R_REDDYX.XYZ
AgentsHTML★ 570⑂ 10 forksMIT

felipebasurto/unslop-ui

AGENT SKILL FOR CLEANING UI FROM AI PATTERNS

An agent skill that removes typical "artificial" patterns from interface and dashboard text: eyebrow labels, decorative numbering, botsplaining, stateless badges, and extra icons. Works subtractively, safe by default, and reports all changes.

// KEY FEATURES

  • Removes 12 AI pattern types: eyebrows, numbering, botsplaining, badges, and more
  • Safe by default: touches only visible text, never breaks logic or data
  • Provides detailed report: what was deleted, kept on purpose, and remaining tasks
  • Works as skill in Claude Code, Cursor, Codex; requires Python 3.9+

// DEEP DIVE: HOW IT WORKS

The skill installs via `npx skills add` and is invoked with `/unslop-ui` targeting a file or folder. The agent scans the target code, finds 12 pattern categories (U1–U12) — from eyebrow labels over headings to decorative pulsing dots — and removes them while preserving `aria-label`, `alt`, visible form labels, and anything tied to real state. The `--surface` flag enables visual layer cleanup: gradients, spacing, heavy weights, and motion.

Unlike linters or formatters, unslop-ui works subtractively: it does not rewrite code but only deletes noise that generative models add. Every run ends with a counts table, a list of rewritten strings, a list of intentionally kept items, and tasks for the human. The skill never triggers automatically — the user always initiates the call.

Requires Python 3.9 or newer, no external packages. Works in Claude Code, Cursor, Codex, and any agent with slash commands or skill invocation. Without `--surface` it does not touch colors, layout, data, or logic. It will not run on a screen with nothing to cut — the verdict step returns the file unchanged.

// USE CASES

Best for: for developers and designers cleaning AI-generated interfaces

  • Clean landing page of eyebrow labels, decorative numbering, and botsplaining
  • Remove fake LIVE statuses, NEW/AI badges, and pulsing dots from dashboard
  • Prepare AI-generated components for production with a single skill invocation

// QUICK START

npx skills add felipebasurto/unslop-ui -g
/unslop-ui src/pages/landing.tsx
/unslop-ui src/pages/dashboard.tsx --surface

Commands are taken from the repository README — check the current version on GitHub before running.

#agent-skills#ai-slop#claude-code#claude-skills#codex#cursor#ui#ux-writing
Open on GitHub →

New repositories every 30 minutes

REDDYX AI scans GitHub 24/7 and ships the best AI/ML/Web3 projects to Telegram.

Join on Telegram
← Full catalog·Full index

// SIMILAR REPOSITORIES

RUNNING DEEPSEEK-V4.1 552B ON A SINGLE PC★ 53445 AGENT SKILLS FOR PRODUCT ENGINEERING★ 611MCP SERVER FOR BROWSER CONTROL BY AI AGENTS★ 612TIMON — AI AGENT ORCHESTRATOR ON LOCAL MACHINE★ 615WHITEBOARD WITH MCP FOR RESEARCH AI AGENTS★ 512CAD SYSTEM WITH AGENTS FOR CADQUERY★ 490

← FULL CATALOG

felipebasurto/unslop-ui — Agent Skill For Cleaning UI From…