agentleFS
KyaniteLabsKyaniteLabs/tastecheck/skills/deslop-ui/SKILL.md

deslop-ui

Skill12 starsChanged 12 days agoMITView on GitHub

Anti-slop UI review for generated or generic frontend work. Use after LLM UI output, or for requests to remove AI tells: purple gradients, pill CTAs, default type, centered heroes, three-card grids, glassmorphism, and template sameness.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 0 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Every finding needs subject, P0–P3, observed evidence, preserved signal, replacement, and after-state proof.

    line 47
  2. Very specific:

    Stop / handoff: Stop before changing a coherent branded choice that is supported by the brief.

    line 72

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “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

Deslop UI

Slop is predictable, uncommitted design, not merely ugly design. Audit surface, structure, and voice against the brief; repair with a specific, verifiable choice.

Repair order and evidence contract

  1. Establish brief, preserved signals, and artifact; separate surface, structural, and verbal findings.
  2. Cite selector, section, phrase, screenshot, or browser observation; assign P0–P3.
  3. Choose a brief-bound direction, prescribe repairs, and reinspect the same views.

n/a requires an absent subject. Stop before changing a coherent brief-supported brand choice; ask for direction when brief or acceptance criteria are absent.

Diagnose cause, then tell

Make two passes: intent (job, audience, hierarchy, retained signals, signature) and behavior (first view, transitions, narrow view; trace code to selector/token or screenshot geometry). Test whether a neutral replacement changes meaning or performance. Preserve a documented functional pill, gradient, table, or system font; repair implementation, not role.

Rank P0 meaning/interaction/accessibility/fact, P1 direction/hierarchy, P2 repeated default, then P3 polish. Every P0–P2 packet names subject, observation+brief, mechanism, preserved signal, concrete replacement+owner, and after-state observation (including narrow view). Make one causal structural repair when the skeleton is generic.

Fast audit and self-check

Render before acting; grep is only a lead. Check these detector/decision pairs:

PlaneDetectorRequired repair decision
surfaceunsupported purple/white gradient, default neutral tokens, equal-weight palettedominant field, purposeful accent, semantic tokens, contrast
type/shapeunexamined default type, timid hierarchy, pill CTA, universal radius/shadowtype stance and role-based scale/radius/elevation
structurecentered hero → one CTA → equal cards, uniform section rhythmreplacement topology, lead/supporting evidence, narrow reading order
decoration/functiondefault glass/blobs/emoji/gradient text, static forms, placeholder copybrief-supported atmosphere or none; route behavior/copy to the owning skill

Every finding needs subject, P0–P3, observed evidence, preserved signal, replacement, and after-state proof. Verify surface, structural stack/rhythm, and verbal claims (hand factual prose to humanize-copy). A named direction states aesthetic, dominant field/accent job, type stance, and signature move. Re-run original views.

Reference files

  • references/anti-patterns.md — full detector catalog and fixes.
  • references/structural-tells.md — structural audit and dated attractors.
  • references/design-direction.md — positive direction method.
  • references/decision-records.md — novel-case ADR rules.

How to deliver

Deliver prioritized repairs and named direction; state what changed, why, what survives, and after-state proof. Respect existing tokens; pair with web-typography, color-system, and micro-motion as needed.

<!-- contract:v1:start -->

Contract (generated)

Canonical detail: contract.json.

  • Route: A rendered or coded interface feels generic and needs a structural, visual, and verbal anti-slop audit. (+1 in contract.json); avoid: The task is a greenfield aesthetic interview before a page exists. (+1 in contract.json)
  • Exclude: Do not replace a coherent brand system merely because it is not your preference. (+2 in contract.json)
  • Stop / handoff: Stop before changing a coherent branded choice that is supported by the brief. (+2 in contract.json); receives [design-system-interview, improve-existing-website, art-direction, responsive-layout] -> sends [web-typography, color-system, humanize-copy, tastecheck-pass, a11y-pass, empty-states, form-ux, micro-motion]
  • Output: prioritized anti-slop repair specification and committed direction
  • Evidence: table_with_evidence with status, reason, remediation, evidence, provenance.
<!-- contract:v1:end -->

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.