Skip to content
Agentmart
Submit a tool

design-rubric

Use when critiquing, reviewing, or refining visual designs, wireframes, or UI layouts — especially for small/medium SPAs.

Skill by els0r

IndexedMIT★ 1
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/els0r-design-rubric

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Use when critiquing, reviewing, or refining visual designs, wireframes, or UI layouts — especially for small/medium SPAs. Applies a six-criterion rubric covering visual clutter, information hierarchy, spacing discipline, data-table rendering, add-flow comprehensibility, and affordance placement. Invoke when the user asks to “review”, “critique”, “audit”, or “refine” a wireframe, mockup, or existing screen, or when they ask whether a design is cluttered, balanced, or readable. Do NOT use for generating new designs from scratch — pair with a generative skill (e.g. `design-crit`, `frontend-design`) for that.

Use cases: design, frontend, data, security

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/design-rubric

What's inside

Skill name
design-rubric
Description
Use when critiquing, reviewing, or refining visual designs, wireframes, or UI layouts — especially for small/medium SPAs. Applies a six-criterion rubric covering visual clutter, information hierarchy, spacing discipline, data-table rendering, add-flow comprehensibility, and affordance placement. Invoke when the user asks to “review”, “critique”, “audit”, or “refine” a wireframe, mockup, or existing screen, or when they ask whether a design is cluttered, balanced, or readable. Do NOT use for generating new designs from scratch — pair with a generative skill (e.g. `design-crit`, `frontend-design`) for that.
License
MIT
Instructions
162 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Design Rubric

A six-criterion rubric for critiquing UI wireframes and layouts.

Philosophy: **strong opinions held loosely**. Opinionated defaults, explicit escape valves. Every critique must be actionable — never “this is cluttered,” always “X and Y compete for attention; demote one.”

## When this skill applies

Trigger when:

- The user asks to review, critique, audit, or refine a wireframe, mockup, or screen.
- The user shares an image, HTML, or sketch of a UI and asks for feedback.
- The user asks a specific rubric question: “is this cluttered?”, “is the hierarchy working?”, “is the spacing off?”.

Do **not** trigger for:

- Generating new designs from scratch (use `design-crit` or `frontend-design`).
- Pure code review without visual concerns.
- Brand / theming decisions in isolation.

Interop: when used alongside `design-crit`, this skill supplies the evaluation rubric for its facet-review loops. `design-crit` generates variants; this skill grades them.

## The rubric

For every screen reviewed, walk the six criteria in order. Name the concrete element that fails and propose one specific fix. If a criterion passes, say so in one sentence and move on.

### 1. Is the UI free of visual clutter?

Signal-to-noise ratio. Every element — borders, shadows, dividers, decorative icons, backgrounds — must justify its presence.

Common violations:

- Borders AND background fill AND shadow on the same card (pick one).
- Dividers between items that whitespace already separates.
- Icons next to labels where the label alone is unambiguous.
- Stroke weights and colors that don’t map to a hierarchy.

Fix pattern: **remove, don’t add**. If a decorative element can disappear without the user being confused, it should.

### 2. Is the most important information visible clearly?

Exactly one primary focal point per screen (or per major region). Squint test: if you blur your eyes until text is unreadable, can you still identify what this screen is *about*?

Common violations:

- Primary metric rendered at the same size as secondary labels.
- CTA competing with a banner, promo, or announcement.
- Header chrome (logo, nav, breadcrumbs) outweighing page content.

Fix pattern: either enlarge/weight the primary, or demote the competitors (lower contrast, smaller size, pushed below the fold).

### 3. Is spacing consistent and principled?

Three-tier spacing model:

**Macro — golden-ratio modular scale, base 8 px.**
Section gaps, card/container padding, heading-to-body spacing, content column widths, vertical rhythm between major blocks.
Scale: `8, 13, 21, 34, 55, 89` px (8 × 1.618ⁿ, rounded — also Fibonacci).

**Micro — 8-px rhythm.**
Icon padding, line-height offsets, form field internal padding, table cell padding, label-to-input gaps, button internal padding.
Scale: `4, 8, 12, 16, 24` px.

**Interaction targets — hard minimums, non-negotiable.**

- Tap targets: ≥ 44 × 44 px.
- Click targets (pointer): ≥ 24 × 24 px.
- Adjacent tap targets: ≥ 8 px gap.

Conflict resolution: **targets > platform conventions > macro scale > micro rhythm**. Golden ratio is a default, not a religion.

Common violations:

- Arbitrary pixel values (`17 px`, `22 px`) with no scale backing.
- Inconsistent gap between similar items (one pair at 16 px, the next at 20 px).
- Micro-spacing inflated to match the macro scale (creates bloat).

### 4. Are tables rendered so they aren’t a wall of text?

Data-dense views fail hardest here. Check:

- **Alignment:** numerics right-aligned, text left-aligned, timestamps monospace or right-aligned.
- **Row separation:** zebra striping OR row dividers, never both. Neither is fine when rows are tall enough.
- **Column count:** more than 6 visible columns usually requires horizontal scroll or column hiding.
- **Primary column emphasis:** the column that identifies the row (name, date) gets weight; metrics are lighter.
- **Empty state:** what shows when there’s no data? A zero-state must exist.
- **Truncation:** long strings truncate with ellipsis AND expose full value on hover/tap.
- **Sticky header:** if more than ~10 rows visible at once.

Fix pattern: if a table feels like a wall, the first move is usually to right-align numerics and demote secondary columns (lighter color, smaller weight) — not to redesign.

### 5. Is the flow for adding X understandable?

For every create/add flow:

- How many clicks/taps from landing to submitted? State it explicitly.
- Is the entry point (the `+` or Add button) visible without scrolling?
- Does the flow have a clear end state (success confirmation, redirect, new item visible)?
- Can the user cancel/back out without losing work — and is that obvious?
- If multi-step: is progress shown (`step 2 of 4`) or disguised as a single action?

Common violations:

- Add button in an unexpected location (top-right on desktop, bottom-center on mobile — pick one per platform).
- Modal stacked on modal.
- Submit button disabled with no explanation of what’s missing.
- Success state that auto-dismisses before the user can read it.

### 6. Does button placement follow intended use?

Placement rules, in priority order:

1. **Primary action** goes where attention lands after reading. LTR forms: bottom-right. Mobile: thumb-reach zone (bottom third).
1. **Destructive actions** separate from constructive ones. Never put Delete and Save adjacent with equal weight.
1. **Button weight mirrors consequence:** primary filled, secondary outlined, tertiary text-only. **One primary per region.**
1. **Cancel / Back** on the opposite side from the primary, or as a subtle link.

Common violations:

From els0r/skills, MIT.

More from els0r/skills

Signals

Stars
1
Forks
0
Open issues
0
Last commit (this folder)
May 20, 2026 (5 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from els0r

Cite this page

design-rubric skill. Agent Mart. https://agentmart.dev/skills/els0r-design-rubric (last updated 2026-10-24)