Skip to content
Agent Orchestration
Skill

/html-explainer

The house standard for any skill that renders an HTML page for a person to read, covering design tokens, typography, components, diagrams, and quality gates so every generated page shares one calm, graphic-first look. Use when a skill's instructions say to render its output per

From plugin
pane
59739 skills4 agents

Claude Code

$ npx -y skills add greenfield-inc/Pane --skill html-explainer --agent claude-code

SKILL.md

html-explainer.SKILL.md
name: html-explainer
description: The house standard for any skill that renders an HTML page for a person to read, covering design tokens, typography, components, diagrams, and quality gates so every generated page shares one calm, graphic-first look. Use when a skill's instructions say to render its output per the html-explainer standards, or when the user asks for an HTML explainer of anything and no more specific skill applies.
argument-hint: "[what to explain, when invoked directly]"
allowed-tools: Read, Grep, Glob, Bash, Write

HTML Explainer

Task: $ARGUMENTS

A page for a person is pictures interrupted by words. The diagram carries the argument and the prose annotates it, the way a whiteboard sketch carries a design review; a section that has no visual either earns its place in prose or belongs folded into depth. Without one standard, every generated page invents its own fonts, colors, and structure, and the reader pays for the drift. This skill is where the look and the bar live. `eli5` renders with it (other pages follow `page`); invoked directly, it renders a one-off explainer of whatever the argument names.

The file

One self-contained `.html` file, tuned entirely for reading: there is no hosting, no publish step, no capability layer to design for. Inline CSS in a single `<style>` block, inline SVG for every drawing, no external requests of any kind: no CDN, no web fonts, no remote images, no scripts unless the page genuinely needs interaction, and then only vanilla inline JS. System font stacks only. Target well under 200KB. Write it to `./tmp/` (or the caller's stated destination), then open it unless the caller says not to. Inside Pane (`PANE_SESSION_ID` is set), run `runpane panels open --file <path> --source agent --yes --json`; it opens a browser tab in split view beside the conversation. Elsewhere use the system browser (`open` on macOS, `xdg-open` on Linux).

Tokens

Copy this block verbatim as the start of the style sheet. Extend it only by adding tokens, never by restyling these.

:root {
  --paper: #FAF8F5; --ink: #1F2328; --ink-soft: #5A5F66;
  --line: #E4DFD7; --panel: #FFFFFF;
  --accent: #0E7569; --accent-soft: #E3F0EE;
  --warn: #B45309; --warn-soft: #F7EBDD;
  --bad: #9F3A38; --bad-soft: #F6E8E7;
  --mono: ui-monospace, "SF Mono", SFMono-Regular, Menlo, Consolas, monospace;
  --serif: Georgia, "Iowan Old Style", "Times New Roman", serif;
  --sans: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, sans-serif;
}
@media (prefers-color-scheme: dark) {
  :root {
    --paper: #15191D; --ink: #E8E6E1; --ink-soft: #9BA1A8;
    --line: #2C3237; --panel: #1C2126;
    --accent: #2FA79A; --accent-soft: #16302D;
    --warn: #D98E3D; --warn-soft: #32271A;
    --bad: #CF6F6C; --bad-soft: #33211F;
  }
}

Drawings

Every page opens with one drawing directly after the masthead: the whiteboard sketch the rest of the page elaborates. More follow wherever a relationship, a flow, or a comparison is the point; drawing first and writing around the drawing is the intended order of work.

  • Inline SVG on the tokens: `--panel` fills, `--line` strokes,

`--accent` for the path that matters, `--warn` and `--bad` where verdicts are part of the picture, `--sans` labels at 12 to 14px.

  • Sketch energy over precision: slightly rounded corners, imperfect

widths, dashed strokes for the tentative and solid for the certain, arrows with real heads. Boxes and arrows over art; three boxes and an arrow beats a mural.

  • Warmth comes from the drawing, never from decoration: no gradients,

no shadows, no icon fonts, no clip art.

  • Legible at page width, `viewBox` set, no fixed pixel widths. A simple

subject gets a simple drawing, never a skipped one.

  • When the user wants a diagram they can edit themselves, that is the

excalidraw-pr-diagrams skill's job, not an inline SVG.

Type and layout

  • Body: `--sans`, 16px, line-height 1.6, `--ink` on `--paper`, one

centered column, `max-width: 72ch`, generous vertical rhythm.

  • `h1`: `--serif`, weight 500, ~2.1rem, `text-wrap: balance`. One per

page. Directly under it, a one-sentence standfirst in `--ink-soft`.

  • Section headers: a mono uppercase eyebrow (`.72rem`, letter-spacing

`.1em`, `--accent`) above an `h2` (~1.25rem). Sections are numbered when order matters and unnumbered when it does not.

  • Code: `--mono` on `--panel` with a `--line` border. Prose lines stay

under 90 characters; nothing scrolls horizontally except inside a `pre` with its own overflow.

Components

The whole vocabulary. A page uses what it needs and invents nothing:

  • `.badge`: mono, uppercase, `.72rem`, soft background. Accent for

identity and success, warn for caution, bad for failure. Badges carry verdicts; prose carries reasons.

  • `.panel`: `--panel` background, `--line` border, 6px radius, padded.

The unit of grouped content; grids of panels for comparisons (before/after, expected/actual).

  • `.callout`: a panel with a 3px left border in accent, warn, or bad.

One-paragraph emphasis, used sparingly.

  • `.card`: a panel with a mono eyebrow title, for repeating items

(a lesson, a promise, a dependency).

  • `details > summary`: for depth the reader opts into. The page must

read complete with every `details` closed.

The bar

Before opening the page, verify all of these; fix rather than ship:

1. The drawings could carry the page alone: a reader who skims only them and the headers leaves oriented. 2. Renders complete with JavaScript disabled and every `details` closed. 3. Nothing is fetched: no `src`, `href`, or `url()` that loads an external resource (stylesheet, script, font, image). URLs quoted as text or code content are fine; a page about an endpoint must be able to print it. 4. Both color schemes hold: readable in light and dark. 5. Every fact on the page came from the caller's material; the page adds structure and pictures, never claims. 6. One `h1`, sections in reading order, no

Read more
@greenfield-inc
Ships withpane

pnpm dlx runpane@latest

59739 skills4 agents
Get the whole plugin

Stars on GitHub

Momentum

stars in total
597
Star history of greenfield-inc/Pane on GitHub, a chart by star-history.com

Other skills on pane.

babysit-pr

babysit-pr

Use after a pull request is open to watch its CI and review-bot comments until it is ready. Fixes real findings, replies with a…

@greenfield-inc@greenfield-incInstall Skill
brief

brief

Use when a problem, idea, or change of intent is worth capturing as a long-form page, or the person asks to write it up. Never…

@greenfield-inc@greenfield-incInstall Skill
bug-intake

bug-intake

Use when the person reports broken behaviour or starts a session to file a bug. Reproduces and reports. Never fixes.

@greenfield-inc@greenfield-incInstall Skill
codebase-design

codebase-design

Shared vocabulary for designing deep modules. Use when the user wants to design or improve a module's interface, find deepening…

@greenfield-inc@greenfield-incInstall Skill
codebase-explorer

codebase-explorer

Explore the codebase to locate files, trace behavior, and document existing patterns with precise file references. Use when the…

@greenfield-inc@greenfield-incInstall Skill
cold-read

cold-read

Give a fresh agent something to walk through with zero context - a PR body, a landing page, a document, a process, instructions,…

@greenfield-inc@greenfield-incInstall Skill

More skills from other plugins.

1password

1password

Set up and use 1Password CLI for sign-in, desktop integration, and reading or injecting secrets.

clawdis12d393k
@steipete@steipeteInstall Skill
computer-use

computer-use

Drives the GUI of a visible local app window through `orca computer`: accessibility tree, clicks, typing, menus, dialogs, and…

orca12d111k
@stablyai@stablyaiInstall Skill
agentmail

agentmail

Use your assigned AgentMail inbox to read email tasks, explicitly send or reply, and check delivery. Provided automatically by…

paperclip12d109k
@paperclipai@paperclipaiInstall Skill
adr-create

adr-create

Create a new Architecture Decision Record with sequential numbering and AgentDB registration

@ruvnet@ruvnetInstall Skill

Questions about html-explainer.

Which coding agents can use html-explainer?
html-explainer installs in Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and Windsurf with the skills CLI, using each agent's flag shown on this page. Flowy's auto-invocation, which fires a skill when your prompt matches, runs in Claude Code only.