agentsclimarketplace

Design fidelity

Skill sukirman1901/frontend-expert/skills/design-fidelity

Spec vs implementation fidelity — Figma/Dev Mode/screenshot compare, spacing tolerance, token remap, close-enough rules. Use when matching a design file, "pixel perfect", "mirip Figma", design QA, or /design with a provided mock.From its SKILL.md

Install
npx -y skills add sukirman1901/frontend-expert --skill design-fidelity

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 10 stars10 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.7 KB, as published. Nobody here has run it

Design Fidelity

Overview

When a design source exists (Figma, screenshot, Spec), compare build to source and close gaps — without inventing a parallel visual system. Prefer project tokens over copying raw hex from the file.

When to use

  • User provides Figma / mock / screenshot and wants implementation match
  • Design QA / “pixel perfect” / “mirip desain”
  • /design audit when a reference image or file is available
  • Remapping Dev Mode values onto the project token scale

When to skip

  • No design source (blank-canvas → frontend-judgment only)
  • Pure copy/token tweak with no layout compare
  • User waived fidelity (“approximate is fine”)

Workflow

  1. Source — note file/frame/screenshot; capture key frames at target breakpoints
  2. Structure first — layout model, hierarchy, type ladder, Auto Layout Fill/Hug (responsive-ui) before pixel nits
  3. Token remap — map Figma fills/strokes/radii/type to project tokens (or scored preset). Do not paste foreign hex into production CSS unless the project has no system and the tree allows it
  4. Compare — spacing, type size/weight, radius, alignment, states (hover/focus if specified)
  5. Tolerance — see checklist; flag only meaningful deltas
  6. Report — Before/After or Spec vs Build table; mark remaining intentional differences

Close-enough rules (do not bikeshed)

PropertyUsually OKFix
Spacing±4px on dense chrome>8px section rhythm drift
Type sizeNearest token stepWrong weight / wrong role (title vs body)
RadiusNearest tokenMixing 3+ one-off radii
ColorToken semantic matchOff-token brand / purple slop
AlignmentOptical nudgeBroken grid / uneven columns

Never claim “100% pixel perfect” without side-by-side evidence.

Checklist

  • Design source identified (link / frame / screenshot)
  • Hierarchy + Auto Layout sizing mapped before polish
  • Colors/type/radius remapped to tokens (or explicit waiver)
  • Primary breakpoints compared (at least phone + desktop when layout differs)
  • States covered if in spec (default / hover / disabled / error)
  • Remaining deltas listed (intentional vs TODO)

Depth

Full compare workflow: references/design-fidelity.md.
Layout map: references/responsive.md (Auto Layout ↔ CSS).
Audit persona: design-reviewer.

Keep looking

Skills are one crate of 627,801. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.