- Home
- Skills
- Prompt & Skill Engineering
- design md creator
Works with the AI Agents you already use
design md creator
Generates observable design rulebooks (design.md) to stop coding agents from producing generic "AI slop" layouts.
Free
design md creator
Example session with this skill installed
I need a design.md for my project 'Nexus'. Here is my palette: #0f172a, #38bdf8, #f8fafc. I want it to look professional and avoid that generic AI dashboard feel.
- Read your context and instructions
- Compiled the design md creator
I have processed your palette and started the interview for Nexus. I derived a Dark Theme as the default given your primary color.
design.md
1. Brand & Identity
- Wordmark: "Nexus" must be written in lowercase bold.
- Placement: Top-left corner of all viewports.
- Constraint: Never place the wordmark over patterned backgrounds; use solid
#0f172aonly.
2. Colors
- Primary (Accent):
#38bdf8(User defined). - Background:
#0f172a(User defined). - Surface:
#1e293b(Derived from accent). - Text:
#f8fafc(User defined). - Border:
#334155(Derived for contrast). - Rule: Accent color is reserved for primary actions only. Do not use for icons, borders, or text.
3. Typography
- Family: System sans-serif stack.
- Scale: Use exactly six steps: 12px, 14px, 16px, 20px, 24px, 32px.
- Constraint: No intermediate sizes. Line length capped at 65 characters for prose.
4. Reader & Decision
- Reader: DevOps Engineer.
- Decision: Identifying which microservice is currently failing.
- Rules:
- Status indicators must be the largest visual element.
- No decorative illustrations.
- Data tables must use mono-spaced digits.
5. Prohibited Patterns (Anti-Slop)
- No Centered Hero: All layouts must be left-aligned.
- No Glassmorphism: Surfaces must be solid colors; no blur effects.
- No Generic Icons: Use Lucide icons; never generate SVG placeholders.
Next steps
- Review the derived Surface and Border colors for contrast.
- Place this
design.mdin your project root. - Instruct your coding agent to "Read design.md before generating any UI."
Connects securely to your tools. The creator never sees your data.
About this skill
Generic AI generated pages often lack brand identity and suffer from repetitive, machine-like layouts. This skill interviews you to extract specific, observable design rules and packages them into a design.md file that other coding agents load to ensure consistent, on-brand output.
What it does
- Observable Rule Creation transforms vague requests like "clean" or "modern" into binary, verifiable design constraints.
- Automated Palette Logic processes hex codes or palette links to derive accessible background, surface, and border roles.
- Anti-Pattern Filtering identifies and bans specific "AI slop" traits through a curated library of common generation mistakes.
- Structural Mapping derives page hierarchy from the specific reader persona and their intended decision point rather than generic templates.
- Primitive Documentation tracks existing component libraries or stylesheets to prevent the agent from hallucinating new CSS.
How it works
- Trigger the interview by providing your brand name or a color palette.
- Answer ten focused questions covering typography, layout, and prohibited patterns.
- Review the color roles generated by the internal palette script which verifies contrast ratios.
- Download the
design.mdfile and include it in the context of your next coding task.
Frameworks & tools
The skill utilizes a Python-based palette processing engine for contrast validation and role derivation. It follows a strict Markdown schema compatible with Cursor, Claude Code, and other LLM-based IDEs.
Why this beats prompting it yourself
Generic prompts fail because they rely on adjectives. This skill enforces observable constraints, ensuring you can look at an output and answer "yes" or "no" to whether a rule was followed. It automates the tedious math of accessibility and color role derivation that manual prompts often skip.
Use cases
- Generating a brand rulebook for a new landing page project.
- Preventing v0 or Claude from producing generic "AI-looking" dashboard layouts.
- Standardizing the UI output of multiple agents working on the same repo.
- Translating a Figma palette into functional system tokens for coding agents.
Known limitations
Does not design individual components or pages. Requires the user to answer the interview questions; skipping questions results in omitted sections. Color processing requires valid hex or image inputs.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Prompt & Skill Engineering.
Produces guidance files only, not actual UI code. Requires active participation in a 10-question interview. Script fails if color contrast thresholds are not met.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 14 days ago
- Free with an Agensi account
- Software
Git