Skip to content
SkillModel-led2,271 tokens in use

canvas-design

Anthropic's skill for posters and static art: Claude writes a design philosophy as an art movement, then expresses it as one crafted PDF or PNG canvas.

197,763 stars on the repositoryanthropics/skills · skills/canvas-design
/plugin install example-skills@anthropic-agent-skills
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing canvas-design costs

59tokens idle
2,271tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts and code blocks. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter15/15
  • Structure21/25
  • Examples3/15
  • Output15/15
  • Size15/15
  • Resources10/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity72
  • Maintenance32
  • Community68
  • Maturity36
  • Momentum40

canvas-design is Anthropic’s agent skill for making posters, artwork and other static visual pieces. It splits the work in two: first Claude invents a visual philosophy, a named aesthetic movement such as “Brutalist Joy” or “Chromatic Silence” written as a four-to-six paragraph manifesto about space, form, color, scale and hierarchy, and saves it as a Markdown file. Then it expresses that philosophy on a canvas as a single-page PDF or PNG that is about 90% visual design and 10% essential text.

The skill asks Claude to weave a subtle reference to the user’s request into the piece rather than illustrate it literally, to keep text sparse and inside the canvas margins with nothing overlapping, and to choose type from a bundled folder of fonts. A final pass refines what is already there instead of adding more, and further pages can be made on request as variations on the same philosophy.

What You Get

  • A two-step method: a named design philosophy in Markdown first, then a canvas that expresses it
  • Five condensed example philosophies, from “Concrete Poetry” to “Geometric Silence”, showing the expected tone and depth
  • Rules for text as a visual element: minimal, usually thin, contained within the margins and never overlapping
  • A folder of TrueType fonts from 29 families, most shipped with their Open Font License text, to pick type from
  • A refinement pass that tightens the existing composition instead of adding new graphics
  • A multi-page option that bundles variations on the same philosophy into one PDF or several PNGs

Common Use Cases

  • Designing a poster for an event, venue or release as a finished PDF or PNG
  • Making an original abstract art piece that carries a subtle reference to a chosen subject
  • Producing a cover or key visual for a book, game or film project that should read as art rather than a cartoon
  • Creating a short multi-page series of related images in one visual language, like spreads in a coffee table book

What Makes It Unique

Where frontend-design shapes web interfaces in code and the pdf skill manipulates existing documents, canvas-design produces a standalone art object. Its distinctive move is writing an aesthetic philosophy before drawing anything, then having Claude treat that manifesto as the brief, with the user’s topic folded in as a hidden reference that only those who know the subject will notice. It leans on a visual language of systematic observation, with repeated marks, layered patterns and sparse clinical labels, and ships its own font library instead of relying on system fonts.

Under The Hood

The skill is a single SKILL.md of instructions with no scripts or reference documents. Its folder also holds the license text and a canvas-fonts directory of TrueType font files, most with an Open Font License text beside them, which Claude is told to search when choosing type. When it loads, Claude writes the design philosophy to a .md file, then writes and runs its own drawing code to render the canvas and saves it as a .pdf or .png file, going back to that code for the refinement pass. The instructions also tell Claude to download any extra fonts it needs, so it may fetch font files from the internet. It calls no other services.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers