unslop
A design skill for Claude Code and Codex. It audits AI-looking UI (purple gradients, glows, emoji labels, uneven spacing), sets a direction from the product, and redesigns the screen in nine ordered passes. It can optionally generate a hero image or render a short 'story' video of the redesign.
What it can touch
| files | write |
|---|---|
| network | cdn.tailwindcss.com, fonts.googleapis.com |
| shell | runs commands |
| secrets | none found |
What it actually does
Almost all of the bundle is design guidance in Markdown (SKILL.md and references/*.md) that tells the agent how to audit and edit the user's UI code, with exact tokens for spacing, color, type, radius and motion. It tells the agent to show the audit and direction to the user and wait for an answer before editing, and to ask before any paid image generation. There are two optional Node scripts. scripts/snapshot.mjs uses Playwright to open a URL the user supplies, fetches that page's stylesheets and images, inlines them into one HTML file, and saves it with a screenshot. scripts/story.mjs starts a temporary HTTP server on 127.0.0.1 that serves the story file's folder. It drives Playwright to capture frames into a .story-frames folder, runs ffmpeg through execFileSync to encode an MP4, then deletes the frames folder. The scripts do not read credentials, have no persistence, contact no hidden hosts, are not obfuscated, and contain no text aimed at the reviewing AI.
What the check found
Nothing above the low level.
Keep an eye on it
This check is of one commit. Watch unslop and we'll email you if a later commit changes what it can touch.
Made this? Add the badge
[](https://plugsday.com/tools/heydamir-unslop)About this check
Checked November 2, 2026 at commit a6f121eee5f2, 21 files. Rules for known red flags, then two Claude reviews, one of them told to assume the author is malicious. A later commit hasn't been checked. How checks work.