canvas-design

Create single-page posters and static visual art as PNG or PDF with a documented design philosophy.

6|1|Updated Sep 3, 2026
One-click install
npx skills add https://github.com/jasonzhu0922-sketch/agentloop --skill canvas-design-jasonzhu0922-sketch
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: canvas-design
Source: https://github.com/jasonzhu0922-sketch/agentloop/tree/main/packages/agentloop-skills/skills/canvas-design
Command: npx skills add https://github.com/jasonzhu0922-sketch/agentloop --skill canvas-design-jasonzhu0922-sketch

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and assets (resource) components.

What problem does it solve? Producing original poster artwork requires both a coherent design concept and a working render pipeline; this Skill enforces a disciplined workflow that turns a brief into a written design philosophy plus a finished single-page PNG or PDF, avoiding generic genre-default visuals. ## Core Features & Use Cases - Structured art direction: Forms three candidate directions across seven axes (emotional register, material language, composition topology, typographic voice, color strategy, image mode, central metaphor) and records the selection as an artDirection contract. - Human-in-the-loop gating: Uses request_human_loop to resolve ambiguous visual direction before writing the philosophy, spec, or artwork when the brief lacks explicit style evidence. - Packaged renderer with schema contract: Runs scripts/render_static_canvas.py to load the current schema and render JSON specs with layout families, composition variants, visual motifs, texture, density, and CJK font glyph checks. - Use Case: Given a brief for a city technology event poster, compare three distinct directions, confirm one with the user, write the philosophy, render a 1800x2700 PNG, and inspect the thumbnail silhouette against the chosen topology. ## Quick Start Use the canvas-design skill to create a poster for my product launch, starting by loading the renderer schema and proposing three distinct visual directions for my approval.

Frequently Asked Questions about canvas-design

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a poster with the canvas-design skill?

Load the renderer schema first, form three candidate directions differing on at least four design axes, select one (using human-in-the-loop if the brief is ambiguous), write the design philosophy, then render a JSON spec through scripts/render_static_canvas.py to produce a PNG or PDF.

What output formats does the canvas poster renderer support?

The renderer produces single-page PNG or PDF files. The output path is set in the JSON spec's output field, and canvas dimensions are configurable through the canvas width and height properties.

When is human-in-the-loop approval required before rendering?

HIL is required when the brief does not specify visual form, metaphor, material, or composition and two or more candidate directions remain viable. It is skipped only when the user describes the visual direction, supplies constraining brand assets, or explicitly delegates the choice.

Does the renderer support Chinese or CJK text in posters?

Yes, CJK text can be used directly in specs. The packaged renderer performs font selection and glyph smoke checks, and bundled OFL fonts are included; a custom renderer must fail clearly if compatible glyphs are unavailable.

What happens if my design concept does not fit the packaged renderer?

If the concept requires photography, detailed illustration, hand lettering, or motifs the schema cannot represent, use a compact custom renderer instead of forcing the direction through the packaged grammar.