Design

Community

ericrisco

design

Use when designing, redesigning, restyling or reviewing a web page or app screen (landing, product site, portfolio, dashboard, docs) or when a UI looks generic or AI-made. Deals a researched direction, takes the idea from the subject, uses real pictures, proves it in a browser. NOT the words (landing-copy), NOT component state (ui-engineering).

Overview

Publisher	ericrisco
Repository	`rsc-harness`
Skill name	`design`
Stars	239
Forks	24
Bundled files	43
License	MIT
Links	Skill source Repository Homepage

Markdown instructions
A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.
Works with any LLM
AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.
43 bundled files
Scripts, templates, and references the model can read while it works. Files are read-only and never executed.
Open source
Published by ericrisco on GitHub. Read the source before you install it.

Installation

Install the Design AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

Open the app and go to Plugins → Skills.
Choose "Install from GitHub".
Paste the skill folder URL below and confirm.
https://github.com/ericrisco/rsc-harness/tree/main/skills/design
Enable the skill in any chat where you want it available.

Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills

git clone --depth 1 https://github.com/ericrisco/rsc-harness.git /tmp/rsc-harness
mkdir -p .claude/skills
cp -r /tmp/rsc-harness/skills/design .claude/skills/design

Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Design in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Design on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Design is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

Loaded only when it is needed
The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.
Switch models mid-chat
Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

design

A model left to design picks the average of every page it has seen. This skill takes those decisions away from the model and makes it look at what it built.

A script deals the direction. You choose among researched directions; you never invent one.
The idea comes from the subject, not from the category.
Real pictures and full content, never empty space dressed as restraint.
A real browser measures the page at three widths. A FAIL means it is not done.
You look at the screenshots, fix the worst, and look again.

The engine is taste-skill (MIT), shipped unmodified in engine/. The modes, the fresh review and the design record adapt impeccable by Paul Bakaus (Apache-2.0). Credits and what changed: THIRD_PARTY.md.

0. Name the surface and its mode

Decide before anything else. The mode is chosen from the surface, not from the product.

Mode	The visitor…	Surfaces	Run
Persuade	decides and acts	landing, marketing, product site, pricing, event, local business	`engine/RUN.md`, the page run
Experience	is inside the work	portfolio, gallery, showcase	`engine/RUN.md`, the page run
Operate	completes a task	dashboard, admin, settings, inbox, editor	`engine/references/app-screens.md`
Read	understands something	docs, guide, article, changelog	the page run, under the Read rules

What each mode allows and refuses: references/modes.md.

1. Brand first, when there is one

Look for an existing identity before dealing anything: 02-DOCS/wiki/brand/, a DESIGN.md in the repo, or a coherent identity already in the code. If one exists, its colours and faces win: pass --keep palette,type to the draw and take only the composition. If none exists, do not stop to interview the user about brand; the draw is how a starting point is made.

2. Run the engine

Follow engine/RUN.md step by step. Read it whole before starting. In its commands, <skill> is this skill's engine/ folder. Nothing in it is optional:

Brief and inventory (eight to twelve things only this client has).
The argument: one claim per section; no claim, no section.
Deal three directions with draw.mjs, choose one against the inventory, write why.
The idea sentence, the competitor test, and one live thing.
The whole copy deck before markup. Never invent proof: customers, quotes, numbers, logos.
Build still: tokens, hero, sections, at least four real pictures (assets.mjs), the live thing.
Gate: lint.mjs, then inspect.mjs at 1440, 768 and 390. Zero FAIL.
Look: open every screenshot, critique with engine/references/critique.md. At most three rounds.
Move: the motion layer, then inspect again.

For a single component or section inside an existing design, skip the draw, inherit the system, and still run the gate.

3. Fresh review

The builder believes its own page. When the gate passes, hand the screenshots to a reviewer with no history of this conversation (a fresh subagent where the harness has one). It returns one of four words: recapture, rebuild, fix or ship. Act on the word. The packet and the checks: references/finish-review.md.

4. Record the system

A new or replaced visual system is written to DESIGN.md at the project root, from what was built, so later edits keep the look. Format: references/design-md.md. Extending an existing surface does not rewrite it.

When the browser cannot run

inspect.mjs and scenes.mjs need Node 22+ and an installed Chrome, Edge or Chromium. Without them: run lint.mjs (Node 18+), take screenshots with any browser tool you have, critique those, and write in the report: "inspector not run: ". Never report a gate as passed that did not run. Without network, draw the pictures and say the open collections could not be searched.

Report

Say briefly: the surface and mode; the direction dealt (name and seed) and any redraw with its reason; the idea sentence and the live thing; the inspector result per width, or that it did not run; the reviewer's word; every found picture with its credit; sample content to replace before launch. Do not claim a check you did not run.

Hand-offs

The words on a conversion page → ../landing-copy/SKILL.md.
Component boundaries, state, loading and error behaviour → ../ui-engineering/SKILL.md.
WCAG conformance work → ../accessibility/SKILL.md.
A full brand system (logo, print colours, usage rules) → ../brand-identity/SKILL.md.
The App Router build → ../nextjs/SKILL.md.
A hero that is a pre-rendered camera flight in video, scrubbed by scroll → ../scroll-film/SKILL.md.

Files

File	What
`engine/RUN.md`	The run, the house rules and the laws of the craft.
`engine/references/`	App screens, type, colour, layout, imagery, motion, scenes, copy, critique, tells, research.
`engine/scripts/`	`draw`, `assets`, `lint`, `inspect`, `scenes`.
`engine/library/`	The data the draw reads. Never read it whole; use the script.
`engine/UPSTREAM.json`	Which upstream versions are shipped here.
references/modes.md	The four modes.
references/finish-review.md	The fresh reviewer.
references/design-md.md	The `DESIGN.md` record.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Design AI skill do?

Why use Design on TypingMind?

Because you install it once and use it with any model. Design is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Design in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/ericrisco/rsc-harness/tree/main/skills/design. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Design?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Design?

As many as you like. As long as a model supports skills, you can use Design with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Design AI skill free?

Yes. It is published on GitHub by ericrisco under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Canvas

openclaw

OrganizationPopular

Present hosted widget documents on a connected macOS panel and control panel visibility or navigation.

Diagram Maker

Create SVG/HTML or Excalidraw diagrams for concepts, architecture, flows, and whiteboards.

Grill With Docs

A relentless interview to sharpen a plan or design, which also creates docs (ADR's and glossary) as we go.

Grill Me

A relentless interview to sharpen a plan or design.

Overview

Installation

Install in TypingMind

Install in another agent

Use it in TypingMind

Works with any AI model

Skill instructions

design

0. Name the surface and its mode

1. Brand first, when there is one

2. Run the engine

3. Fresh review

4. Record the system

When the browser cannot run

Report

Hand-offs

Files

Bundled files

Frequently asked questions

Related AI Skills

Set up your own AI workspace now