Skip to content
SkillModel-led1,536 tokens in use

Better Colors

Builds and audits color systems: palette ramps in OKLCH, semantic token names, format conversion and measured contrast on the pair that renders on screen.

12,497 stars on the repository MIT Licensejakubkrehel/skills · skills/better-colors
npx skills add jakubkrehel/skills
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Better Colors costs

26tokens idle
1,536tokens in use
9,110tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. 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, its code blocks and the commands it names in inline code. 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
  • Frontmatter8/15
  • Structure14/25
  • Examples0/15
  • Output10/15
  • Size15/15
  • Resources13/15

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 activity96
  • Maintenance52
  • Community48
  • Maturity24
  • Momentum40

Better Colors is an agent skill for building and auditing color systems. It generates ramps with a color library, maps each step to a role, names primitive and semantic tokens, converts between notations and measures contrast on the rendered foreground and background pair.

It reuses the project’s own tokens and notation, and says what it measured. A contrast value is never reported unless it was computed, and a failing pair is reported with its measured value and the threshold it misses.

What You Get

  • Palette ramps that hold one hue end to end, peak in vividness mid-ramp and step more finely at the light end, built in the project’s notation
  • A two-tier token scheme with primitives such as blue-600 and semantic tokens such as color-text-secondary
  • Contrast measured on the background a text renders on, in light and dark, with the pair, the value and the threshold it misses
  • Rules for one color one meaning, one filled primary action per view and gradient interpolation space
  • A before-you-finish table of code patterns, such as raw color literals in components or text-white on a 500 fill, each with its fix

Common Use Cases

  • Generating a brand accent ramp with matching dark-mode values and checking every pair
  • Renaming a tangle of color variables into primitive and semantic tokens
  • Converting a hex palette to oklch, or the reverse, without changing the colors
  • Auditing a design for status colors that sit too close to the accent hue

What Makes It Unique

It separates findings from project choices. A broken role mapping, a failing pair and an out-of-gamut step are findings, while notation, a tinted neutral and a gradient’s interpolation space are decisions it leaves alone. It also leaves failing colors unchanged until asked, because the fix is a design decision, and it hands the question of which pairs must pass to the accessibility skill.

Under The Hood

The skill is a SKILL.md with reference files for color formats, color usage, contrast, palette generation, palette structure, token naming and reporting. It ships no scripts. The agent reads your token files and stylesheets, may run a color library from your project to compute ramps and contrast, and in a browser reads the background rendered behind text in both themes. It reports by default and edits files only when you ask for changes.

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