Skip to content
SkillMixed7,208 tokens in use

Unslop

Turns vibecoded, AI-looking UI into a designed interface: audits the tells, sets a direction from your product, then fixes the screen in nine ordered passes.

19,221 stars on the repository CC-BY-4.0heydamir/unslop · skills/unslop
claude plugin install unslop --marketplace heydamir/unslop
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Unslop costs

154tokens idle
7,208tokens in use
16,889tokens, up to

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Up to adds the files its instructions link. Counts are approximate, made offline with cl100k_base.

Model dependence Mixed

Script-ledMixedModel-led

Scripts do part of the work and the model's judgement does the rest.

This describes the skill's structure, how much of it is scripts. It does not say how well it works. Scripts: scripts/snapshot.mjs, scripts/story-path.mjs, scripts/story.mjs.

Touches

shelldetectednetworkdetectedwrites filesdetected

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter15/15
  • Structure19/25
  • Examples3/15
  • Output10/15
  • Size11/15
  • Resources9/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Good
  • Reaches the network−10

    does this on the machine, and the frontmatter does not say so

    scripts/snapshot.mjs, line 53
  • Writes files−8

    does this on the machine, and the frontmatter does not say so

    scripts/snapshot.mjs, line 191
  • Runs shell commands−5

    does this on the machine, and the frontmatter does not say so

    scripts/story.mjs, line 4

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity92
  • Maintenance52
  • Community56
  • Maturity4
  • Momentum40

Unslop is a redline-style design pass for screens that look machine-made: purple gradients, glows, emoji labels, shouty copy and uneven spacing. It audits the screen, writes a five-line direction taken from what the product does, and waits for your answer before editing.

It then works through spacing, color, icons, copy, type, layout, details, visuals and motion, in that order, on React, Tailwind, HTML and CSS. It can also snapshot a live page from a link and render a short story video of the screen cleaning itself up.

What You Get

  • An audit listing each AI tell and where it lives in the code
  • A five-line design direction derived from the product
  • The screen rewritten pass by pass, with exact spacing, color, type and motion values from a token file
  • A report of what changed per pass, plus any image prompts it wrote
  • Optionally a square story video under ten seconds showing the screen unslopping itself

Common Use Cases

  • Clean up a dashboard generated by an AI coding tool before showing it to customers
  • Redesign a landing page hero so it has a real scene and a product preview
  • Polish a mostly designed screen by fixing the few remaining tells
  • Rebuild a live site’s page from its link as static HTML and hand back a redesign
  • Make a before and after video for social media

What Makes It Unique

Most UI polish skills list rules. Unslop sets the order of work (spacing first, visuals and motion last), makes the agent write a direction from the product and stop for approval, and adds a template check at the end: if swapping in another product’s name would still fit, push the direction harder. It also has a polish mode for screens that are already mostly good and says so when a screen needs no change.

Under The Hood

The agent reads SKILL.md and reference files for signs, tokens, visuals, effects and right-to-left rules, then edits your project’s source files. To work from a link it runs a bundled Node script that opens the page in Playwright, scrolls through it and saves an HTML copy and a screenshot. For the story video it runs another Node script that renders the page frame by frame with Playwright and encodes it with ffmpeg. When an image tool is available it can send a prompt to an image generation service to make a hero visual. It writes files in your project and an unslop folder, and the agent may run a dev server and take screenshots to check its work.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers