agentleFS
whanoftheworldwhanoftheworld/graphic-skills/skills/color-system/SKILL.md

color-system

Skill0 starsChanged 34 days agoMITView on GitHub

Choose the brand palette from meaning, category, and contrast, build the tonal scales, and prove every text pair passes.

7 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Files
SKILL.md

Color system

Color is the fastest signal a brand sends and the easiest to get wrong without numbers. This skill chooses the palette with reasons, builds scales, and checks each pairing by script.

Load the brand-context, design-craft, and design-grilling skills (the Skill tool, or read their SKILL.md), and read color.md. Reply in the user's language. The brand memory needs a decided Direction (a chosen route, or "existing identity" kept at intake); without one, tell the human to run /design-routes.

1. Gather the evidence

From the brand memory: personality, the chosen route's palette sketch, color items in Taste evidence (read each image for hue, saturation, and temperature), constraints from the intake (colors to keep or avoid), and the logo if one is decided. Research the competitors' dominant colors and write the category color map: the colors to avoid blending into and the open space.

2. Choose the roles

Propose, with a one-line reason each, the brand primary, secondary, accent, and the neutral family tint. Offer two alternatives for the primary that differ on one axis (hue, or saturation) so the human can compare by eye: render brand/explore/palette.html with each option on a sample post, a button row, and body text, and look at it rendered. Ask which feels like the brand.

3. Build and prove

For each chosen color run design-craft's palette.mjs "#hex" --json <name>, and put the scales into brand/tokens.json (see TOKENS.md in brand-context: create the file from the template when missing, merge when present), with the exact brand hex pinned to its nearest step. Add semantic colors (success, warning, danger, info) tuned to sit beside the brand colors, and a surface and text family from the neutrals.

Declare every text-on-background pair the brand will use (body on base surface, on raised surface, on brand primary, accent on dark, each button state) under $extensions["graphic-skills"].pairs, with need set to text, large, or ui. Run design-craft's contrast.mjs --tokens brand/tokens.json. Adjust steps until every pair passes, and record the final ratios in the Color table.

4. Extend

  • Dark mode for screen brands: derive, do not invert (see color.md), and re-run the pairs.
  • Print: give the nearest CMYK or Pantone for each brand color, marked approximate and needing a proof, and define the one-color version.
  • Grayscale test: view the palette desaturated and confirm hierarchy survives without hue.

The skill is done when every role has a hex and a scale in the tokens, every declared pair passes its required ratio (or has a named exception the human accepted), the human chose from the rendered palette page, and the Color section and Decision log are written, including the alternatives rejected. Set Color to decided.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.