Nanosaur2-670M
well9472 · Kit
Makes illustrations from text prompts with a small 670M-parameter diffusion model, run through ComfyUI. Built as a low-budget research model for people who wan…
Gives Claude Code skills for deciding interface, motion and copy, auditing a screen for generic AI design before a pull request, and cutting video. For front-end developers and designers.
visant-skills-public is an open-source project on GitHub in the Developer category, listed by Mindket. It is free. This is version 1, last updated on . Install it from GitHub as its README says. Get Mindket’s setup notes on your Mindket account, as a zip.
npx skills add pedrojaques99/visant-skills-public # pick from the list npx skills add pedrojaques99/visant-skills-public --skill visant-killer # just one npx skills add pedrojaques99/visant-skills-public --skill visant-design visant-killer visant-frontend visant-motion visant-refine visant-copy # the interface set
visant-skills-public is a set of Claude Code skills from Visant. Their topics are interface, motion and copy decisions, checking a screen before it goes into a pull request, and video editing. Each skill first checks what your repo already has, such as gates, primitives and tokens, and only then acts. The checks are real scripts, and npm test runs every rule against fixtures both ways.
It suits front-end developers and designers who use Claude Code and want interface, motion and copy work held to clear rules.
| Industry | Who uses it | How they use it | Example |
|---|---|---|---|
| Software products | Front-end developers | Audit a screen before the pull request | Scan a pricing card and get findings on palette classes, copy tics and visual noise |
| Product design | Designers who work in code | Decide each interface element by the business variable it moves | An element that moves no business variable gets removed |
| Marketing | Writers of product copy | Lint copy for signs it was written by AI | Catch an em dash or a hype word in landing page text |
| Video production | Editors of product demos | Cut footage of a recorded website | Turn a recorded website into a rough cut for Premiere |
Teams with no web front end can skip it, and so can anyone who does not use agent skills.
Claude Code is the only agent it was tested on. Support depends on the type of project.
| Use case | Who | What they get back |
|---|---|---|
| Anti-slop audit | A front-end developer | Findings by gate with file and line, then a judged score |
| Motion tuning | A developer | Easing and duration tokens for CSS and framer-motion |
| Metrics and charts | A developer | Charts, gauges, sparklines, sliders and metric tiles that share one structure |
| Portuguese copy | A writer | Brazilian Portuguese copy in a spoken voice, checked by copy-lint |
You need Claude Code and, for most skills, Node.js 20 or newer.
npx skills add pedrojaques99/visant-skills-public # pick from the list npx skills add pedrojaques99/visant-skills-public --skill visant-killer # just one npx skills add pedrojaques99/visant-skills-public --skill visant-design visant-killer visant-frontend visant-motion visant-refine visant-copy # the interface set
Ask Claude Code in English or Portuguese. These requests all work, and the second one means it looks like AI made it.
audit this screen tá com cara de IA make this smoother edit this lesson
The audit skill also runs from a terminal against a source folder, using the command below.
node visant-killer/scripts/killer-scan.mjs demo/src
| Tool | What it does |
|---|---|
| visant-design | The place to start. It reads your request and loads the right playbook |
| visant-killer | Audits a UI for generic AI design. Detectors act as a hard gate, then a counter-argument pass and a score follow |
| visant-frontend | Guides interface decisions. An element stays only if it names the business variable it moves |
| visant-motion | Gives easing and duration tokens for CSS and framer-motion, plus motion-lint and measurement from a URL |
| visant-refine | Covers metric tiles, gauges, charts, sparklines and sliders, all built on one common anatomy |
| visant-copy | Writes Brazilian Portuguese copy in a conversational voice, and copy-lint flags AI giveaways in Portuguese and English text |
You open a React project and say audit this screen. On the sample pricing card, the scan runs four gates and flags a Tailwind palette class that bypasses the project's tokens. It also flags an em dash and a hype word in the copy, an italic word in the headline, and a tightly spaced display serif. After that the skill loads the screen in a browser, challenges its own conclusion and gives a score.
A skill is a folder of instructions that Claude Code loads when your request fits. These ones guide and check interface work. A good first try is to install the set, open a React or Next screen and ask Claude Code to audit it.
The detector scripts form a gate. A confirmed finding counts as a defect, not a taste call, so the agent lists it and asks before editing. The code detectors refuse non-React projects with exit code 2, so they never report a zero they did not read. The author calibrated the detectors on three public Next.js projects until what remained was defensible.
# visant-skills-public Gives Claude Code skills for deciding interface, motion and copy, auditing a screen for generic AI design before a pull request, and cutting video. For front-end developers and designers. Made by pedrojaques99 on GitHub. These notes are Mindket's, written from its README; the project's own files and license are in its repository. ## About What it is: visant-skills-public is a set of Claude Code skills from Visant. Their topics are interface, motion and copy decisions, checking a screen before it goes into a pull request, and video editing. Each skill first checks what your repo already has, such as gates, primitives and tokens, and only then acts. The checks are real scripts, and npm test runs every rule against fixtures both ways. Who it suits: It suits front-end developers and designers who use Claude Code and want interface, motion and copy work held to clear rules. ```table Industry | Who uses it | How they use it | Example Software products | Front-end developers | Audit a screen before the pull request | Scan a pricing card and get findings on palette classes, copy tics and visual noise Product design | Designers who work in code | Decide each interface element by the business variable it moves | An element that moves no business variable gets removed Marketing | Writers of product copy | Lint copy for signs it was written by AI | Catch an em dash or a hype word in landing page text Video production | Editors of product demos | Cut footage of a recorded website | Turn a recorded website into a rough cut for Premiere ``` Teams with no web front end can skip it, and so can anyone who does not use agent skills. Where to use it: Claude Code is the only agent it was tested on. Support depends on the type of project. - Claude Code, where the skills load when your request matches - React and Next projects with .tsx and .jsx files, where every tool works - Vue, Svelte, Astro and HTML projects, for the judgment skills, the open-screen scans and copy-lint - Other agents that read SKILL.md, which may work but are untested ```table Use case | Who | What they get back Anti-slop audit | A front-end developer | Findings by gate with file and line, then a judged score Motion tuning | A developer | Easing and duration tokens for CSS and framer-motion Metrics and charts | A developer | Charts, gauges, sparklines, sliders and metric tiles that share one structure Portuguese copy | A writer | Brazilian Portuguese copy in a spoken voice, checked by copy-lint ``` How to set it up: You need Claude Code and, for most skills, Node.js 20 or newer. - Claude Code - Node.js 20 or newer, for every skill except visant-frontend - Optional: Playwright for screen checks, and axe-core for target size and name checks 1. Run the first command below to pick skills from a list. 2. To add one skill, run the second command. To add the interface set, run the third. 3. You can also place a skill's folder in the ~/.claude/skills/ directory yourself. ```bash npx skills add pedrojaques99/visant-skills-public # pick from the list npx skills add pedrojaques99/visant-skills-public --skill visant-killer # just one npx skills add pedrojaques99/visant-skills-public --skill visant-design visant-killer visant-frontend visant-motion visant-refine visant-copy # the interface set ``` How to use it: Ask Claude Code in English or Portuguese. These requests all work, and the second one means it looks like AI made it. ```text audit this screen tá com cara de IA make this smoother edit this lesson ``` The audit skill also runs from a terminal against a source folder, using the command below. ```bash node visant-killer/scripts/killer-scan.mjs demo/src ``` ```table Tool | What it does visant-design | The place to start. It reads your request and loads the right playbook visant-killer | Audits a UI for generic AI design. Detectors act as a hard gate, then a counter-argument pass and a score follow visant-frontend | Guides interface decisions. An element stays only if it names the business variable it moves visant-motion | Gives easing and duration tokens for CSS and framer-motion, plus motion-lint and measurement from a URL visant-refine | Covers metric tiles, gauges, charts, sparklines and sliders, all built on one common anatomy visant-copy | Writes Brazilian Portuguese copy in a conversational voice, and copy-lint flags AI giveaways in Portuguese and English text ``` Example: You open a React project and say audit this screen. On the sample pricing card, the scan runs four gates and flags a Tailwind palette class that bypasses the project's tokens. It also flags an em dash and a hype word in the copy, an italic word in the headline, and a tightly spaced display serif. After that the skill loads the screen in a browser, challenges its own conclusion and gives a score. For newcomers and experts: A skill is a folder of instructions that Claude Code loads when your request fits. These ones guide and check interface work. A good first try is to install the set, open a React or Next screen and ask Claude Code to audit it. The detector scripts form a gate. A confirmed finding counts as a defect, not a taste call, so the agent lists it and asks before editing. The code detectors refuse non-React projects with exit code 2, so they never report a zero they did not read. The author calibrated the detectors on three public Next.js projects until what remained was defensible. ```flow You ask Claude Code to audit a screen The skill finds your repo's gates and tokens Detector scripts scan the code and run the gates The skill judges what detectors miss and scores it It lists findings and asks before editing ``` Good to know: - It was tested on Claude Code only, and other agents that read SKILL.md are untested - The killer scan and the code detectors refuse Vue, Svelte, Astro and HTML projects with exit code 2 - Most skills need Node.js 20 or newer, and visant-frontend needs nothing - Playwright and axe-core are optional extras for screen checks - Commands in the skill docs use a skill-dir placeholder for the folder that holds SKILL.md - The copy skill targets Brazilian Portuguese, and copy-lint also checks English ## On GitHub, 7 November 2026 - Stars: 84 - Forks: 2 - Open issues and pull requests: 1 - Language: JavaScript - License: MIT - Topics: claude-code, claude-skills, design-system, ui-audit ## Links - Code: https://github.com/pedrojaques99/visant-skills-public - Releases: https://github.com/pedrojaques99/visant-skills-public/releases - Issues: https://github.com/pedrojaques99/visant-skills-public/issues - Star history: https://www.star-history.com/#pedrojaques99/visant-skills-public&Date ## Get the code ``` git clone https://github.com/pedrojaques99/visant-skills-public.git ``` Then follow the setup in its README. Nothing in this kit runs by itself.
README.md 7 KB
Listed GitHub projects in the same category that share the most topics with this one.
| Product | Stars | Activity | Language | License | Last change |
|---|---|---|---|---|---|
| visant-skills-public This page | 84 | Active | JavaScript | MIT | |
| Taste SkillShares design-system · claude-code | 114,346 | Active | JavaScript | MIT | |
| Claude-Mem by thedotmackShares claude-skills · claude-code | 116,594 | Active | TypeScript | Apache-2.0 | |
| ArchifyShares claude-skills · claude-code | 94,756 | Active | JavaScript | Apache-2.0 | |
| pstackShares claude-skills · claude-code | 4,132 | — | JavaScript | MIT | — |
| Scenario Agent SkillsShares claude-skills · claude-code | 3,546 | Active | Python | MIT |
well9472 · Kit
Makes illustrations from text prompts with a small 670M-parameter diffusion model, run through ComfyUI. Built as a low-budget research model for people who wan…
Waqf-AI · Kit
Reads handwritten Arabic text from images and returns it as text. A fine-tuned PaddleOCR-VL model for people digitizing Arabic handwriting.
Wan-AI · Kit
Makes video clips from a text prompt or an image plus text, at up to 720P and 24 frames per second, in the Diffusers format for a single consumer GPU.
thedotmack · Kit
Alex Newman’s memory plugin for Claude Code and other coding agents. It records what the agent does in each session, turns that into short searchable notes and…
Leonxlnx · Skill
Leonxlnx’s Agent Skill that gives Claude Code and Codex an art direction, pictures from open collections and a real-browser check, so the websites and app scre…
tt-a1i · Skill
Generates interactive diagrams and explanatory pages from codebases or textual plans. It allows AI agents to visualize workflows, architecture, and complex log…