Skip to content
← All AI workflows

Brand System: check brand colors against production tokens

Use Shahar Shavit's Brand System skill to draft brand documents and matching tokens. Compare documented colors with production CSS and review contrast warnings. These checks do not cover all brand rules or establish accessibility conformance.

From the source

Public repository page from Shahar Shavit
Shahar Shavit · Captured 2026-10-14Read the original

How it works

  • The original skill links brand documents, CSS and token JSON with color-consistency and contrast scripts. The scripts are reusable source, not an audit run on a project.
  • The source inspection found that the contrast script blocks failed required body-text pairs but can warn on other failures. A passing script does not establish WCAG conformance.
  • Color consistency does not prove typography, motion, layout or component behavior follows the brand book. Rendered and interactive states need manual review.
  • The recorded repository license is MIT. Model, rendering and design-review cost are unverified. Skill execution permits file and shell actions and needs restricted project access.

Workflow guide

  1. Record authoritative, approved brand choices. Inspect the original skill, templates and scripts before granting project access.

  2. Draft the brand book, printable version, CSS tokens and token JSON in an authorized isolated working branch. Exclude secrets.

  3. Run the color consistency check against selected production CSS files after execution is approved.

  4. Run the contrast audit for each fixed theme and intended foreground/background pair. Record required failures and warning-only results separately.

  5. Review unexplained mismatches, rendered components, typography, motion, layout and interactive states with a designer.

  6. Record manual defects and review time. Obtain separate approval for code merges and public brand publication.

Sources and downloads

Discuss this workflow →

Book a call

A free 30 minute call. Tell us where you are with AI, what your IT allows, and what management needs to see. We'll tell you honestly what we would do first.