agentleFS
0xrohan100xrohan10/skills/product-design-evolve/SKILL.md

product-design-evolve

Skill2 starsChanged 4 months agoMITView on GitHub

Evolve an existing product's visual language into a more distinctive design system while retaining its recognizable colours, fonts, and product identity. Use when the user wants a design-system or distinctive redesign of an existing UI; not for greenfield branding.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Keep the invariants. Rewrite the elastic.

    line 8
  2. Distinctive:

    Commit to one thesis:

    line 28
  3. Distinctive:

    Leave a running development server running.

    line 65
  4. Distinctive:

    Use the repository's existing checks and verify the actual interface in a browser or appropriate app runtime.

    line 67

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Evolve

Keep the invariants. Rewrite the elastic. Leave debt behind.

Classify the DNA

Inspect the running product and its implementation. Capture screenshots at representative desktop and mobile viewports, then trace each visible choice to tokens, styles, assets, and components.

Classify every visible choice:

  • Invariant — primary brand colours, typefaces, logos, product meaning, accessibility, established interaction expectations. Preserve these unless the user changes the brief.
  • Elastic — colour roles and proportions, type scale, spacing rhythm, grid, density, shape, borders, elevation, imagery, motion, composition. This is the creative material.
  • Debt — accidental inconsistency, duplicated values, one-off components, inaccessible combinations, broken responsiveness, styling with no product rationale.

A token name is not evidence that a value is visibly important. Compare implementation with the rendered product.

DNA is classified when every visible choice on the representative viewports is invariant, elastic, or debt, each with evidence from both the render and its source.

Write the thesis

Derive direction from the product's purpose, content, and strongest existing cues. When the brief leaves room, explore a few interpretations that change composition or interaction character, not decoration.

Commit to one thesis:

  • the feeling the interface should create
  • which invariants anchor it to the product
  • the one or two deliberate tensions that make it memorable
  • what the system will refuse

Whimsy comes from a repeatable idea — unusual rhythm, expressive type scale, tactile interaction, illustration, controlled asymmetry.

The thesis is set when one is chosen and its anchors and tensions are named. If alternatives were explored, they are recorded.

Build a vertical slice

Implement the smallest representative surface that exercises the system under real content and interaction. Prefer an important existing screen that contains typography, navigation, controls, content, and responsive behavior.

From that slice, establish the project's natural source of truth for:

  • semantic colour roles and accessible states
  • typography roles using the existing font families
  • spacing, grid, breakpoints, radii, borders, and elevation
  • core controls and content primitives
  • iconography, imagery, motion, and reduced-motion behavior where they serve the thesis

Reuse the project's architecture and component conventions. Introduce a token or primitive only when it represents a recurring decision. New work lands in production UI. No parallel component library, speculative variants, or disconnected showcase.

The slice is done when every system decision it exercises has a source-of-truth location, and the slice uses those locations.

Subtract

Review the rendered slice for slop and remove anything without a named job: extra cards, pills, ornamental copy, gradients, glows, uniform rounding, fake metrics, needless icons, motion on every element. Prefer hierarchy, spacing, proportion, imagery, and a few intentional moments.

Native or established controls win when a custom treatment does not improve meaning or experience.

Subtraction is done when every remaining flourish has a named job.

Verify

Leave a running development server running. If a restart is required, ask the user.

Use the repository's existing checks and verify the actual interface in a browser or appropriate app runtime.

Deliver only when:

  • Invariants — main colours and fonts remain recognizable; the result reads as an evolution of the same product
  • Thesis — tokens, components, layout, imagery, and motion express one thesis
  • Slice — the representative surface works with realistic content, long and short text, empty or error states it exposes, and existing interactions
  • Viewports — chosen mobile and desktop compositions are deliberate, with no overflow, clipping, or accidental collapse
  • Access — contrast, keyboard use, focus, semantics, targets, and reduced motion remain sound
  • Subtraction — every prominent flourish supports hierarchy, meaning, brand, or delight
  • Diff — existing behavior and unrelated screens remain intact; every changed line belongs to the system or its slice

Deliver the implemented system, the thesis, the preserved invariants, the source-of-truth locations, and visual proof from the representative viewports.

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.