Learn-skills.dev html-visual
Generate an interactive single-file HTML visualization (mockup, wireframe, ERD, flowchart, chart, slides, architecture diagram, dashboard, timeline, mindmap, kanban, table).
install
source · Clone the upstream repo
git clone https://github.com/NeverSight/learn-skills.dev
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/NeverSight/learn-skills.dev "$T" && mkdir -p ~/.claude/skills && cp -r "$T/data/skills-md/2ykwang/agent-skills/html-visual" ~/.claude/skills/neversight-learn-skills-dev-html-visual && rm -rf "$T"
manifest:
data/skills-md/2ykwang/agent-skills/html-visual/SKILL.mdsource content
Input
$ARGUMENTS
Instructions
Visualize the user's request as an interactive single HTML file. First, read
references/common-rules.md to review common rules.
Type Determination
Determine the type from the first argument (
$0).
| Argument | Type | Filename Pattern |
|---|---|---|
| UI mockup (high-fidelity) | |
| Wireframe (low-fidelity, hand-drawn style) | |
| ERD | |
| Flowchart / Sequence diagram | |
| Data chart | |
| Presentation | |
| Architecture diagram | |
| Composite dashboard | |
| Timeline / Gantt chart | |
| Mindmap | |
| Kanban board | |
| Interactive data table | |
No match: Infer the type from the request content. If unable to infer, ask the user. When inferred, use
visual-{name}.html as the filename.
rule: Extract the core noun from the request and convert to kebab-case. e.g., "user login form" → {name}
login-form, "payment flow" → payment-flow.
Input Handling
- File path provided: Read and analyze the file, then visualize.
e.g.,
→ Analyze the Prisma schema to auto-generate ERD/html-visual erd schema.prisma - Existing HTML modification: Read and modify the existing file. Do not recreate from scratch.
- Natural language only: Infer the type, then generate.
Context Gathering
- Description is sufficient: Generate immediately (e.g., "simple login form mockup")
- Project code reference needed: Read code/schema/API first (e.g., "our project's ERD", "current payment flow")
- Criterion: If the request contains project context references like "our", "current", "project's", read the code first.
Type-Specific Guides
mockup
- Device frame: Actual device frame shape for mobile/tablet UI
- Multiple screens: Side-by-side layout + screen labels
- Placeholder data: Realistic data matching project context
- Tab/swipe for screen transitions
wireframe
- Hand-drawn (sketch) style: Slightly irregular lines, hand-drawn feel
- Black-and-white or grayscale. Minimal color
- Text areas shown as gray blocks (no "Lorem ipsum")
- Focus on layout and information structure, exclude visual details
erd
- Entity boxes with attribute lists. Distinguish PK/FK
- Relationship lines: 1:1, 1:N, N:M notation. Auto-track on node drag
- Include relationship type legend
flow
- Node types: Start/End (circle), Process (rectangle), Decision (diamond)
- Directional arrows. Auto-track on node drag
- Display branch conditions on connection lines
chart
- Auto-select appropriate chart type for the data (bar, line, pie, scatter, etc.)
- Axis labels + units, hover tooltips, legend
- Use Chart.js or D3.js
slides
- Reveal.js CDN-based
- Slide transition animations
- Code block highlighting (highlight.js)
- Speaker notes support (toggle with S key)
arch
- Separate system components by layer/zone (Frontend / Backend / DB / External)
- Label communication lines with protocols (HTTP, gRPC, pub/sub, etc.)
- Zoom/pan support
- D3.js force-directed or direct SVG generation
dashboard
- Arrange multiple charts/metrics in grid layout
- KPI cards at the top (numbers + change rates)
- Cross-chart interaction: Click one → filter others
timeline
- Horizontal or vertical timeline
- Event nodes + date labels
- Zoom/scroll for period navigation
- Use Mermaid gantt or D3.js
mindmap
- Radial expansion from center node
- Node collapse/expand
- Use Mermaid mindmap or direct SVG generation
kanban
- Columns: TODO / In Progress / Done (customizable)
- Drag and drop cards between columns
- Display labels/tags on cards
table
- Sort by clicking column headers (ascending/descending)
- Search/filter at the top
- Pagination or virtual scroll
- Cell highlight, row selection
Procedure
- Identify type + target. Ask if ambiguous. Read the file if a path is provided.
- Context gathering decision. Determine if project context is needed. If so, read relevant code/docs.
- Read
. Review common principles, aesthetics, CDN, and error prevention rules.references/common-rules.md - Read
. Start from the base HTML template.references/html-boilerplate.md - Generate HTML following the type-specific guide.
- Validate: Review the generated HTML.
- No smart quotes (curly quotes) in HTML attributes
- No unclosed tags
- No overlapping nodes/elements
- If issues found, fix and re-validate
- Instruct to
.open {filename}