- name
SKILL.md
Le fichier du skill reste dans le dépôt de l'éditeur.
Collez la commande copiée dans un terminal. Elle installe ce SKILL.md dans cet agent.
npx skills add BlackBeltTechnology/pi-agent-dashboard --skill anti-slop-frontend -a cursor -yCollez la commande copiée dans un terminal. Elle installe ce SKILL.md dans cet agent.
npx skills add BlackBeltTechnology/pi-agent-dashboard --skill anti-slop-frontend -a claude-code -yCollez la commande copiée dans un terminal. Elle installe ce SKILL.md dans cet agent.
npx skills add BlackBeltTechnology/pi-agent-dashboard --skill anti-slop-frontend -a codex -yCollez la commande copiée dans un terminal. Elle installe ce SKILL.md dans cet agent.
npx skills add BlackBeltTechnology/pi-agent-dashboard --skill anti-slop-frontend -a grok -yEnvoyez le message copié dans un chat Grok Bot. Le Bot récupère ce SKILL.md sur GitHub et l'ajoute aux skills que partagent tous vos Bots.
Install this skill: https://github.com/BlackBeltTechnology/pi-agent-dashboard/tree/HEAD/packages/anti-slop/.pi/skills/anti-slop-frontendDe l'éditeur
A flat, mechanical checklist of the concrete signatures an undirected model emits when it tries to "look designed." Every rule here is countable or binary - you can mechanically verify pass/fail, not argue taste. That is the whole point: "it looks better" is not a check; "eyebrow count > ceil(sections/3)" is.
frontend-mockup-loopDifferent jobs, intentionally separate:
When both are present: the loop's a11y floor and cite-a-source rule win. This skill feeds concrete failing items into the loop's FIX step. It never overrides a gate, and a tell here is never a reason to violate a cited rule.
Authority: advisory. Findings feed the frontend-mockup-loop FIX step and never override a WCAG-AA or severity gate. When a recommendation conflicts with a gate result, the gate result stands and the recommendation is recorded as overridden.
Honesty note: these rules are curated taste, hardened into countable form. They are good defaults, not laws of nature. Every rule has an override path: when the brief explicitly asks for the "banned" thing, it is allowed - execute it with intent, not by accident.
Before reviewing or generating, state ONE line:
Design Read: <profile> · <audience> · <mood> · VARIANCE <n> / MOTION <n> / DENSITY <n>
Example: Design Read: marketing · technical buyers · calm, precise · VARIANCE 4 / MOTION 3 / DENSITY 5.
Audience and mood are declared context, not checks. The profile is the switch that decides which rule groups fire:
product-ui: dashboards, admin panels, data tables, editors, wizards, any
app screen.marketing: an existing landing / portfolio / about / docs-home page.new-site: a site with no shipped design yet.Never infer the profile silently. An undeclared profile is how marketing rules leak into dense UI. If the brief is ambiguous, ask exactly one question. A review or generation reported without a Design Read fails Verification.
State these before reviewing or generating. They gate which rules fire and how hard.
VARIANCE (1-10) - 1 = perfect symmetry, 10 = artsy chaosMOTION (1-10) - 1 = static, 10 = cinematic/physicsDENSITY (1-10) - 1 = art-gallery airy, 10 = cockpit/packed-dataInfer from the brief; don't silently use a baseline. Dashboards/data UI live high on DENSITY and low on VARIANCE/MOTION. Landing/portfolio live the opposite.
These fire regardless of surface kind. A dense admin panel is just as guilty of AI-purple and Inter-everywhere as a landing page.
#000000 / pure #ffffff. Off-black (zinc-950) and off-white;
pure values kill depth.— and zero –-as-separator anywhere visible. Headlines, labels,
pills, body, quotes, attribution, captions, buttons, alt text. No "sparingly."
Replace with a period, comma, colon, parentheses, line break, or a spaced
hyphen -. Ranges use a plain hyphen (2018-2026, €40-80k).—/–. Any hit = fail.99.99%, 50%, 1234567 → organic values
(47.2%, real-looking phone formats). Fake-precise engineering specs
(5.8mm, 4.1×) are banned unless from real data or labeled mock.<div>s is the single biggest tell. Use a real screenshot, a
generated image, a real mini component preview, or nothing.picsum.photos/seed/{descriptive}/{w}/{h} or
generated assets, never dead Unsplash URLs.MOTION > 4, the page actually moves; if
you can't ship working motion, drop the dial and ship clean static.window.addEventListener('scroll', …) and React-state
scroll/rAF loops. Use scroll-driven CSS, IntersectionObserver, or a motion
library's scroll primitives.MOTION 3.z-* / z-index value comes from one
documented layer scale (e.g. base, sticky, dropdown, overlay, modal, toast).
Mechanical check: count distinct z-index values in the diff; any value not in
the project's scale = fail. z-[9999] / z-50 sprinkled to "win" a stacking
fight is the tell.pointer-events-none layers,
never on a scrolling container (continuous repaint).transform and opacity. Animating top / left /
width / height = fail.<div>s to
place one element is flagged; flatten with grid/flex on the parent.Skip Part B entirely for dashboards, data tables, wizards, editors, product UI. These rules govern hero-driven marketing pages, where the model's worst templating habits live.
pt-24 desktop (content must not float mid-viewport).V0.6, BETA, INVITE-ONLY) unless the brief is a launch.text-[11px] uppercase tracking-[0.18em]).uppercase tracking micro-labels above headlines; fail if > ceil(sections/3).00 / INDEX, 001 · Capabilities, 06 · how it works). The section's position already categorizes it.BRAND. MOTION. SPATIAL.).LIS 14:23 · 18°C) unless the
brief is genuinely place- or timezone-focused.Scroll, ↓ scroll, Scroll to explore).v1.4.2, Build 0048, last sync 4s ago) on
marketing pages - those are devtool fixtures.· is rationed (max 1 per metadata line; not the default
separator for everything).Field study no. 12 · Ines Caetano) on stock/placeholder images.border-t + border-b on every row of a long list/spec table. Long
lists (>5 items) use a real component (grouped chunks, card grid, tabs,
scroll-snap), not a hairline-per-row <ul>.marketing / new-site only)A change must hold in every theme the project ships, not only the default. Each item is pass/fail. Human visual judgment of the screenshots is a non-gating note: record it, it never fails the review.
Count #hex, rgb(, rgba(, hsl(, hsla( on added lines only (the
+ lines of the change's diff) in shipped UI source
(*.tsx|*.jsx|*.css|*.html|*.vue|*.svelte). Any hit = fail; use a theme
token instead.
Excluded: tests, fixtures, stories, *.svg, and the token-definition files
themselves (the files that declare the tokens).
Diff-scoped command (adjust the base ref):
git diff -U0 origin/main...HEAD -- '*.tsx' '*.jsx' '*.css' '*.html' '*.vue' '*.svelte' \
':!**/__tests__/**' ':!**/*.test.*' ':!**/*.spec.*' ':!**/fixtures/**' ':!**/*.stories.*' \
| grep -E '^\+[^+]' | grep -nE '#[0-9a-fA-F]{3,8}\b|rgba?\(|hsla?\('
Also exclude your token-definition files with extra ':!<path>' entries.
product-ui)product-ui: three screenshots: the default theme in dark, the default
theme in light, and one non-default palette. A missing non-default palette
screenshot = fail.marketing: screenshots in light and dark, both recorded in the review.new-site: one screenshot per shipped mode.The rules below can be checked by string-search, not judgment. Run them last.
— / –-as-separator in rendered output (A3).addEventListener('scroll'); reduced-motion present (A7).If a box can't be honestly ticked, it's a flagged item - feed it to the fix step.
Adapted from Leonxlnx/taste-skill
(design-taste-frontend, MIT) at commit 18dfc92 (2026-10-08); the
per-section map lives in packages/anti-slop/UPSTREAM.md. This is a distillation of its countable rules:
stack-coupling (Next RSC / Motion / GSAP / next/font) removed, rules re-scoped
into universal vs marketing-only, and reframed as an advisory catalog that pairs
with - never overrides - frontend-mockup-loop's cite-a-source loop and hard
gates. Original taste, original author's eye; full prose corpus and GSAP code
skeletons live in the upstream repo.
A mechanical, countable anti-slop checklist for AI-generated frontend. Catches the specific signatures an undirected model defaults to: AI-purple glows, Inter-everywhere, em-dashes, div-based fake scr
---name: anti-slop-frontenddescription: A mechanical, countable anti-slop checklist for AI-generated frontend. Catches the specific signatures an…---# packages/anti-slop/.pi/skills/anti-slop-frontend/SKILL.mdVue d'ensemble
A mechanical, countable anti-slop checklist for AI-generated frontend. Catches the specific signatures an undirected model defaults to: AI-purple glows, Inter-everywhere, em-dashes, div-based fake screenshots, eyebrow-on-every-section, beige+brass \"premium\" palettes, generic Jane Doe / Acme data. Advisory layer that pairs with frontend-mockup-loop (which owns the cite-a-source loop plus WCAG gates) but works standalone in any React/Tailwind/HTML project. Triggers: \"does this look AI-generated\", \"anti-slop pass\", \"remove the AI tells\", \"why does this look templated\", \"design review for slop\".
FAQ
Téléchargez packages/anti-slop/.pi/skills/anti-slop-frontend/SKILL.md auprès de l'éditeur, ou copiez le chemin et ouvrez le dépôt. PlugMCP répertorie le skill et n'héberge pas le fichier.
L'éditeur le maintient sur https://github.com/BlackBeltTechnology/pi-agent-dashboard.
Claude, Cursor, Codex et Grok Build peuvent lire un SKILL.md que vous ajoutez depuis le dépôt de l'éditeur. Grok Bot l'installe lorsque vous envoyez dans un chat le lien du dossier GitHub du skill.