Html Draft logo

Html Draft

Community
serejaris
html-draft

Use when user wants a standalone HTML diagram in flat engineering blueprint style — architecture diagrams, system flows, technical spec sheets, component maps. Generates one HTML file using Tailwind v4 (browser CDN) for layout and D3 v7 (CDN) for SVG diagrams. User-invoked only — do NOT auto-trigger. Triggers on "/html-draft", "сделай blueprint", "технический чертёж", "архитектурная схема", "инженерная схема", "blueprint diagram", "engineering blueprint", "technical spec sheet", "architecture diagram", "system flow diagram".

Overview

Publisherserejaris
Repositorypersonal-corp-os
Skill namehtml-draft
Stars
232
Forks
26
Bundled files
2
LicenseMIT
Links
  • 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.

  • 2 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by serejaris on GitHub. Read the source before you install it.

Installation

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

1

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.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/serejaris/personal-corp-os/tree/main/skills/html-draft
  4. 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.
2

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/serejaris/personal-corp-os.git /tmp/personal-corp-os
mkdir -p .claude/skills
cp -r /tmp/personal-corp-os/skills/html-draft .claude/skills/html-draft
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Html Draft 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 Html Draft 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 Html Draft 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.

html-draft — Flat Engineering Blueprint Diagrams

Generate one HTML page that renders a technical diagram in a strict flat-blueprint aesthetic — the look of a printed engineering specification sheet, not a marketing landing.

Stack: Tailwind v4 via @tailwindcss/browser CDN for layout + utilities, D3 v7 via jsDelivr CDN for SVG-based diagrams (nodes, connectors, layouts, animations).

Use this when the user wants an architecture diagram, system flow, technical spec sheet, or component map as a standalone HTML artifact (suitable for slides, reports, exports).

Don't use this for:

  • Inline schemas inside markdown documents — use a mermaid renderer instead
  • Newspaper / reading-first single-column pages with monospace ink-on-cream feel
  • Multi-section interactive explainers with pill navigation

Design philosophy

Precise. Objective. High data-ink ratio (Tufte). Every pixel earns its place; nothing decorative. The stack is modern (Tailwind + D3) but the output looks like a printed engineering doc.

Visual rules

Flat, outlined, monochrome

  • No drop shadows, gradients, glassmorphism, blur, rounded buttons
  • 1px or 2px solid borders define structure
  • White content blocks on a light-gray canvas
  • Accent: black, or a single semantic color (red for error, etc.) used sparingly
  • Do not import a Tailwind component library — pure utility classes only

Design tokens (declared once via @theme)

css
@theme {
  --color-c-bg: #f8fafc;          /* page background — slate-50 */
  --color-c-canvas: #ffffff;      /* diagram canvas */
  --color-c-border: #cbd5e1;      /* slate-300 */
  --color-c-text-main: #0f172a;   /* slate-900 */
  --color-c-text-sub: #64748b;    /* slate-500 */
  --color-c-accent: #b91c1c;      /* red-700 — semantic only */
  --font-ui: system-ui, -apple-system, 'Segoe UI', sans-serif;
  --font-mono: 'SF Mono', Monaco, Consolas, monospace;
}

Tokens become Tailwind utilities automatically: bg-c-canvas, border-c-border, text-c-text-sub, font-mono.

Typography

  • Headings, labels: sans-serif (font-ui)
  • Data, paths, code, IDs, version strings: font-mono
  • Never link Google Fonts — the system stack already covers both roles

Layout

  • Whole diagram lives in a .diagram-canvas — bordered box with generous padding (p-8 or more)
  • Header: title + UPPERCASE subtitle, separated from body by a 1px bottom border
  • Strict alignment via grid / flex utilities; no eyeballing

Connectors

  • Thin straight or orthogonal lines (1px solid)
  • Dashed lines for abstract / logical relationships, never structural ones
  • D3-rendered SVG for non-orthogonal arrows; Tailwind border-t / border-l for orthogonal CSS connectors

Icons & badges

  • Icons: simple stroke SVG (no fills, no detail) drawn via D3 or inline <svg>
  • Badges: outlined or solid black/gray block, small uppercase mono text

Hard requirements

  1. Tailwind v4 via browser CDN — version-pinned https://cdn.jsdelivr.net/npm/@tailwindcss/browser@4
  2. D3 v7 via jsDelivr CDN — version-pinned https://cdn.jsdelivr.net/npm/d3@7
  3. Return only the HTML — no markdown wrapper, no commentary outside the file
  4. Complete document — <!DOCTYPE html> through </html>
  5. Design tokens declared in a single <style type="text/tailwindcss"> @theme block — no scattered custom CSS
  6. Custom CSS minimal — only what Tailwind utilities cannot express (e.g. SVG marker definitions, complex pseudo-elements)
  7. No external fonts (no Google Fonts, no Adobe Fonts) — only Tailwind + D3 CDNs

Output template

html
<!DOCTYPE html>
<html lang="en">
  <head>
    <meta charset="UTF-8" />
    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
    <title>[Diagram Title]</title>
    <script src="https://cdn.jsdelivr.net/npm/@tailwindcss/browser@4"></script>
    <style type="text/tailwindcss">
      @theme {
        --color-c-bg: #f8fafc;
        --color-c-canvas: #ffffff;
        --color-c-border: #cbd5e1;
        --color-c-border-strong: #94a3b8;
        --color-c-text-main: #0f172a;
        --color-c-text-sub: #64748b;
        --color-c-accent: #b91c1c;
        --font-ui: system-ui, -apple-system, 'Segoe UI', sans-serif;
        --font-mono: 'SF Mono', Monaco, Consolas, monospace;
      }

      body {
        font-family: var(--font-ui);
      }

      .mono {
        font-family: var(--font-mono);
      }
    </style>
  </head>
  <body class="bg-c-bg text-c-text-main p-10">
    <div class="max-w-[1200px] mx-auto bg-c-canvas border-2 border-c-border-strong p-8">
      <header class="border-b border-c-border pb-4 mb-6 flex items-end justify-between">
        <div>
          <h1 class="text-2xl font-semibold">[Title]</h1>
          <p class="mono text-[11px] uppercase tracking-widest text-c-text-sub mt-1">
            [SUBTITLE]
          </p>
        </div>
        <div class="mono text-[11px] text-c-text-sub text-right">
          DOC-[ID]<br/>REV A
        </div>
      </header>

      <!-- Diagram content: Tailwind grid for spec sheets, D3 SVG for flows -->
      <section class="grid grid-cols-2 border border-c-border">
        <!-- spec cells, see snippets below -->
      </section>

      <!-- D3 mount point for SVG diagrams -->
      <svg id="d3-diagram" class="w-full border border-c-border mt-6" height="400"></svg>
    </div>

    <script src="https://cdn.jsdelivr.net/npm/d3@7"></script>
    <script>
      // D3 diagram rendering — see "D3 patterns" section below
    </script>
  </body>
</html>

Reusable component snippets

Node / box

html
<div class="bg-c-canvas border border-c-border p-3">
  <div class="text-[10px] uppercase tracking-wide text-c-text-sub">label</div>
  <div class="mono text-sm">value</div>
</div>

Badge

html
<span class="inline-block mono text-[10px] uppercase px-1.5 py-0.5 border border-c-text-main">
  ACTIVE
</span>

<span class="inline-block mono text-[10px] uppercase px-1.5 py-0.5 bg-c-text-main text-c-canvas">
  SCHEDULED
</span>

<span class="inline-block mono text-[10px] uppercase px-1.5 py-0.5 bg-c-accent text-c-canvas">
  OVERDUE
</span>

Connector (orthogonal, CSS)

html
<div class="border-t border-c-border"></div>
<div class="border-t border-dashed border-c-border"></div>

Spec grid cell

html
<div class="p-4 border-r border-b border-c-border last:border-r-0">
  <div class="text-[10px] uppercase tracking-wide text-c-text-sub mb-1">label</div>
  <div class="mono text-sm">value</div>
</div>

D3 patterns

Use D3 when geometry is non-orthogonal, computed, or large enough that hand-placing nodes is unmaintainable.

Pattern 1 — explicit nodes + links (architecture diagrams)

javascript
const svg = d3.select('#d3-diagram');
const nodes = [
  { id: 'api',    x: 100, y: 100, label: 'API'    },
  { id: 'worker', x: 400, y: 100, label: 'Worker' },
  { id: 'db',     x: 250, y: 280, label: 'DB'     },
];
const links = [
  { source: 'api',    target: 'worker', style: 'solid'  },
  { source: 'api',    target: 'db',     style: 'solid'  },
  { source: 'worker', target: 'db',     style: 'dashed' },
];
const byId = Object.fromEntries(nodes.map(n => [n.id, n]));

// arrow marker
svg.append('defs').append('marker')
  .attr('id', 'arrow').attr('viewBox', '0 -5 10 10')
  .attr('refX', 8).attr('refY', 0).attr('markerWidth', 6).attr('markerHeight', 6)
  .attr('orient', 'auto')
  .append('path').attr('d', 'M0,-4L8,0L0,4').attr('fill', '#0f172a');

// links
svg.selectAll('line').data(links).enter().append('line')
  .attr('x1', d => byId[d.source].x).attr('y1', d => byId[d.source].y)
  .attr('x2', d => byId[d.target].x).attr('y2', d => byId[d.target].y)
  .attr('stroke', '#0f172a').attr('stroke-width', 1)
  .attr('stroke-dasharray', d => d.style === 'dashed' ? '4 3' : null)
  .attr('marker-end', 'url(#arrow)');

// nodes
const g = svg.selectAll('g.node').data(nodes).enter().append('g')
  .attr('transform', d => `translate(${d.x - 50}, ${d.y - 18})`);
g.append('rect').attr('width', 100).attr('height', 36)
  .attr('fill', '#fff').attr('stroke', '#0f172a').attr('stroke-width', 1);
g.append('text').attr('x', 50).attr('y', 22)
  .attr('text-anchor', 'middle').attr('font-size', 12)
  .attr('font-family', 'system-ui').text(d => d.label);

Pattern 2 — tree layout (hierarchical structures)

d3.hierarchy() + d3.tree() for parent/child trees (component maps, org charts). Render with the same flat node style; never use the default rounded D3 examples.

Pattern 3 — DAG / flow

d3-dag (optional) or manual topological layout. For < 15 nodes, hand-place coordinates — it's faster and tighter than a layout algorithm.

Pattern 4 — sankey / flow volumes

d3-sankey plugin (https://cdn.jsdelivr.net/npm/d3-sankey@0.12) when volumes matter. Keep ribbons grayscale; one accent only for the watched flow.

What D3 must not do here

  • No force-directed simulations bouncing around — diagrams are static engineering docs
  • No smooth zoom/pan unless the user asks — extra interactivity adds noise
  • No tooltips / hover popups unless the user asks
  • No colorful palettes — the visual rules above still bind

Composition guide

  • Architecture diagram: services as D3 rect nodes, data flow as solid SVG lines with arrow markers, dependencies as dashed
  • System flow: linear stages left-to-right or top-to-bottom; decision points as outlined diamonds (D3 polygons); use Tailwind grid for non-flow sections of the same page
  • Spec sheet: Tailwind grid of labeled cells, each with a mono value and a sans-serif label; status badge top-right of each cell
  • Component map: nested boxes in HTML (Tailwind) for top level; D3 hierarchy for deep trees; badges on each node

Quality bar

  1. Every text item earns its space — no decorative copy
  2. Alignment is strict — no off-grid placement
  3. Mono and sans roles never bleed (don't put labels in mono or data in sans)
  4. Color usage stays monochrome unless one semantic accent is justified
  5. CDN scripts pinned to specific major versions (@tailwindcss/browser@4, d3@7)
  6. D3 code is readable — named variables, no one-letter chaining beyond what's idiomatic
  7. The page renders correctly on first paint even before D3 mounts (no layout jump)

When input is incomplete

  • No content → ask for the diagram type (architecture / flow / spec sheet / component map) and the items to render
  • Vague subject → propose a node list and ask for confirmation before rendering
  • User asks for a different library (Mermaid inside, Recharts, Chart.js) → push back: this skill is Tailwind + D3 only; suggest mermaid-diagrams skill if the user wants Mermaid
  • User asks for interactivity (tooltips, drag, zoom) → confirm explicitly; default is static print-style

Source

Methodology adapted from QoderWork's drafter-diagram skill (flat-engineering-blueprint visual system), restacked on Tailwind v4 + D3 v7.

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 Html Draft AI skill do?

Use when user wants a standalone HTML diagram in flat engineering blueprint style — architecture diagrams, system flows, technical spec sheets, component maps. Generates one HTML file using Tailwind v4 (browser CDN) for layout and D3 v7 (CDN) for SVG diagrams. User-invoked only — do NOT auto-trigger. Triggers on "/html-draft", "сделай blueprint", "технический чертёж", "архитектурная схема", "инженерная схема", "blueprint diagram", "engineering blueprint", "technical spec sheet", "architecture diagram", "system flow diagram".

Why use Html Draft on TypingMind?

Because you install it once and use it with any model. Html Draft 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 Html Draft in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/serejaris/personal-corp-os/tree/main/skills/html-draft. 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 Html Draft?

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 Html Draft?

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

Is the Html Draft AI skill free?

Yes. It is published on GitHub by serejaris 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

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇