agentsclimarketplace

Preflight

Skill tommylower/cortex/design/review/preflight

Public skill library, workflows, tools, and references for AI-assisted development.

Install
npx -y skills add tommylower/cortex --skill preflight

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

  • 1 stars1 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.

What its author says it does

Copied from the file, not written here

Preflight design review before shipping. Use for static UI/a11y checks, visual consistency, AI-slop detection, or a final component/page ship check. Triggers: preflight, design review, accessibility audit, WCAG, UI polish, design critique, AI slop check.

SKILL.md

3.4 KB, as published. Nobody here has run it

preflight

final design audit before shipping. reviews code for accessibility issues, visual inconsistencies, and AI-generated pattern tells, then offers to fix them.

usage

run /preflight on a component or page to get a scored review with line numbers, code snippets, fix suggestions, and references.

what it checks

accessibility

  • images without alt text
  • icon-only buttons missing aria-labels
  • form inputs without labels
  • non-semantic click handlers
  • missing keyboard handlers
  • missing visible focus states or outline-none without replacement
  • unlabeled async updates, toasts, validation, or pending regions
  • color-only information
  • touch targets under 44x44px
  • skipped heading levels
  • contrast ratio below 4.5:1

visual consistency

  • inconsistent spacing values
  • overflow and alignment issues
  • z-index conflicts
  • mixed font families and weights
  • line height issues
  • missing font fallbacks
  • missing global font smoothing (-webkit-font-smoothing: antialiased + -moz-osx-font-smoothing: grayscale)
  • thin weights (below 400) at 14px or smaller — antialiased rendering leaves them too faint
  • URL/state mismatch for filters, tabs, pagination, or expanded panels
  • missing safe-area handling for fixed mobile surfaces

AI pattern detection

flags common tells that make interfaces look AI-generated:

layout tells

  • side-tab borders (thick colored border on one side of a card)
  • nested cards (cards inside cards)
  • everything centered (all text center-aligned)
  • icon-tile-stack (small rounded-square icons stacked above headings)
  • monotonous spacing (same spacing value used everywhere)

typography tells

  • gradient text (background-clip: text with gradients)
  • overused fonts (Inter, Roboto, Open Sans, Lato, Montserrat, Arial)
  • single font (only one font family on the whole page)
  • flat type hierarchy (font sizes too similar between levels)

color and effect tells

  • AI color palette (purple/violet gradients, cyan-on-dark)
  • dark glow (dark bg + colored box-shadow)
  • bounce/elastic easing (feels dated)
  • pure black/white backgrounds (no subtle tinting)
  • gray text on colored backgrounds

quality

  • text lines exceeding 75 characters
  • cramped padding (text too close to container edges)
  • tight leading (line-height below 1.3x)
  • justified text without hyphenation
  • body text below 12px
  • long passages in all caps
  • letter-spacing over 0.05em on body text
  • animating width/height/padding/margin (performance)
  • transition: all
  • blocking paste in inputs
  • hardcoded date, number, or currency formatting instead of Intl
  • images without explicit dimensions when dimensions are knowable

forms

  • inputs without name, appropriate type, or autocomplete intent
  • errors not linked to fields with aria-describedby
  • disabled submit buttons that do not explain why
  • submit states that resize the button or lose the accessible label
  • critical errors shown only in toasts

output

each issue includes:

  • line number and code snippet
  • what's wrong and why it matters
  • fix suggestion
  • WCAG reference where applicable

outputs a score out of 100. run after building components, before shipping.

Keep looking

Skills are one crate of 328,083. 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.