Security analysis

A100/100

Pro scans all 5 files and shows the line behind each finding

Scanned November 7, 2026

npx -y skills add shouvik12/living-canvas --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of living-canvas?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for living-canvas
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/shouvik12-living-canvas/badge)](https://www.skillsdirectory.com/skills/shouvik12-living-canvas)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: living-canvas
description: "Turn anything (GitHub repo, URL, doc, concept, topic, one sentence) into a beautiful animated canvas: README-ready animated SVG and/or interactive canvas page."
---

# Living Canvas

Input can be anything under the sun. Output is a moving piece of art that captures its essence, and that a stranger understands at a glance.

## 1. Understand the subject
- Repo/URL: read README, file tree, languages, tagline. Doc: skim headings. Sentence/topic: use it directly.
- Distill into: ONE core transformation or idea, a mood, 3-5 key visual elements, and a tagline (max 8 words).
- Find the subject's own concrete example (a before/after, a real input and output, a number). Use it instead of inventing an abstract one.

## 2. Clarity first: the still-frame test
- A viewer who has never seen the subject must get the idea from ONE still frame, with no README and no explanation.
- Tell it as a short story, left to right: problem → what the subject does → result. Two to three labeled panels or stages work well.
- Label every stage in plain words (e.g. "1 · WITHOUT X", "2 · X ASKS", "3 · WITH X"), and put a one-line caption under each.
- Add a plain-language subtitle under the title that says what the subject does.
- Show real-looking content (readable code, actual text, actual numbers), not abstract bars or blobs.
- Never use an unlabeled symbol (ladder, funnel, rings) that only makes sense if you already know the project.

## 3. Pick a visual metaphor that fits
- Converters, compressors, pipelines: things flowing into a glowing core and emerging clean.
- AI, data, networks: particles, constellations, orbiting nodes.
- Games: generative terrain, parallax, motion toward the horizon.
- Nature, emotion, abstract topics: weather, light, water, ink, organic noise.
- Tools/CLIs: typing terminal lines that resolve into a result.
Never default to generic gradient blobs. The metaphor must say something true about the subject, and it serves the story from step 2. It never replaces the labels.

## 4. Choose output format
- GitHub README: animated SVG only (SMIL + CSS keyframes, no JS, no external fonts/images, viewBox ~900x300, under 100KB).
- Otherwise: one self-contained HTML page with <canvas> + requestAnimationFrame, fullscreen, responsive, one rewarding interaction (cursor/click). Publish as an artifact when available.
- If unclear, produce both.

## 5. Craft rules
- Dark rich background, 2-3 accent colors from the subject.
- Motion follows the story: highlight each stage in turn (dim the others) rather than adding decoration. Effects (particles, bursts, beams) only where they make a step clearer.
- Seamless loop: shared cycle durations, staggered begin offsets.
- Layer depth: ambience, midground motion, foreground hero + title.
- Smooth, readable motion; no harsh flicker. Respect prefers-reduced-motion, and in that mode show every stage fully visible so the still frame tells the whole story.
- Title in clean type, tagline in small letter-spaced caps.
- No copyrighted characters, others' logos, or real people.

## 6. Check before delivering
- Ask: "If someone saw only a screenshot, could they say what this project does?" If not, add labels or captions and simplify before sending.

## 7. Deliver
Send the file(s). For SVG, give the README embed line. Offer one tweak axis (palette, speed, metaphor).

Files in this skill

  • SKILL.md3.3 KB
  • assets/banner.svg4.3 KB
  • examples/devtv-canvas.svg6.8 KB
  • examples/trooper-canvas.svg6.8 KB
  • living-canvas-banner.svg5.4 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.